Report of the Secretary's Task Force on Black & Minority Health U.S. Department of Health and Human Services U.S. NATIONAL LIBRARY OF MEDICINE Volume IV: Cardiovascular and Cerebrovascular Disease Part 2 Report of the Secretary's Task Force on Black & Minority Health U.S. Department of Health and Human Services January 1986 Cxjfl 3(bCb SUBCOMMITTEE ON CARDIOVASCULAR AND CEREBROVASCULAR DISEASES TABLE OF CONTENTS PART 2 Supporting Papers 1. Shiriki K. Kumanyika, Daniel D. Savage: Ischemic Heart Disease Risk Factors in Black Americans ......... 229 2. Lucile L. Adams, Laurence 0. Watkins, Lewis H. Kuller, Daniel D. Savage, Richard Donahue, Ronald E. LaPorte: Relationship of Social Class to Coronary Disease Risk Factors in Blacks: Implications of Social Mobility for Risk Factor Change................285 3. Hector F. Myers: Coronary Heart Disease in Black Populations: Current Research, Treatment, and Prevention Needs ...... 303 4. Melford J. Henderson, Daniel D. Savage: Prevalence and Incidence of Ischemic Heart Disease in United States' Black and White Populations.............347 5. Helen P. Hazuda: Differences in Socioeconomic Status and Acculturation among Mexican Americans and Risk of Cardiovascular Disease ............... 367 6. Shiriki K. Kumanyika, Daniel D. Savage: Ischemic Heart Disease Risk Factors in Hispanic Americans ........ 393 7. Shiriki K. Kumanyika, Daniel D. Savage: Ischemic Heart Disease Risk Factors in Asian/Pacific Islander Americans .... 415 8. Shiriki K. Kumanyika, Daniel D. Savage: Ischemic Heart Disease Risk Factors in American Indians and Alaska Natives....................44 5 9. Lewis H. Kuller: Stroke Report 477 Ischemic Heart Disease Risk Factors in Black Americans Shiriki K. Kumanyika, Ph.D., M.P.H. Assistant Professor Department of Epidemiology Johns Hopkins School of Hygiene and Public Health Baltimore, Maryland Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland TABLE OF CONTENTS 1.0 CIGARETTE SMOKING ............... 232 1.1 Prevalence.................232 1.1.1 Data Sources 1.1.2 Black/white differences in prevalence of current and heavy smoking: 1965-1982 1.1.2.1 Current Smoking: Men 1.1.2.2 Heavy Smoking: Men 1.1.2.3 Current Smoking: Women 1.1.2.4 Heavy Smoking: Women 1.2 Impact..................239 1.2.1 Data Sources 1.2.2 Findings 1.3 Implications................2^4 1.4 Recommendations...............2^5 1.5 References ................2^6 2.0 CHOLESTEROL................248 2.1 Data Sources................248 2.2 Findings.................250 2.2.1 Total and lipoprotein cholesterol comparisons among black and white men 2.2.1.1 Total cholesterol: men 2.2.1.2 Lipoprotein cholesterol: men 2.2.2 Total and lipoprotein cholesterol comparisons among black and white women 2.2.2.1 Total cholesterol: women 2.2.2.2 Lipoprotein cholesterol: women 2.2.3 Basis of lipoprotein cholesterol differences in blacks and whites 2.2.4 Impact of cholesterol-associated risk on CHD disparities between blacks and whites 2.2.5 Intervention studies 229 2.3 Recommendation ................ ° 2.4 References.................263 3.0 ELEVATED BLOOD PRESSURE AND HYPERTENSION ....... 267 3.1 Prevalence.................2"' 3.1.1 Data Sources 3.1.2 Findings 3.1.2.1 Elevated Blood Pressure and Hypertension in Black and White Men, 1960-1980 3.1.2.2 Demographic Patterns : Men 3.1.2.3 Prevalence of Elevated Blood Pressure and Hypertension in Black and White Women, 1960-1980. 3.1.2.4 Demographic Patterns : Women 3.2 Left Ventricular Hypertrophy ........... 275 3.3 References.................279 4.0 CONCLUSION ...............282 230 INTRODUCTION The Task Force on Black and Minority Health Subcommittee on Cardiovascular and Cerebrovascular Diseases is attempting to understand and document the extent and determinants of racial disparities in cardiovascular and cerebrovascular diseases. This paper focuses on black Americans and on the ischemic heart disease (IHD) component of cardiovascular morbidity and mortality. IHD prevalence and incidence are addressed in a separate paper (Henderson and Savage). Data and issues relating to standard and other IHD risk factors are considered here. The specific objectives of this paper are to summarize and evaluate the available information on IHD risk factors in black men and women and to discuss implications of the data for the DHHS Task Force on Black and Minority Health. The scope of this paper is broad due to the need to cover several different risk factors in both black men and women and the need to scan a large literature on cardiovascular disease in whites for the relatively limited information on minority populations. For practical purposes related to the division of responsibilities among consultants to the Task Force, subjects which would be coherently considered together have been separated. Thus, IHD outcomes are discussed in relative isolation from cerebrovascular disease and diabetes. Also, an attempt is made to examine separately risk factors which occur in clusters and which interact. Some important factors such as obesity have already been incorportated into the subcommittee report by Dr. Watkins and, to avoid duplication of effort, have not been discussed here at length. The assumption is that a complete multivariate picture will emerge from a synthesis of several papers, by the Subcommittee and Task Force members. Based on the impression of Henderson and Savage from their review of prevalence and incidence of IHD in black Americans, the context for consideration of IHD risk among blacks has been as follows: * In black men, up to between ages 55 and 65, the prevalence- incidence of IHD appears to be similar to that in white men. * After about age 60-65 (the range of the average life span of black men), black men have a substantially lower prevalence/ incidence of IHD than white men (average life span 65 to 71 years). * IHD prevalence/incidence in black women is similar to or greater than that of white women, with no crossover. Cigarette smoking, cholesterol, hypertension, and diabetes are the standard risk factors for IHD. References describing prevalence, impact, and intervention aspects of three of these determinants have been selected and evaluated for implications related to the Task Force Mission. Diabetes' risks have been addressed by the Diabetes Subcommittee of the 231 Task Force. Left ventricular hypertrophy (LVH) (which is highly prevalent in blacks relative to whites) is discussed as a putative independent IHD risk factor. Emphasis has been given to recent references (1980 to present), with attention to time trends, in order to make the scope manageable and also to provide sufficiently timely perspectives for action. 1.0 CIGARETTE SMOKING AMONG BLACK AMERICANS 1.1 PREVALENCE 1.1.1 Data Sources National probability estimates of the prevalence of cigarette smoking among black adult men and women during and between 1965-1983 are available from the National Health Interview Survey (NHIS), the National Health Examination Surveys (NHANES), and the Centers for Disease Control Behavioral Risk Factor Survey (CDC-BRFS)(1-9). NHIS and NHANES data are from in-person interviews. CDC data are from a telephone survey. All of these sources provide estimates of rates of current smoking for the years covered. In addition, NHIS sources provide data on number of cigarettes smoked per day for current smokers (mean or categorical distribution), percentages of former smokers, smokers who have attempted to stop smoking, length of time since stopping, and tar and nicotine levels. NHANES provides percentages of heavy smokers, defined as 25 cigarettes or more per day. All sources provide comparison data for whites. However, NHIS and NHANES include Mexican-Americans in the white category. The treatment of age in published tables varies. NHIS reports use 20-24, 25-34, and 35-44 years, with 45-54 and 55-64 years either separate or combined as 45-64. NHANES data are reported for decades between 25 and 74 years. Only age-adjusted data were available for some black-white comparisons in both NHIS and NHANES. CDC data are reported for the categories 25-44, 45-64 and 65+ years. Time trends can be examined within NHIS and NHANES data, but only on a cross-sectional basis. No direct comparisons across data bases are appropriate due to the different sampling and data collection protocols. All smoking data in these surveys are self-reported (i.e., they have not been validated by biochemical tests). In addition, estimates of smoking status of black and white men and women in several large, widely-based samples were identified: * American Cancer Society data on smoking status of 1 million men and women (black n=25,000) over age 40, screened in 25 states in 1959 for prospective follow-up (10). " American Health Foundation data on smoking status by education and black or white race among 23,953 (black n=3628) men and women in nine U.S. cities (major east, west, north, south population centers), 1970-75 and 1976-80 (cross-sectional). Seventy percent or more of this population was over 45 years of age (range 20-80). The population consisted of hospitalized men and women diagnosed with nontobacco-related conditions and selected to be age-matched controls of adult cancer cases. Data for 232 the four race-sex groups are presented for each of the two time periods within two educational categories (high school or less and college or more) (11). * Multiple Risk Factor Intervention Trial (MRFIT) data on smoking status of 348,874 (black n=23,490) men ages 35-57 screened at 22 clinical centers in 18 U.S. cities during 1973-75 (12). * MRFIT data on smoking status of 12,866 (black n=931) men ages 35-57 with risk scores in the upper 10-15% of the Framingham distribution who were screened during 1973-75 and randomized to a specialized intervention or to control. Smoking status at screening and after intervention (thiocyanate adjusted) are presented (13). * Hypertension Detection and Follow-Up Program (HDFP) data on smoking status by systolic and diastolic blood pressure levels for 59,610 (black n=18,127) men and women ages 30-69 years screened during 1973-74 in 14 U.S. communities. Data for each sex-race group are presented for several blood pressure categories and by medication use with four age categories and age-adjustment (14). 1.1.2 Black/White Differences in Prevalence of Current and Heavy Smoking: 1965 - 1982 1.1.2.1 Current Smoking: Men Age-adjusted national estimates of prevalence of current smokers indicated an 8-9% higher prevalence of smoking in black compared to white men (age >=20 years) in 1965, 1976, and 1980--although rates for both black and white men had declined by approximately 14% (4). The direction of the difference in smoking rates between black and white men was the same in all data sources reviewed, i.e., rates for black men were higher than those for white men. However, the size of the difference varied from 2% in the American Cancer Society study (10) to 17-18% in college-educated hospitalized cancer controls (11) and in the HDFP screenees (14). In the NHIS and NHANES reports (6,7), the size of the black-white differential in male smoking rates remained relatively constant between 1970 and 1980 (less than 1% change in the difference between the two rates) (Table S-l). In the hospitalized cancer control population (11) with a high school education or less, the difference in black and white male smoking rates increased by 5% (black rates dropped less than white rates). In contrast, smoking rates in college educated men in this population showed a 6.5% narrowing of the black/white difference: rates among college-educated black men dropped more than those of the white men. Longitudinal data on the MRFIT Usual Care group showed a slight increase in the black/white difference (from 5.7% to 7.4%) between the time of screening 1973-75 and follow-up six years later. 233 Table S-l. Selected comparisons of self-reported smoking rates for black and white men, 1965-1980 (% current smokers in population subgroup). REF/ AGE SOURCE RANGE 1965 | 1971-75 | 1976 | 1976-80 | 1980 |CHANGE (6) >=20 BM 59.6 NHIS years WM 51.3 50.1 41.0 44.9 37.1 -14.7 -14.2 (7) 25-74 BM NHANES WM 55 43 50.7 39.6 -4.7 -4.0 (11) 20-80 hospitalized BM <=high sch WM >=college BM WM 59, 47 54 35 55.1 37.8 36.5 25.3 -4.3 -9.7 -15.7 -10.2 (12) 35-57 BM MRFIT WM screenees 50.0 35.9 (14) 30-69 BM HDFP meds WM 49.2 30.6 no meds BM WM 57.1 39.2 Age-specific rates in men show 45-65, and higher to be the highest higher prevalence group in data for difference for 35 changing pattern tabulations of national probability estimates of smoking the lowest rates in men over 65, next lowest among men rates among men 20-44. Rates in 25-34 year-old men tend Comparison of black and white rates indicate that the of smoking in black vs. white men is seen in every age 1965, 1976, and 1980. A relatively small black/white -44 year-old men in 1980 (44.2% vs. 42.2%) may indicate a for this birth cohort (6). Age-specific data for blacks but not whites are given in one of the available NHANES reports (7). Race-specific data from the CDC-BRFS for the 25-44, 45-64, and 65+ age-groups are included in Carter Center Report (8) as the basis for calculating attributable risks associated with smoking. These data (see below) do not fit the overall picture indicated by NHIS and 234 NHANES regarding smoking prevalence differences between black and white men--in either the size of the black/white difference or the prevalence of smoking in black men over 65 (NHIS 27.9% in 1980 (6) and NHANES II 26.9 in 1976-80 (7) as compared to 39.2 in the CDC data for 65+ black men. Table S-2. Percent of current smokers in the NHIS (1980) and CDC-BRFS (1982) U.S. population samples, black and white men ages 25 years and over NHIS CDC 25-34 35-44 45-64 65+ 25-44 45-64 65+ BM 52.0 44.2 48.8 27.9 39.4 38.3 39.2 WM 42.0 42.4 40.6 16.6 37.8 36.5 16.4 At the present writing, two important aspects of the CDC data are unclear: 1) the socioeconomic distribution of the sample (given that all respondents were contacted by telephone; and 2) the upper age limit of the sample compared to NHIS (actually the upper limit of NHIS is unclear; only NHANES data clearly designate the upper age limit). The definition of "current smoking" is not likely to differ greatly between surveys. If the CDC data are comparable to that from NHANES or NHIS, then the CDC prevalence rates can be interpreted as the more current data as compared to the pre-1981 estimates from the other two sources. As indicated above, with the exception of black males over 65 years, the CDC rates are lower than those reported for men of similar ages in 1980. Thus, the CDC rates are plausible as evidence of a continued decline in smoking rates in both races and a narrowing of the gap between black and white men under age 65. The high rate of smoking in black men over age 65 in the CDC data is difficult to interpret. The possible limitations of the CDC data should be kept in mind when evaluating the attributable risk calculations, presented later in this section, which are based on these smoking prevalence rates. 1.1.2.2 Heavy Smoking : Men NHIS age-adjusted estimates of the percent of adult smokers (age >=20 years) who smoke 25 or more cigarettes per day show an increase between 1965 and 1980 in both black and white males. The prevalence of heavy smoking is substantially higher among white men compared to black men and the increase in heavy smoking rates between 1965 and 1980 is steeper for white men (Table S-3). 235 Table S-3. Percentages of smokers or population sub-group smoking 25 or more cigarettes per day, black and white men, 1965-1980. REF/ AGE | 1965 | 1971-75 | 1976 | 1976-80 | 1980 |CHANGE SOURCE RANGE | | | I (6) >=20 BM 8, .6 10, .8 13.8 +5.8 NHIS * WM 26 .0 33, .3 37.3 +11.3 (7) 25-74 BM 7.5 8.5 +1.0 NHANES # WM 16.9 16.6 -0.3 (7a) 35-74 BM 5.6 9.6 +4.0 NHANES// (12) 35-57 WM BM 17.1 18.5 17.5 +0.4 MRFIT WM 26.9 screenees % of current smokers # % of population subgroup NHANES estimates of heavy smoking in the 25-74 year-old population --expressed per 100 population (% of population subgroup rather than as a proportion of smokers only)--indicate no appreciable change in the proportion of heavy smokers in the population. Thus, the increase in the percent of heavy smoking in men appears to be relative to the overall decline in the number of smokers. The decline in smoking among men has been among light (<15 cigarettes per day) and moderate (15-24 cigarettes per day) smokers. The NHANES estimates of heavy smoking in the 35-74 year old segment of the population do suggest an increase in the number of black men who are heavy smokers. Smoking rates in this age-group of men declined only slightly (48.7% to 46.5% from 1971-75 to 1976-80) (7a). Age-specific data indicate that 35-64 year-old smokers are more likely to be heavy smokers than the other age-groups among both black and white men (6). Among white men, all age-groups of smokers showed a consistent increase in the percent of heavy smokers between 1965 and 1980, with the steepest rise in the 35-44 year-old age-group. Among black male smokers, although no age-groups show a decline in rates, the age-specific picture of changes in heavy smoking among black men is much less consistent than for whites and is complicated by inadequate cell sizes for stable estimates in the 20-24 and 65+ year age-groups in 1965 and 1976. The age-adjusted 236 cross-sectional data are probably misleading as to trends in heavy smoking for any age-group or birth cohort of black men. In the NHANES data which do give estimates for all decades between 25 and 74, 25-34 year-olds show a 7% decrease while 35-44 and 55-64 year-olds show 8 and 7% increases. 1.1.2.3 Current Smoking: Women With the exception of MRFIT, all of the data sources described include data for women. Smoking patterns for women are very different from those of men. Rates are lower, time trends are different, and consistent black-white differentials are not observed. Data collected before 1970 (6,10) and after 1976 indicate an equivalent or slightly lower prevalence of smoking in black vs. white women (6,7,11) (Table S-4). Between 1971 and 1976 the prevalence of smoking in black women was higher in all samples examined (6,7,11,14). Smoking rates of black women have decreased to a greater extent than those of white women since 1976 in the NHIS and NHANES samples and in both the lower and higher educated segments of the American Health Foundation population of hospitalized cancer controls (11). Table S-4. Selected comparisons of self-reported smoking rates for black and white women, 1965-1980 (% current smokers in population subgroup). REF/ SOURCE AGE RANGE | 1965 | 1971-75 | 1976 | 1976-80 | 1980 |CHANGE (6) >=20 BF 32.7 NHIS WF 34.5 34.7 32.4 30.6 -2.1 30.0 -4.5 (7) 25-74 BF NHANES WF 46.2 34.0 31.6 33.0 -14.6 - 1.0 (11) 20-80 hospitalized BF <=high sch WF >=college BF WF 39.3 33.1 42.3 34.1 25.0 -14.3 28.0 - 5.1 26.5 -15.8 28.9 - 5.2 (14) HDFP 30-69 BF meds WF 29.8 23.8 no meds BF WF 38.6 33.8 237 The age-specific data for women indicate substantially lower prevalence of smoking in women age 65 years and older compared with younger women (6). The percent of current smokers among white women (but not for black women) age 65 years and over increased sharply between 1965 and 1980. NHANES data, which separate 45-54 and 55-64 year-old women, show a substantial .break towards lower smoking prevalence rates in black women over age 55 years. CDC and NHIS age-specific data on smoking prevalence in women are shown in Table S-5. As seen in the data for men, CDC rates are lower and may be evidence of a continuing downward trend. No striking black/white differences in prevalence are evident in either data set. If the data sets are comparable, then smoking trends in black and white women age 65 years and over are changing: rates in black women in this age-group may be increasing and those in this age-group of whites decreasing. Table S-5. Percent of current smokers in the NHIS (1980) and CDC-BRFS (1982) national probability samples, black and white women ages 25 and over NHIS CDC 25-34 35-44 45-64 65+ 25-44 45-64 65+ BF 34.2 36.5 34.3 9.4 30.8 24.4 13.6 WF 31.6 35.6 30.6 17.4 33.4 31.3 11.2 1.1.2.4 Heavy Smoking: Women Black/white differences in the prevalence of heavy smoking are evident among women and are of the same nature as those observed for men. Substantially more white than black women smokers report use of 25 or more cigarettes per day. The percentage of heavy smokers among white women smokers increased to the same extent as for white men in the NHIS data for 1965-80. A much lesser increase was observed in black women over these three surveys. NHANES estimates of the percentage of heavy smokers among 25-74 year-old women do not show an appreciable change for either black or white women (see Table S-6). 238 Table S-6. Percentages of smokers or population sub-group smoking 25 or more cigarettes per day, black and white women, 1965-1980. REF/ AGE | 1965 | 1971-75 | 1976 | 1976-80 | 1980 |CHANGE SOURCE RANGE | | | | | I (6) >=20 BF 4, ,6 5, .6 8.6 + 4.0 NHIS * WF 13, .9 20 .9 25.2 +11.3 (7) 25-74 BF 3.0 3.9 +0.9 NHANES # WF 7.4 8.6 +1.2 (7a) 35-74 BF 1.1 3.5 +2.4 NHANES# WF 6.0 8.2 +2.2 * % of current smokers # % of population subgroup Age-specific data for heavy smoking among women indicate that women smokers age 20 or less or over age 65 are least likely to fall into the heavy smoking category (6). The increase in percent of white women smokers who are heavy smokers between 1965 and 1980 is seen consistently in each age-group of white women (6). Trends across surveys cannot be evaluated in the NHIS data for black women heavy smokers due to inadequate numbers in several categories. NHANES age-group-specific percentages of heavy smokers are available for black women ages 25-54 for 1971-75 and ages 25-74 for 1976-80 (7). Apparent differences in smoking trends of different birth cohorts render general statements about time trends in heavy smoking among black women misleading—a situation similar to that observed in the heavy smoking data for black men. 1.2 Impact 1.2.1 Data Sources Four sources for estimates of the impact of smoking on coronary heart disease in nationally-based samples were identified: * The American Cancer Society prospective study (10) presents CHD mortality ratios for black men and women. Mortality ratios for blacks were calculated using deaths for whites in the comparable sex-age-group (5 year age-group at entry) as the denominator of "expected" deaths and deaths observed in blacks as the numerator. Actual death rates are not presented. Dose-response effects of smoking were examined by calculating the ratio of deaths for each of three smoking categories to deaths among nonsmokers, within race. Mortality ratios of black smokers with and without a history 239 of disease were also presented, using rates for white nonsmokers with no major disease history as the expected rates. The analysis is based on 642 and 487 CHD deaths in 22,000 black males and females 40 years and older during a 12 year period of follow-up between 1960 and 1972. Smoking categories were: never smoked, 1-9, 10-13, and 20+ cigarettes per day. CHD was defined by ICD-7 420.0, 420.1 and 420.2. * Kleinman, Feldman, and Monk (National Center for Health Statistics) estimated the extent to which the 20% decline in smoking between 1965 and 1976 could account for the decrease in CHD mortality for men and women 35-64 years old during the same time period (3). NHIS population percentages in 4 smoking categories were used to estimate expected percent change in CHD death rates by each of four dose-response models and for each of four risk models. Smoking categories were: 0, <15, 15-24, and >=25 cigarettes per day--recognizing that the larger percentage of former smokers in the nonsmoking category in 1976 vs. 1965 would lead to some degree of overestimation of effects. The dose-response models were: constant risk at any dose; logarithmic (risk increases rapidly in the lower dose range and then levels off); linear (similar increase in risk with each additional cigarette); and exponential (higher risk associated with each additional cigarette). Risk models were derived from the Veterans', American Cancer Society, British Physicians, and Framingham studies. CHD was defined by ICD-8 codes 410-413 (1968) extrapolating backwards to cover 1965, 1966, and 1967. " Five year cerebrovascular, CHD, CVD and all-cause mortality (6018 deaths) among 23,940 black and 325,384 white male MRFIT screenees (12) was subjected to logistic regression analyses to determine associations with several risk factors, including cigarette smoking. CVD was defined by ICD-9 codes 390-459; CHD by codes 410 to 414; cerebrovascular disease by codes 430-438. Smoking categories were: nonsmokers, 1-15, 16-25, 26-35 or >35 cigarettes per day at the time of screening. * The Carter Center Report includes estimates of the CHD deaths, years of life lost due to CHD death, and CHD cases attributable to smoking in the U.S. population and the proportion of CHD deaths and cases in persons 25 years or older that could be prevented if cigarette smoking were eliminated (8). Cigarette smoking is defined by the CDC 1982 data on percent of current smokers, described earlier. Population attributable risk estimates were derived from the Risk Factor Update Project logistic regression equation (Breslow et al., cited in 8). This equation is similar to the Framingham equation but is based on pooled equations for several major CVD studies and permits estimates for blacks and whites. CHD death was defined by ICDA codes 410-414 using 1980 death rates. Each of these analyses carries several assumptions and can only be interpreted in light of the risk model, morbidity and mortality data, and risk factor patterns which apply. For example, the MRFIT and Carter Center analyses of smoking effects hold several competing risk factors, and some risk factor interactions constant. The American Cancer Society and NCHS analyses do not. The ACS and MRFIT populations are volunteers for health risk-related studies and are prospective. Both the NCHS (3) and Carter 240 Center (8) reports provide smoking risk estimates for the U.S. population but they use smoking prevalence data from two different and not necessarily comparable data bases. 1.2.2 Findings The American Cancer Society study (10) demonstrates a significantly higher risk of 12-year CHD mortality among black men and women smokers under age 65, compared to white men and women nonsmokers in the same age range at entry. Higher CHD mortality associated with smoking was seen in black men and women both with and without a major disease history. The prevalence of smoking was somewhat higher in white men than black men, consistent with the higher CHD mortality among the white men (i.e., the black-white CHD mortality ratio was less than one for men in all age groups). However, certain characteristics of the study sample suggest that it was not representative of blacks in the general population. For example, the prevalence of smoking was lower among black men compared to white men in this sample and the prevalence of major disease was not substantially greater in the black men, as might have been expected. The sample was geographically broad, but consisted of friends, relatives, and neighbors of ACS volunteers. The authors point out that educational levels in the sample are higher than those reported in census figures and that inner city persons are underrepresented. CHD mortality among black women was greater than for white women for women ages 40 through 59 at entry. Somewhat more white women than black women smoked and were heavy smokers. The prevalence of major disease, as defined in this study, was 12% higher among the black women (32% vs. 20%). More of the black women were 30% or more overweight (17% vs. 4%). The MRFIT screenee population (12) is probably healthier than average due both to the use of employment sites for screening and to the fact that the highest risk men have been excluded due to enrollment in the randomized trial. Individual effects of smoking are considered, i.e., after adjustment for other risk categories (age, serum cholesterol concentration, diastolic blood pressure level). Smoking is more prevalent in the black men, but they are less heavy smokers. Smoking was associated with a significant, similar excess risk of CHD in black and white men in this population. This association was not dependent on the presence of heart attack or diabetes; removal of such men from the analysis did not alter the result. Dose response effects of smoking on CHD mortality were positive and generally similar in blacks and whites, allowing for some instability of cause-specific mortality rates among the relatively small numbers contributing to rates for black men in the heavy smoking categories. CHD mortality was higher in smokers than nonsmokers among black men with and without hypertension and with and without elevated serum cholesterol. Kleinman et al. (3) determined the percentage decline in CHD deaths for 35-64 year-olds in the U.S. population between 1965 and 1976 which could be attributed to a decline in smoking over the same period. 241 Estimates are only presented for men, since smoking rates for women did not decline significantly in the time period addressed. Although four models are used to calculate expected decline in deaths associated with changes in smoking patterns, estimates based on the "constant risk" model lend themselves to the most straightforward interpretation. Estimates with this model are based simply on the percentage of current smokers, without regard to average number of cigarettes smoked. In black males, changes in smoking prevalence could account for 23, 23, and 21 percent of the CHD mortality decline for the 35-44, 45-54 and 55-64 year-old age-groups, respectively. Comparison percentages for white males were 40, 50, and 16%. Thus, changes in smoking had less of an impact on the decline in CHD mortality in black as compared to white men 35-54 and slightly more of an impact in black men 55-64 years old. Results based on the other risk models, which take the number of cigarettes smoked into account, show similar results. The factors other than smoking which contributed differentially to the decline in CHD mortality among black and white men are uncertain. The Carter Center Report (8) provides estimates of the number and percent of CHD deaths and cases which would theoretically be eliminated if smoking prevalence were reduced to zero. These estimates are conservative in that they assume that no other risk factors would change in a favorable direction. The CDC-based smoking prevalence data used in these calculations are in Tables S-2 and S-5. In general, the population attributable risk estimates (8, Table 17) closely parallel the prevalence data and to that extent show a high black-white ratio only among men over age 65, for whom the CDC data show a questionably high prevalence rate. For both white and black men between 25 and 65, smoking-related preventable CHD deaths are estimated at 33 to 37% of total CHD deaths. For CHD cases, percentages of preventable events are approximately 10% lower. A higher proportion of preventable black compared to white male CHD cases is suggested in the 45-64 year age-group as well as in men 65 years and over. (Note, the picture for smoking-related preventable cerebrovascular disease deaths is quite similar to that for CHD in this analysis, except that estimates of preventable cases among 45-64 men are the same for blacks and whites.) Other "CHD impact" variables analyzed were proportional number of cases and deaths attributable to smoking, years of life lost, number of hospital and restricted activity days and medical costs which can be attributed to smoking. These estimates are summarized in Table S-7 for men. 242 Table S-7 Population attributable risk fractions for smoking for several CHD variables in non-white and white men: estimates for the U.S. population (8) Table No Variable Non-white White (percent of cases) 23 CHD deaths 33 21 24 CHD-related lost yrs. of life 35 34 25 CHD cases 28 17 26 CHD-related hospital days 27 16 27 CHD-related days of rest, activ. 28 17 28 CHD-related medical expend. 27 16 All of these estimates show a proportionately greater impact on CHD of smoking in non-white compared to white men, with the exception of years of life lost. Estimates of the percent of CHD deaths in women that would be prevented if smoking in women were eliminated show a much lesser impact than in men, but very similar within-age percentages in black and white women: BW - 8.2, 6.7, and 3.7 % in the 25-44, 45-64 and 65+ age-groups, respectively; 8.8, 8.3, and 3.1% in white women. Referring back to the CDC prevalence data in Table S-5, the impact of smoking elimination in women is about a third of the prevalence of smoking, whereas in men the impact was approximately equal to the smoking prevalence. Data for women for the other CHD variables are summarized in Table S-8. Smoking attributable CHD risks in women are much lower than those for men and there are no racial differences. As pointed out in the Carter Center report, the impact is least for deaths and higher for variables which reflect the relatively young age of occurrence of smoking related CHD deaths and cases. Table S-8. Population attributable risk fractions for smoking for several CHD variables in non-white and white women: estimates for the U.S. population (8) Table No Variable Non-white White (percent of cases) 23 CHD deaths 4 4 24 CHD-related lost yrs. of life 7 8 25 CHD cases 12 10 26 CHD-related hospital days 10 9 27 CHD-related days of rest, activ. 12 10 28 CHD-related medical expend. 10 10 243 Confidence in the estimates for the 65+ age male population is limited by the departure of smoking prevalence figures in black men from the pattern usually seen in the 65+ age-group. In men 25-64, based on data indicating roughly equivalent smoking prevalence in black and white men, the theoretical benefits of eliminating smoking in blacks are substantial. The population impact of smoking on CHD morbidity and mortality in black men may be slightly greater than for white men in the 45-64 age-group. Smoking has a lesser impact on CHD in women compared to men. When risks associated with smoking alone are considered, CHD impact is not different in black compared to white women. However, smoking enhances a more unfavorable multivariate risk picture for blacks than for whites. 1.3 Implications That the basic association between smoking and CHD holds for blacks as well as whites is apparent from the studies available, although the degree of risk appears to be somewhat less in black compared to white men, and similar in black and white women. The dynamic state of population smoking behaviors over time and the substantial changes in IHD risk that occur with smoking cessation severely limit the utility of cross-sectional data. In addition, the sampling approach once appropriate to descriptions of the U.S. population is outdated. Since race has been recognized as an important proxy health variable, sampling approaches should be designed to permit racial subgroup comparisons as well as descriptions of U.S. population parameters in the aggregate. Such sampling should include awareness of the need to partition race and socioeconomic status effects. For example, the extent of socioeconomic differences among blacks is demonstrated in the hospitalized cancer control data from the American Health Foundation (11): both high school and college educated black men smoked more than white men, in both time periods studied (see Table S-l). However, the decrease in smoking prevalence among college educated black men was more than three times that in high school educated black men and nearly 60% greater than that in college educated white. Data aggregated on either educational level or race would not convey this difference, but the difference has important implications for intervention and provides important clues to patterns which might be hidden in longitudinal data which classify blacks and whites only on smoking status at entry. The longitudinal data available are outdated in terms of the current smoking prevalence and, as implied above, in terms of the actual smoking status of persons initially classified as smokers during the year or months immediately prior to fatal or morbid events. Consideration of potential benefits of intervening on smoking in black men and women should include the relative likelihood of success per unit of effort compared to the other risk reduction possibilities. Only minimal literature was identified relating to perspectives for intervention on smoking in the black population. NCHS data indicate that in 1978, black 244 men and women, although lighter smokers, are more likely to smoke cigarettes with high tar and nicotine content, that fewer blacks are successful when they attempt to stop smoking although more blacks had tried to quit at some time (2). Sex difference in attempts to stop smoking were more pronounced in blacks than whites. Similar percentages of men and women had stopped smoking or attempting to in the year prior to interview (28.5 and 31.5%). Among black men and women, substantially more women than men had attempted to quit (44.2% vs. 34.1%). In the MRFIT randomized trial, thiocyanate-adjusted percentages of black and white men in the special intervention group who had stopped smoking at each annual visit were quite similar (13, Table 8). It is interesting to note that the unadjusted percentages among the black men were 6 to 15% higher than those of whites before adjustment, implying that reporting bias may have been greater in the black men. However, the success at smoking cessation in the special intervention group of black men was equal to that of whites (35.9 and 35.5 of white and black special intervention men stopped smoking). Both blacks and whites in the usual care group were less successful in smoking cessation, and whites did somewhat better than blacks (22.2 vs. 18.9% stopped smoking). A study of factors influencing trial and adoption of smoking among children and adolescents (ages 8-17) in Bogalusa, Louisiana reported significant race and sex differences in the social learning which leads to eventual smoking (15). The number of "smoking models" in the child's environment was found to predict whether a child would try smoking, but the pattern of influence was different in black compared to white children. 1.4 Recommendation A broader base of short-range prevalence and trend data on smoking behaviors in blacks is needed, followed by specialized interventions to accomplish permanent smoking cessation in blacks. Although smoking cessation may be associated with some weight gain, there will still be a net benefit of the smoking cessation. Smoking differentials between blacks and whites do not parallel morbidity-mortality patterns. However, smoking is a risk factor for both CHD and for cancer. The need to reduce smoking prevalence among blacks should not be neglected. 245 1.5 References 1. Bonham, GS. Use habits of cigarettes, coffee, aspirin, and sleeping pills, United States, 1976. Vital and health statistics. Series 10. Data from the National Health Survey; No. 131. DHEW Publication No. (PHS) 80-1559). October 1979. 2. Moss AJ. Changes in cigarette smoking and current smoking practices among adults. United States, 1978. National Center for Health Statistics. Advancedata. No. 52. September 19, 1979. 3. Kleinman JC, Feldman JJ, Monk MA. The effects of changes in smoking habits on coronary heart disease mortality. Am J Pub Health 1979;69:795-800. 4. Schoenborn CA, Danchik KM. Health practices among adults, United States, 1977. National Center for Health Statistics. Advancedata No. 64. November 4, 1980. 5. National Center for Health Statistics. Health United States 1980, Tables 27 (pp 162-63) and 28 (pp 164-65). DHHS Pub. No. (PHS) 81-1232. Public Health Service. Washington. U.S. Government Printing Office, Dec. 1980. 6. National Center for Health Statistics. Health United States, 1983, Tables 30 (pp. 129-130) and 31 (pp. 131-132). DHHS Pub. No. (PHS) 84-1232. Public Health Service. Washington. U.S. Government Printing Office, Dec. 1983. 7. Rowland MB, Fullwood R. Coronary heart disease risk factor trends in blacks between the first and second National Health and Nutrition Examination Surveys, United States, 1971-1980. Am Heart J 1984;108:771-778. 7a. Rowland M, Fulwood R, Kleinman JC. Changes in heart disease risk factors, in reference 7, pp. 13-17. 8. Haynes SG, White CC, Tolsma D, Newman JM, McGee D. Closing the gap for cardiovascular disease. Carter Center Report August 27, 1984 (unpublished manuscript courtesy of Dr. Haynes) Table 16. 9. Bradstock MK, Marks JS, Forman M, Gentry EM, Hagelin GC, Trowbridge FL. Behavioral risk factor surveillance. 1981-83. MMWR 1984;33:lss-4ss. (reference on methodology of CDC Survey) 10. Garfinkel L. Cigarette smoking and coronary heart disease in blacks. Comparison to whites in a prospective study. Am Heart J 1984;108:802-806. 246 11. Covey LS, Mushinski MH, Wynder EL. Smoking habits in a hospitalized population. Am J Pub Health 1983;73:1293-1297. 12. Neaton JD, Kuller LH, Wentworth D, Borhani NO. Total and cardiovascular mortality in relation to cigarette smoking, serum cholesterol concentration, and diastolic blood pressure among black and white males followed up for five years. Am Heart J;1984;108:759-770. 13. Connett JE, Stamler J. Responses of black and white men to the special intervention program of the Multiple Risk Factor Intervention Trial (MRFIT) Am Heart J 1984;108: 839-850. 14. Hypertension Detection and Follow-up Program. Baseline characteristics of the enumerated, screened, and hypertensive participants. The Hypertension Detection and Follow-Up Program Cooperative Group. Hypertension 1983;5:IV-1 - IV-295. 15. Hunter SM, Baugh JG, Webber LS, Sklov MC, Berenson GS. Social learning effects on trial and adoption of cigarette smoking in children. The Bogalusa Heart Study. Prev. Med. 1982;11:29-42. 247 2.0 CHOLESTEROL 2.1 Data sources The adequacy of the available data on cholesterol levels in black men and women is markedly influenced by the increasing interest in lipoprotein cholesterol components rather than total cholesterol. In particular, high-density lipoprotein cholesterol has been the focus of considerable recent interest both as a possible protective factor against IHD and as one for which black persons may have an advantage relative to their white age-sex peers (1-4). For total cholesterol, the Health Examination Survey (NHES, 1960-62) and National Health and Nutrition Examination Survey (NHANES I 1971-74 and NHANES II 1976-80) provide national probability estimates of distributions and means by age, sex, and race for black and white children and adults (5-7,9). Analyses of 1971-74 serum cholesterol levels by socioeconomic characteristics have also been published (8). The NHES and NHANES data provide representative estimates of mean serum cholesterol levels and the prevalence of elevated cholesterol (using arbitrary cut points) for the time periods they cover and can, to some extent, be used to evaluate changes in the cholesterol levels of blacks and whites over time. Cycle II of the National Health and Nutrition Examination Survey will provide data on 1976-80 levels of high-density lipoprotein cholesterol (HDL) for a national probability sample of U.S. black and white men and women ages 20-74. Estimates of mean HDL levels for race-sex groups in 6 age categories have been made. However, release of these data has been delayed due to a need to resolve methodological issues which influence interpretation of NHANES HDL values relative to published values from other large studies. These issues are currently being evaluated by the National Center for Health Statistics and the National Heart, Lung, and Blood Institute. For the purposes of this report to the Task Force on Black and Minority Health, Robert Murphy, Director of the NCHS Division of Health Examination Statistics, has authorized inclusion of certain provisional statements on black-white differences in mean HDL levels. These statements, included under "Findings" (section 2.2), were provided by Robinson Fulwood, Health Statistician (communication to Shiriki Kumanyika), with certain caveats. Sample sizes for the NHANES II HDL values in blacks are small. Numbers of intended and actual HDL values in each age-group of black men and women are given in Table C-l. 248 Table C-l. Sample sizes for estimates of HDL values in black men and women ages 20-74, United States 1976-80. Black Men Black Women Age Group Intended" Actual// Intended* Actual// 20-24 79 65 94 77 25-34 139 110 145 116 35-44 70 53 103 81 45-54 62 45 100 79 55-64 129 91 135 98 65-74 128 92 152 115 Total 607 456 729 566 * Intended = number of respondents for whom total cholesterol was determined # Actual = number of respondents for whom an adequate quantity of blood remained after determinations of total cholesterol and trigylcerides to permit determination of HDL-cholesterol as well For whites, HDL values are available for approximately 3800 men and 4300 women. The smallest sample in any age-group for white men is 472 (581 intended); for white women 506 (624 intended). The sample sizes for blacks are adequate for stable estimation of mean HDL cholesterol levels in each sex-age-group according to the criteria applied by the Division of Health Examination Statistics. However, most of the age-sex samples of HDL values for blacks do not meet optimum criteria for estimation of standard errors. Thus tests of statistical significance of black-white differences must be interpreted cautiously. All of the sample sizes meet the criterion of a minimum of 25 for standard error estimation. However, only two of the groups (25-34 year-old black men and 65-74 year-old black women) in the total cholesterol data (i.e., the numbers shown in Table C-l as "intended") met a second criterion of having observations in both of the paired probability sampling units for at least 12 of the 32 pseudostrata. Since the first criterion is met in the HDL data, standard errors will be published, but with asterisks indicating that full confidence can not necessarily be placed in the starred standard error values or in related statistical tests. In this summary for the Task Force, emphasis has been given to published studies which present data comparing lipoprotein rather than only total cholesterol levels in blacks and whites, and primarily in the adult population at IHD risk. (If needed for reference, HDL comparisons in black and white neonates, preschool, school-age and adolescent children are summarized in a 1984 review by Glueck et al.(2)). Studies associating either total or HDL cholesterol with IHD risk have also been reviewed. 249 Data sources identified in addition to the NCHS population estimates can be categorized according to the relevant issues, as follows: a) comparisons of lipoprotein cholesterol levels in black and white men, women, or families (4,11-18); b) cholesterol comparisons between blacks and whites where the socioeconomic status of the black and white subjects is considered (4,11,13,14,17,19,33); c) studies related to possible explanations for HDL differences in blacks and whites, including data on lipoprotein cholesterol in Caribbean and African blacks (4,20-27); d) data comparing the effectiveness of dietary intervention on cholesterol levels in blacks and whites (28,29); e) estimates of the relative impact of. elevated serum cholesterol on IHD risk in blacks compared to whites (30-32). The ascertainment of cholesterol levels is not equivalent across studies. Methodological differences include whether or not subjects had fasted before blood was drawn, whether cholesterol was measured in plasma or serum, whether samples were frozen before analysis and, if so, the time and conditions of storage, and analytic procedures. Although the effect of these factors on cholesterol determinations is not large, the small differences introduced complicate the interpretation of small differences in group means which, if real, may have important risk implications. Thus, comparisons across studies are made here only with reference to the relative consistency of racial differences observed in different data sets. 2.2. Findings 2.2.1 Total and lipoprotein cholesterol comparison among black and white men 2.2.1.1 Total cholesterol: men Representative national data permitting comparisons of total cholesterol levels between black and white men within the three time periods of the NCHS examination surveys are shown in Table C-2. Data from the NHES are not strictly comparable with the NHANES data(6). However, comparison of cholesterol distributions in NHANES I and II is appropriate for estimation of time trends. 250 Table C-2. Age-specific comparisons of mean serum cholesterol in national probability samples of U.S. black and white men in three time periods between 1960 and 1980. NHES NHANES I NHANES II 1960-62* 1971-75// 1976-80! ( mgms per deciliter ) AGE BM WM BM WM BM WM 25-34 195.0 207.1 214.2 209.9 35-44 217.2 228.1 224.7 231.5 207.1 206.8 45-54 226.6 231.4 237.1 239.9 55-64 229.5 233.6 243.8 239.9 227.8 228.4 65-74 224.0 229.9 237.3 236.4 217.5 221.8 age 210.5 218.8** 225.7 217.5!! 215.4 216.7!! adjusted * Ref 5, Table 6 ** for ages 18-74; age-adjusted estimates not published for 25-74 year-old adults // age-specific data for 1971-74, from ref 6, table 4 ! ref 30, table 15a; data for ages <65 grouped by 20 year spans !! age-adjusted data for 25-74 year-olds taken from ref 9, table VI In 1960-62, age-specific and overall serum cholesterol levels of black men were consistently less than those of white men. Age-specific data for 1971-75 and 1976-80 do not show consistent black-white differences. The data in Table C-3 indicate a decrease in the prevalence of elevated serum cholesterol in black men between 1971-75 and 1976-80 but this decrease was not statistically significant. The prevalence of elevated serum cholesterol was eight percent higher in black than white men in 1971-75, decreased to 3% higher in 1976-80. The largest decrease in cholesterol levels among black men was among men 55 or older. 251 Table C-3. Prevalence of Serum Cholesterol >260 mg/dl in black and white men, 1971-75 and 1976-80: age specific rates for black men ages 25-74; age- adjusted rates for black and white men ages 25-74. NHANES I 1971-75 NHANES II 1976-80 rate per 100 population AGE Black Men 25-34* 10.9 9.3 35-44 24.3 23.6 45-54 28.7 25.3 55-64 32.3 24.2 65-74 28.0 18.7 25-74// BM:23.2 WM:15.3 BM: 19.3 WM:16.3 * Ref 9, Table IV // Ref 9, Table V A consistent socioeconomic influence on cholesterol in black men has not been reported. • An NCHS analysis of cholesterol levels in 1971-74 (8) suggests that the impression of socioeconomic effects is related to the SES variable used. Cholesterol levels in men tended to increase with increasing levels of annual family income. An inverse effect of education on serum cholesterol levels was observed in white but not black men. A discrepancy in SES effects on cholesterol in black and white men was observed in a MRFIT screenee population in California (33). The prevalence of elevated cholesterol (defined as >= 260 mg/dl) among white men was lower at higher SES levels, using categories of a 77 point education-occupation scale as the SES measure. Cholesterol levels of the black screenees were different at different SES levels, but the direction of these differences was not consistent (33). There were noteworthy black-white differences in the prevalence of elevated serum cholesterol in two of the six SES categories (lowest and second highest categories; prevalence higher in whites in both groups). Prevalence rates in the other categories were similar (33, Table 3). The pattern seen may have been somewhat affected by differing distributions of the black and white men within the 35-57 year age range (52% of black men vs. 61% of white men were 45-57 years old). Total cholesterol levels of the black men in the Framingham Minority Study population were lower than those of the white men in every age-group studied (age range 20-69) and the age-adjusted means in black and white men were significantly different (11). In this sample, 56% of black compared to 44% of white men under age 50 were college-educated. Sixty-four percent of black compared to 22% of white men over age 50 were college-educated. 252 2.2.1.2 Lipoprotein cholesterol: men HDL levels are consistently higher in black men than in white men in the 20-74 year-old cross-section of NHANES II respondents. The size of the differences is relatively similar in the age-groups between 20 and 54 (7,6,8,and 6 mg/dl in each successive age-group). Levels in white men are the same at ages 55-64 and 65-74, but levels in black men are higher in the 55-64 year-olds than in younger or older black men. Thus the black-white difference is highest in the 55-64 year age-group (10 mg per dl). The difference in 65-74 year-old men is 6 mg/dl, similar to that seen before age 55. The statistical significance of these differences has not yet been tested. The relationship of the HDL distributions to the respective total cholesterol distributions should be considered in interpreting these HDL data. As summarized by Glueck et al. (2), studies in black and white male children and adults suggest that HDL levels in black males are higher than those of whites from early childhood on and that the size of the difference increases after maturation. Maturational changes in lipid profiles occur in males of both races (decline in HDL, increases in low-density lipoprotein (LDL) cholesterol and triglycerides) but appear to follow a less atherogenic course in black vs. white males. The Evans County Heart Study investigators compared HDL and LDL cholesterol levels of 110 randomly-selected black men with those of white men individually matched on age and total cholesterol level (17). Men with evidence of CHD were excluded. With total cholesterol levels held constant (by matching) mean HDL cholesterol levels were significantly higher in the black men (by 11 mg/dl) and LDL cholesterol levels significantly lower (by 6 mg/dl). The differences were not statistically explained by occupation, education, social class, or smoking habits but were to some extent explained by body mass differences between the black and white men. Consistency of these differences across age cannot be determined due to the age-matching analytical approach. HDL and LDL cholesterol levels of black and white men in the upper fertile of the Quetelet index were not different (17, Figure 8). More recent studies in the Evans County population further specify apolipoprotein differences between black and white men (4). Apo A-I levels were significantly higher and Apo C-II significantly lower in black than in white men, suggesting that black men in Evans County have a greater proportion of the (lighter, cholesterol-rich) HDL-2 subfraction. Alcohol consumption rather than race was a significant variable in levels of Apo A-II. The differences in levels of Apo C-II but not Apo A-I were still statistically significant after adjustment for age, Quetelet index, smoking, and alcohol consumption. The Lipid Research Clinics (LRC) Program has contributed several studies which consistently indicate a more favorable lipid profile in black compared to white men (the LRC protocol excludes men using antidiabetic, antiuricemic, antihypertensive, or lipid lowering medications). The Cincinnati LRC Population Study compared lipoprotein cholesterol levels of 253 43 pairs of black and white men, ages 20 to 60, matched on age and total plasma cholesterol. HDL levels were significantly higher and LDL/HDL ratios significantly lower in black men (15). A substudy in 9 pairs of hypercholesterolemic men (total plasma cholesterol >250 mg/dl) yielded the same results (15). The Princeton School District Study of the Cincinnati LRC compared HDL and LDL levels in black and white men ages 20-59. This study also indicated significant racial differences (higher HDL and lower LDL in black men) (12). Age-group (20-39 or 40-59) did not affect relative HDL and LDL levels, although total plasma cholesterol was higher in the older men. A study of socioeconomic factors and CHD risk in the same population indicated that HDL and LDL differences between blacks and whites remain after adjustment for education and occupation (13). Black-white differences in cholesterol in men age 20-44 years were analyzed in pooled screening data from five LRC centers (14). Higher HDL and lower LDL were observed for black men when the distributions of these variables were compared by pooling across age with covariance adjustment for age, total cholesterol, and Quetelet Index. Similar differences were observed when comparisons were made between pairs of black and white men matched for age and total cholesterol. Stepwise covariance adjustment for possible confounding factors indicated that with age and total cholesterol effects statistically removed, Quetelet index had a small negative confounding effect on black-white differences in HDL-cholesterol levels (i.e., the difference increased from 8.3 to 9.5 mg/dl when Quetelet index was added as a covariate). Triglycerides had a positive confounding effect (the difference decreased from 9.5 to 7.9 mg/dl with the further addition of triglycerides to the equation). In a collaborative LRC report combining data on men ages 20-79 from ten study centers, a greater proportion (17% vs. 3%) of black men than white men with high total cholesterol were of the hyperalpha type (16). This was found using either of two definitions of hypercholesterolemia (race-age-sex specific 95th percentile or an age-specific cut point). Although the reported black-white differences in HDL and LDL levels were not statistically explained by socioeconomic status variables in the above studies, the limited range of socioeconomic status among black subjects may give an incomplete picture on this point. Black men in the Evans County cohort were of low socioeconomic status whereas the white subjects include both low and high SES persons (32). In the Cincinnati Princeton School District Study (13), some black households were in the highest education and income categories but the numbers of these were too small to permit stratum specific comparisons. Thus, although an overall black-white difference may still remain after differences in the black-white SES distribution have been accounted for, it is not certain that the high SES black men have the advantageous lipid profiles relative to their white counterparts. 254 The Framingham Minority Study (11) is unique in having compared blacks and whites in a sample which included a high proportion of college-educated black men. HDL-cholesterol levels of black men in this study were not higher than those of the white men. In fact, the mean HDL levels of the black men were significantly lower than those of the white men, although the ratio of mean HDL cholesterol to mean total cholesterol (which was also significantly lower in black men) was similar. The Framinghman Minority Study finding is supported by a report from the Bogalusa study. Hunter and coworkers reported that although alpha-lipoprotein levels of black children were higher and beta-lipoprotein levels lower than those of white children in most parental-education categories, an opposite pattern was observed among children whose parents had a post-graduate education (19). When the parental SES variable was blue vs. white collar occupation, black children but not white children had SES-related differences in mean alpha-lipoprotein levels. Racial differences in alpha-lipoprotein levels decreased from 64 mg/dl (black children higher) among children of blue collar workers to only 6.9 mg/dl (black children higher) among children of white collar workers. The relative difference in alpha-lipoprotein levels between children of blue vs. white collar parents was much larger than that for total cholesterol, beta- or pre-beta-lipoprotein levels. In summary, when one assesses differences in total cholesterol between blacks and whites, HDL-cholesterol levels should be considered. The total cholesterol/HDL cholesterol ratio is needed for useful HDL comparisons in data where subjects have not been matched on cholesterol levels or HDL levels. Significant declines in the prevalence of elevated serum cholesterol levels are not evident in the NHANES data for either black or white men. Socioeconomic aspects of cholesterol-related risk may vary with the indicator used, are probably different for black and white men, and are not monotonic in black men. More data are needed on this point. With the exception of a study which included a large percentage of high SES black men, higher HDL cholesterol levels in black compared to white men are a consistent finding, across several studies and at several ages. Apparently, the protection theoretically afforded by HDL cholesterol does not effectively counter other risk factors in black men under age 60 years, since prevalence/incidence of IHD appears similar in younger black and white men. 2.2.2. Total and lipoprotein cholesterol comparisons among black and white women. 2.2.2.1 Total cholesterol: women National probability estimates of mean serum cholesterol and elevated serum cholesterol in black and white women in the United States are shown in Tables C-4 and C-5. In 1960-62, overall mean cholesterol levels of black women were lower than those of white women. Levels of black and white women tended to diverge with increasing age, with the largest difference in women 55-64. For NHANES I, the age-specific data indicate slightly higher cholesterol 255 levels in black vs. white women under age 65, although the overall age-adjusted means for the two racial groups are identical. Cholesterol levels of black and white women changed only slightly between 1971-75 and 1976-80. Black women aged 45-64 years had slightly lower mean levels than white women. Table C-4. Age-specific comparisons of mean serum cholesterol in national probability samples of U.S. black and white women in three time periods between 1960 and 1980. NHES NHANES I NHANES II 1960-62* 1971-75// 1976-80! milligrams per deciliter AGE BF WF 25-34 197, ,9 198, .0 35-44 213, .1 214, .2 45-54 232, ,2 237, ,6 55-64 239, .0 264, .9 65-74 258 .2 267, .4 age 217, .8 224, , 1* adjusted BF WF BF WF 207.3 202.8 216.9 216.4 198.3 198, .3 244.7 242.5 260.8 256.4 237.8 239 ,7 259.7 261.8 242.6 246, .3 221.2 221.4!! 219.2 219 .9! ! * Ref 5, Table 6 ** for ages 18-74; age-adjusted estimates not published for 25-74 year-old adults // age-specific data for 1971-74, from Ref 6, Table 5 ! Ref 30, Table 15b; data for ages <65 grouped by 20 year spans !! age-adjusted data for 25-74 year-olds taken from Ref 9, Table VI Table C-5 indicates inconsistent trends in the prevalence of elevated serum cholesterol (using a 260 mg/dl cut point) among different age-groups of black women between NHANES I and NHANES II. The largest change was a decrease among the oldest group of women; however, prevalence rates for women age 35-64 years increased slightly. The similarity of age-adjusted rates for black women in NHANES I and NHANES II may be misleading because age adjustment obscures age-specific trends. 256 Table C-5. Prevalence of Serum Cholesterol >260 mg/dl in black and white women, 1971-75 and 1976-80: age specific rates for black women aged 25-74; age- adjusted rates for black and white women aged 25-74. NHANES I 1971-75 NHANES II 1976-80 AGE rate per 100 population Black Women 25-34* 35-44 45-54 55-64 65-74 8.7 7.7 23.5 29.3 43.5 6.0 13.3 25.0 32.2 29.9 25-74// BF:19.6 WF:20.2 *Ref 9, Table IV BF: 19.0 WM:19.5 // Ref 9, Table V The NCHS analysis of socioeconomic influences on serum cholesterol levels of NHANES I adult female respondents indicated an inverse effect of both education and income (8). This inverse effect in women was in contrast to the weakly positive effect observed in men. The income effect was significant in white but not black women. In the Framingham Minority Study (11), total cholesterol levels of the black women were lower than those for white women for all age-groups studied (range 20-69) except women aged 50 to 59 years. Sixty-five percent of the black vs. 32% of the white women under age 50 years in the Framingham Minority Study sample were college-educated. Among the women over age 50, 27% of the black vs. 14% of the white women were college-educated. 2.2.2.2 Lipoprotein cholesterol: women In the NHANES II data, high-density lipoprotein cholesterol levels are consistently higher in black women than in white women although the differences are smaller than those seen in men. As in men, the largest black-white difference in HDL levels among women is in the 55-64 year age group. Differences between ages 20 and 54 are 4, 2, 3, and 1 mg/dl in successive age-groups. In the 55-64 year-old women, HDL levels are 6 mg/dl higher in black women than white women, due primarily to an increase in the black women. HDL levels are somewhat lower for both black and white women in the 65-74 year-old age-group, but the black-white difference decreases to 4 mg/dl due to a larger drop in the HDL levels of the black women. Thus, black-white differences in lipoprotein cholesterol for women parallel'those for men but are of a lesser magnitude. This parallel picture of slightly, but not substantially higher HDL and lower LDL levels among black compared to white women is reported in several studies 257 described earlier for men (12,14-17). In the Cincinnati LRC Princeton School Study, HDL differences in black compared to white women were larger in women not using exogenous sex steroid hormones than in women who were using them and were statistically significant in the nonhormone using group (12). In the collaborative LRC Prevalence Study, the increase in age- total cholesterol-adjusted HDL differences in black and white women when the Quetelet index was added to the equation was greater than the increase in men. When the Quetelet index was in the model, the higher mean HDL cholesterol levels in black women were statistically significant (further addition of trigylcerides decreased the difference to nonsignificant levels) (14). HDL-cholesterol differences in women in the Framingham Minority Study also parallel those in men, i.e., HDL and total cholesterol levels in black women were significantly lower than in white women (with the exception of higher total cholesterol levels in black than in white women age 50-59). However, HDL/total cholesterol ratios were similar in black and white women. The mean Quetelet index of black and white women in this sample was not different. In summary, distributions of serum cholesterol and lipoprotein cholesterol are similar in black and white women. Data regarding socioeconomic influences on cholesterol in women are limited. NHANES I data suggest that levels of cholesterol increase as SES decreases. If so, then the larger proportion of black than white women who are of low SES are at risk for elevated cholesterol. Black-white differences in levels of HDL cholesterol which would be expected on the basis of findings in men are suggested in the data for women but are usually very small and not significant. When the entire adult age spectrum is considered, serum cholesterol levels appear to have remained stable in both black and white women between 1971-75 and 1976-80. However, a higher prevalence of elevated cholesterol in 35-64 year-old black women in NHANES II compared to NHANES I is suggested by the data. This may be a noteworthy finding, if valid. Any trends towards increased risk in black women should be targeted for intervention, due to the excess of other risk factors and IHD prevalence in this group. 2.2.3 Basis of lipoprotein cholesterol differences in blacks and whites Gartside et al. (22) examined the extent and direction of black-white differences in several factors associated with HDL-levels including: intakes of protein, fat, carbohydrate, saturated fat, oleic acid, lineoleic acid and cholesterol (total and per kilogram of body weight); Quetelet index, self-reported maximum and minimum weight and weight at age 25; alcohol intake; cigarette smoking,; leisure time exercise and habitual physical activity; self-reported prevalence of diabetes; and hypertension. Of the factors found to differ between blacks and whites in NHANES II, the substantially higher relative weight of black compared to white women appeared to be the most likely explanation for the loss of the HDL advantage for black women. Substantial differences in obesity between 258 black and white men were not evident in the national population data and most of the studies reviewed. The greater prevalence of diabetes and the greater tendency to be treated with HDL-lowering antihypertensive medications among blacks than whites would work in the same direction for women. Higher levels of habitual activity in black than in white men at some ages would tend to enhance differences (22). As noted above, adjustment for the Quetelet index has been more often reported to alter the observed black-white differences in HDL-cholesterol than does controlling for other possible confounders (14,17). The environmental influence on lipoprotein patterns is also suggested by the absence of significant black-white differences at birth (2) and by a possible disappearance of racial differences among high SES groups. It might be said that a genetic basis for lipoprotein profile differences between blacks and whites is suggested by studies of blacks in other countries. For example, in a study of men from 13 countries, the highest HDL to total cholesterol ratios were observed among African men (21). In his review of CHD in developing countries, Watkins cites several studies indicating that lipoprotein profiles among African and Caribbean men are favorable to low CHD risk (20). These findings clearly do not separate genetic from environmental influences on differential lipid patterns. Gartside et al. postulate that genetic selection has favored blacks with relatively higher HDL related to the protective role of HDL-cholesterol against sleeping sickness (22). In this respect, it is also interesting to note a recent report by Dai et al. of a possible direct or indirect positive association between HDL levels and plasma testosterone in a sample of 255 MRFIT participants (23). Studies by Ross et al. have postulated that higher testosterone levels in black men might explain their higher prevalence of prostatic cancer compared to white men (24). 2.2.4 Impact of cholesterol-associated risk on CHD disparities between blacks and whites The importance of black-white differences in lipoprotein profiles may be much more in the utility of racial contrasts for etiologic analyses than in the potential for reducing racial disparities in CHD incidence and prevalence. The Cincinnati LRC group concluded (based on familial aggregation studies) that the relative contribution of genetic compared to environmental factors to HDL levels may be greater in blacks than whites (2). This line of reasoning is compatible with the lesser consistency and strength of association between lipid levels and measures of obesity in black compared to white children (25) and adults (26) and with lesser effects of cigarette smoking and oral contraceptives on HDL in black compared to white children and adolescents (27). This might suggest that interventions directed specifically at the HDL component of cholesterol would yield a relatively lesser benefit for blacks. This is a testable hypothesis. 259 Berenson and coworkers have pointed out that the proneness of blacks to other CHD risk factors may override the degree of "subtle protection afforded by high alpha-lipoprotein levels (25). Considering that the overall lipid picture is better for black men and no worse for black women compared to their age-sex peers, the similar rates of CHD in black men compared to white men before age 55 or 60 and in black women at all ages indirectly supports this conclusion. Analyses in the Cooperative Lipoprotein Phenotyping Study indicate that HDL levels were lower in CHD cases compared to control in black, white and Japanese men (18). However, a tabulation of CHD prevalence rates by HDL cholesterol level implied a threshold effect rather than a gradient. Prevalence rates did not differ markedly among men with HDL-cholesterol levels between 45 and 75 mg/dl. The striking prevalence differences were between men with HDL cholesterol levels less than 35 or 25 mg/dl (18). The HDL levels in most of the studies of black-white differences report mean cholesterols in the intermediate rather than low range, possibly because persons at highest risk have been excluded by the study protocols. However, this does not advise against efforts to prevent or reverse trends toward elevated total cholesterol in the black population. The NHANES data suggest that the prevalence of elevated serum cholesterol in black men may be higher than that of white men (Table C-3) using 260 mg/dl as a cut point. The impression from multivariate analyses is that, with other factors held constant, the impact of elevated serum cholesterol on CHD in blacks is not substantially different from that in whites (31,32). In addition, when higher prevalence of hypertension and smoking (black men) are factored in, any added risk factor multiplies the risk of IHD. The Carter Center findings on the potential benefits of reducing serum cholesterol levels in the population can be summarized as follows (see section 1.2.1 (data sources on the impact of smoking on IHD risk) for a description of the Carter Center analytic approach). In men, the proportion of CHD deaths that could be prevented if borderline (240-259 mg/dl) and elevated (>= 260 mg/dl) cholesterol levels were decreased to 219 mg/dl (considered to be normal) is slightly greater for black than white men (14.1 vs. 9.0 percent) in the 25-44 year age range. The comparable proportions for black and white men ages 45-64 and 65 plus are approximately eight and three percent, respectively, in both racial groups. Proportionate reductions in CHD cases are 11 and 7% for black and white men ages 25-44, 12 and 10 percent for 45-64 year-old men and 8% in black and white men ages 65 and over (30). In women, the greatest reductions in CHD mortality associated with reduction of borderline and elevated cholesterol levels to below 220 mg/dl would be among women over age 45, approximately 18% in both black and white women. Below age 45, reduction in CHD deaths among women is a third of that after age 45 and is the same (6%) for black and white women. The potential benefits of such cholesterol reductions on CHD cases among women is more striking: 18% among women under age 45 and 38-39% among women over age 45, again, with the same level of benefit in women of both races (30) 260 The Carter Center report (30) also presents estimates for the theoretical possibility of reducing cholesterol levels above 260 mg/dl to below that level. Percentage reductions follow similar patterns as those described above, but are much lower. For example, reductions in CHD cases among women over age 45 would be 18-22 percent compared to the 38-39 percent which would occur if "normal" were defined as below 220 mg/dl. Carter Center population attributable risk fractions for cholesterol on CHD are summarized in Table C-6 for non-white and white men and women. The earlier stated pattern of greater potential benefits in women but with no substantial racial difference is also reflected in these estimates. Table C-6. Population attributable risk fractions for cholesterol on several CHD variables in non-white and white men and women: estimates for the U.S. population (30) Table CHD MEN WOMEN No* Variable Non-white White Non-white White (percent 1 of cases) 23 deaths 5 1 5 1 17 19 24 yrs.lost life 10 8 I 15 16 25 cases 11 9 1 35 38 26 hospital days 10 9 1 37 38 27 restrict, act. 11 9 1 35 38 28 medical expend. 11 9 1 38 38 in Reference 30 2.2.5 Intervention Studies The limited evidence identified on the effectiveness of dietary intervention on cholesterol levels of hyperlipidemic black women and men is quite encouraging. A report from the Chicago Heart Association (29) indicated that an educational intervention was more successful with black men and women than with whites. Most of the participants in this study were black, primarily from the inner city. Six to nine months after entry into the program, the black participants, who had higher cholesterol levels at baseline, had made slightly greater changes in fat consumption and had achieved greater reductions in cholesterol levels (8.1 vs. 4.5%, p<0.05) than white participants. The majority of participants (60%) were women; dietary adherence among women was better than among men. Nutritional knowledge scores of blacks were lower at baseline but were equivalent to those of whites after one, six, and nine months of intervention. In the MRFIT study, plasma cholesterol levels of black and white men decreased to a similar extent. Improvements of special intervention compared to usual care black men were slightly greater than for whites (28). 261 2.3 Recommendat ion The evidence of equivalent or better lipoprotein profiles in blacks does not justify inattention to cholesterol levels in blacks. Efforts targeted to blacks should be included in the general effort to shift the distribution of cholesterol levels in the population downward. This should not, however, take priority over targeted interventions on hypertension, obesity, LVH, and smoking in blacks. 262 2.4 References 1. Gillum RF, Grant CT. Coronary heart disease in black populations. II. Risk factors. Am Heart J;1982:104:852-864. 2. Glueck CJ, Gartside P, Laskarzewski PM, Khoury P, Tyroler HA. High-density lipoprotein cholesterol in blacks and whites. Potential ramifications for coronary heart disease. Am Heart J. 1984;108:815-826. 3. Curry CL, Oliver J, Mumtaz FB. Coronary artery disease in blacks. Risk factors. Am Heart J 1984;108:653-657. 4. Heiss G, Schonfeld G, Johnson JL, Heyden S, Hames CG, Tyroler HA. Black-white differences in plasma levels of apolipoproteins. The Evans County Heart Study. Am Heart J 1984;108:807-814. 5. National Center for Health Statistics. Serum cholesterol levels of adults. United States 1960-62. Vital and Health Statistics. PHS Publication No. 1000. Series 11. No.22 March 1967. Washington, D.C. 6. National Center for Health Statistics. Total serum cholesterol levels of adults 18-74 years. United States 1971-74, by Sidney Abraham, Clifford Johnson, and Margaret Carroll. Vital and Health Statistics. Series 11, Data from the National Health Survey; No 205. DHEW Publication No. (PHS) 78-1652. April 1978. 7. National Center for Health Statistics. Total serum cholesterol levels of children 4-17 years, United States, 1971-74, by Sidney Abraham, Clifford Johnson, and Margaret Carroll. Vital and Health Statistics. Series 11. No. 207. DHEW Pub No. (PHS) 78-1655. 1978 8. National Center for Health Statistics. Serum cholesterol levels of persons 4-74 years of age by socioeconomic charac- teristics, United States 1971-74, by Robinson Fulwood, Sidney Abraham, and Clifford Johnson. Vital and Health Statistics Series 11, No. 217. DHEW Publication No. (PHS) 80-1667. 9. Rowland ML, Fulwood R. Coronary heart disease risk factor trends in blacks between the first and second National Health and Nutrition Examination Surveys, United States, 1971-1980. Am Heart J 1984;108:771-779. 10. Namboodiri KK, Green PP, Kaplan EB, Tyroler HA, Morrison JA, Chase GA, Elston RC, Rifkind BM, Glueck CJ. Familial aggregation of high density lipoprotein cholesterol. The Collaborative Lipid Research Clinics Program Family Study. Arteriosclerosis 1983;3:616-626. 263 11. Wilson PWF, Savage DD, Castelli WP, Garrison RJ, Donahue RP, Feinleib M. HDL-cholesterol in a sample of black adults. The Framingham Minority Study. Metabolism 1983;32:328-332. 12. Morrison JA, Khoury P, Mellies M, Kelly K, Horvitz R, Glueck CJ. Lipid and lipoprotein distributions in black adults. The Cincinnati Lipid Research Clinics Princeton School Study. JAMA 1981;245:939-942. 13. Khoury P, Morrison JA, Laskarzewski P, Kelly K, Mellies MJ, King P, Larsen R, Glueck CJ. Relationships of education and occupation to coronary heart disease risk factors in schoolchildren and adults. The Princeton School District Study. Am J Epidemiol 1981;113:378-395. 14. Tyroler HA, Glueck CJ, Christensen B, Kwiterovich PO. Plasma high-density lipoprotein cholesterol comparisons in black and white populations. Circ 1980;62:IV-99 - IV-107. 15. Morrison JA, de Groot I, Kelly KA, Mellies MJ, Khoury P, Edwards BK, Lewis D, Lewis A, Fiorelli M, Heiss G, Tyroler HA, Gleuck CJ. Black-white differences in plasma lipids and lipoproteins in adults. The Cincinnati Lipid Research Clinic Population Study. Prev Med 1979;8:34-39. 16. Morrison JA, Khoury P, Laskarzewski P, Gartside P, Moore M, Heiss G, Glueck CJ. Hyperalphalipoproteinemia in hyper- cholesterolemic adults and children. Trans Assoc Am Phys 1980;93:230-243. 17. Tyroler HA, Hames CG, Krishan I, Heyden S, Cooper C, Cassell JC. Black-white differences in serum lipids and lipop- rotein in Evans County. Prev Med 1975;4:541-549. 18. Castelli WP, Doyle JT, Gordon T, Hames CG, Hjortland MC, Hulley SB, Kagan A, Zukel WJ. HDL cholesterol and other lipids in coronary heart disease. The Cooperative Lipoprotein phenotyping study. Circ 1977;55:767-772. 19. Hunter SM, Frerichs RR, Webber LS, Berenson GS. Social status and cardiovascular disease risk factor variables in children. The Bogalusa Heart Study. J Chron Dis 1979; 32:441-449. 20. Watkins LO. Coronary heart disease and coronary disease risk factors in black populations in underdeveloped countries. The case for primordial prevention. Am Heart J 1984;108: 850-862. 264 21. Knuiman JT, West CE, Burema J. Serum total and high density lipoprotein cholesterol concentration and body mass index in adult men from 13 countries. Am J Epidemiol 1982;116: 631-642. 22. Gartside PS, Khoury P, Glueck CJ. Determinants of high-density lipoprotein cholesterol in blacks and whites. The second National Health and Nutrition Examination Survey. Am Heart J 1984;108:641-653 23. Dai WS, Gutai JP, Kuller LH, Laporte RE, Falvo-Gerrard L, Caggiula A. Relation between plasma high-density lipoprotein cholesteroland sex hormone concentrations in men. Am J Cardiol 1984;53:1259-1263. 24. Ross RK, Paganini-Hill A, Henderson BE. The etiology of prostate cancer. What does the epidemiology suggest? The Prostate 1983;4:333-343. 25. Berenson GS, Webber LS, Srinivasan SR, Cresanta JL, Frank GC, Farris RP. Black-white contrasts as determinants of cardiovascular risk in childhood. Precursors of coronary artery and primary hypertensive disease. Am Heart J 1984;108:672-683. 26. Khoury P, Morrison JA, Mellies MJ, Gleuck CJ. Weight change since age 18 in 30-to-55-year-old whites and blacks. Assoc- iations with lipid values, lipoprotein levels, and blood pressure. JAMA 1983;250:3179-3187. 27. Webber LS, Hunter SM, Baugh JG, Srinivasan Sr, Sklov MC, Berenson GS. The interaction of cigarette smoking, oral contraceptive use, and cardiovascular risk factor variables in children. The Bogalusa Heart Study. Am J Pub Health 1982;72:266-274. 28. Connett JE, Stamler J. Responses of black and white males to the special intervention program of the Multiple Risk Factor Intervention Trial. Am Heart J 1984;108:839-849. 29. Mojonnier ML, Hall Y, Berkson DM, Robinson E, Wethers B, Pannbacker B, Moss D, Pardo E, Stamler J, Shekelle RB, Raynor W. Experience in changing food habits of hyperlipidemic men and women. J Am Dietet A 1980;77:140-148. 30. Haynes SG, White CC, Tolsma D, Newman JM, McGee D. Closing the gap for cardiovascular disease. Carter Center Report. August 27, 1984. 265 31. Neaton JD, Kuller LH, Wentworth D, Borhani NO. Total and cardiovascular mortality in relation to cigarette smoking, serum cholesterol concentration, and diastolic blood pressure among black and white males followed up for five years. Am Heart J 1984;108:759-770. 32. Tyroler HA, Knowles MG, Wing SB, Logue EE, Davis CE, Heiss G, Heyden S, Hames CG. Ischemic heart disease risk factors and twenty- year mortality in middle age Evans County black males. Am Heart J 1984;108:738-746. 33. Kraus JF, Borhani NO, Franti CE. Socioeconomic status, ethnicity, and risk of coronary heart disease Am J Epidemiol 1980;111:407-414. 266 3.0 ELEVATED BLOOD PRESSURE AND HYPERTENSION 3.1 Prevalence 3.1.1 Data Sources The NCHS Health Examination and Health and Nutrition Examination Surveys (NHES and NHANES) provide probability estimates for several blood pressure variables for black and white adults in the U.S. population for three time periods (1960-62, 1971-75, and 1976-80). Measurements and definitions in NHES and NHANES I and II are sufficiently comparable to permit valid estimates of trends in hypertension prevalence among blacks and whites over the 1960-1980 period. Several NCHS reports (1-4) have served as the basis for the following discussion of prevalence of elevated blood pressure and hypertension for race-sex groups in the age range 25-74 years. When considering the black-white comparisons which follow, the larger standard errors of the estimates for blacks should be kept in mind. In addition, for 1971-75, published tables are available for estimates of blood pressure variables by geographic region, urban/rural residence and socioeconomic status. These data are of interest because of possible variation in hypertension patterns in blacks and whites in demographic sub-categories in the U.S. Emphasis has been given to variables for which tabulations for all three surveys are available by race and sex. These variables are as follows: elevated blood pressure, (defined by NCHS as systolic blood pressure (SBP) of at least 160 and/or diastolic blood pressure (DBP) of at least 95), hypertension (elevated blood pressure and/or on medication), previously undiagnosed hypertension, treated hypertension, and controlled hypertension. Although the cut points for designation of elevated blood pressure are somewhat arbitrary, the ones used by NCHS are useful for comparing NHES and NHANES data. There is no indication that different conclusions about black-white disparities would be drawn if variables reflecting the entire blood pressure distributions were to be used. NCHS blood pressure estimates reported here are based on three readings taken at specified points in the examination protocols (the NHANES I and NHANES II measurements each include one reading with the respondent supine; other readings were with respondents sitting). Data for the different time periods have been age-adjusted (by NCHS) to the U.S. population at the midpoint of NHANES II (1978) for comparison purposes. Now that substantial progress in blood pressure treatment and control among both blacks and whites has been documented, the changes in patterns of hypertension in blacks, compared to whites, over time deserve special attention (i.e., have there been decreases in the black-white disparities on hypertension variables?). Also, trends in the prevalences of target organ damage and hypertension-related end points deserve careful assessment. 267 Prevalence is a function of incidence and duration. That is, the percentage of people in the population who will be found to have hypertension at a given point in time is influenced both by the percentage of new cases of hypertension at that point (incidence) and by the percentage of people with hypertension who are surviving (duration). If the rate at which new cases of hypertension develop remains constant, the proportion of hypertensives in the population will increase if hypertensives are effectively treated and thereby live for longer time periods. Effective treatment of hypertension should be reflected in a decrease in the prevalence of elevated blood pressure. However, the survival of hypertensives is not only a function of blood pressure level but also of the degree of end organ damage and comorbidity. These may differ among controlled hypertensives with similar blood pressures. For example, among hypertensive men in the Hypertension Detection and Follow-up Program randomized trial, black men had substantially higher mortality than white men (black-white ratios of 5-year-life table all-cause mortality were 1.8 and 1.9 in stepped and referred care groups)(6). The variable which most differentiated the black and white men at baseline was prevalence of ECG-LVH. In addition, more of the black men were on antihypertensive medication at baseline and more had a history of diabetes or stroke (6). The incidence of hypertension is difficult to ascertain. Long-term follow-up of large populations, with careful attention to the initial definitions of who is normotensive at the outset, are needed. Attempts to ascertain hypertension incidence among blacks have been summarized in a recent review by Oni (5). In cross-sectional data used for time trend analyses, incident and prevalent cases among the newly-identified hypertensives are difficult to separate on the basis of blood pressure measurement alone. Newly identified cases of elevated blood pressure may have been of longstanding duration. Patterns of hypertension incidence over time are essentialy unknown. At present, there is no basis for assuming that the incidence of hypertension has decreased, since the major efforts have been at the level of secondary, rather than primary prevention. These secondary prevention efforts are expected to prolong the lives of people with hypertension, i.e., to increase the duration of hypertension, and might be expected to result in increased prevalence over time. If similar or decreased prevalence is seen, then the possibility that treated hypertensives are dying at similar or greater rates should be considered. This could be due to increased case-fatality among hypertensives (perhaps associated with a particular type of treatment), or to coincidental increases in overall mortality of persons with hypertension due to other causes. Another likely possibility is that hypertensives, especially mild hypertensives, may go off of medication after a long period of control and not have increases in blood pressure for several months. In a cross-sectional survey, such individuals would not be classified as hypertensive by either the elevated blood pressure or the "on medication" criteria. 268 3.1.2 Findings 3.1.2.1 Elevated Blood Pressure and Hypertension in Black and White Men, 1960-1980 Prevalence rates of elevated blood pressure for black and white men in 1976-80 (the most current national probability sample estimates available) are compared in Table H-1. In general, elevated blood pressure is observed with greater frequency in older age-groups. Elevated blood pressure was more common in black than in white men surveyed in 1976-80 in all age groups. However, the slope of increased prevalence with age is different for black and white men. Large racial differences in rates of elevated blood pressure were observed among the 35-44 and 55-64 year-old men (but only at these ages). The black-white prevalence differences among the 35-44 year-olds probably reflect the younger age of onset of hypertension among blacks compared to whites; white rates have "caught up" in the 45-54 year-old group. The disparity among the older men (55-64 years of age) may reflect somewhat lower proportions of effectively treated hypertensives in black compared to white men at these ages (see Table H-3). The inconsistency in the pattern of race difference by age suggests that comparisons based on age-adjusted data may be inadequate for some purposes. Table H-1. Age-specific and overall prevalence of elevated blood pressure in black and white men ages 25-74, United States, 1976-1980.* rate per 100 population with elevated blood pressure** Age-group 25-34 35-44 45-54 55-64 65-74 Total Black Men 11.7 22.3 23.0 39.2 27.5 22.4 White Men 8.4 10.6 21.2 22.3 24.5 15.9 Black minus White 3.3 11.7 1.8 16.9 3.0 6.5 * from Reference 1, Table 5 ** defined as SBP >=160 mmHg and/or DBP> >=95 mmHg As shown in Table H-2, the overall prevalence of elevated blood pressure has declined substantially in both black and white men ages 25-74. The decline among black men (35.3%) was larger than for white men (11.8%), lowering the black-white ratio from 2.0 in 1971-75 to 1.45 in 1976-80. The highest prevalence of elevated blood pressure among black men was in the 55-64 year age-group in both surveys. The percent decrease in prevalence 269 was relatively consistent across age-groups among black men--slightly lower than the overall 35.3% in the 25-34 year-old men and slightly higher among the 35-44 year-old men. Table H-2. Percent decline in prevalence of elevated blood pressure between 1971-75 and 1976-80, black men in age-groups between 25 and 75 and age-adjusted for black and white men, U.S. rates per 100 population 1971-75 1976-80 % change* 25-34 16.4 11.7 - 28.7 35-44 37.7 22.3 - 40.8 45-54 34.7 23.0 - 33.7 55-64 59.9 39.2 - 34.5 65-74 43.7 27.5 - 37.1 25-74 Black 35.7 23.1 - 35.3 25-74 White// 18.0 15.9 - 11.7 * calculated from rates in reference 4, table II; elevated blood pressure defined as SBP >=160 and/or DBP >=95 mmHg. // from Reference 4, Table V On the surface at least, the data in Table H-2 indicate progress in hypertension detection and control between 1971 and 1980 which benefitted black men to a relatively greater extent than white men. However, the data on hypertension prevalence are only partially helpful in confirming this impression of progress in reducing the black-white hypertension disparities (Table H-3). As noted earlier, better coverage of black hypertensives theoretically implies improved survival and increased prevalence (assuming no decrease in incidence). The proportion of hypertensives among black men ages 25-74 increased between NHES and the NHANES I augmentation survey (1974-75) (NHANES la), but then decreased such that prevalence in 1976-80 was lower than that observed in 1960-62. Although this difference was apparently not statistically significant, it is in contrast to what the NCHS report identifies as a significant increase in hypertension prevalence among white men over the 20 year period (1). The meaning of this trend in prevalence is unclear. Trends in hypertension awareness, treatment, and control among black and white men are also summarized in Table H-3. The interpretation of these variables is relatively straightforward and shows considerable improvement for men of both races, particularly for black men, in awareness and use of medication. The percent of black hypertensive men who had not been diagnosed previously dropped sharply between NHES and NHANES la and declined further between NHANES la and NHANES II. Improvements in 270 awareness among white men were similar, but of a lesser degree. The proportion of unaware hypertensives was nearly 15% greater for black than for white men in 1960-62 and nearly 5% less in 1976-80. A similar cross-over was observed for the percentage of black compared to white hypertensives on medication between NHES and NHANES II: a slightly smaller percentage of hypertensive black than white men were on medication in 1960-62; a slightly higher percentage of black compared to white men were on medication in 1976-80. Levels of awareness were 60% or greater among both black and white men; medication use was lower--around 40%. However, although the extent of control tripled for black men and nearly doubled for white men during this period, hypertensive black men on medication were less likely than white men to have adequately controlled blood pressure in all three survey periods. Overall levels of control were still quite low in men (on medication) of both races in 1976-80. Table H-3. Rates of hypertension prevalence, awareness, treatment, and control, black and white men, ages 25-74, U.S., 1960-62, 1974-75, and 1976-80* percent of hypertensives in population// percent of hypertens ives not previously diagnosed percent of hypertensives on medication percent of hypertensives on medication whose blood pressure was controlled//// 1960-62 1974-75 1976-80 BM 31.8 37.1 28.3 WM 16.3 21.4 21.2 BM 70.5 41.0 35.7 WM 56.7 42.3 40.6 BM 18.5 24.0 40.9 WM 22.4 25.9 38.3 BM 5.0 12.7 16.1 WM 11.8 15.1 20.9 * Reference 1, Table 7; data are age adjusted to the U.S. population at the midpoint of NHANES II (1976-80). // defined as either SBP >=160, DBP >=95, or taking antihypertensive medication //// hypertensives on medication and with blood pressures below the levels considered as elevated range (below the cut points for elevated blood pressure) 271 3.1.2.2. Demographic Patterns : Men Mean blood pressure levels of black men in 1971-75 were consistently higher than those of white men in all four regions of the country. Mean blood pressure levels among black men are highest in the Northeast and lowest in the West (3, Table 17; age-adjusted rates). Similarly, black men had higher mean blood pressure levels than white men in both urban and rural areas. The mean blood pressures of black men in rural areas were higher than for black men in urban areas (3, Table 18; age-adjusted rates). Tabulations by annual family income and level of education indicated higher blood pressures among the black compared to white men at each level of these variables (3, Tables 19 and 20; age-adjusted rates), with the exception that black men with less than 5 years of education had mean blood pressures considerably below the rest of the distribution. (Note that these are trends in the data and not necessarily statistically significant differences.) 3.1.2.3 Prevalence of Elevated Blood Pressure and Hypertension in Black and White Women, 1960-1980. Age-specific and overall prevalence rates of elevated blood pressure in black and white women in 1976-80 are shown in Table H-4. Prevalence increased with age in women of both races and was 1.7 to 3 times higher in black than white women in every age-group. Black-white differentials are larger in women over 45 years of age than among younger women (this is also the age-group when blood pressure levels of women catch up with and then exceed those of men). Table H-4. Age-specific and overall prevalence of elevated blood pressure in black and white women ages 25-74, United States, 1976-1980.* rate per 100 women with elevated blood pressure** Age Group 25-34 35-44 45-1 54 55-64 65-74 Total Black Women 4.3 17.6 37, ,3 36.4 43.4 23.2 White Women 2.3 6.5 12, ,1 18.3 26.3 11.4 Black minus White 2.0 11.0 25, .2 18.1 17.1 11.8 from Reference 1, Table 5 defined as SBP >=160 mmHg and/or DBP >=95 mmHg Table H-5 indicates decreases in the prevalence of elevated blood pressure among black women in each age-group, but with a considerable age-time interaction (i.e., very different levels of decrease in different 272 age-groups). The decrease among black women 25-74 years is approximately representative of the level of decrease among 35-44 and 55-64 year-old women. A very large decrease (65%) was observed among the youngest group of women. Minimal decreases were observed in the other two age-groups. Unlike black men, who showed proportionately larger overall decreases than white men in 1971-1980, the overall decreases in black and white women were of the same order. Thus, the ratio of elevated blood pressure prevalence in black and white women was constant at 2.2. Table H-5. Percent decline in prevalence of elevated blood pressure between 1971-75 and 1976-80, black women in age-groups between 25 and 75 and age-adjusted for black and white women, U.S. rates per 100 population 1971-75 1976-80 % change 25-34 12.4 4.3 - 65. 3* 35-44 23.9 17.6 - 26.3 45-54 39.4 37.6 - 4.6 55-64 46.0 36.4 - 20.9 65-74 46.7 43.4 - 7.1 25-74 Black 30.5 24.4 - 20.0 25-74 White// 14.2 11.1 - 21.8 * calculated from rates in Reference 2, Table II; elevated blood pressure defined as SBP >=160 and/or DBP >=95 mmHg. // from Reference 4, Table 5 Data on hypertension prevalence (Table H-6) show higher prevalence in black compared to white women, with essentially no changes in either group over the three survey periods. The 20% decline in the prevalence of elevated blood pressure is not accompanied by a change in hypertension prevalence (at least not in the overall age-adjusted data). Thus, as discussed earlier, no conclusion is possible regarding changes in the survival of women with hypertension over time. Awareness of hypertension increased in women of both races and to approximately the same degree. More black than white women with hypertension had been previously informed of their condition in NHES and NHANES II. In 1976-80 less than 15% of hypertensive black women and 25% of white women were previously unaware of their condition. Approximately 60% of both black and white women hypertensives were on medication in the 1976-80 survey. There was no racial difference in the proprortion of the women on medication whose blood pressures were adequately controlled in either NHES or NHANES II. However, the proportion of control approximately doubled over this time period. The 273 proportion of hypertensive women whose medications were adequately control- ling their blood pressure levels was approximately twice that for men. Table H-6. Rates of hypertension prevalence, awareness, treatment, and control, black and white women, ages 25-74, U.S.,1960-62, 1974-75, and 1976-80* percent of hypertensives in population// percent of hypertensives not previously diagnosed percent of hypertensives on medication percent of hypertens ives on medication whose blood pressure was controlled//// 1960-62 1974-75 1976-80 BW 39.8 35.5 39.8 WW 20.4 19.6 20.0 BW 35.1 28.9 14.5 WW 43.9 29.7 25.2 BW 48.1 36.4 60.6 WW 38.2 48.5 58.6 BW 20.2 22.3 38.3 WM 21.9 28.1 40.3 * Reference 1, Table 7; data are age adjusted to the U.S. population at the midpoint of NHANES II (1976-80). // defined as either SBP >=160, DBP >=95, or taking antihypertensive medication //// hypertensives on medication and with blood pressures below the levels considered as elevated range below the cut points for elevated blood pressure 3.1.2.4 Demographic Patterns : Women As described for men, blood pressures of black women in 1971-75 were higher than those of white women in all four regions, in all types of residential areas and at all income and education levels (3, Table 17-20; age-adjusted rates). Regional patterns of blood pressure are different in black men and women. Whereas blood pressures among black men in the Northeast were highest, the highest levels among black women were in the South and West. Rural black women had higher blood pressures than those in urban areas. In tabulation by educational levels, black women with less than 5 years of education were noticeably different from the rest of the blood pressure distribution, but blood pressure levels were higher in this group of women--not lower, as observed in men (note that these are trends from tables and not necessarily significant). 274 3.2 Left Ventricular Hypertrophy (LVH) Left ventricular hypertrophy (LVH), a putative independent risk factor for IHD, is a lethal marker (7,8). LVH in the general adult population is primarily due to prolonged blood pressure elevations (7). It was an antecedent finding in 45% of all cardiovascular deaths in the Framingham Study (9). The five-year mortality rate in Framingham men who developed ECG-LVH was 35% compared to approximately 10 to 15% or less expected otherwise (9). For women, 20% of those who developed ECG-LVH were dead within five years. For those aged 65 years and older these mortality rates were 50% and 35%, respectively. Kannel has suggested that the excess mortality associated with ECG-LVH may not be attributable to associated hypertension since ECG-LVH carried three times the risk of hypertension without the ECG finding (9). On the other hand, the degree of hypertension may not have been comparable in those with and without ECG-LVH since the clinic blood pressures used to define hypertension in the Framingham study may not have adequately reflected degree of hypertension (e.g. by not adequately reflecting variations in 24-hour blood pressure or duration of hypertension). The mortality rate associated with possible LVH was one-half that of definite ECG-LVH. Chest x-ray LVH was associated with one-third the cardiovascular mortality rate of ECG-LVH (9). Risk of cardiovascular morbidity is also markedly increased in those with ECG-LVH. In age 40 year-old Framingham men with systolic blood pressures of 195 and positive status on glucose intolerance and cigarette smoking, ECG-LVH raises the eight-year probability of cardiovascular disease from 459 per 1000 to 708 per 1000 (10). ECG-LVH confers a 2 to 9-fold increased risk of stroke, cardiac failure, coronary disease and peripheral arterial disease (9). In the HDFP, age-adjusted five-year total mortality in the stratum with mild hypertension at baseline (stratum I) was approximately double in white and black Referred Care males who were LVH positive compared to LVH negative by ECG (11). The prevalence of ECG-LVH is greater in blacks than whites for both sexes (8). These prevalence differences are clearly demonstrated in baseline data for 10,940 HDFP participants, as shown below (Table P-l). These baseline prevalence patterns were evident after adjustment for age and blood pressure level (12). The prevalence of ECG-LVH was progressively higher in HDFP Strata II and III (baseline diastolic blood pressure 105-114 and 115+ mmHg) vs. stratum I (baseline DBP 90-104 mm Hg) (13). Percentages in stratum III were more than 3 times those in stratum I, suggesting a dose response relationship. Race-sex specific baseline ECG-LVH prevalence data by blood pressure strata have also been published (13,14). 275 White men 1892 2 .4 Black men 1064 8 .5 White women 1185 1. .8 Black women 1344 7. .9 Table P-l. Baseline prevalence of ECG-LVH by race-sex and randomization group for HDFP participants (ages (30-69 years, n=10,940) * Percent of Subgroup n Stepped Care n Referred Care 1861 3.1 1084 9.0 1156 1.7 1354 7.8 * from Reference 13, Table 2.3--"...based on combined R-wave and ST-T segment changes: tall R-wave (Minnesota Code 3.1) and major ST segment depression (Minnesota code 4.1-4.3) or major T-wave inversion (Minnesota code 5.1-5.3)" The reasons for the black excess of ECG-LVH are unclear. Since the differences are observed within each stratum of baseline blood pressure among the HDFP participants, these racial differences for given levels of blood pressure may be indicative of the inadequacy of our usual measures of blood pressure (e.g., one to three blood pressures in a clinic setting)(7,8). Earlier onset and longer duration of blood pressure in blacks are likely contributors (8). Duration of blood pressure elevations is difficult to ascertain. Assessment of a possible association with ECG-LVH is further complicated by duration and type of antihypertensive treatment (7,15-17). Regression of ECG-LVH with antihypertensive treatment was observed in the HDFP (18). An additional or alternative explanation for black-white differences in LVH prevalence is that the left ventricular response to a given blood pressure elevation is greater in blacks than in whites (8,19,20). Echocardiography (echo) detects LVH earlier than either ECG or chest x-ray (15,21). Echo has the greatest sensitivity and specificity of the three diagnostic tools (7,15,21). Thus, the prevalence of echo-LVH is higher than LVH by ECG in a given population (7). The proportion of echo-LVH cases which progress to ECG-LVH and the proportion of the latter which survive to be ascertained have not been documented. The prognostic significance of LVH by echo rather than ECG criteria has been uncertain. However, recent data from Framingham have documented a significant association of echo-LVH with all-cause mortality, independent of standard risk factors (22). In that study, approximately twice as many of the men and women about age 70 at time of echo measurement) who died (compared to those alive) in a two-year period after examination had echo LVH. However, only 7% of the women who died and none of the men had LVH by ECG. Data on the prevalence of echo-LVH are limited. Reference data for whites have been reported from Framingham (7). There are no population-based echo-LVH reference data for blacks. Small-scale studies in American and African blacks have documented the presence of 276 significantly higher LV wall thickness, interventricular septal thickness, LV mass or mass index and lower LV fractional shortening in hypertensives vs. normotensive controls (23,25), and proportionate to the level of blood pressure. These patterns are reportedly similar uo those observed in whites and in Japanese (8). Dunn et al. (20) observed higher LV mass index (LV mass per square meter of body surface area) in 30 black hypertensives compared to age-sex-blood pressure matched controls. Posterior wall thickness was correlated to blood pressure and total peripheral resistance in the black but not white subjects in that study. Hammond et al. (19) similarly found greater echo target organ changes in blacks matched for blood pressure levels with whites. Savage et al. (23) found that echo findings were similar between black and white hypertensive subjects carefully matched for several possible confounders: age, sex, treatment status, level and known duration of hypertension, and renin subgroup. If the origin of black-white echo-LVH differences is related to blood pressure variables, then the matching protocol and inadequate power (n=35 blacks and 35 whites) in this study may have obscured these differences. Hammond et al. (19) have reported results of echo-LVH comparisons in an occupationally-based sample of black and white normotensive (n=75) and hypertensive (n=132) men and women in New York City. The hypertensives in this study included newly identified hypertensives defined as (DBP>=95 mmHg, SBP >=160 mmHg or both, sustained on three occasions over a three week period and not previously taking antihypertensive medications. Hypertensives also included persons receiving antihypertensive medications before joining the study, regardless of blood pressure level at screening. No echo variables were different between black and white normotensives, suggesting that the echo baseline in blacks and whites was the same. Among the hypertensives, LV mass index was similar in blacks and whites but relative wall thickness, cardiac output and total peripheral resistance were higher in blacks. Blood pressures of the black and white hypertensives were similar. Differences observed may be explained in various ways including either different durations (and/or actually different severity of hypertension) or differential effects of blood pressure elevations. In the New York City study (19), progression of LVH in the newly diagnosed, untreated hypertensives was not differentiated from possible regression of left ventricular measurements towards normal in the hypertensive subjects on medications at the time of screening. Differences in echo findings in the two groups were not observed at the time of study but may have been present at one time. Three studies (16,17,26) documenting regression of echo-LVH in hypertensives with antihypertensive treatment have included blacks. One of these studies (16) reported regression very early in therapy (at one month) which was sustained during treatment over an 18 month follow-up period. It is noteworthy, however, that methyldopa but not hydrochlorothiazide, was associated with regression of echo-LVH. Diuretics (e.g. hydrochlorothiazide) are often recommended as the preferred 277 treatment for blacks (27). Much more information is needed regarding the significance of regression of echo-LVH (and other changes) in response to various antihypertensives. If differing hypertensive regimens are recommended for black and white hypertensives on a racial basis (27) the potential significance of such information is obvious. More information is also needed about how representative clinic blood pressures are. This is suggested by the 10 year prospective study of Perloff et al. (28), showing prognostic significance of ambulatory blood pressure independent of clinic blood pressures. In summary, if blood pressure elevation of long duration is related to LVH (which is true (7)) and if LVH then becomes established as an independent IHD risk factor (which appears to be true (22)), then the attributable risks of LVH in blacks may be greater than in whites. As noted above, the onset of blood pressure elevations in blacks occurs at an earlier age than in whites. The prevalence of such elevations is also greater in blacks. Finally, early (echo) LVH and late (ECG) LVH may be greater in blacks than in whites with the same apparent level of blood pressure elevation. ACKNOWLEDGEMENTS We gratefully acknowledge the editorial assistance of Sandra J. Anderson and Elisabeth Pitt. 278 3.3 References 1. Rowland M, Roberts J. Blood pressure levels and hypertension in persons ages 6-74 years. United States 1976-80. NCHS. Vital and Health Statistics. AdvanceData No. 84, October, 1982. 2. National Center for Health Statistics. Blood pressure levels of persons 6-74 years, 1971-1974, Roberts J., Maurer K. Vital and Health Statistics. Series 11. No. 203. DHEW Pub. (HRA) 78-1648. September, 1977. 3. National Center for Health Statistics. Hypertension in adults 25-74 years of age, United States 1971-75, Roberts J, Rowland, M. Vital and Health Statistics. Series 11. No. 221. PHHS Pub. No. (PHS) 81-1671. April, 1981. 4. Rowland ML, Fulwood R. Coronary heart disease risk factor trends in blacks between the first and second National Health and Nutrition Examination Surveys, United States 1971-1980. Am Heart J 1984; 108: 771-779. 5. Oni A. Intraracial factors in blood pressure variations among the black population. J Natl Med Assoc 1984; 76:594-603. 6. Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. II. Mortality by race, sex, and age. JAMA 1979;242:2572-2577. 7. Savage DD, Abbott RD, Padgett S, Anderson SJ, Garrison RJ. Epidemiologic features of left ventricular hypertrophy in normotensive and hypertensive subjects. In Cardiac Left Ventricular Hypertrophy, HEDJ ter Keurs, JJ Schipperheyn, eds. Martinus Nijhoff. 1983, pp. 3-15. 8. Savage DD. Echocardiographic assessment of cardiac anatomy and function in black and white hypertensive subjects. In Textbook of Ethnic Medicine, R Williams ed (in press). 9. Kannel WB. Left ventricular hypertrophy in hypertension: prognostic and pathogenetic implications. The Framingham Study. In The Heart in Hypertension, Strauer BE ed. 1981, p.123. 10. Kannel WB. An overview of the risk factors for cardiovascular disease. In Prevention of Coronary Heart Disease. Practical Management of the Risk Factors, Kaplan NM, Stamler J, eds. Philadelphia: W.B. Saunders Company. 1983, Chapter 1, pp. 1-19. 11. Tyroler HA. Overview of risk factors for coronary heart disease in black populations. Am Heart J 1984;108:658-660. 279 12. The HDFP Cooperative Group, presented by Prineas R. Sex and race differences in end organ damage among 10,940 hypertensives. Am J Cardiol 1978:41:402(abstract). 13. Taylor JO, Borhani NO, Entwisle G, Farber M, Hawkins CM, on behalf of the HDFP Cooperative Group. Summary of the baseline characteristics of the hypertensive participants. Hypertension 1983;5(part II, no.6):IV-44 - IV-50. 14. Hypertension Detection and Follow-up Program Cooperative Group. Five-year findings of the Hypertension Detection and Follow-up Program. Mortality by race-sex and blood pressure level. A further analysis. J Comm Health 1984;9: 314-327. 15. Savage DD, Devereux RB. Echocardiography and hypertension. Primary Cardiology. 1981;7:137-149. 16. Wollam GL, Hall WD, Porter VD, Douglas MB, Unger DJ, Blumstein BA, Costonis GA, Knudtson ML, Felner JM, Schlant RC. Time course of regression of left ventricular hypertrophy in treated hypertensive patients. Am J Med 1984;75:100-110. 17. Schlant RC, Felner JM, Blumstein BA, Wollan GL, Hall WD, Shulman NB, Heynsfield SB, Gilbert CA, Tuttle EB. Echo- cardiographic documentation of regression of left ventricular hypertrophy in patients treated for essential hypertension. Eur Heart J 1982;3(supp A):171-175. 18. The Hypertension Detection and Follow-up Program Cooperative Group. Regression of left ventricular hypertrophy (LVH) with anti-hypertensive therapy. Circ 1981;64(Supp IV); IV-322 (abstract). 19. Hammond IW, Alderman MH, Devereux RB, Lutas EM, Laragh JH. Contrast in cardiac anatomy and function between black and white patients with hypertension J Natl Med Assoc 1984;76:247-255. 20. Dunn FC, Chandraratna P, de Carvalho JGR, Basta LL, Frohlich ED. Pathophysiologic assessment of hypertensive heart disease with echocardiography. Am J Cardiol 1977;39:789-795. 21. Savage DD, Drayer JIM, Henry WL, Matthews EC Jr, Ware JH, Gardin JM, Cohen ER, Epstein SE, Laragh JH. Echocardiographic assessment of cardiac anatomy and function in hypertensive subjects. Circulation, 1979;59:623-32. 280 22. Savage DD, Garrison RJ, Castelli WP, Kannel WB, Anderson SJ, Feinleib M. Echocardiographic data from Framingham and newer studies of echocardiography. Echocardiographic left ventricular hypertrophy in the general population is associated with increased 2-year mortality, independent of standard coronary risk factors. AHA Council on Epidemiology (invited paper), 1985. 23. Savage DD, Henry Wl, Mitchell JR, Taylor AA, Gardin JM, Drayer JIM, Laragh JH. Echocardiographic comparison of black and white hypertensive subjects. J Natl Med Assoc 1979;71:709-712. 24. Schlant RC, Felner JM, Heynsfield SB, Gilbert CA, Tuttle EB, Blumenstein BA. Echocardiographic studies of left ventricular anatomy and function in essential hypertension. Cardiovas Med 1977;2:477-491. 25. Adesanya CO, Sanderson JE, Verheihen Ir PJY, Brinkman AW: Echocardiographic assessment and systolic time interval measurements in the evaluation of severe hypertension in Nigerian Africans. Aust NZ J Med 1981;11:364-369. 26. Devereux RB, Savage DD, Sachs I, Laragh JH. Long-term effects of hypertension on the heart. Clin Res 1980;28: 330A (abstract). 27. The 1984 report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. The Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1984;144:1045-1057. 28. Perloff D, Sokolow M, Cowan R. The prognostic value of ambulatory blood pressures. JAMA 1983;249:2792-2798. SUPPLEMENTAL REFERENCES S-l. Folkow B, Nordlander MIL, Strauer B-E, Wikstrand J eds. Pathophysiology and clinical implications of early structural changes. Hypertension 1984;6:(Suppl 3)iii-l - iii-187. S-2. Messerli RH, Schlant RC, eds, Proceedings of a Symposium: Left Ventricular Hypertrophy in Essential Hypertension. Am J Med 1983;75:1-120. S-3. Panidis IP, Kotler MN, Ren J-F, Mintz GS, Ross J, Kalman P. Development and regression of left ventricular hyertrophy. J Am Coll Cardiol 1983;3:1309-1320. S-4. Tarazi RC. Regression of left ventricular hypertrophy: partial answers for persistant questions (editorial). J Am Coll Cardiol 1983;3:1349-1351. 281 4.0 CONCLUSION Our review suggests that intensified efforts to understand and reduce black-white differentials in hypertension (and, possibly, left ventricular hypertrophy) and obesity (black women) offer the greatest opportunity among the standard and putative independent cardiovascular risk factors for reduction of cardiovascular morbidity-mortality disparities between blacks and whites. These intensified efforts should of course be in the context of continued vigorous efforts to reduce all of the cardiovascular risk factors in both blacks and whites. 282 Relationship of Social Class to Coronary Disease Risk Factors in Blacks: Implications of Social Mobility for Risk Factor Change Lucile L. Adams, Ph.D. Department of Epidemiology Graduate School of Public Health University of Pittsburgh Pittsburgh, Pennsylvania Laurence O. Watkins, M.D., M.P.H. Section of Cardiology Medical College of Georgia Augusta, Georgia Lewis H. Kuller, M.D., Dr.P.H. Chairman, Department of Epidemiology Graduate School of Public Health University of Pittsburgh Pittsburgh, Pennsylvania Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland Richard Donahue, Ph.D. National Heart, Lung, and Blood Institute Bethesda, Maryland Ronald E. LaPorte, Ph.D. Department of Epidemiology Graduate School of Public Health University of Pittsburgh Pittsburgh, Pennsylvania Coronary heart disease (CHD) is the leading cause of death among United States blacks, mortality rates being similar in black and white men and greater in black women than in white women (1). National survey data from 1962 revealed similar CHD prevalence rates for blacks and whites (2). Small studies in two southern states have revealed lower CHD incidence in black Chan in white men in the period 1960-1974, and greater CHD incidence in black than in white women (3,4). Among both men and women, black rates of admission to the hospital with the diagnosis of acute myocardial infarction are lower than those of whites(5). These data on the CHD experience of blacks are conflicting, but it is clear that compared to whites, blacks have a greater prevalence of CHD risk factors including hypertension, cigarette smoking, and among middle-aged women, obesity and diabetes mellitus (6). It is somewhat paradoxical that black men in particular do not have higher CHD rates than whites. This paradox would be intensified if, as has been hypothesized, there is an inverse relationship between social class and risk of CHD. A 1968 review of this subject by Antonovsky (7), and a more recent one by Jenkins (8), have discussed the conflicting data with regard to CHD in white populations. Some studies suggest higher risk of CHD with high social status, while others indicate a lower risk. An explanation for these diametrically opposed results has been proposed by Morgenstern (9). Guided by Cassel, he suggested that the association of CHD risk with social status might have been altered during the period of rapid social change in urban populations in the U.S. in the period 1930-1950. In this formulation, among individuals and groups which came to enjoy the benefits of rapid industrial growth, those of high socioeconomic status (SES) initially asssumed high CHD risk status, but as they adapted more rapidly to the new material conditions by altering lifestyles, they became at significantly lower risk than their low SES counterparts. Such links as exist between SES and CHD risk may be mediated by causal links between SES and lifestyle, and between lifestyle or lifestyle changes and biological risk factors. In this hypothetical framework, low or high SES may be linked to CHD risk because of an association with adverse risk factor distributions, or SES may modify the effect of the major biological risk factors on CHD incidence. In an investigation of the data on CHD prevalence (1960), incidence (1960-1967), and mortality (1960-1974) in the white adult population of Evans County, Georgia, Morgenstern found evidence of a change in the predicted direction for the association between SES and CHD in men (9) That is, the prevalence of CHD was initially higher in white men of high SES, but during follow-up, the incidence of CHD was lower in the cohort of 35-54 year old men of high SES. In addition, there was evidence that the changing relation between SES and CHD risk was partly due to differential change in certain biological risk factors, especially hypertension. That is, there was greater adverse risk factor change among younger low SES men. In addition, there was no evidence that SES modified the effect of the major biological risk factors on CHD. Because of systematic discrimination, the black population of the U.S. has been largely of low SES (10). In the last 20 years, there have been rapid 285 socioeconomic changes in certain segments of the black population (11), and some black families have experienced upward mobility, either within a generation or between successive generations. In view of such changes, the implications of Morgenstern's hypothesis and observations for the associations betwen SES and CHD incidence and between SES and CHD risk factor distributions in blacks are important, but no systematic examination of this subject has been done (12-14). This review will summarize the data on 1) the extent and characteristics of socioeconomic change in blacks since 1960; 2) The relation between SES and CHD incidence and mortality in blacks; 3) the cross-sectional associations between SES and related factors and the prevalence (and incidence) of CHD risk factors; and 4) the risk factor characteristics of black population samples from the higher end of the socioeconomic spectrum. The latter will be discussed as examples of black populations which may have experienced upward mobility, either recent or remote, and which might make manifest the impact of such mobility on CHD risk factors. Socioeconomic Status Changes in Blacks In the period 1960-1981, there were increases in the percentage of black adults above age 25 years who completed high school or four years of college (15). By 1981, the median number of years of schooling for black males and females exceeded 12, and the black-white difference was less than one year. During these two decades, the median duration of formal education had increased by 4.4 years for black males, and 3.5 years for black females, but by only 1.9 and 1.3 years for white males and females, respectively. These changes in educational attainment were associated with changes in the occupational profile of blacks. There has been an increase in the percentage of blacks who are white-collar workers, and a decline in the percentage of farm workers. Between 1972 and 1981, the percentage of blacks employed in professional, technical, managerial or administrative jobs increased from 11.2% to 15.7% (16). However, despite the broader range of occupations now held by blacks, especially black women, the majority of blacks, especially men, remain blue-collar or service workers. Blacks remain more likely to be unemployed (17). The black unemployment rate in 1982 was 17.3%, compared to 8.6% for whites. This 2:1 ratio has remained virtually unchanged since 1960. Blacks who are unemployed are likely to remain so for longer periods of time. Family incomes remain substantially lower among blacks than whites (15). This is so, particularly for families headed by women. It is noteworthy that the financial rewards of education are dissimilar for blacks and whites. Only 18.1% of white families in which the head of the household has at least four years of college earn less than $20,000 per year. The comparable number for black families is 34.7%. Similarly, 35.5% of white families whose heads of household have four years or more of college earn more than $40,000 per year, while only 18.1% of similar black families have such incomes. The disparity is more striking, if it is noted that of all black families in which the householder has at least four years of college, 23.7% earn less than $15,000 per year, a proportion similar to that for white families in which the 286 householder has four years of high school (26.1%). Income disparities between blacks and whites appear to be attributable to differences in pay-rates for black and white men holding the same types of jobs (18), and more to restriction in the job opportunities open to blacks once they have completed their education, than to differences in educational attainment. These considerations serve to emphasize that a relatively small proportion of a black population can be considered "upwardly mobile". Social mobility refers to the movement of an individual from one social position to another different from the one to which he/she was accustomed. For the individual, there are different types of social mobility and the classification depends on the reference point (14). Intergenerational mobility takes into account the social position of offspring with respect to that of the parents. This can be compared to intragenerational or career mobility which takes account of changes experienced by the same individual within his/her lifetime. As a consequence of recent social change, some blacks have experienced intergenerational mobility, and others intragenerational mobility, or even both. In the course of this report we shall examine data on groups of blacks in whom upward mobility is presumed to have occurred in current or recent generations, by virtue of their superior educational attainment and social position compared to the general black population. One of the samples of blacks which shall be considered "upwardly mobile" for the purpose of this review is constituted of college students. This appears reasonable, since data from a national study of male college students which examined occupational expectations according to race and socioeconomic background reveal that among blacks, the level of education of the father is an important determinant of career aspirations, college performance and occupational expectations (19). Professional orientations of black college students may reflect the orientations and experiences of high-achievement parents. Socioeconomic Status, Social Mobility and CHD Incidence. As a prelude to consideration of the limited data on blacks, observations on white U.S. populations will be reviewed. The observations in Evans County, Georgia have been noted briefly (3). In the 1960 prevalence study, high-SES white men had a CHD prevalence rate 2 or 3 times that of lower SES white men. The 1960-1967 incidence study showed that high SES was associated with higher CHD incidence for white men over age 55 years, but was inversely related to CHD incidence in white men 35-54 years of age. In this study, SES was assessed by use of the McGuire-White index which takes account of occupation, education, and source of income of the head of each household. Studies in white populations which have used education as the marker of SES have suggested that education is associated with lower CHD incidence. For example, Hinkle et al (20) observed that the incidence of new CHD events and deaths was approximately 30% lower among college-educated men than among non- college-educated men. On the other hand, the association between CHD 287 incidence and occupation may be different from that observed for education. In a 1957 study of CHD incidence in North Dakota, Syrae et al (21) found that the incidence of CHD was higher among white-collar workers than among blue- collar or agricultural workers. An analysis which took account of the occupation (agricultural or non-agricultural) of each subject's father, indicated that the excess risk was confined to those white-collar workers from an agricultural (i.e., rural) background (22). Studies of career or intragenerational occupational mobility among whites have also detected higher rates of coronary disease among the mobile than among the nonraobile. Syrae et al (22) made such observations on coronary incidence in North Dakota, and similar cross-sectional observations on CHD prevalence in an urban area in California (23). In the North Dakota Study (22), Syme et al noted significantly higher CHD incidence rates in the group of men who had made four or more major job changes since age 18 years (compared to those with one or no job changes), and the group of men who had made two or more major cross-country moves since age 18 years (compared to those who had made one or no such moves). The CHD incidence rates were significantly higher in the "highly mobile" group (the entire group of white- collar workers with paternal agricultural background, and all men with high occupational or geographic mobility) than in the stable group (farmers with paternal agricultural background) or the remaining "moderately mobile" group. The analyses suggested that the contribution of sociocultural mobility t~ CHD incidence was independent of cigarette smoking, blood presssure, and body- weight. The necessity for such analyses is illustrated by a report from the Tecumseh Community Health Study (24). In men and women, ages 35-69 years, residential mobility and urban residence, particularly in childhood, were related not only to the prevalence of CHD, but also to hypertension and cigarette smoking. Kaplan et al (25), in the initial Evans County report, noted that the prevalence of CHD in white men was higher in upwardly mobile men from the lower social classes than among those who were not mobile. The observation discordant with those of Syme et al (22) was that upward mobility among the upper social classes was associated with CHD prevalence rates lower than those among the nonmobile. Other data on the relationship of sociocultural mobility and CHD incidence have been obtained from a follow-up study of 13,728 male Harvard University students. Gillum and Paffenbarger (26) observed that individuals who exhibited intergenerational mobility (that is, their social class was eventually higher than that of their fathers) had a 1.5 times higher risk of fatal CHD and MI than those who did not experience such mobility. A single study has examined the relationship between SES and CHD incidence in blacks in and around Charleston, South Carolina, and one recent analysis may cast some light on the relationship between SES and CHD mortality in urban blacks in Los Angeles, California. In the Charleston Heart Study (4), a special cohort of high-SES black men was assembled by peer nomination. Over 14 years of follow-up (1960-1974), the incidence of CHD in this group of men was less than half of that in the predominantly low-SES men in the random population sample. In this study, SES was assessed on the basis of occupation 288 and income. The risk characteristics of the high-SES group have not been specifically reported, though one publication from the Charleston Heart Study has revealed that the incidence of hypertension was 3-4 times as high in low- SES as in high-SES men (27). This particular observation is similar to the Evans County finding in white men. Frerichs and coworkers examined the association betwen income and cardiovascular disease mortality rates in 1979-1981 in Los Angeles County (28). In this study, they determined the median family income in each census tract, assigned this value to the resident population and to individual residents who died, and aggregated census tracts into five income groups. The age-adjusted mortality rate from Diseases of the Heart (International Classification of Disease Codes 390-398, 402, 404-429) was inversely related to income. The contribution of ischemic heart disease (ICD 410-414) to the age and sex-adjusted mortality rate for diseases of the heart was 62.6% for the entire county, 53.7% for blacks, and 63.4% for whites. The age-adjusted ischemic heart disease mortality rate of black men was 18% lower than that recorded for white men, 223.9 compared to 274/100,000, while the rates for women were virtually identical, 158.1 and 158.5/100,000. The inverse association between mortality from diseases of the heart and income was common to both blacks and non-blacks. Compared to black men in the higher income group (_> $28,500), black men in the lowest income group (_< $13,600) had a 53% excess risk of death from diseases of the heart. The risk differential for women was 33%. Such analyses were not performed for deaths from ischemic heart disease, but since they constitute the majority of the deaths in question, a similar relationship to income is likely. No study of intragenerational or intergenerational mobility (social, cultural, occupational or geographic) and CHD incidence in blacks has been reported (14). SES and CHD Risk Factor Associations in Blacks Hypertension Among both blacks and whites, there is a distinct inverse association between SES and blood pressure. For example, Syrae et al (29) assessing SES in 1968-1969 on the basis of education and occupation, detected a higher prevalence of hypertension in blacks of low SES. Similarly, as noted earlier, in Charleston County, South Carolina, the incidence of hypertension in 1960- 1974 in black men was inversely related to social class (27). In a black Baltimore population (30), there was also an inverse association between income and the incidence of hypertension in 1973-1977. Over this 3-4 year period, the incidence of hypertension in sons of professionals was approximately one quarter of that observed in sons of laborers. The relationship between the prevalence of hypertension and the duration of formal education was also striking in the screening conducted in 1973-1974 for the Hypertension Detection and Follow-Up Program (31). I~ blacks with less than 10 years of education, the prevalence of hypertension was 43.9%, while in those who had completed college, the prevalence was 27.7%. For whites, the corresponding rates were 23.1% and 13.5%. Mean systolic and 289 diastolic blood pressures in adults ages 18 to 74 years examined by NHANES II (1976-1980) were inversely related to the duration of formal education of examinees in all race and sex groups, the association being more marked for women than for men (32). The prevalence of definite hypertension (systolic blood pressure 160 mmHg or greater or diastolic blood pressure 95 mmHg or greater, or taking antihypertensive medication) showed a slight, not statistically significant, decline between 1960 and 1980 in black adults (32), coincident with their increased educational attainment. In Bogalusa, Louisiana, among black children ages 5-14 years examined in 1973-1974, those whose parents had a postgraduate education had lower systolic blood pressures than did similar white children. In addition, there were no racial differences in systolic and diastolic blood pressure between children of white-collar workers (33). These findings are in contrast to the significantly higher blood pressure levels observed among black children after statistical control for education and occupation, when the entire group of children in the community sample was used as the basis for racial comparisons. Elevated Serum Cholesterol In NHANES I (1971-1975), serum cholesterol levels were significantly higher in the lowest socioeconomic class (on the basis of education and income) among whites, but not among blacks (34). There was no clear relationship of cholesterol level to education among black men ages 18-74 years, but among women, serum cholesterol level was generally lower among those of higher income. Across income levels, black women had generally lower mean serum cholesterol levels than did white women. Examination of the impact of parental social status on risk factor variables in 5-14 year old children in Bogalusa, Louisiana in 1973-1974 revealed that black children had significantly higher total serum cholesterol levels and alpha lipoprotein (HDL-cholesterol) after account was taken of parental education and occupation (33). However, among children whose parents had a postgraduate education, blacks tended to have lower total serum cholesterol and HDL-cholesterol as well as higher pre-beta lipoprotein cholesterol (VLDL) than white children. Among children of white-collar workers, there were no statistically significant differences between the groups for total cholesterol and HDL-cholesterol. High Density Lipoprotein Cholesterol There is limited information concerning the relationship of SES to HDL- cholesterol in blacks. Most of the studies which have evaluated HDL- cholesterol levels in blacks have examined low-SES poulations. Among U.S. adults, black males have higher HDL-cholesterol levels than white males, while the levels for black females are comparable to those of white females (35). In the Lipid Research Clinics Program (1971-1976), a direct association was detected between SES (measured by education) and HDL-cholesterol in white subjects. The black sample in this study was too small to permit such an evaluation. 290 The Framingham Minority Study evaluated a sample of well-educated blacks in Framingham, Massachusetts in 1980 (36). It is remarkable that the age- adjusted HDL-cholesterol levels observed in this population are lower than those of whites, and lower than those reported for other U.S. black populations (37). The educational attainment of this black population was such that 56% had completed four years of college. Wilson et al (36) speculated that blacks of middle-and upper-SES (a correlate of relatively high educational attainment) might have lipoprotein profiles different from those reported in lower-SES groups. Data on men recruited into the Multiple Risk Factor Intervention Trial (1973-1976) are in accord with this hypothesis (38). In this study, mean HDL-cholesterol levels of black men were significantly higher than those of white men, but in a multiple regression analysis which took account of age, cigarettes smoked per day, diastolic blood pressure, number of alcoholic drinks consumed per week, and body mass index, socioeconomic status was inversely related to HDL-cholesterol concentrations in black men, but directly related to HDL-cholesterol concentrations in white men. The black-white differences in HDL-cholesterol levels were largest among those of lower SES and were smaller at higher SES levels. Cigarette Smoking There are relatively few data available on the association between SES and cigarette smoking in blacks. Data from the 1970 Health Interview Survey indicate that within industries, substantially higher percentages of individuals in the lower occupational ranks (where blacks are overrepresented) smoke (39). In the Princeton School District Study (40), there was an inverse relationship between SES (assessed by education and occupation of the head of the household) and the prevalence of cigarette smoking among both black children and adults. Comparison of data from two successive National Health and Nutrition Examination Surveys (NHANES I and II, 1971-1975 and 1976-1980) (41), indicates that the prevalence of smoking declined in black men and women between the surveys, more among women than among men. The proportion of black adults who were heavy smokers (25 or more cigarettes per day) did not change significantly, indicating that the decrease in the prevalence of smoking was due to a reduction in the ranks of light or moderate smokers. Data derived from interviews performed between 1970 and 1980 of hospitalized black adult patients yield some information on the prevalence of cigarette smoking in two educational strata, high school or less, and college or more (42). They indicate that the prevalence of cigarette smoking was lower among college- educated black men than among black men with less education, and there was a more marked increase in the proportion of former cigarette smokers among the college-educated men in the second five-year period. Among black women, the prevalence of cigarette smoking was similar in the two educational strata in both periods, 1970-1975 and 1976-1980, though the prevalence had declined in both groups by about 15% during the period between 1976-1980. These data suggest that smoking cessation is more likely among those with more education. This is consistent with data from the 1970 Health Interview Survey which revealed that smokers in the higher occupational ranks are more likely to 291 cease smoking (39). Also, other survey data (43), indicate that black smokers who earn more than $15,000 per year, and those who report regular medical care are more interested in quitting smoking and report enjoying smoking less than do those with lower income and less access to health care. Obesity The relation of SES and obesity in blacks is largely unexplored. Physical Activity There is a paucity of literature that evaluates physical activity levels in blacks. The limited information from one available study suggests no difference in physical activity levels between blacks and whites. In the 1977 National Health Interview Survey (44), respondents were asked to rate their own level of physical activity relative to other persons of their age: more active, about as active, or less active. Self-perceived levels of physical activity were similar in blacks and whites. A race-specific analysis has not been reported, but self-perceived level of physical activity exhibited a positive relationship to income. The likelihood of a direct association between nonwork-related physical activity and SES is also suggested by the results of a statewide Massachusetts Survey (probably predominantly white) in which reported participation in sports and exercise increased linearly with both education and income (45). CHD Risk Factor Characteristics of Upwardly Mobile/High SES Blacks Two population samples examined within the last decade provide relatively complete data on CHD risk factors in blacks at the upper end of the SES spectrum. The Framingham Minority Study (36) examined a random sample of 45 black men and 55 black women, residing in Framingham, Massachusetts in 1980. The educational characteristics and HDL-cholesterol data of this sample have already been discussed. The age range of the participants was 20-69 years and the mean age was 42 years for both sexes. A younger sample of blacks who may be upwardly mobile is represented by University of Pittsburgh freshmen recruited during the 1982-1983 and 1983-1984 academic years (46). This sample included 285 black students (134 men and 151 women; mean age of 18.0 +0.6 years for each sex). Their parents were well-educated compared to the general black population. They represented a relatively homogeneous population with a median family income of $21,000 which is similar to $20,500 for the general white population, and far in excess of the $11,600 median income for the general black population. Mean blood pressures of Framingham Minority subjects were considerably lower than those of the general U.S. population (47). However, the small sample size precludes definitive statements in this regard. The Pittsburgh freshmen had lower systolic blood pressures than those observed for men and women of comparable age in the First National Health and Nutrition Examination Survey (NHANES I) (47). In the Framingham Minority Study, 13% of the men had hypertension and 18% of the women. In the Pittsburgh sample, 6.0% of the men 292 and 4.0% of the women had either SBP _^ 140 mm Hg or DBP > 90 mm Hg. Comparable figures are 15.0% for men and 3.2% for women ages 18-24 in the general U.S. population. The prevalence of cigarette smoking in the Framingham Minority study was 42.2% for men and 30.9% for women. Comparble data reported in the NHANES II study were 46.5% and 29.9% for black men and women, respectively (41). This is similar to the prevalence of smoking in the U.S. white population but lower than that of U.S. blacks described in the NHANES I (48). Relatively few of the Pittsburgh Freshmen were smokers, 7.5% of men and 10.7% of women. Framingham Minority Study men had a mean body mass index similar to that of the U.S. general population. However, the 90th and 95th percentiles of body mass index for black Framingham women were somewhat lower than those of U.S. black women (49) suggesting a lower prevalence of obesity in this population and a greater similarity to the white female population. The HDL-cholesterol levels of the Pittsburgh black men were significantly greater than those of a white Pittsburgh freshmen comparison group (50). This reflected primarily differences in the HDL2 fraction. However, for women there was no difference in HDL levels between black and white Pittsburgh freshmen. In addition, self-reported physical activity levels paralleled the HDL-chol and HDL2~chol results. That is, black females and white females had similar self-reported physical activity levels while black males had a significantly higher self-reported activity level than white males (51). Socioeconomic Status and Cardiovascular Health Knowledge The data presented so far on associations of SES with behaviors which affect cardiovascular risk suggest that cigarette smoking is more likely among blacks of low-SES, and smoking cessation and high levels of regular leisure- time physical activity are more likely among blacks of high-SES. The few available studies of knowledge of cardiovascular risk factors among blacks suggest relative lack of knowledge among low-SES blacks regarding the association between certain dietary practices, including intake of salt and saturated fat, on cardiovascular risk profiles (52,53). A 1982 telephone survey of a national probability sample of 1000 subjects performed by the Food and Drug Administration and the National Heart, Lung and Blood Institute revealed that respondents who were poorly educated, of low income or resident in the South had less than average awareness of diet-health relationships. In particular, educational level was positively correlated with concern about consumption of fats and cholesterol. In addition, food purchasing, food preparation, and food consumption habits among low-SES blacks are intimately connected with a culture of poverty (54). The available data suggest the need for focused and behaviorally-oriented cardiovascular health education programs in socioeconomically deprived black populations. 293 Conclusions There are limited data on the relationship between social class and cardiovascular disease incidence and mortality in blacks. However, these data suggest that the risk of coronary disease is higher among low-SES blacks than among high-SES blacks. The data on the relationship between SES and the prevalence of CHD risk factors in blacks also suggest an association of low- SES with higher prevalence of hypertension, obesity, and cigarette smoking. Since the excess prevalence of diabete.s in middle-aged black women appears to be related to obesity (55), this may reflect some influence of SES. High-SES appears to be associated with a lower prevalence of hypertension, a greater likelihood of smoking cessation, and greater leisure-time physical activity. Only in the case of HDL-cholesterol does there appear to be an association between high-SES and a potentially adverse level of the risk factor. No study to date has examined the impact of "upward mobility" (operationalized as described earlier) on CHD risk factors in blacks. In the case of HDL-cholesterol, this appears to be an important area for future investigation. If, as has been suggested, HDL-cholesterol might account for the lower CHD incidence in black men observed in some studies, it would be important to take explicit account of social class and social mobility differences among blacks in future cohort or surveillance studies. In addition, if socioeconomic improvement is associated with a lower incidence or prevalence of risk factors such as hypertension, obesity, diabetes mellitus, and cigarette smoking, it is conceivable that factors which impede such improvement among blacks might contribute to persistently higher levels of these risk factors than in whites. However, it is probably inappropriate to focus on economic factors to the exclusion of cultural factors. The recent (1979-1982) evaluation by Stern et al (56) of CHD risk factors among Mexican Americans in three San Antonio neighborhoods, (labelled "traditional", "transitional" and "suburban") revealed an inverse relationship between socioeconomic status and the "obesity-related" CHD risk factors, diabetes mellitus, hypertriglyceridemia and low HDL-cholesterol levels in women. Among men, such a relationship existed only for diabetes, and was less marked. On the other hand, total cholesterol levels and low density lipoprotein cholesterol levels were higher in more affluent men, but not in women. The observation that "obesity-related" risk factors were higher in Mexican Americans than in Anglos of similar SES in the same neighborhoods suggested that cultural factors might be more important than socioeconomic ones. Dietary, physical activity, cigarette smoking, and alcohol consumption habits are among the possible culturally-based determinants of CHD risk. The issue of socioeconomic and sociocultural change and its effect on CHD risk factors and CHD incidence is an important one for blacks. Research should be initiated in this area. Simultaneously, interventions should be directed to fostering the possible beneficial effects of socioeconomic and sociocultural change and blunting possible adverse effects. 294 REFERENCES 1. Gillum RF. Coronary Heart Disease in Black Populations I. Mortality and Morbidity. Am Heart J, 104:839-851, 1982. 2. Gordon T, Ganst CC. Coronary Heart Disease in Adults, United States 1960-1962, National Center for Health Statistics, Series 11, No. 10, Washington D.C., U.S. Government Printing Office, 1965. 3. Cassel J, Heyden S, Bartel AG, Kaplan BH, Tyroler HA, Cornoni JC, Hames CG. Incidence of Coronary Heart Disease by Ethnic Group, Social Class, and Sex. Archives of Internal Medicine, 128:901-906, 1971. 4. Keil JE, Loadholt B, Weinrich MC, Sandifer SH, Boyle E, Jr. Incidence of Coronary Heart Disease in Blacks in Charleston, South Carolina. American Heart Journal, 108:Part IV:779-786, 1984. 5. Watkins LO. Epidemiology of Coronary Heart Disease in Black Populations: Methodologic Proposals. American Heart Journal, 108:Part 11:635-640, 1984. 6. Gillum RF, Grant CT. Coronary Heart Disease in Black Populations II. Risk Factors. American Heart Journal, 104:852-864, 1982. 7. Antonovsky A. Social Class and the Major Cardiovascular Diseases, Journal of Chronic Diseases, 21:65-106, 1968. 8. Jenkins CD. Psychosocial Risk Factors for Coronary Heart Disease. Acta Medica Scandinavica, Supplement 660:123-136, 1982. 9. Morgenstern H. The Changing Association Between Social Status and Coronary Heart Disease in a Rural Population. Social Science and Medicine, 14A:191-201, 1980. 10. The Social and Economic Status of the Black Population in the United States: An Historical View, 1790-1978, Current Population Reports, Special Studies, Series P-23, No. 30, United States Department of Commerce, Bureau of the Census. 11. Featherman D, Hauser R. Changes in the Socioeconomic Stratification of the Races, 1962-1973. American Journal of Sociology, 82:621-651, 1976. 12. Kasl SV. Social and Psychologic Factors and the Etiology and Coronary Heart Disease in Black Populations: An Exploration of Research Needs. American Heart Journal, 108:Part 11:660-669, 1984. 13. James SA. Socioeconomic Influences on Coronary Heart Disease in Black Populations. American Heart Journal, 108:Part 11:669-672, 1984. 14. James SA. Coronary Heart Disease in Black Americans: Suggestions for Research on Psychosocial Factors. American Heart Journal, 108:Part 11:833-838, 1984. 295 15. A Dream Deferred: The Economic Status of Black Americans, A Working Paper by the Center for the Study of Social Policy, Washington D.C., July 1983. 16. America's Black Population: 1970-1982, A Statistical View, Special Publication PI0/P0P-83:1, U.S. Department of Commerce, Bureau of the Census, July, 1983. 17. Mare RD, Winship C. The Paradox of Lessening Racial Inequality and Joblessness Among Black Youth: Enrollment, Enlistment, and Employment, 1964-1981. American Sociological Review, 49:39-55, 1984. 18. Reimers CW. Sources of the Family Income Differentials Among Hispanics, Blacks and White Non Hispanics. American Journal of Sociology, 89:889- 903, 1984. 19. Hogan DP, Pazul M. The Career Strategies of Black Men. Social Forces, 59:1217-1228, 1981. 20. Hinkle LE, Whitney LH, Lehman EW. Occupation, Education and Coronary Heart Disease. Science, 161:238-246, 1968. 21. Syme SL, Hyman MM, Enterline PE. Some Social and Cultural Factors Associated with the Occurrence of Coronary Heart Disease. Journal of Chronic Disease, 17:277-289, 1964. 22. Syme SL, Hyman MM, Enterline PE. Cultural Mobility and The Occurrence of Coronary Heart Disease. Journal of Health and Human Behavior, 6:178-189, 1965. 23. Syme SL, Borhani NO, Buechley RW. Cultural Mobility and Coronary Heart Disease in an Urban Area. American Journal of Epidemiology, 82:334-346, 1966. 24. Metzner HL, Harburg E. Lamphiear DE. Residential Mobility and Urban- Rural Residence within Life Stages Related to Health Risk and Chronic Diseasse in Tecumseh, Michigan. Journal of Chronic Diseases, 35:359-374, 1982. 25. Kaplan BH, Cassel JC, Tyroler HA, Cornoni JC, Kleinbaum DG, Hames CG. Occupational Mobility and Coronary Heart Disease. Archives of Internal Medicine, 128:938-942, 1971. 26. Gillum RF, Paffenbarger RS Jr. Chronic Disease in Formal College Students: XVII Sociocultural Mobility as a Precursor of Coronary Heart Disease and Hypertension, American Journal of Epidemiology, 108:289-298, 1978. 296 27. Keil JE, Tyroler HA, Sandifer SH, Boyle E Jr. Hypertension: Effects of Social Class and Racial Admixture: The Results of A Cohort Study in the Black Population of Charleston, South Carolina. American Journal of Public Health, 67:634-639, 1977. 28. Frerichs RR, Chapman JM, Nourjah P, Maes EF. Cardiovascular Diseases in Los Angeles, 1979-1981, American Heart Association, Greater Los Angeles Affiliate, Inc. Los Angeles, 1984. 29. Syme LS, Oakes TW, Friedman GD, Feldman R, Siegelaub AB, Collins M. Social Class and Racial Differences in Blood Pressure. American Journal of Public Health, 64:619-620, 1974. 30. Dischinger PC, Apostolides AY, Entwisle G, Hebel JR. Hypertension Incidence in an Inner City Black Population. Journal of Chronic Diseases, 34:405-413, 1981. 31. Hypertension Detection and Follow-Up Program Cooperative Group. Race, Education and Prevalence of Hypertension. American Journal of Epidemiology, 106:351-361, 1977. 32. Blood Pressure Levels in Persons 18-74 Years of Age in 1976-1980, and Trends in Blood Pressure from 1960 to 1980 in the United States. Data from the National Health Survey, Series 11, (forthcoming) U.S. Department of Health and Human Services, 33. Hunter SM, Frerichs RR, Webber LS, Berenson GS. Social Status and Cardiovascular Disease Risk Factor Variables in Children: The Bogalusa Heart Study. Journal of Chronic Diseases, 32:441-449, 1979. 34. National Center for Health Statistics: Serum Cholesterol Levels of Persons 4-74 Years of Age by Socioeconomic Characteristics, United States, 1971-74. U.S. Vital and Health Statistics, Series 11, No. 217, U.S. DHEW Publication No. (PHS) 80-1667, Washington D.C., 1980, U.S. Government Printing Office. 35. Tyroler HA, Glueck CJ, Christensen B, Kwiterovich PO Jr. Plasma High Density Lipoprotein Cholesterol Comparisons in Black and White Populations: The Lipid Research Clinics Program Prevalence Study. Circulation, 62 (supplement IV): IV-99 - IV-107, 1980. 36. Wilson P, Savage DD, Castelli WP, Garrison RJ, Donahue RP, Feinlieb M. HDL-Cholesterol in a Sample of Black Adults. The Framingham Mnority Study. Metabolism, 36:328-332, 1983. 37, Morrison JA, Khoury P, Mellies M, Kelly K, Howitz R, Glueck CJ. Lipid and Lipoprotein Distribution in Black Adults: The Cincinnati Lipid Research Clinics Princeton School Study. JAMA, 245:939-942, 1981. 297 38. Watkins LO, Neaton JD, Kuller LH for the MRFIT Research Group. High Density Lipoprotein Cholesterol (HDL-C) and Coronary Heart Disease Incidence in Black and White MRFIT Usual Care Men. Circulation, 1985:71- 417A. Presented at American Heart Association 25th Annual Cardiovascular Epidemiology Conference, Tucson, AZ, March 1985. 39. Sterling TD, Weinkam JJ. Smoking Patterns by Occupation, Industry, Sex, and Race. Archives of Environmental Health, 313-317, 1978. 40. Khoury PR, Morrison JA, Laskarzewski P, Kelly K, Mellies MJ, King P, Larsen R, Glueck CJ. Relationships of Education and Occupation to Coronary Heart Disease Risk Factors in School Children and Adults: The Princeton School District Study. American Journal of Epidemiology, 113:378-395, 1981. 41. Rowland ML, Fulwood R. Coronary Heart Disease Risk Factor Trends in Blacks Between the First and Second National Health and Nutrition Examination Surveys, United States, 1971-1980. American Heart Journal, 108:771-779, 1984. 42. Covey LS, Mushinski MH, Wynder EL. Smoking Habits in a Hospitalized Population. American Journal of Public Health, 73:1293-1297, 1983. 43. EVAXX, Inc. A Study of Smoking Behavior Among Black Americans, Presentation at Workshop on Smoking Prevention and Cessation for National Cancer Institute Division of Research Centers and Community Activities, March, 1983. 44. Health Practices Among Adults: United States, 1977, Advance Data from Vital and Health Statistics of the National Center for Health Statistics. U.S. Department of Health and Human Services. Office of Health Research, Statistics, and Technology, PHS 64, November 4, 1980. 45. Lambert CA, Netherton DR, Finison LJ, Hyde JN, Spaight SJ. Risk Factors and Life Style: A Statewide Health-Interview Survey. Massachusetts Department of Public Helath. New England Journal of Medicine, 306:1048- 1051, 1982. 46. Adams LL, Kuller LH. Blood Pressure Correlates and Family (Parental) History of Hypertension In A Predominantly Middle Class Black College Population. (In Press). 47. National Center for Health Statistics. Blood Pressure of Persons 6-74 years in the United States 1971-1974. U.S. Vital and Health Statistics. Series 11, No. 203, Washington, D.C, 1977, U.S. Government Printing Office. 298 48. The Prevalence of Cigarette Smoking In the Health and Nutrition Examination Survey, 1971-1975. Advance Data, September 19, p.5, 1979. U.S. Bureau of the Census, Statistical Abstract of the United States: 1984 (104th Edition). Washington, D.C, pp. 126-127, 1983. 49. National Center for Health Statistics, W.R. Harlan, M.D., A.L. Hill, Ph.D., R.P. Schraouder, et al. Dietary intake and cardiovascular Risk Factors, Part I, Blood Pressure Correlates. Vital and Health Statistics. Series 11, No. 226, DHHS Pub. No. (PHS) 83-1676. Public Health Service. Washington. U.S. Government Printing Office, February 1983. 50. Adams LL, LaPorte RE, Watkins LO, et al. The Association of Lipoprotein Cholesterol with Vitamin A. Cancer, (In Press). 51. Adams L, LaPorte R, Kuller L, Savage D, Ohayon J. Sex Pattern Differences in HDL-Cholesterol: Blacks Versus Whites. Circulation, 68:111-180, 1983; (abstract). 52. Williams PB. Assessing Awareness of Coronary Disease Risk Factors in The Black Community. Urban Health, 8(9):34-37, 1979. 53. Heimbach JT. Cardiovascular Disease and Diet: The Public View. Public Health Reports, 100:5-12, 1985. 54. Payne ZA. Diet and Cardiovascular Disease: The Morbid Connection. Journal of the Medical Association of Georgia, Heart Page, 74:49-52, 1985. 55. Bonham GS, Brock DB. The Relationship of Diabetes With Race, Sex, and Obesity. American Journal of Clinical Nutrition, 41:776-783, 1985. 56. Stern MP, Rosenthal M, Haffner SM, Hazuda HP, Franco LJ. Sex Differences in the Effects of Sociocultural Status on Diabetes and Cardiovascular Risk Factors In Mexican Americans. American Journal of Epidemiology, 120:834-851, 1984. 299 Coronary Heart Disease in Black Populations: Current Research, Treatment, and Prevention Needs Hector F. Myers, Ph.D. Department of Psychology University of California at Los Angeles ABSTRACT This paper summarizes the existing knowledge base contained in the collection of papers and reports from two national meetings on coronary heart disease (CHD) in black populations. Evidence is presented that suggests that CHD morbidity and mortality rates are roughly comparable between blacks and whites, but blacks appear to be at greater risk of sudden death than whites due to the higher prevalence of hypertension, obesity, diabetes, and greater socioeconomic disadvantage. The quality and availability of sound epidemiologic data on CHD in blacks are generally criticized, and specific recommendations for improvements are offered. An extensive review of available evidence on risk factor status and trends in black and white populations is also provided, and speculations about possible racial differences in operative risk factors and function, and in protective factors are made. Arguments are provided for the salience of socioeconomic and sociocultural factors in CHD in blacks, and an analysis of how these factors might interact to affect CHD risk status, risk management, and CHD morbidity and mortality differences is offered. Consideration of the possible role of differences in the quality, effectiveness, and accessibility of health care between blacks and whites as possible contributors to the observed CHD differences is given. Provocative evidence of differential patterns of medical care of CHD patients as a function of race and socioeconomic status is presented. Finally, the need to support CHD prevention activities is discussed, and specific programmatic recommendations for future research on coronary heart disease in black populations are offered. 303 INTRODUCTION The American Heart Journal recently published a special supplement issue that compiled the papers and plenary group reports from two national meetings on coronary heart disease in black populations: a symposium sponsored by the American Heart Association Council on Epidemiology, and a workinq conference sponsored by the National Heart, Lung, and Blood Institute (NHLBI) (American Heart Journal (AHJ), Sept., 1984, Vol. 108, No. 3, Part 2). These meetings and the subsequent publication were catalyzed by the recommendations of the Black Health Care Providers Task Force on High Blood Pressure Education and Control (BHCPTF) and by the recent publication of two major review papers on coronary heart disease in blacks (Gillum, 1982; Gillum & Grant, 1982). Both factors stimulated the interest and concern of medical and social scientists in exploring the range of issues and questions raised about coronary heart disease in black populations. To this end, the papers in the special AHJ issue reviewed and discussed the major theoretical and clinical issues regarding the incidence and prevalence of CHD, the operative risk factors for CHD, and the possible contributory role of current medical attitudes and practices in CHD-related deaths among blacks. The working conference participants also offered a series of specific recommendations for further research needs and priorities. This report provides an integrative summary of the state of our current knowledge about CHD in blacks as reviewed in the AHJ proceedings, and organizes the sugqested research priorities into coherent programmatic recommendations to guide future research on coronary heart disease in this population. Four major substantive questions emerge as salient concerns in the recent reviews on CHD in black populations. First, what conclusions can be drawn from our present epidemiologic evidence on the incidence and prevalence of coronary heart disease in black Americans as compared to white Americans? Included here are considerations of CHD morbidity and mortality, and sudden death, as well as epidemiologic and other evidence on angina pectoris, myocardial infarction (MI), and coronary atherosclerosis as integral to the CHD spectrum. The second major question is: What are the specific risk factors and possible resistant or protective factors for CHD for blacks, and to what extent do blacks and whites differ in the nature and relative importance of these factors? Consideration is given here to whether or not the CHD risk factors identified for whites are the same and have the same degree of importance for blacks, and whether there may be genetic and/or socioenvironmental factors that increase the resistance of blacks to coronary heart disease. The third major question is: What role do socioeconomic, sociopsychological, and sociocultural factors play in the observed patterns of CHD morbidity and mortality in blacks? Consideration is given here to the range of social factors that exert detrimental effects on the social and health status of black Americans. Particular attention is paid to the conceptualization and measurement of these factors, as well as to the specification of their diffprential impact on 304 CHD risk and risk management, on CHD morbidity, on medical treatment of CHD, and, ultimately, on CHD mortality in blacks. The fourth major question addresses the health care system. Specifically, to what extent do physicians' attitudes and beliefs, and clinical practices in health-care institutions in treatinq coronary heart disease in blacks result in either differential access or differential quality of medical care for blacks? Furthermore, to what extent do these practices ultimately contribute to the disproportionately higher MI case fatality rates for blacks? CONCEPTUAL PERSPECTIVE Underlying these major concerns about CHD in black populations is a conception of hypothesized roles that various factors play in determining risk for CHD, in impacting on the course of the disease, and ultimately on the outcome of this disease in black populations. Included are biological factors, and psychosocial and sociocultural factors, all of which interact to differing degrees and at different points in the coronary heart disease process. Current conceptual models confer priority on biologic processes (i.e., genetic and metabolic) and on health behaviors as primary contributors to CHD risk, morbidity, and mortality. A family history of high blood pressure or of CHD, elevated blood pressure, presence of ECG abnormalities, obesity, diabetes, and high cholesterol levels are all acknowledged biologic risk factors. Similarly, personal health habits including smoking, limited physical exercise, heavy alcohol consumption, and such psychosocial factors as disadvantaged socioeconomic status, sustained disturbing emotions (e.g., anger), a time-oressured, aggressive behavioral type (Type A), and high life stresses have all been associated with increased CHD risk (Figure 1, Kasl, 1984; McDonald et al., 1984; Gillum & Grant, 1982). Finally, a tendency to ianore, deny, or minimize the significance of CHD symptoms (e.g., chest pains), and to delay seeking medical care has also been associated with increased risk for CHD-related deaths (Gillum, 1982). These factors are viewed as interacting to affect CHD morbidity, with the psychosocial factors operating primarily as mediators of the biological predispositions. However, several questions have been raised about whether some of the psychosocial factors might not operate as independent precursors. Smoking has been identified as an accepted risk factor for CHD and there is preliminary evidence that Type A behavior is a potential risk factor for CHD; these are examples of factors that operate biobehaviorally to increase CHD risk (Kleinman et al., 1979; Coronary Prone Behavior Review Panel, 1981). Similarly, questions have been raised about socioeconomic disadvantage which appears to be a social system precursor of a variety of pathogenic outcomes including CHD (James, 1984a). In addition to their direct and indirect effect on CHD risk, psychosocial factors also seem to affect disease course by influencing risk behavior management and the recognition and management of symptoms. 305 This is particularly true of such factors as socioeconomic status, cultural beliefs about health and illness, symptom tolerance, heal inn beliefs and practices, and help-seeking behaviors (Jackson, 1981). Individual help-seeking behavior, compliance behaviors, and attitudes toward the sick role also contribute to the course of CHD. Symptom recognition and management are also likely to be affected by the prevailing beliefs and treatment practices of health care providers. These include the knowledge and beliefs of individual physicians about CHD in particular populations, the pattern of usual care provided to CHD victims (e.g., diagnostic and treatment practices), and the host of factors that affect the pattern of care likely to be provided (e.g., financial resources of the patients, the technical resources of the providers, etc.) (Davis et al., 1981; Yelin et al., 1983; Brook et al., 1983). Analyses of health care system factors should also include the availability and response times of emergency services, and the beliefs and behaviors of these personnel when responding to CHD crises in low-income, black communities. Although comparatively less is known about the contribution of health care system factors in CHD morbidity and mortality, there is preliminary evidence suggesting that certain patient characteristics such as race, ability to pay, gender, and age may contribute to differences in the quality of medical care received, and to the pattern of health compliance behaviors observed (Shapiro et al., 1970; Jackson, 1981; Yelin et al., 1983). This interaction of differential treatment and differential pre- and post-hospital patient behaviors probably contributes ultimately to differential CHD mortality rates in the affected qroups. In sum, at the core of our concerns about CHD in black populations is the need to identify the major biologic, sociopsychologic, sociocultural, and health care system factors that may be causally implicated in this disease. There is a related need to specify more carefully to what degree and at what point in the CHD process these factors make their contributions, and whether the contributions are direct or indirect. In the specific case of CHD in blacks, we must actively entertain the possibility that not only is the pattern of occurrence of CHD somewhat different, but that the factors causally implicated and their pattern of contribution at each point in the CHD process in blacks may differ significantly from that reported for whites. The justification for this hypothesis rests in the suggestive evidence reported in the most recent review on CHD risk factors (Curry et al., 1984; Tyroler, 1984; McDonald et al., 1984). WHAT DO WE PRESENTLY KNOW ABOUT CHD IN BLACKS? 1. Epidemiologic Evidence There is general consensus that our present epidemiologic data on coronary heart disease in blacks are inadequate. Specific criticisms have been levelled at the practice of the National Center for Health Statistics and other national health data collection agencies of collapsing all health statistics on blacks into the "non-white" category. This practice confounds the health trends of the various 306 ethnic minorities and, in so doing, obscures important differences between them. The national health surveys such as the National Health and Nutrition Examination Surveys (NHANES) I and II, and some of the major cardiovascular disease surveillance and treatment studies, when they have not excluded blacks from their samples (as in the Framingham Heart Study), have used small, nonrepresentative samples -- inadequate indicators of national black population trends. In the Multiple Risk Factor Intervention Trial (MRFIT), for example, only 6.5% of their randomized sample was black. In the Hypertension Detection and Follow-up Program (HDFP), however, the investigators, recognizing the importance of hypertension in the black population, oversampled and thus enrolled a sample that was 26% black, in the study, as a whole. In the hypertensive group of HDFP, 44% were black. Questions have also been raised about how representative the largely rural, southern, black samples in the Evans County and Charleston Heart Studies are of black populations elsewhere in the nation. Several investigators have also questioned the scientific basis for our rather simplistic classification of groups by race, the tendency of many to infer (inappropriately) that observed racial group differences are due to biologic differences between the groups (Cooper, 1984), and the tendency of many biomedical investigators to ignore the biologic, socioeconomic, and psychosocial heterogeneity of blacks (Watkins, 1984a; James, 1984a, b; Myers, 1982). Our present knowledge suggests that the U.S. black population is neither biologically nor socioeconomically homogeneous and, as such, failure to take adequate account of this heterogeneity in study samples is likely to produce nongeneralizable findings, or may result in incorrect conclusions about the contribution and significance of race as a biologic factor in explanations of observed differences between blacks and whites. This error of assuming homogeneity in study cohorts is probably also committed with respect to whites and other ethnic groups. As a result of these conceptual and methodologic problems, there are ongoing disagreements about what conclusions can be drawn about coronary heart disease prevalence and mortality rates, and about the incidence of myocardial infarction in blacks. Gillum (1982) and Watkins (1984a) both note contradictory findings when comparing vital statistics with mortality data from more carefully designed epidemiologic studies that included large samples of blacks. For example, according to the vital statistics, the age-adjusted mortality rates from CHD in blacks exceeded those of whites, especially between the age-groups 35-44 years, 45-55 years, and 55-64 years. Similarly, the CHD mortality rate for black women exceeds that for white women (Report of the Working Group on Atherosclerosis, 1981). However, when these data are compared with the mortality rates from the MRFIT screenee population, the American Cancer Society (ACS), and the Evans County studies, the CHD mortality rates for black men range from 78% to 90% that of white men. Only the excess CHD mortality rate for black females (i.e., 107% in the ACS study) is consistent with the national mortality data. This picture is further complicated when the data on sudden death are considered. The results of the Charleston Heart Study show that, between 1960-1975, black males were three times and black females 1.5 307 times more likely to die suddenly (i.e., within 1 hour of onset of symptoms) than were white males and females (Keil et al., 1984). On the other hand, the Baltimore sudden death study revealed no black-white differences in sudden death (Kuller, Perper & Cooper, 1975). Questions were raised by Watkins (1984a) about whether these contradictory findings are due to differences in death certification and/or to differences in access to medical care. Although many studies also report a lower prevalence of CHD in blacks, other studies report that younger blacks and black women show evidence of more CHD symptoms than do whites (Gillum, 1982; Langford et al., 1984). Among MRFIT screenees followed for 5 years, blacks appeared to be at greater risk for cerebrovascular accidents than whites, given the higher overall prevalence of essential hypertension (Neaton et al., 1984) and, in the Charleston Heart Study, blacks had a higher MI case-fatality rate than their white counterparts (Table 2, Keil et al., 1984). Furthermore, although CHD-related mortality has decreased significantly in both blacks and whites since 1968, there is still a large disparity in black-white CHD mortality rates because of the persisting excess CHD mortality rates for black females (Figure 1, Gillum & Liu, 1984). Important age, gender, socioeconomic status, and geographic differences are also evident and must be considered in interpreting these findings (Gillum & Liu, 1984; Savage et al., 1984; Watkins, 1984a; Figures 1-4, Leaverton et al., 1984; Strong et al., 1984; Keil et al., 1984). These somewhat contradictory findings, along with the limited data sources and the methodological limitations of many of the studies reviewed, make it unwise to draw firm conclusions about the true prevalence of CHD and its consequences in black populations. However, it is clear that CHD is far from uncommon among blacks, and that blacks are not immune to its effects. Rather, what we see is preliminary evidence of gender differences in the pattern of CHD morbidity and mortality when different races are compared. For example, if we look at overall CHD mortality rates, then both black and white men have significantly higher rates than both black and white women, but black women have significantly higher mortality rates than white women (Figure 1, Gillum & Liu, 1984). On the other hand, if we consider sudden death rates, than black men and women appear to be at greater risk than their white counterparts (Gillum, 1982; Keil et al., 1984). 2. Risk Factor Differences Several hypotheses have been proposed to explain the possible reasons for these racial differences in CHD. The two most promisinq hypotheses are (1) that there are racial differences in the operative risk factors for CHD, and (2) that blacks may have some biologic and/or psychosocial protective factors which may reduce their CHD risk. An alternative hypothesis is that the standard CHD risk factors may be of different degree of importance (i.e., different risk functions) (Currv et al., 1984; Table 4, Figures 4 & 5, Tyroler, 1984; McDonald et al., 1984). The available evidence also identifies several differences between blacks and whites such as in the prevalence of hypertension, in metabolic or biochemical factors, and in psychosocial characteristics 308 that suggest that different factors may potentiate coronary heart disease in these group. a. Biological Risk Factors A strong family history of coronary heart disease is an accepted risk factor for both blacks and whites. However, a family history of essential hypertension may not be as strong a predictor of coronary heart disease for blacks given the greater overall prevalence of hypertension in black populations (Curry et al., 1984; Neaton et al., 1984). For example, in the five-year follow-up of 361,620 men screened for the MRFIT study, although elevated blood pressure was a strong predictor of CHD mortality in whites, high blood pressure appeared to be a better predictor of cerebrovascular disease and CVD mortality than of CHD mortality in these black men (Neaton et al., 1984). This does not mean, however, that hypertension in blacks is not a major CHD risk factor. The evidence is incontrovertible that there is a high population-attributable risk of hypertension for all-cause mortality and for CHD mortality in blacks (See table 2 and Figure 1, Tyroler et al., 1984; Figures 1 & 2, Neaton et al., 1984; Haywood, 1984). For these reasons, aggressive prevention and treatment of hypertension must be continued (Tyroler, 1984). Similarly, the presence of such ECG abnormalities as minor S-T segment depressions and T-wave abnormalities may not have the same diagnostic significance for blacks as for whites (Bartel et al., 1971) given the greater prevalence of these abnormalities in elderly U.S. black populations (Riley et al., 1973) and even in black societies in Africa and the Caribbean (Watkins, 1984b) where there is low CHD prevalence. However, when abnormalities indicative of left ventricular hypertrophy (LVH) are present in conjunction with other major risk factors, then total mortality risk is likely to be even higher for blacks than for whites (Table 1, Tyroler, 1984). On the other hand, obesity and diabetes mellitus, which are not as prevalent among whites, may be more significant risk factors for CHD among blacks, especially for black women (Curry et al., 1984; Tables 6 & 7, Cooper et al., 1984; McDonald et al., 1984). For lipid profiles, the situation is somewhat more complex. Typically, blacks tend to have higher HDL-cholesterol (HDL = high-density lipoprotein) and apolipoprotein Al levels, and lower levels of LDL-cholesterol (LDL = low-density lipoprotein), VLDL-cholesterol (VLDL = yery low-density lipoprotein), apolipoprotein CI I, and triglycerides than whites (Gartside et al., 1984; Figures 1-4, Glueck et al., 1984; Figures 3-5, Heiss et al., 1984; Rowland & Fulwood, 1984; Neaton et al., 1984). This combination of higher HDL- and lower LDL-cholesterol levels for equivalent total cholesterol level may confer some protection on blacks. Unfortunately, there is some preliminary evidence from the MRFIT study suggesting that an unexpected by-product of aggressive pharmacotherapeutic intervention in hypertension is a significant reduction in HDL-cholesterol level especially when propranolol is added to diuretic regimens. Concurrent changes in diet and weight may have also contributed to these changes in the lipoproteins (Lasser et al., 1984; Currv et al., 1984). Tyroler (1984), however, argues against 309 overinterpreting these results as directly relevant to the larqe population of black hypertensives, since these findings were obtained in a select sample of upper middle-class white men. Whatever reduction, if any, in degree of protection against CHD that may result from diuretic- and propranolol-induced HDL-cholesterol level changes is not comparable in magnitude to the significant benefits derived from aggressive blood pressure control in black communities. McDonald et al. (1984) also point to several apparent biochemical risk factor differences between blacks and whites that include differences in sodium/potassium ratios in urine, in plasma renin and epinephrine levels, in hematocrit values, in differential rates of G6PD enzyme deficiency, and in possible differences in transport of these electrolytes via the sodium/potassium pump in the red blood cells. All of these differences may be related to blood pressure differences between the groups. However, as noted by Watkins (in press), the Evans County, Georgia study found that black-white differences in blood pressure were not due to higher sodium intake in blacks (Grim et al., 1980). Blaustein (1984) and Trevisan et al. (1984) also discount disturbances in the sodium transport mechanism as accounting for racial differences in hypertensive populations. b. Psychosocial Risk Factors When we consider the host of psychosocial risk factors implicated in coronary heart disease, the picture is even more complex. Psychosocial risk factors for CHD that are generally accepted or suggested include cigarette smoking and heavy alcohol consumption, social isolation and inadequate social support, and sustained disturbinq emotions (e.g., anxiety, hostility, etc.), especially as part of the Type A behavior pattern (Coronary Prone Behavior Review Panel, 1981; McDonald et al., 1984). In the case of blacks, these factors do not appear to operate as directly or to have the same demonstrated importance as they do for whites. For example, socioeconomic disadvantage, including experiences of racism and discrimination, appears to make a more significant risk contribution for blacks than for whites (Tyroler et al., 1984; James, 1984a). On the other hand, risk behaviors such as smoking, alcohol consumption, and limited physical activity appear to be pathogenic to blacks and whites to the same degree. However, although more blacks smoke, black smokers typically smoke fewer cigarettes than white smokers (See Table 6, Garfinkel, 1984). Similarly, the relationship between cigarette smoking and pattern of CHD and risk appears to be different for blacks and whites. Neaton et al. (1984) note that, in the MRFIT screenee follow-up studies, different 5-year, age-adjusted, CHD mortality rates were obtained for black and white men as a function of the number of cigarettes smoked (Figure 8). They also noted that black, 5-year, cerebrovascular mortality rates were significantly higher than for whites, for a comparable number of cigarettes smoked (Figure 9). In addition, there are complex interrelationships between cigarette smoking, cholesterol levels, and blood pressure that impact on CHD mortality; these appear to be discordant between blacks and whites 310 (Neaton et al., 1984; Garfinkel, 1984). The relative risk (RR) of CHD death over 5 years, in smokers compared to nonsmokers, was relatively constant in subgroups of black and white MRFIT screenees categorized according to serum cholesterol level (250 mg/dl) and diastolic blood pressure (DBP), both)90 mm Hg and<90 mm Hg. The RRs lay in the ranqe 1.62 - 1.83. In contrast, for men with a serum cholesterol level of 250 mg/dl, the corresponding RRs lay in the range 2.23 - 3.72, and were highest for black males with a DBP of 90 mm Ha (RR = 3.72) and for white males with a DBP of 90 mm Hg (RR - 2.87). However, the highest mortality rates were observed among smokers with elevated serum cholesterol levels, in black normotensive men, and in white hypertensive men (Table 7, Neaton et al., 1984). Also, blacks as a group tend to be more physically active than whites (Rowland & Fulton, 1984), although this difference may well be due to socioeconomic status-related differences in the proportion of white- vs. blue-collar workers in these groups. Research on the role of Type A behavior pattern and related effects of strong emotions on CHD risk, although extensive for whites (Review Panel, 1981) is extremely limited for blacks (James, 1984b). The absence of relevant research on this question is a major lacuna in the literature. The related work on suppressed anger (Diamond, 1983) and, most recently, on the interaction of low education and the John Henry-type coping pattern in hypertensive black men (James, 1984c) suggests that these psychological processes need to be seriously considered as possible risk factors in CHD in blacks as well. Similar attention should be given to occupational stresses (House, 1974; James et al., 1984d), and to social mobility (Gillum & Paffenbarger, 1978) as additional, potential risk factors for CHD, especially qiven the unique pressures faced by increasing numbers of black men and women who are entering professional and managerial occupations. At the broader socioecologic level, blacks face disproportionately greater CHD risks due to higher life stresses (Harburg et al., 1973) and to lower medical care access, utilization, and, possibly, lower quality care (Yelin, Kramer & Epstein, 1983). In addition to the consideration of risk factors, there is also the need to explore the role of possible biologic and psychosocial protectors against CHD. Given the higher prevalence of hypertension and comparable prevalence of other CHD risk factors in blacks, it is surprising that the prevalence of CHD and CHD mortality rates in blacks is not even higher (Gillum, 1982). One possible explanation for the current lower-than-expected rates in blacks is the presence of factors that confer some degree of protection against CHD on blacks. We have already reviewed the evidence of possible biologic protection in the lipid profile that combines higher HDL-cholesterol and apolipoprotein Al levels with total cholesterol level, which is more common in blacks than in whites (Heiss et al., 1984; Glueck et al., 1984). It would also be useful to explore the possible protective role that social supports (Broadhead et al., 1983), psychological hardiness (Kobasa, Maddi & Kahn, 1982), a high sense of personal, social, and cultural coherence (Antonovski, 1979), low "John Henryism" (James et al., 1984c), and other 311 adaptive stress-coping styles (Billings & Moos, 1981) may play in reducing CHD risk in blacks. In sum, these data point to the need to pursue actively this question of possible racial differences in CHD risk factor levels and risk functions, as well as in protective factors. In this process, a multidimensional perspective that includes biological, psychological, and social risks, as well as possible protectors, should be considered. 3. The Salience of Socioeconomic & Sociocultural Factors A major recurrent theme in the debate over black/white differences in the pathogenesis of coronary heart disease is the apparent differential significance of socioeconomic status (SES) and sociocultural factors. There seems to be some consensus of opinion that, for blacks, social status and related sociocultural influences probably contribute in some as-yet-undefined way to the observed pattern of CHD risks and outcomes (Myers et al., 1984; James, 1984a, b; Kasl, 1984). This consensus appears to be based on a collection of theoretical formulations, clinical observations, and on provocative empirical evidence. The absence of more conclusive data seems to be due to deficits in the conceptualization, specification, and measurement of social status variables; to inconsistencies in the consideration of these variables in studies of CHD; and to the frequent omission of these variables, or to the confounding of these variables, with race in studies of CHD in multiethnic populations. A related point is that investigators typically look for main effects of SES and related variables rather than testing the hypothesis that these factors may operate interactively with other factors. The hypothesis that these variables may be important to different degrees in different populations, and may operate at different points in the CHD process is also typically ignored. An alternative conceptualization of social status and of this selective effects' hypothesis was articulated in the Working Group III report (Myers et al., 1984). The contributors suggested that socioeconomic and sociocultural factors should be subsumed under the generic heading of social status, and that this construct should be defined multidimensionally. Sociostructural and sociopsychological characteristics were distinguished and speculations about their contributions to CHD were made. Sociostructural features of social status include such factors as race, education, occupation, income, marital status, and social mobility. These factors define social status through the social meanings, institutions, and practices that govern life opportunities and obstacles for blacks. The sociopsychological features of social status, on the other hand, include the host of individual differences in experiences and personal attributes, resources, and liabilities, all of which determine individual vulnerability or resistance to objective external social conditions. These include personality characteristics, life stresses, coping styles, the availability and use of social supports, patterns of help-seeking, and characteristic levels of anxiety, happiness, anger, and other similarly strong emotions. 312 In the assessment of these factors, careful attention must be given to distinguishing objective, concrete events and experiences that blacks face from the subjective interpretations and meanings conferred on these experiences. The latter include consideration of both individual and collective group meanings (McDonald et al., 1984; James, 1984a; Kasl, 1984; Myers, 1982). In addition, consideration should be given to the analysis of main effects and interactions of social status as predictors of CHD outcomes. Most of the available evidence seems to provide more support for an interactional contribution of these variables toward predicting illness outcomes (Harburg et al., 1973; Kasl, 1984). Finally, there is growing justification for considering separately the influence of social status on the presence of factors that either increase or decrease CHD risk status; on factors that influence the management of risk (i.e., early symptom identification, intervention, and disease prevention); on factors that influence the level of CHD morbidity observed; and on factors that ultimately affect the level and pattern of CHD mortality observed in the black population. Social status factors may also exert their impact through such health care system factors as the timing and availability of medical services, the accuracy of clinical diagnosis, and the appropriateness and effectiveness of the treatment received (Myers et al., 1984; Cooper et al., 1981; James, 1984a; Kasl, 1984). Figure 1 in Myers et al. (1984) describes these hypothesized influences in a simplified diagram. 4. Racial Differences in the Diagnosis, Treatment, & Course of Coronary Heart Disease The final substantive concern is whether the observed black/white differences in CHD mortality might be due, in part, to differences in the natural history and treatment of CHD. This concern over possible differential health care system effects as a function of race and social class is based on anecdotal, clinical, and preliminary research reports that suggest that there are important differences between blacks and whites in rates of access and utilization of preventive and emergency medical services (Jackson, 1981; Yelin et al., 1983). Also, reports of differences in risk status at the point at which treatment is initiated (Haywood, 1984a, b; Francis et al., 1984; Curry et al., 1984), and differential use of coronary arteriography and coronary artery bypass grafting (CABG) procedures with black and white CHD patients (Oberman & Cutter, 1984) have been noted. Physicians have also reported greater resistance to, and lower compliance with, medical regimens among black patients: This is believed to contribute toward diminishing the effectiveness of treatment (Jackson, 1981). Data are inadequate with respect to secondary prevention efforts (Oberman & Cutter, 1984), but those studies that used standardized diagnostic and intervention procedures (e.g., MRFIT & HDFP) show that blacks often derive similar or greater benefit than whites from such intensive approaches to CHD risk reduction and treatment (Tables 6 & 7, Rowland & Fulwood, 1984; Table 5, Connett & Stamler, 1984; Table 5, Langford et al., 1984). 313 The data on surgical outcomes for black CHD patients are also inadequate (Haywood, 1984). However, preliminary evidence suggests that blacks show no greater operative mortality risk to myocardial revascularization procedures when hypertension, diabetes, and prevalence of hyperlipidemia are taken into account (Table I, Sterling et al., 1984). It is apparent from these reviews that no strong conclusions about black/white differences in the natural course of CHD are justified at this time, given the paucity of good data. However, there is enough evidence to justify the suspicion that some racial and SES differences in CHD course do exist, and that these differences are probably attributable to differences in level of risk, pattern of health care utilization, and in the pattern of health care system characteristics. These factors probably interact to produce a more pathogenic process, high CHD mortality, and higher CHD-related sudden death rates in black Americans. RECOMMENDATIONS FOR FUTURE RESEARCH The present evidence supports the contention of Gillum and others (Gillum, 1982a; Gillum & Grant, 1982) that CHD is a major health problem for black Americans. Further, though there is growing and provocative evidence suggesting that blacks and whites may not differ significantly in overall prevalence and incidence rates for CHD, these groups do appear to differ in overall CHD mortality and sudden death rates. Furthermore, low-income blacks, and black women in particular, appear to be at greater mortality risk than their male, white, and upper-SES counterparts. In addition, the evidence suggests that many of these racial differences may well be due to differences in biologic and psychosocial risk factors, as well as to differences in the usual diagnostic and treatment practices for suspected CHD. Because of such significant deficits in the available data on CHD in blacks, these statements should be treated as tentative and viewed more as interesting hypotheses to be pursued vigorously rather than as firm conclusions. The significance of these issues to the health of the nation, however, clearly justifies an aggressive approach toward answering major questions about CHD in black populations. The chief areas of deficit and the appropriate recommendations for future research are organized programrnatically as follows: 1. The most glaring lacuna results from the limited availability of epidemiologic and clinical data on CHD in blacks. This deficit can be corrected by a two-stage process consisting of: a. first, existing NCHS data bases should be improved by separating the data on blacks from the "non-white" category, and by including large, nationally representative samples of blacks in all existing or proposed national health studies (e.g., NHANES, the Coronary Artery Risk Development Study, and the Community CHD Surveillance Studies); and 314 b. second, population-based prospective studies of black cohorts similar to the Framingham Heart Study should be funded. In all of these studies great care should be exercised in the selection of appropriately representative black samples, with consideration given to such sources of group variability as urban vs. rural residency, geographic location, socioeconomic status, and national origin. The reliability and appropriateness of the assessment and diagnostic tools and procedures used with this population must be demonstrated. 2. The second major research priority should be a series of studies addressing the issue of differential CHD risk pattern and functions in blacks and whites. This priority could be addressed by: a. a more complete risk profile analysis of the black participants of the HDFP and MRFIT studies should be supported. This is a cost-effective strategy that avoids the high start-up costs typically associated with new research efforts with large samples; and b. the development of multi-center, multi-disciplinary, case-control studies focused specifically on such issues as sudden death in the community, hospital admission and discharge of patients with diagnoses compatible with CHD, emergency room visits for chest pain and related complaints, and surveillance of the offspring of indexed cases should be supported. The latter is probably more cost-effective than separate studies, and would take better advantage of existing research teams in geographically diverse areas with access to different black populations. Risk factor studies should be based on a multifactorial, biobehavioral perspective that includes possible biological, psychological, behavioral, and social sources of differences in risk factors and protective factors. Consideration of possible differences in both risk factor level and risk function should also be included. 3. Finally, there is considerable justification for supporting studies that address the interface of patient and community characteristics on the one hand, and health care system characteristics and practices on the other, as predictors of CHD risk, natural course, and outcome. Priority should be placed on studies that focus on physician beliefs and attitudes, and the usual diagnostic and intervention practices used to treat black patients with suspected or confirmed CHD. Attention should also be given to how social status and associated sociopsychologic factors influence the host of CHD-related health beliefs and behaviors of blacks and how these, in turn, interact with the characteristics of community health care providers to affect the observed pattern of coronary heart disease morbidity, natural course, and mortality in black 315 populations. (See Appendix for a summary list of the specific research recommendations made). Underlying all of these recommendations is the need to give priority to sound conceptualization and methodological rigor. Most studies on CHD in blacks, as previously noted, suffer from a variety of conceptual and methodological deficiencies which must be avoided in future studies. Central among these are problems in the specification and adequacy of measurement of the variables of interest, and inadequate size and representativeness of the sample of blacks studied. Particular attention to the issues of socioeconomic status and to gender must be given in all of these studies. In addition to aggressive support for research on the causes of CHD in blacks, there is need also to increase support for prevention efforts directed at reducing CHD risk factors in blacks. Ultimately, little will be gained by knowing how prevalent CHD is in black populations or what factors contribute to its occurrence if we do not also act aggressively to prevent it. Results from the NHANES I & II studies (Rowland & Fullwood, 1984), from the HDFP study (Langford et al., 1984), and from the MRFIT study (Connett & Stamler, 1984) show significant reductions in CHD risk-related factors in blacks through aggressive programs of education and treatment. Such efforts must be continued and should include greater black representation in future samples. In sum, the present evidence points to the significance of CHD as a major health problem for black Americans. Further, these data argue for the need for more aggressive and focused support for studies targeted at the major questions about CHD epidemiology, risk status, and patient-health care system transactions that might account for the observed CHD differences between black and white populations. Support for targeted primary prevention programs in black populations is also justified. This paper was prepared under personal services contract No. 263SGX13736 from the Secretary's Task Force on Black and Minority Health, Department of Health and Human Services. Dr. Myers is also a Scholar-In-Residence at the Fanon Research & Development Center, Charles R. Drew Postgraduate Medical School, Los Angeles. The author wishes to express his appreciation to Dr. Laurence 0. Watkins and Elisabeth Pitt for their careful review and suggested revisions. 316 APPENDIX SUMMARY RECOMMENDATIONS FROM: The National Heart, Lung, and Blood Institute's Working Conference on Coronary Heart Disease in Black Populations Bethesda, Maryland September 29-30, 1983 PURPOSE: To review current knowledge about coronary heart disease in blacks and identify areas for priority research effort. The areas recommended included: Prospective studies of the incidence of coronary heart disease in blacks. Research to determine why black women apparently have higher coronary heart disease mortality rates than white women. Research to determine the impact of antihypertensive treatment on coronary heart disease in blacks. Research on ways to modify, in school-age populations, the development of eating patterns and physical activity behavior that may result in obesity in later life. Research on variables such as risk factor patterns associated with coronary heart disease treatment, includina coronary artery bypass surgery and angioplasty and its outcome in blacks compared with whites. • Studies of beliefs, awareness, and pre-hospital behavior that might delay appropriate diagnosis and treatment for individuals in the black community who have symptoms of coronary heart disease. • Studies of risk factors for coronary heart disease in blacks that include investigation of the impact of socioeconomic status and ethnic diversity. • Special efforts be made to aid minority researchers and those in minority research settings to be competitive in seeking research funding. 317 SUMMARY RECOMMENDATIONS FROM: The National Heart, Lung, and Blood Institute's Working Group to Discuss Strategies for Minimizing Coronary Heart Disease in Black Populations February 21-22, 1984 PURPOSE: To develop research strategies to answer priority issues related to the gaps in our knowledge concerning coronary heart disease in blacks. Five task groups were convened to focus on priority issues developed by the previous conference. Task Group I focused on the problem of a paucity of epidemiologic data available on the occurrence of coronary heart disease among blacks and on the uncertain quality of mortality and morbidity data available for blacks. • They recommended further development of NHLBI research in community and cohort surveillance of coronary heart disease risk factors and events in black Americans. They recommended a specific, prospective, community-based cohort study with oversampling of blacks in one geographic area which would ensure an adequate number and diversity of blacks to assess prevalence of coronary heart disease, and physiologic and psychosocial risk factors of coronary heart disease. • Longitudinal follow-up for at least five years would then provide natural history data for an adequate number of blacks. The Task Group suggested that cooperation with the NHANES III study (to be carried out by the National Center for Health Statistics) would be desirable. The phase one validation and pilot studies, the definitive study, and the final data analysis would require an estimated eight years. • In addition, they recommended cohort studies among black professionals and among industrial populations or medical insurance groups. • This Task Group also recommended a mortality data follow-back survey to focus on the comparability of the accuracy of cause of death, as reported on death certificates and in hospital records, for blacks and whites, with particular concern for the sensitivity and specificity of coronary heart disease ascertainment. The goal would be to determine causes of noncomparability and to try to develop programs to reduce the problems. 318 Task Group II focused on the research strategies for ascertaining the distribution of known and suspected risk factors for coronary heart disease (biochemical and psychosocial) for race-sex-socioeconomic status subgroups of the population and their relative contribution over time to the development of coronary heart disease in these subgroups." • They recommended that a large, multicenter, prospective study in diverse settings be initiated to determine differences in the prevalence of coronary heart disease and associated risk factors between blacks and whites. This study would include prospective follow-up to determine the relative contribution of these risk factors to coronary heart disease events over time. Risk factors to be measured would include the standard risk factors, health habits, and psychosocial environmental factors. Cooperative efforts with the National Center for Health Statistics were also recommended. Various instruments for measurement of risk factors, especially psychosocial variables, would need to be validated in blacks. (This recommendation is compatible with the similar recommendation from Task Group I.) Task Group III focused their attention on possible research efforts to determine the effects of various antihypertensive therapies on coronary heart disease events in high-risk populations (particularly blacks) and to determine whether the metabolic, hemodynamic, and side-effects of these treatments are different in blacks as compared with whites. • The task aroup recommended a multicenter, randomized, double-blind, clinical trial as the best type of study to address these issues. Phase I pilot studies, as well as the phase II definitive trial, were recommended. The latter would be at least five years in duration. Task Group IV focused on the issue of obesity as a risk factor for coronary heart disease in black females. • They recommended that research is needed to determine the relative importance of obesity as a risk factor for coronary heart disease in black females. Case-control studies to obtain estimates of relative risk were recommended initially and then prospective epidemiologic and intervention studies. • Methodologic studies to develop tools for assessing coronary heart disease risks in blacks compared with whites were also recommended. • Studies were recommended to assess the value of weight reduction on modification of other coronary heart disease risk factors in black and white women, stratified by different levels of obesity. Methodologic studies may be needed to identify effective means for achieving weight loss in different subgroups. 319 Research was recommended to determine factors associated with and potentially responsible for the observed divergence in development of obesity after adolescence in black compared with white females. Task Group V focused on the problem of apparent greater incidence of pre-hospital and sudden death and the lower hospital survival rates of blacks following heart attack. This task group recommended survey efforts to learn what level of awareness is present among adult blacks regarding risk factors and symptoms of coronary heart disease, and to investigate and assess the appropriateness of pre-hospital behavior by blacks, in response to such symptoms. The goal of this research is to obtain information to assist in implementing interventions that will reduce mortality and morbidity resulting from CHD in black populations. Pilot studies and pretests of measurement instruments may be necessary. This task group also recommended a study to assess the extent of knowledge, attitudes, and practices of health providers in diagnosing and treating coronary heart disease in black populations. Such a study would also develop appropriate educational interventions. 320 REFERENCES 1. Autonovski A: Health, Stress & Coping. San Francisco, Jossey-Bass, 1979. 2. Berenson GS, Webber LS, Srinivasan SR, Cresanta JL, Frank GC, and Farris RP. Black-White contrasts as determinants of cardiovascular risk in childhood: Precursors of coronary artery and primary hypertensive diseases. American Heart J 108(3;2):672-683, 1984. 3. Billings AG and Moos RH: The role of coping responses and social resources in attenuating the impact of stressful life events. J of Behavioral Medicine 4:139-157, 1981. 4. Blaustein MP: Sodium transport and hypertension: Where are we going? Hypertension 6:445-453, 1984. 5. Broadhead WE, Kaplan BH, James SA, Wagner EH, Schoenbach VJ, Grimson R, Heyden S, Tibblin G, and Gehlbach SH: The enidemioloqic evidence for a relationship between social support and health. American J of Epidemiology 117:521, 1983. 6. Brook RH, Ware JE, Rogers WH, Keeler EB, Davies AR, Donald A, Goldberg GA, Lohr N, Masthay C, and Newhouse JP: Does free care improve adult health? Results from a randomized controlled trial. New England J of Medicine 309(23):1426-1434, 1983. 7. Connett JE and Stamler J: Responses of black and white males to the special intervention program of the Multiple Risk Factor Intervention Trial. American Heart J 108(3;2):839-848, 1984. 8. Cooper R: Cardiovascular mortality among blacks, hypertension control and the Reagan budget. J of the National Medical Association 73(11):1019-1020, 1981. 9. Cooper R: A note on the biologic concept of race and its application in epidemiologic research. American Heart J 108 (3;2):715-723, 1984. 10. Cooper R, Liu K, Stamler J, Schoenberger JA, Shekelle RR, Collette P, Shekelle S: Prevalence of diabetes/hyperglycemia and associated cardiovascular risk factors in blacks and whites: Chicago Heart Association Detection Project in Industry. American Heart J 108 (3;2):827-833, 1984. 11. Curry CL, Oliver J, Mumtaz FB: Coronary artery disease in blacks: Risk~factors. American Heart J 108(3;2):653-657, 1984. 12. Davis, Gold, and Makue: Access to health care for the poor: Does the gap remain? Annual Review of Public Health 2:159-182, 1981. 13. Diamond EL: The role of anger and hostility in essential hypertension and coronary heart disease. Psychology Bulletin 92 (2)_:410-433, 1982. 14. Francis C, et al.: Summary of Workshop IV: Working group on natural history, prevention, and medical and surgical treatment. American Heart J 108(3;2):711-715, 1984. 15. Garfinkel L: Cigarette smoking and coronary heart disease in blacks: Comparison to whites in a prospective study. American Heart Journal 108(3;2):802-807, 1984. 16. Gartside PS, Khoury P, and Glueck CJ: Determinants of hiah-density lipoprotein cholesterol in blacks and whites: The second National Health and Nutrition Examination Survey. American Heart J 108 (3;2):641-653, 1984. 17. Gillum RF: Coronary heart disease in black populations. I: Mortality and morbidity. American Heart J 104(4;1):839-851, 1982. 321 18. Gillum RF and Grant CT: Coronary heart disease in black populations. II: Risk factors. American Heart Journal 104 (4;l):852-864, 1982. 19. Gillum RF and Liu KC: Coronary heart disease mortality in United States' blacks, 1940-1978: Trends and unanswered questions. American Heart J 108(3;2):728-732, 1984. 20. Gillum RF and Paffenbarger RS: Chronic disease in former college students. XVII: Sociocultural mobility as a precursor of coronary heart disease and hypertension. American J of Epidemioloqv 108:289, 1978. 21. Glueck CJ, Gartside P, Laskarzewski PM, Khoury P, Tyroler HA: High-density lipoprotein cholesterol in blacks and whites: Potential ramifications for coronary heart disease. American Heart J 108(3;2):815-827, 1984. 22. Harburg E, Erfurt JC, Chape LS, Hauestein LS, Schull WJ, and Schork MA: Socioecological stressor areas and black-white blood pressure: Detroit. J of Chronic Diseases 26:595, 1973. 23. Haywood LJ: Coronary heart disease mortality/morbidity and risk in blacks. I: Clinical manifestations and diagnostic criteria: The experience with the Beta-blocker Heart Attack Trial. American Heart J 108(3;2):787-794, 1984a. 24. Haywood LJ: Coronary heart disease mortality/morbidity and risk in blacks. II: Access to medical care. American Heart J 108 (3;2):794-797, 1984b. 25. Haywood LJ: Issues in the natural history and treatment of coronary heart disease in black populations: Medical management. American Heart J 108(3;2):683-687, 1984c. 26. Heiss G, Schonfeld G, Johnson JL, Heyden S, Hames CG, Tyroler HA: Black-white differences in plasma levels of apolipoproteins: The Evans County Heart Study. American Heart J 108(3;2):807-814, 1984. 27. House JS: Occupational stress and coronary heart disease: A review and theoretical integration. J of Health & Social Behavior 1^:12, 1974. 28. Jackson JJ: Urban black Americans. In A. Harwood (Ed.), Ethnicity & Medical Care. Cambridge, MA, Harvard University Press, 1981, p 37. 29. James SA: Socioeconomic influences on coronary heart disease in black populations. American Heart J 108(3;2):669-672, 1984a. 30. James SA: Coronary heart disease in black Americans: Suggestions for research on psychosocial factors. American Heart J 108 (3;2):833-838, 1984b. 31. James SA, Hartnett SA, and Kalsbeek W: John Henryism and blood pressure differences among black men. J of Behavioral Medicine 6:259, 1983c. 32. James SA, LaCroix AZ, Kleinbaum DG, and Strogatz DS: John Henryism and blood pressure differences among black men. II: The role of occupational stressors. J of Behavioral Medicine 6:257, 1984d. 33. Kasl SV: Social and psychologic factors in the etiology of coronary heart disease in black populations: An exploration of research needs. American Heart J 108(3;2):660-669, 1984. 34. Keil JE, Loadholt CB, Weinrich MC, Sandifer H, and Boyle E: Incidence of coronary heart disease in blacks in Charleston, South Carolina. American Heart J 108(3;2):779-786, 1984. 322 35. Kleinman JC, Feldman JJ, and Monk MA: The effects of changes in smoking habits on coronary heart disease mortality. American J of Public Health 6:795, 1979'. 36. Kobasa, Maddi SR, and Kahn S: Hardiness and health. J of Personality and Social Psychology 42:168-177, 1982. 37. Kuller L, Perper J, and Cooper M: Demographic characteristics and trends in atherosclerotic heart disease mortality: Sudden death and myocardial infarction. Circulation 52 (Suppl. IH):1, 1975. 38. Langford HG, Oberman A, Borhani NO, Entwisle G, and Tung B: Black-white comparison of indices of coronary heart disease and myocardial infarction in the stepped-care cohort of the Hypertension Detection and Follow-up Program. American Heart J 108 (3;2):797-801, 1984. 39. Lasser NL, Grandits G, Caggiula W, Cutler JA, Grimm RH, Kuller LH, Sherwin RW, and Stamler J: Effects of antihypertensive therapy on plasma lipids and lipoproteins in the Multiple Risk Factor Intervention Trial. The American J of Medicine 76 (II-A):52-66, 1984. 40. Leaverton PE, Feinleib M, and Thorn T: Coronary heart disease mortality in United States' blacks, 1968-1978: Interstate variation. American Heart J 108(3;2):732-737, 1984. 41. McDonald R, et al.: Summary of Workshop II: Working group on risk factors. American Heart J 108(3;2):703-706, 1984. 42. Myers HF: Stress, ethnicity and social class: A model for research with black populations. In E.E. Jones and S.J. Korchin (Eds.), Minority Mental Health. New York, Praeger Press, 1982, p 118. 43. Myers HF, et al.: Summary of Workshop III: Working group on socioeconomic and sociocultural influences. American Heart J 108 (3;2):706-710, 1984. 44. Neaton JD, Kuller LH, Wentworth D, and Borhani NO: Total and cardiovascular mortality in relation to cigarette smoking, serum cholesterol concentration, and diastolic blood pressure among black and white males followed up for five years. American Heart J 108 (3;2):759-769, 1984. 45. Oberman A, and Cutter G: Issues in the natural history and treatment of coronary heart disease in black populations: Suraical treatment. American Heart J 108(3;2):688-694, 1984. 46. Rowland ML, and Fulwood R: Coronary heart disease risk factor trends in blacks between the first and second National Health and Nutrition Examination Surveys, United States, 1971-1980. American Heart J 108(3;2):771-779, 1984. 47. Savage D, et al.: Summary of Workshop I: Working group on epidemiology. American Heart J 108(3;2):699-703, 1984. 48. Shapiro S, Weinblatt, Frank CW, and Sager RV: Social factors in the prognosis of men following first myocardial infarction. Milbank'Memorial Fund Quarterly 48:37-50, 1970. 49. Sterling R, Graeber GM, Albus RA, Burton NA, Lough FC, and Fleming AW: Results of myocardial revascularization in black males. American Heart J 108(3;2):695-699, 1984. 50. Strong JP, Oalman MC, Newman WP, Tracy RE, Malcom GT, Johnson WD, McMahan LH, Rock WA, and Guzman MA: Coronary heart disease in young black and white males in New Orleans: Community Pathology Study. American Heart J 108(3;2):747-759, 1984. 323 51. The Review Panel on Coronary-Prone Behavior and Coronary Heart Disease. Coronary-prone behavior and coronary heart disease: A critical review. 'Circulation 63(6):1199-1215, 1981. 52. Trevisan M, Ostrow D, Cooper RS, Sempos C, and Stamler J: Sex and race differences in sodium-lithium counter-transport and red cell sodium concentration. American J of Epidemiology 120(4):537-541, 1984. 53. Tyroler HA: Overview of risk factors for coronary heart disease in black populations. American Heart J 108(3;2):658-660, 1984. 54. Tyroler HA, Knowles MG, Wing SB, Logue EE, Davis CE, Heiss G, Heyden S, Hames CG: Ischemic heart disease risk factors and twenty-year mortality in middle-age Evans County black males. American Heart J 108(3;2):738-747, 1984. 55. Watkins LO: Epidemiology of coronary heart disease in black populations: Methodologic proposals. American Heart J 108(3;2):635-640, 1984a. 56. Watkins LO: Coronary heart disease and coronary disease risk factors in black populations in underdeveloped countries: The case for primordial prevention. American Heart J 108(3;2):850-862, 1984b. 57. Watkins LO: The epidemiology of cardiovascular disease in black populations. In R.A. Williams (Ed.), Textbook of Ethnic Medicine. New York, McGraw-Hill, in press. 58. Yelin EH, Kramer IS, and Epstein WV: Is health care use equivalent across social groups? A diagnosis-based study. American J of Public Health 73(5):563-571, 1983. 324 Fig. 1: S.V. Kasl. p. 661 I. Asynptonatic status, risk factor(s) absent II. Asymptcnatic status, risk faetor(s; present III. Subclinical disease susceptible to detection i • su I IV. Initial tymptan experience l~ V. Initial event (diagnostic criteria net) l~ VI. Course of disease (repeat episodes, residual disability, etc.) Via. Natural course VTTb. Treatment experience I* VII. Mortality (oase fatality) Fig. 1. Developmental schema for coronary heart dis- ease: the possible role of race and psychosocial factors. 325 Table 2: Keil et al, 1984 T«bl« II. Incidence* of CHD manifestations by race and sex, 1960-1961 to 1974-1975 Black males Black males , high Black fema les White fema les (N - 322) White males (TV '601) SES (N - 101) (N - 440) (N = 71b Manifestation n Rate SE n Rate SE n Rate SE n Rate SE n Rate SE 11 CHD 43 131.7 18.3 114 188.4 15.5 4 61.2 34.5 72 161.0 17.1 84 113.8 11.3 CHD living 16 50.7 12.6 60 95.2 11.6 2 26.0 26.6 42 100.9 14.8 50 67.4 9.2 AMI 9 28.5 9.9 51 80.5 10.8 2 26.4 26.9 15 36.3 9.4 27 36.3 6.9 Angina pectoris 7 22.2 9.1 8 12.8 4.8 0 0 27 63.6 12.1 22 29.8 6.4 HD death 27 79.8 14.2 54 93.8 11.7 2 38.3 37.4 30 62.2 10.5 34 46.3 7.1 AMI 12 35.8 10.2 35 60.5 9.7 0 0 17 35.6 8.4 18 24.4 5.5 Sudden death 11 32.2 9.3 6 10.2 4.6 0 0 6 13.0 5.8 6 8.2 3.7 ASHD 4 11.7 7.3 13 23.2 6.4 2 44.0 43.0 7 13.8 5.7 10 13.8 4.4 ASHD - Atherosclerotic heart disease. 'Rates per thousand, age adjusted by indirect method. Rates for subheadings not additive. 326 FIG. 1: GILLUM £ LIU , p. 729 1000 _ 500- F 400 1940 Fig. 1. Age-adjusted CHD mortality rates per 100,000 population for U.S. nonwhites aged 34 to 74 years, from 1940 to 1978. WM = White males; NM = nonwhite males; WF - white females; NF = nonwhite females. 327 FIGS. 1 & 2: LEAVERTON, FEINLIEB & THQM, p.p. 733-73*< 3 LOWFST Fig. 1. U.S. age-adjusted CHD mortality for white males aged 35 to 74 years. IKS-1972 1978 Fig. 2. U.S. age-adjusted CHD mortality for white females aged 35 to 74 years. 328 FIGS. 3 & A: LEAVERTON, FEINLIEB & THQM, p.p. 735-736 W68-1972 1968-1972 1978 1978 LOWEST NO OATA HIGHEST SECOND Third LOWEST NO OATA Fig. 3. U.S. age-adjusted CHD mortality for black males aged 35 to 74 years (34 states with data). Fig. 4. U.S. age-adjusted CHD mortality for black females aged 35 to 74 years (34 states with data). 329 TABLE k: TYROLER et. al., p. 7^1 Table IV. Multivariate association of risk indicators with time to death (PH) and 20-year cumulative risk of death (LRF) in males aged 40 to 64 years (all-cause mortality) in Evans County Blacks Whites LSS HSS PH LRF PH LRF PH LRF Intercept -5.653 4.054 -6.943 Age 0.075* 0.103' 0.059* 0.086* 0.059* 0.076+ SBP 0.016* 0.022* 0.015* 0.019' 0.012+ 0.017+ Cholesterol^ -0.020 -0.041 -0.033} -0.05SJ 0.0661 0.0921 Cholesterol2^ 0.005 0.010 0.009+ 0.015J -0.0131 -0.0181 Smoking current 0.547+ 0.492j| 0.701 + 1.071+ 0.862' 1.195+ Smoking past -0.037 -0.115 0.54 2} 0.909§ 0.347 0.489 Quetelet index*; -0.480 0.399 -2.903 J -3.513§ -3.676t -6.0131 Quetelet index:^ 0.053 -0.081 0.369 J 0.429§ 0.428+ 0.7361 Deaths/population 129/294 126/276 88/236 XJ 73 64 68 55 48 47 •p < 0.001. tO.001

0 1 1 1 1 P6) 1 C7S-S0) HI 111-151 IV (•6-91) V <>ti) Diastolic Blood Pressure Ouintile (mm Hg) ■ lack Men: wrui. u.n 6 3.602 7.418 3.724 65242 3.955 61.351 4.708 65.051 7.491 66.322 Fig. 1. Five-year age-adjusted total mortality rate (per 1000) by diastolic blood pressure level by race. • Btack Men O WNte Men £| a t> • < Is I < ■ • > n 2 ^ 1 ' ' No. Bi.o Man. Mo. wrvi. Umn. 0'— 1 1.76) 3602 67.418 i «i rv (76-60) <»1-6S) C8f-»0 Diastolic Blood Pressure Ouintile (mmHg) 3.734 3955 «-7n0 65.242 61.251 65.051 V M1I 7,491 66.322 Fig. 2. Five-year age-adjusted CHD mortality rate (per 1000) by diastolic blood pressure level by race. 332 TABLE 1: H. A. TYROLER , p. 659 Table I. Five-year mortality rates in black and white males by education and presence of LVH at baseline for ail HDFP stratum I participants and those not receiving medication at baseline: Referred-care males ages 40 to 69 years, with entry diastolic blood pressure of 90 to 104 mm Hg* % Mortality (Deaths) N Crude Age -adjust edf LVH-t LVH+ LVH- LVH+ LVH- LVH+ All stratum I participants WM > HS (18) 446 WM «= HS (26) 402 WM < HS (42) 335 BM < HS (56) 344 Participants not receiving medication at baseline WM>HS '(13)347 WM «= HS (21) 315 WM < HS (22) 260 BM < HS (41) 268 ( D 8 4.0 12.5 4.5 10.2 ( 2)10 6.5 20.0 7.0 21.9 ( 7)16 12.5 43.8 10.6 28.6 (11)32 16.3 34.4 15.4 32.3 ne ( D 8 3.8 12.5 4.4 10.2 ( 0) 7 6.7 0.0 7.4 0.0 ( 5)12 8.5 41.7 7.2 12.4 ( 6) 18 15.3 33.3 14.9 33.2 WM = White males; BM * black males: HS «= high school education completed. •From TyToler HA: Race, education, and 5-year morulity in HDFP stratum I referred-care males. In Gross F, Strasser T, editors: Mild hypertension: Recent advances. New York. 19S3, Raven Press. ♦Age-adjusted rates by direct method. Standard population is age decade distribution of all white males aged 40 to 69 years with entry diastolic blood pressure of 90 to 104 mm Hg. JLVH determined by ECG. LVH+ denned as major LVH by Minnesota Code; LVH- denned as all others.. 333 TA3LES 6 & 7 COOPER et. al., p.o . 830-331 Table Vi. Mean values of cardiovascular risk factors in black and white males and females, by diagnostic category: CHA Detection Project' Risk factor Diagnostic category White males Black males White females Black females Systolic BP (mm Hg)t DM/H 144.6 147.9 137.0 146.9 Non-DM/H 138.5 141.8 131.8 134.3 Diastolic BP (mm Hg)t DM/H 84.9 89.9 81.1 88.8 Non-DM/H 81.6 84.6 77.4 80.3 Serum cholesterol (mg/dl) DM/H 210.0 211.9 215.3 209.7 Non-DM/H 205.7 201.9 208.5 205.6 No. of cigarettes smoked/day DM/H 11.6 10.6 8.0 9.2 Non-DM/H 9.5 8.9 7.1 5.7 Relative weight DM/H 1.25 1.30 1.20 1.32 Non-DM/H 1.22 1.22 1.18 1.26 BP « Blood pressure; DM ■ diabetes mellitus; H •Age standardized, ages 2? to 64 years. ♦ Excludes those treated for hypertension. hyperglycemia. Table VII. All-cause death rates (per 1000) in black and white males with diabetes or hyperglycemia vs those with normoglycemia: CHA Detection Project* Black/ White Black white Diagnostic group males males ratio Diabetes or hyperglycemia 96.0 (240)+ 95.7 ( 9) 1.00 Normoglycemia 52.3 (618) 62.7 (45) 1.20 Ratio of diabetes and 1.84 1.53 hyperglycemia to normoglycemia •Age standardized, ages 25 to 64 years. ♦ Number of deaths. 334 FIGS. 1 - 3: GLUECK et. al., p.p. 8l6t 818 •Or While 9 Black o* Black 0 6-17. n-358 6-17. n«326 6-17. n«1l9 6-17. n-124 32 <1 26 18 5 10 50 00 95 Birth Birth TH TH TH TH TH Fig. 1. Mean levels of HDL cholesterol (HDLC) in healthy, full-term, black and white neonates. (Data from Glueck et al.11) Percentile distribution of HDL cholesterol in black and white males and females ages 6 to 17 years: The Princeton School District Prevalence Study. (Data from Morrison et al.15) •s •5 35 »S 20-59. ft' 20-59 n- 134 ?C3 Bock o White o Black O*. 20-59. n»76 While O*. 20-59. n«306 6 10 25 60 75 90 95 TM TH TH TH TH TM TH Fig. 2. HDL cholesterol distribution in black and white males and females, ages 20 to 59 years: The Cincinnati LRC's Princeton School District Prevalence Study. (Data from Morrison et al.17) 100 90 BO 70 60 so 40 30 20 10 0 N-BO M'44 N*B6 N-56 N-673 N-30 nnn i ES0 EHldlci. holci COlDICI. HOLCJ E23lDLC|. HOlCt.#Blacka more likely to have ILDLC1. HOLC) predominant N-433 N-30 f*7^ accounting for their hjpei Cho'e»t«'oiemia than whites lor both Children and aduiti. p • .02 JESSE r*xr WHITE BLACK MALES 6-19 YRS WHITE BLACK FEMALES 6-19 YRS WHITE BLACK MALES 20-79 YRS WHITE BLACK FEMALES 20-79 YRS Fig 3. PercenUge of hypercholesterolemic subjects (total cholesterol on two consecutive visits ^ the 90th percentile) having LDL and/or HDL cholesterol primarily accounting for their hypercholesterolemia. These subjects were arbitiarily categorized by having LDL > the age-sex-race-specific 95th percentile and HDL > the age-sex-race-specific 95th percentile (at visit 2); LDL > the 95th percentile and HDL <95th percentile; LDL < the 95th percentile and HDL 2: th? 95th percentile; and LDL < the 95th percentile and HDL < the 95th percentile. (Data from Morrison et al.:') 335 FIG. A: GLUECK et. a).t p.819 Black Excess HDLC, mg/dl 0 2 4 6.8 n = 434 n=459 n = 174 n=188 "I---1---1---1---1---I---1---1 L • ry-'-TKfflrfrffltG.rls n = 434 n = 45 n = 174 n = 188 E^BaBBBSSatBoys Women gaS^EESSBBi t Men Covariance Adjusted Means Quetelet Calories White Black White Black 1.88 1.93 1939t 1642 1.90 1.95 2364 2156 2.39t 2.61 1746* 1464 2.65 2.72 2520 2450 Means covariance adjusted for age and education of the head of the household t, p<.01 'J tAgM** ^lif 'ifi%'4ttipi *; tGirls tBoys Women g^gT^BSatMen Means covariance adjusted for age and occupation of the head of the household t, p<.01 1.88 1.94 1920t 1669 1.90 1.95 2345 2186 2.39* 2.61 1746* 1451 2.65 2.71 2515 2440 Fig. 4. Covariance-adjusted mean levels of HDL cholesterol, Quetelet index, and calories per day in randomly recalled, blacks and whites. Means were covariance adjusted, separately for age and education of head of household and for age and occupation of head of household. Mean excess of HDL cholesterol (blacks above whites) is displayed. (Data from Khoury et al.1) 336 FIGS. 3-5: HEISS et. al., p.p. 809-810 -20 -15 -10 -5 0 5 mg/dl C-TOTAL (p-19) C-HDL 250 mg/dl Smokers Serum cholesterol <250 mg/dl Serum cholesterol >250 mg/dl Diastolic BF ^90 mm Hg Nonsmoker Serum cholesterol <250 mg/dl Serum cholesterol >250 mg/dl 6mokcrs Serum cholesterol <250 mg/dl Serum cholesterol >250 mg/dl Total No. of men Black 6.094 1.056 6.215 925 3.669 929 3.771 831 23.490 White 127.864 21.182 72.635 14,905 46.016 13,419 22.010 7,353 325,384 Age-adjusted rate (5 yr) Black 1.67 6.86 6.21 11.OS 2.88 4.92 6.42 9.04 4.83 White 2.40 6.12 5.62 10.78 3.66 9.88 11.06 17.49_ 5.22 Black/u-hiie ratio 0.70 1.12 1.10 1.03 0.75 0.50 0.58 0.52 0.93 339 FIG. 1: MYERS et. al. , p. 708 FIGURE 1. CONCEPTUAL MODEL OF THE ROLE OF SOCIAL STATUS IN CHD RISK, CHD MORBIDITY AND MORTALITY SOCIOSTRUCTURAL VARIABLES © SOCIAL STATUS © PERSONAL, SOCIOPSYCHOLOGICAL VARIABLES © • ETHNICITY • EDUCATION • INCOME • OCCUPATION • GENOER • AGE • NATIONALITY • SOCIAL MOBILITY • SOCIAL POSITION • SOCIO ENVIRONMENTAL FACTORS • BEHAVIORAL CHARACTERISTICS • PSYCHOLOGICAL TRAITS* STATES MORTALITY ® TABLES: 6 & 7: ftQV'UND & FULWOOD, p.p. 776-777 Table VI. Mean systolic and diastolic blood pressure and mean serum cholesterol level of adults 25 to 74 years old b> race, sex, and survey period, showing standafd errors and age-adjusted means: United States, 1971-1975 and 1976-1980't 1971-1975 1976-1980 1971-197 1976-1 no Age-adjusted meant Age-adjusted meanf Risk factor and race-sex group Mean SE Mean SE SE SE Systolic blood pressure (mm Hg) White males 132.8 0.5 130.5 0.6 132.8 0.5 130.6 0.7 White females 130.1 0.5 125.6 0.7 129.4 0.4 125.1 0.7 Black males 140.0 1.6 133.2 1.1 140.4 1.7 133.7 1.1 Black females 137.8 1.4 130.6 1.2 138.6 1.4 131.7 1.1 Diastolic blood pressure (mm Hg) White males 83.7 0.3 82.3 0.5 83.6 0.3 82.2 0.5 White females 80.4 0.4 78.1 0.5 80.1 0.3 78.0 0.5 Black males 89.9 1.0 85.3 0.8 89.8 1.0 85.5 0.8 Black females 87.0 0.9 82.3 0.6 87.1 0.8 82.6 0.6 Serum cholesterol level (mg/dl) White males 218.0 1.1 216.6 1.2 217.5 1.0 216.7 1.1 White females 222.8 1.1 220.7 1.2 221.4 1.0 219.9 1.2 Black males 225.5 4.5 215.0 3.3 225.7 4.9 215.4 3.0 Black females 219.4 2.9 217.1 3.2 221.2 2.6 219.2 2.7 SE = Standard error of the mean. •Data from the National Health and Nutrition Examination Survey. Division of Health Examination Statistics, National Center for Health Sutistics, HyatLsville. Md. tApe adjusted by direct method to the total U.S. population as estimated at the midpoint of the 1976-1980 NHANES. fable VII. Age-adjusted percent distribution among risk factor groups with standard error of age-adjusted percent for selected race-sex groups ages 25 to 74 years by survey period: United States, 1971-1975 and 1976-1980*______________ Risk factor groupst None} One§ Two or more \\ 1971-1975 1976-1980 1971 ■ 1975 1976-1980 1971-1975 1976-1980 Racc-scx group % SE % SE % SE % SE % SE % SE White males White females Black males Black females 390 45.4 23.4 28.0 1.3 1.1 3.2 2.4 42.5 47.0 29.3 41.4 1.1 0.9 2.0 2.3 46.2 41.9 44.3 49.2 1.3 1.0 3.9 2.5 44.5 42.9 50.9 43.7 1.1 0.8 2.3 2.0 14.8 12.7 32.3 22.8 0.8 07 4.1 1.8 13.0 0.6 10.1 0.5 19.8 2.1 14.9 1.2 cs, •L^om^he^tional Health and Nutrition Examination Survey. Division of Health Examination Statistics. National Center for Health Statisti rAyr.djuesied t direct method to the total U.S. population estimated at the midpoint of the 1976-1980 NHANES tlSten "of the following: systolic blood pressure of at least 100 mm Hg and/or d.aslol.c blood pressure of at least 95 mm Hg. serum cholesterol level of at least 260 mg/dl. of smoker. jOnly one of the above. rjVo or more of the above. 341 TABLE 5: COHMETT & STAMLER, d. 8^3 Ta&le V. Mean values at sixth annual visit and percent changes in risk factors from baseline, by study group and race SI group UC group White (N ^5338) Black (N - 416) White (N «= 5227) Black (N = 411) Sixth ann. Sixth ann. Sixth ann. Sixth ann. Risk factors visit % Change visit % Change visit % Change visit % Change DBP (mm Hg) 80.4 -11.4 81.3 -13.6 83.5 -7.9 85.6 -8.1 SBP (mm Hg) 121.4 -10.3 122.3 -11.7 126.5 -6.4 129.4 -5.8 Percent smokers* 33.3 -46.2 38.0 -43.0 43.8 -29.0 42.9 -22.5 Serum cholesterol (mg/dl) 235.8 -7.3 231.0 -6.0 240.5 -5.6 237.3 -3.7 Plasma cholesterol (mg/dl) 228.5 -6.4 224.4 -6.5 233.3 -4.5 230.2 -4.3 Plasma HDL cholesterol (mg/dl) 41.3 -2.6 46.0 -6.5 41.4 -2.6 47.6 -4.8 Plasma LDL cholesterol (mg/dl) 148.7 -8.6 148.3 -8.5 152.8 -6.4 153.8 -5.4 Plasma triglycerides (mg/d#l) 200.9 +1.7 153.9 +7.2 203.0 +3.3 147.1 +4.0 Weight (lb) 187.7 -0.7 192.3 -0.2 190.1 +0.7 190.1 +0.7 HDL - High-density lipoprotein; LDL ■ low-density lipoprotein. •Thioeyanau-adjusted rates, as defined by Neaton et al.' 342 Table 5: Langford et al, p. 800 Table V. Five-year incidence of angina, ECG MI, history of MI, and MI by RQ for SC participants in HDFP Angina ECG Ml Sa mple size Events Rate Sample size Events Rate White males 1799 89 4.9 1868 12 O.fi Black males 983 54 5.5 1056 15 1.4 White females 1078 103 9.6 1172 3 0.3 Black females 1211 79 6.5 1329 10 0.8 Total 5071 325 6.4 5425 40 0.8 343 Table 1: R. Sterling et al, p. 696 Table I. Operative mortality (1970-1982) Black male patients Other WRAMC patients Year No. /c No. % 1970-1977* 1978-1982 1970-1982 2/14 1/40 3/54 14 2.5 5.6 22/407 33/916 55/1323 5.4 3.6 4.2 WRAMC - Walter Reed Army Medical Center. 'Prior to use of cardioplegia. 344 Prevalence and Incidence of Ischemic Heart Disease in United States' Black and White Populations Melford J. Henderson, M.A., M.P.H. Research Fellow Association of Black Cardiologists Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland A major deficiency of epidemiologic research on ischemic heart disease (IHD) in the United States has been the limited amount of reliable information published on blacks (1). Published information on the prevalence and incidence of IHD in blacks is incomplete and, at least on the surface, inconsistent (1). Reliability of data on prevalence of IHD in blacks has been compromised by the inclusion of relatively small numbers of blacks in population-based studies and/or selection bias in the reported study groups. Varying diagnostic criteria and study methods have contributed to the apparent confusion. The purpose of this paper is first, to review current information on the prevalence and incidence of IHD in U.S. black and white populations; second, to document the inadequacy of available information; and third, to suggest implications of these data for the DHHS Task Force on Black and Minority Health. Definition of Ischemic Heart Disease Ischemic heart disease in this review includes all of the disorders of the heart that result from inadequate supply of oxygen to the myocardium. Manifestations are: * myocardial infarction (MI), or death of myocardial tissue resulting from significant reduction or interruption of coronary blood flow; * angina pectoris (AP), a varied symptom complex (most commonly including substernal chest discomfort) caused by transient ischemia that falls short of producing infarction; * arteriosclerotic heart disease, defined as obstruction of coronary vessels by atherosclerotic plaques leading to functional and structural changes in the myocardium with or without symptoms; * sudden death, defined as death within seconds to several minutes (for some studies, several hours) after the onset of symptoms. Gillum (2) emphasized that IHD is a major cause of illness in both U.S. black and white populations, and is the leading cause of death. National data derived from death certificates indicate that it is the cause of death in 30-40% of all adult blacks and whites (2). Death from IHD occurred at a lower rate in black men than white men in 1940, increased until it exceeded the white death rate in 1968, and then declined to a level comparable to that in white men in 1976, but still exceeded the 1940 rate (2). Black women appear to have had IHD mortality rates similar to or higher than those of white women (2), but comparisons of morbidity and mortality are difficult because absolute numbers of cases of IHD in studies involving black women are small (2). The 1960-62 National Health Examination Survey (NHES) conducted by the National Center for Health Statistics (NCHS) indicated little difference between U.S. blacks and whites in prevalence of IHD (3). In a random population sample, black and white persons were equally likely to have arteriosclerotic heart disease, but the sampling error was too large for meaningful, detailed comparison. Electrocardiographic (ECG) data from two national probability samples (NHANES 1 and 2) are currently being analyzed at NCHS. The data consist of ECGs from nearly 17,000 subjects (age range, 347 25-74 years) with a total of about 2000 blacks in the two studies, combined (Tables 1 and 2). These data sets are deceptively large. The number of black men and women in individual age groups is small, as is the percentage with MI-associated ECG changes. For example, the percent with Ml-associated Q-wave Minnesota codes (1-1-1 to 1-1-7)(4) in the white men ranged from less than 2% in those age 44-55 years to about 3% in those age 65-74 years. For black men, the percentages with such codes were similar to those of whites in younger men but substantially lower (about 1/2 the white male rate) in those over age 65 years. In contrast to those for black men, the prevalence estimates for white men are relatively stable. For example, the addition of two "cases" in the 55-64 year age group would double the prevalence of Ml-associated ECG findings in black men. The small numbers of blacks in the sample and low number of such ECG findings in black and white men in the general population do not allow comparisons of prevalence. Differences in ECG criteria for Ml-associated changes has added to the confusion in various studies. The use of a standardized code (e.g., the Minnesota code) contributes to the comparability of various epidemiologic studies. However, even this has not resolved all criteria-related issues. For example, use of the Minnesota coding system for major Q wave findings (e.g., codes 1-1-1 to 1-1-7) allows comparison of these studies with others. However, it is possible that this would leave undetected a group of individuals who have clinically accepted ECG evidence of MI but do not quite^meet the criteria for these codes. The addition of other Minnesota codes (e.g., 1-2-1 to 1-2-5, 1-2-7, 1-2-8 and various ST and T wave codes) increases sensitivity but at the expense of specificity. If these various ECG manifestations have the same meaning in blacks and whites, these technical considerations would not necessarily invalidate comparisons. However, there is some evidence that the meaning of at least some of the ECG markers (e.g., some ST-T changes) may be different in blacks and whites (4a). Some preliminary comparisons can be made between blacks and whites in the NHANES studies despite the indicated limitations of the data. The prevalence of Ml-associated ECG changes is similar in 45 to 64 year old black and white men but greater in older white men than in black men. This might be related to the 5 to 7 year shorter life span of black men (average life span of about 65 years) such that the group that lives to 65 includes a larger proportion of those who have not succumbed to IHD at an earlier age. Differential case-fatality (discussed below) could also be involved. The prevalence of IHD in black women was greater than or equal to that of white women (0% in the young to about 1% over age 64) with no cross-over. The 1972 Health Interview Survey (5) indicated higher age-specific self-reported prevalence of heart disease in nonwhites (predominantly blacks) than in whites in both age groups 17-44 and 45-64 years. However, nonwhites reported the diagnosis of coronary heart disease less often than whites in ages 45-64 years (15.5 vs. 36.8 per 1000). In some studies, prevalence of IHD was estimated from surveys or baseline examinations of population-based cohorts. In 1960, the Evans County, Georgia study (6) surveyed nearly all residents in this rural county 40-74 years old and 50% of those aged 15-39 years; this resulted in 3102 participants. Nearly three times as many white men as black men 348 reported syndromes of IHD (Table 3A), with fewer black men reporting IHD at all ages. When black and white women were compared, there was no significant racial difference in prevalence (14 vs. 18 per 1000). In the Charleston, SC study, 2275 blacks and whites were examined in 1960/61 (7). Age-adjusted IHD rates were computed indirectly using standard age-specific rates for the total study sample. In the cohort intake of 1960/61, white men of all ages had greater IHD prevalence than black men (Table 3B). Prevalence in black women was identical to that in white women (26.6 per 1000). Black women aged 75 years and over had a greater IHD prevalence than black men (though not statistically significant). Thus, the Charleston and Evans County studies reported similar findings on the order of prevalence of IHD (Tables 3A and 3B). White men were more likely to have IHD than black men. White and black women had almost identical prevalence rates. In the Hypertension Detection and Follow-Up Program (HDFP)(8), hypertensive participants were evaluated by ECG for evidence of MI and were administered the Rose Questionnaire to detect symptoms of AP. AP was found to be more prevalent in black men than white men (7.7 vs. 5.1 per 100) and more prevalent in black women than white women (9.9 vs. 8.6 per 100). Compared to those without AP, those subjects with AP had twice the likelihood of death during the ensuing five years. This applied to all race-sex groups except black women. The prevalence of MI at baseline as determined by the Rose Questionnaire was higher in whites (5.9 per 100) than blacks (4.6 per 100), and higher in white women (7.2 per 100) than black women (4.2 per 100). Incidence of IHD in Black and White Populations The study of incidence requires thorough assessment of the population at risk, thorough tabulation of cases, and recognition of all indicated age and sex differences. Completed population studies yield too few incidence cases of acute myocardial infarction (AMI) and do not permit adequate assessment of comparative incidence among men and women in black and white populations. Prior to 1970, studies that compared IHD incidence in blacks and whites lacked population-based age-specific data. They usually involved only one hospital or group of hospitals. For example, Mihaley and Whiteman (9) studied AMI at Harlem Hospital in 1950-54, but during that time only 152 such cases were admitted. Whites, accounting for 5.8% of admissions, constituted 13.1% of infarctions. The investigators maintained that AMI was as common in blacks as in whites, but they did not have adequate data to support their conclusion. Data from NCHS Hospital Discharge Surveys in 1972, 1975, 1978, and 1981 (10) indicate that men and women over age 45 maintained nearly constant hospital discharge rates for AMI. During 1981, rates for black men were about 45% of total white rates, while rates for black women were about 70% of white rates (ages 45-64). However, these data may be not reflect true incidence because they do not take into account subjects who died suddenly before hospital admission. Few major studies involving the epidemiology of IHD yield data on racial incidence. Ongoing prospective studies, such as those conducted in Framingham, MA (11) and by the Hospital Insurance Plan of New York (HIP) 349 (12), involve predominantly white samples. The HIP study of insured persons found the annual age-adjusted rate of incidence of first MI in nonwhite (predominantly black) men to be half the rate in white men and the case-fatality rate to be higher in nonwhites (47.5%) than in whites (35.3%). However, the total number of nonwhite cases was too small to yield statistically significant conclusions. Additional studies of interest are those carried out in Nashville, TN (14); Baltimore, MD (15); Newark, NJ (16); and Columbia (17,17a) and rural Pee Dee (17a), SC. These studies reported incidence of hospitalization for IHD in blacks and whites by age and sex (Table 7). In the Nashville study (14), 1967-68, hospital surveillance was used to detect all cases of AMI and sudden coronary death in persons ages 35-74 years. Although only 20% of the population group was black, the total number of subjects was large (167,000). Annual incidence of MI was substantially lower for black men than for white men at each age (Table 4). White men had more than double the rate of black men. Annual incidence of AMI was also greater in white women than in black women (Table 4). In this study, black men had greater incidence than black women. In the 1970-72 Baltimore study (15), the number of nonfatal Q wave ("transmural AMIs") was determined from a review of hospital records in the Baltimore area. One-third of the 500,000 population sample was black. The incidence of "transmural" AMI in white men ages 25-64 years was more than twice the incidence in black men; but white women had only slightly greater incidence than black women. Black and white women had similar nonfatal MI rates, and in Nashville and Baltimore the occurrence of nonfatal MI was twice as frequent as sudden death. The Newark study (16) involved review of 89.5% of all 1973 hospital admissions (15,124) in that city. Age-decade specific MI hospitalization rates were reported for seven age groups (Table 5) . When analyzed by decade, data revealed no statistically significant difference in incidence between white men and black men, but the incidence was slightly higher for white men at every decade except 40-49. AMI rates for black women were slightly higher than for white women at every decade, but these differences were not statistically significant. The Columbia, S.C. study (17) in 1978 examined age-adjusted incidence of AMI. The population included patients with prior heart disease and MI, as well as those whose sudden death was associated with MI. Both proven and suspected MI incidence were included, and in each age group over age 35 years, white men had higher incidence of MI than did black men. Black and white women under age 65 years showed no consistent difference in incidence, but black women ages 66-75 years had somewhat lower incidence. Overall frequency was greatest in white men and decreased in the following order: black men, white women, black women. These studies differ in various ways making a firm conclusion regarding the relative incidence of IHD in blacks and whites impossible. Inadequacy of the available incidence data is related to several problems including: * the limited number of studies that have been carried out; * the inadequate numbers of blacks in the populations studied; * the use of varying standards for age-adjustment; * the use of varying diagnostic criteria; * varying lengths of follow-up. 350 09 13 ht po O P 13 rt cr rt po CT rt O [fl 13 tJ CT P- 0 o o p rt H po Cfl 09 rt p «. p. P- H O CD Cfl Cfl 13 P- o p Cfl CD M CD Cfl P h-' rt w Hi H 0 3 rt CT CD < po p Cfl o o p p rt Hi Cfl P l-h P- < CD rt p- 3 CD Cfl rt PJ CD H PO rt CD P- P CL ph rt £ O P CT 0 P CD CT rt p h-> p- H CD CD H rt n po po P CT p4 rt CL Cfl CD po CD H Cfl < p- po H h-> po CD cr p- rt !?d CD H- P P Cfl S rt p4 P CD P- CL P rt v< o O £ Hi CD M P- P H- O P P- n o o CL P- P- rt CD in 0 po 3 P CD CL CL V P- P' Hi po Hi O CD H CD P O 0 3 CD CD CL CL Cfl P O Cfl CD /O P CD Cfl rt P- O P P po P- H CD CL Hi po CD cr £ P-1 CD O P- l-h W rt H P- 13 5 O 3 3 13 O CC o a rt O P- P H CD 13 0 H rt CD p» P- 3 o p o OOP O P- rt CL CT < CD Cfl P 13 • O CD CD H P» P- p> o o • Hi CL > £ t3 SIC CD PH H H CD CT p>*< a O CD o m rt o n cd ^ Q H 3 0 H- H P £ CD CD X hj po CD 3 P- H P CD CD CD a CD rt CD CD CL CD P V CD rt > Cfl O 3 rt rt po P crvj CL CD << H P Cfl CT 0 £ CD CL rt P" po rt > CD po CT H P H po po P CD CT CT X po h-> £ po a cd tr 3 p. 13 rt rt rt p^ p* O CD CD CD rt 3 ? P* (D H' P- CD P P CL Cfl O 00 p. pf p- po W P CD P O P- CL £ CD O P 3 O CD CD P Cfl cr rt p < CL PO P Cfl p p 3 cr CD po H H Cfl CD O Cfl Hi P- PO CL OQ p • P rt P- O i-h 13 H p. w p3 n *< cd po P P- Cfl rt P rt P a s a p. p-v< Hi P- Hi rt p. CD CD P p, Cl CD rt p- P P* O O po po CD P rt 7? O P CX CD P- H Hi - Hi CD W H P CD 09 P OP rt CD Cfl rt 3 CL CD CD P P - O CD po P O CL Hi P PO > P 3 - £ o Cfl rt o o P P O rt Hi P- *< P - CC o cr a p* h-1 po po h o CD Cfl 3 rr rt cd p. o P rt p CD - PO Cfl P CL CD O P Po Cfl P- CD 3 Cfl rt £ CT CD CD M CD n CT Cfl po 3 H po h-' pJ CD P-1 Cfl • rt O P po vO Oi CD O CD P Po P p- a VD ON P» CTv po N) P • CL l_i hrj VD H o o VD 3 po rt P P4 CL P- Cfl H- P O O H P- P- CHX3 CD P- P P O Po CD pJ p^ H- rt CD 3 CD P po P CL £ 0 3 CD P P' P- 09 P" CD Cfl rt Cfl rt P4 P4 O CD P cr po rt Cfl po O P O P- a CD p o n cd o P P rt < po P Cfl o o p' o H VI rt Cfl O rt Hi P CL CTV5 P' po < CD CD P Cfl P P- rt p P'OO CD po cr pj p- Hi 09 rt p" Cfl O O cr p* 4> p* p- i PO OQ V- O P4 CD ?? CD P0 Cfl H 3 rt 5 ?: CD P" H H- CD rt CD CD Cfl P • H O *5 O O 3 Hi CD Hi O PO Pi rt po Cfl o CD P- O O CD Hi O O rt P P4 O n cr (D H ? po p- P O 09 O JT p- rt CD 3 H Hi CD O P Po P 09 P /-s CD CL LO I OJ PO cr • a CD Ox «-i- rt p £ 13 w CD CD rt CD H CD p a H O po O rt p- CD O Cfl H rt O - Hi PH Cn a o Hi 13 O CD H H Cfl rt O P' P CD Cfl OJ P» po O P N> CL 13 po CD H 3 Cfl p o o p p- Cfl PJ 5 po po H Cfl 09 CD P- H P P- O rt O H- P4 po O P1 H p« rt P < O h-' < CD CL Po Cfl Cfl CD Cfl Cfl 3 CD P rt po P CL Hi O po P CL P O p Hi po rt po CD O 0 3 3 P P p- rt > H I-H09 • H O P 13 cfl P' p- rt CD Cfl C/l po m 3 KA 13 w p» • CD O H hi ^r CD cr p- PJ H po o > 7? S PH 3 po po p- p CD CL Cfl PH W a P1 H po po O rt 7? CD Cfl O P CD rr P p CL H CD CL P0 P CL 0 p CD 13 CD H P- O CL I CD CJ\ Cfl N> rt O P 00 P o p- CL CD P o CD Hi O H P' H- rt CD 3 CD P CD X o CD CD CL CD CL po p CL rt P" CD P- P O P- CL CD P o CD o cr Hi PJ PO n o w ?r a ca ?: o o O £T o H rt Cfl rt P CL P- CD Cfl po Cfl Cfl CD Cfl Cfl CD a p- p o tO CL N> CD -J P Ln O CD a p- p PO £ s Cfl P' 0 Cfl p- 3 CD rt CD Cfl CD P Cfl Cfl CD CL P- P rt cr CD PO o ?r Cfl O po Pi Cfl p o H- a CD P O CD O Hi w C3 H- P n o p4 o H rt C/l rt P CL P- CD Cfl P P CL CD H PO 09 CD CJi rt /- O P rt CD P4 • CD " < 3 pO PH H H- £ po cr rt p- P- O o cr P rt w sr p CD 09 P- 09 H CD Cfl Hi rt P- CD P CL CL p- cr p *< 09 Cfl 3 CD CL P O H CD 09 PO PO H P' CL h- cr P p- 09 Cfl rt > O S M PH P- CD P- Cfl P ^ O CL P Po h-- rt i< po rt po O Cfl Hi o P- H p p- CL rt CD CL H CD P- Hi PO P- P Hi P- O rt h CD a 3 P- PH pO -• 09 P rt O cr w CD p- 09 CD 3 > 13 2 P^ PH O - CD rt cl cr CD PO P n O O rt \-> cr p CD 3 h cr p- O Po PO rt Cfl cr p- CD 09 O po H o cr 7? P- Cfl < Cfl 3 . po cr o p- o . rt P pj H vj - P Cfl = CD 13 CL O cfl rt ca ?r P- o cr w CD CD O 2 H PH P- Z rt CD H P- Po Cfl rt P CL P- CD Cfl P Cfl CD CL W n Q po p a CD p N V! 3 CD po CD Cfl X O 3 > po rt P rt P- Po Hi O cd ?r Cfl rt H H O P- Po n p- cr o P po o < P- CD CL cd cr P CD O CD CD P - Cfl £ p- p CD po 0 Cfl < PO p CD Pi rt 0 Cfl rt H Hi CD po rt M P P P4 Cfl 0 po S 3 p P HH P 1 H o s; cfl PO cr p P1 £ P- 09 3 po rt 09 < CD CD S CD Cfl Cfl H cr cr CD p rt w n rt cd vr CD p- co Z CD P po O S ?r P Cfl rt rt O P CL cr^ CD a P- CD P Hi PO P- CL P CD CD ►o a p PO = rt rt CD H po P- P P Cfl 3 3 P po M ?r po p- p^ P : 09 H cr cr p- CD po o Z "V PO Cfl Cfl cr po < H P- CD 3 13 PO H P' Cfl O p O rt o hi cr hh CD cr ph cl p^ p- PO C Hi O P- Hi ?r rt p. cr o po o P P P H- CL rt rt vi 5 CL P4 P- P- p. po p rt 09 cd p a O P- Cfl CA po P rt 09 CT P- P (_.. o o CD Cfl O O P- rt P cfl s: 09 PO < > CD Cfl CL rt O Po P cr cr CD po Cfl 09 rt CD H rt H CD O rt < po P4 po rt po CD rt CD p- H DQ hO HOP P- P P- P O p- a CL po CD C po CD Cfl rt H Hi P- CD H CL H P O CD P 3 P Hi 3 O H O* CD CD CD CD t-' £> H Cfl O P CD hh CD O S P hh cr > O CD 2 O CD O Cfl P- Hi 13 O P p- p CD P H O - P po 13 rt p pi i rt O Po rt Hi CL CD i 3 H cr p- p O Cfl Cfl Cfl Cfl 13 P- O cr p- rt CD 3 CD P CD rt P PO O CD Hi H cr o CD 3 rt £ rt cd cr CD CD p z s: CD cr £ P- po rt H CD 7? po po P P cl a cr w h-» po po P-" O rt ?r p- PH p. O Hi C h-' Cfl CD O 3 po CD P 5 cl cr PO W rt po rt CD p. Cfl 3 en O H h-> CD P- ?r Cfl CD rt p-- P *< CL P- rt CD O Cfl p Cfl p CD O cl i cr O h-- p. po po P h-> O h-> 7? rt CD Cfl cr cl CD CL = p W P CD CD O rt P rt po rt O H W po po (-• P O Cfl 09 O 3 H 7? P CD CD H po H Po rt h-' CD CC = H CD Po S H l-H rt • rt p po Cfl po h-> O po O CL 3 CD CD O Cfl po P Cfl rt 09 cr ca cr o rt po P Cfl • rt P a H p- cr cd CD Cfl Cfl CD po cr O O 3 > JC Cfl CD P 13 P O 3 P- • rt CD rt P* P po CD P- H 3 po P- P o a o P- a o P Po h-> rt po CD rt CC o O £ Hi V- po CD PO H < o cr cd CD rt po H O < Cfl u H CD p 09 3 CT09 Po CD CD V- CD Cfl P rt O P- P rt rt Cfl p rt CD P CL CL P- P- CD Hi Cfl hh CD Hi H O CD P P P O CL CD L/l to z CD z Vi 0 £ 0 rt p- po p- Ln z CL fV £ Cfl po •s o CL P■ < cr cr ^J cr /-~> o Hi 0 H 3 Cfl cr hH O Co 3 m tn o 0 H p X po CD H cr P« O p Cfl P Ln CD p- P O vs 09 cr 0 CD CJ\ Cfl O p1 o p- po 0 O PO cr PH rt vj P < CL hH < < cr Pi po 0 po Cfl po p- P- 0 < ^ £ Cfl 0 3 3 CD p- 3 rt S^ • Cfl po 09 09 p p Cfl p- P 1 CD CD po Co CA rt p 3 cr H O rt P po p" CL CD PO CD •s CD 13 rt 13 CD 13 O vi P" CD P1 p CD CA PO CL 0 P4 H XI PO P P CD 0 CD rt P- < Cfl rt CD P- H p CD Cfl cr H H PO >—> P rt ?: CD H 0 N> p. 7? CD CD Cfl CL CL 1 P VJ 0 0 CD P P Cfl Cfl H PO CA H P p- H u cr 0 rt CL PO rt p> 7? cr Cfl CD O PO Cfl P- rt rO rt Po H > Hi Hi CL 3 I-H CA Cfl PO CL < CA po po p-" cr CD 13 po CD CD ti- P- PO CD CD a Cfl 3 Cfl CA > CD fr« Co € H M cr Ln Co P- >~*\ 13 P 13 rt rt n n PO CD P- P rt pj CD 13 H O h-1 H CL erce O Cfl 09 Co Cfl /"> CD Hi CL/C ^^ h-1 O Cfl rt O CD P1 H rt P £ P" Co P- p- CD > o 0 rt CA P Cfl p- CD CD 13 P- po 7? po z O CD 5 H Ui CL rt ^ 3 cr 1 ^J CD PO a CD Ln H P- rt 3 h-' *n p p p- 3 0 rt H p- P o O Cfl 0 rt 3 po P x 0 CD Ln CD cr £ 3 CD po OO Co < i—' p- H ^-/ p- P Co Co vi p p 0 Pi cr P P 09 £ Cfl CL pj H Cfl P o O O 13 Cfl rt Po Hi 13 ^5 Ln H vi po cr o P P o rt P- p- P- p CD P o pj rt cr rt rt P 0 p-1 H >• H 0 0 CD po P- 0 CL 0 H 0 p1 cr O po O ^ M < P H - C ^> rt P CD rt 09 Hi 09 p- CD < vi Vi v; H Co PO 0 p p rt ti pj CL p Cfl 3 CD < O rt 13 P- H &5 CD W CD rt rt P PO H CL •5 vj CL 0 a < Cfl Co C o p' 0 p 09 Cfl rt p- CD CD CD H 13 p- 3 rt cr 3 rt rt X n Co cr CL 3 Cfl P- P- CA P- p 13 CD Cfl h-' 3 Cfl Cfl po € j»r - H 13 H p- O H h) P H CD po 5 pj CD p- v; 13 CD cr O O n /~\ 5 pi P CD O cr cr < O po P Cfl p po p • 0 P- CD rt rt CA 0 0 H Hi P P- O p- < H PO h-> • rt Cfl CD CL PO Hi h-1 z cr p. rt H P 0 p P- Hi 3 09 o CL rt 0 hh CL Po P P 3 3 3 m P P CD P- po rt CL p po CD Cfl CD P p- P P P s: rt p- rt 09 z rt CL O 09 p^ P- CD o PO P a a P1 CD CD PH o Cfl 13 Cfl PO P- CD 09 P' CL CL o CL CA Cfl CD H rt v- O PO CD CD P CL P CD Co rt CD p CL rt CA vi Vi O P P Q CD Cfl rt cr CD O po P po po CD p- P- a4 P P- H Po CD < 09 £ CL CL PO CD 09 Cfl H cr P CA rt P- CD ■* • M s CL H P1 H Cfl rt £ po rt P1 p Cfl po P ■c p' CL 0 CD P Hi CD CD cr p' po rt 09 PO rt rt i Cfl CL 0 € £ CD po P- CD cr rt cr cr P- rt O 3 a « P 09 cr po CL p Cfl C H H p- rt rt Cfl rt cr Cfl CD rt Cfl P O CD hh PO 3 CD CD H 0 P' O CD po CD Cfl rt CD O p- 0 CD Hi 3 O 0 ^J rt cr O cr p• CL CL Hi Co P- p PH 0 H H CD I-H P Cfl rt O cr po cr CL NO Cfl z po 13 s rt 7? P 0 0 P 3 3 CO O CD Co CL po H cr C P- 1 Cfl P rt CL P P CD CD P CD 0 CD po z p' H hh p- O rt CD p H M CD Cfl Hi p H CD 09 Cfl rt pi CD 0 O CD po pj P' CD cr CD cr o rt CL CL p CD p- rt H P a ^s € CL P- CL p po P > CD VJ CC O CD 7? P' h-1 po O CA 0 cr CD rt p- cr rt Co P- rt 0 P Cfl £ Cfl 0 o CD po O cr 3 /-•> CD CL CL p' 3 CD p- rt C3 P" H Po P- o * Cfl Hi p" cr CL Cfl 0 h-1 PO cr H Cfl rt CD po cr a 3 N3 > p- P- H cr H vi CD ■P- Po P- z cr i-h OO PO P cr CO rt 3 rt < 7? Z 13 H po w CD CD O 3 09 CD CD O *.. H CD p- 4> p 09 7? CD P Ln rt PO P- S—' PO O H CD /—\ H P- CD P- CD CD 3 P- CD O o P Hi 0 P 3 CL 7? CL Cfl P cr cr cr rt Cfl ^S CT09 P CL K Cfl po (-• CD Hi P < 3 H O rt XI 7? CA n rt o CA Cfl 0 z O cr ti- CD P crv- z rt Cfl PO O CD £ *w< p PO po cr u P1 Hi ON CL P- CD Hi CD CD H po P Cfl CD cr H 3 rt Cfl CD H CL 0 er 09 P- PO 7? 0 po rt < CD CD P- O CD p rt ca p' 3 H \s 0 P1 H P rt h-1 CD po O PO CD t CD P- 3 CD H P Cfl P p a Cfl rt P p- CL 7? hj ?: P Cfl ►o CL S po rt O CD • z po VJ 0 rt po P JO H po ti- >—• CL CL PO Hi CD O CD O H CD cr P- CL 0 CD P- CD CD CD P p- cr o cr Cfl X! CL rt 3 cr H p1 3 O c I—1 Cfl er h-1 0 H 0 Hi z P rt u £ Cfl PO < 3 P- H po P H rt CL PO P O P- 7? CD rt P CD CD cr po CD P- 0 v; H CD CD CD CD P 7? Hi CD po 13 cr P rt p- CD CD rt 09 3 CD cr P1 0 CD 09 CD > Po p- rt CL Co rt H cr CA Cfl CA Hi cr H cr Cfl CD s O cr h-' Cfl Cfl 3 cr O CD O p cr •5 p- CD P P CL CL 09 P' rt ^ rt rt p- rt rt n 0 (—• C3 Cfl rt CD 0 cr CD cr o Hi P1 po 3 H C3 Co p rt CD O P < O PO po cr CD rt CD po cr CA P- 0 0 cr H CD CD rt cr P Cfl < X CC p- po 0 CD po CL vi CD CD O Z Cfl p o cr h 3 o CD vi 1—' O rt £ CL CD pj Co O P Hi PO Cfl P CD rt X p- 0 PH p■ P 0 CD • Cfl P- CL > po CD 3 CD CD p- H VJ P po 13 cr p" Cfl P- p XI P H 3 rt 13 a p- P' p' TJ CD Cfl p H Hi po H P- PO Cfl Cfl cr •3 Hi P CL Vj O Cfl H 0 0 Hi rt rt p P Cfl 3 P1 CD cr p- ^ Hi O rt h-' p CD Cfl rt P1 Hi o cr O 13 /'~N cr CD Co 3 O CD Hi p- PO Hi CD Cfl CD 0 p- vi 0 CD po rt hH CD Co Co rt o H po CD CL Cfl rt Cfl CD P- > PO CD O CD po rt X < O M < 0 H P^ P po CA H H CT13 H 3 0 CA P- P CA P rt CD h-' P- h-• CD rt cr rt Hi P1 H H 0 Cfl P1 A p- Hi CA CD H PO 0 09 a H O O p- po po w p rt H a *£ rt P- cr P-1 rt u Cfl CL P CD po p- rt 09 CD p CD ti- O p1 rt cr CL CD CD o a4 P- 3 Cfl rt Cfl > rt Co Cfl p- CD v: CD O P O rt O 0 CD rt CL Cfl O O cr CD CL p 0 er • i—1 s P- r—> po \-> >Vj < 13 po 3 CD rt CD O - H P 09 Hi cr rt K H 0 P CLV- z Hi po 0 rt po po po O CD PI P Z (_i po PO £ 09 po rt p- p' 3 PH Cfl CD H P CL CD Hi CD cr p- P- H CL CD po •5 rt P po 3 13 H P Ln >• P P- O po H po H P1 cr rt po CA < P po CD CA P £ Cfl p po Vi PO ^-s Cfl CD Cfl CD 09 Cfl po Hi 13 CD ^w' Cfl a4 0 rt 7? 3 Cfl 0 3 po z CD po h-' P P- cr PO po rt P- rt t—■ cr CD 09 O rt P1 cr P1 Vi cr H O Co y cr X rt Cfl p• p v: P p O o p1 CD VJ P po rt P- rt CD CL P p- rt" rt rt CD cr N> < P1 3 po rt 7? H H h-' P P po 3 O cri3 P CD /^\ z Cfl < P- H CD H CD a O rt CL H 0 cr X Vw/ p- 13 P a4 cr CD CD PO PO cr a cr O rt O CD cr p- 09 >• C3 CD CL Cfl Cfl cr CD < CD O 3 P vi O CD p- po CD O CL pj PO rt CL 0 PO CL H P P O P vi (_>. 7? H Hi P 09 CD O *£ CD p 0 CD cr p- rt u 3 Hi P po CL H h-' P Hi p1 Cfl O rt rt h-' po (t CL K rt CD po cr cr X cr Cfl > po CO 0 CL Cfl cr P- P P P- 13 3 Hi Cfl p p- P CD 0 CD 3 CL cr V! O /—\ 0 3 P cr rt p- CD o p1 CD ^ hi rt Hi CL13 CD W CA p- CD Hi PO po CD CD 3 rt po P- CL H CL Cfl CD O P- CL 3 3 P p« rt CD Cfl p> cro9 H CD po X ?r cr rt CD p- CD n X! a P- p H P H p 0 09 P CD rt 09 CD V. Hi 09 Hi p- po P- O CD O H> CA P 3 po po cr 13 O " P- CD P rt p Q 3 P CD CL P- » CD H 13 P Cfl CL O po H PO P- H P P1 Hi H P ^ P O P CD Hi rt 7? P CA H P po O CD p- CL Cfl P po Cfl 0 rt P O rt rt P 0 Z Hi CD CA • CL PO H 7? H 0 Cfl po P PO CL P1 0 CD H p OO O O" o p* vi Cfl p- T3 rt P P4 cr CD p po 3 CL CD rt rt CD p Co £ H Hi P •s H rt rt CD CA cr Cfl CD rt o cr >vj P O CD CD rt CD H O 13 Cfl CD CL13 Cfl PO P- H *2 O H 0 CL P" £ P4 H O £ a cr < CD P" PO ■ rt rt cr p- vj Cfl € Cfl p hf po < Cfl p' P Hi CD cr po 2 P1 z O P- po a4 H CO O CD CA po rt Co P- P< Co a 3 . CD po 0 > rt Cfl P CD P po z Cfl < vi Hi P p- H rt 0 Co cr w 3 rt rt p■ Cfl po 3 P- P cr p- 3 P- rt CD w 0 < P PJ s p- O 3 H Cfl rt CD P 0 p- CD 0 rt CD P- H H p• CD CD CD po 3 09 h-" PO Cfl H P Co 7? p n P CD rt CL p. h-1 O P po JC CD rt s CL Hi h-' £ H CD CD Cfl Hi CD CD PO P- P • O CD Cfl CD rt P1 p £ cr CD rt CD cr rt o rt H H cr p P Cfl cr CL H po CL P' po CD Cfl 13 P- CL O P 3 7? P rt cr PO CL cr po CD CL CD pJ Co cr cr CD p. Cfl CD P- H CD CD VJ P rt O 0 O 7? Hi s-\ CD H CL CD P p- Cfl CA vi Cfl O p- CA CD rt 0 O O h-J rt a4 7? P » Cfl 0 p- cr P po Hi a PO ■f> P H 3 O P- CD o rt w P1 • po p cr CL po H < po 3 cr P CD p po CL rt o £ H O cr CD < 7? O CL CD CD p po Co rt p CA Hi rt CD po pj cr CD P p- CD rt 09 CL PO Cfl p P- rt > O O ^5 /—s CD P CD P- PO CD 0 (—> H O 0 CD rt 0 P- rt CD H O p po H CD P cr rt O cr 3 rt LO H P CL po Cfl 3 p rt 09 rt Cfl cr H 3 p po CD 09 cr O PO CL H P- CD cr O P CD CD Cfl < M rt rt 09 CD rt rt CD Hi P- CD vi CD CD p ^J H p1 pj P rt cr p PO p' P Cfl $~5 O cr Po P' cr P P Co 3 P' p- cr 0 rt 3 H> vi CD 13 cr PO Cfl CD p• rt cr PO O 7? cr v- rt po h-1 po P 09 CD ^ rt Co P1 O H CD vi CD h-1 CL CD CD P Hi cr (D 0 Hi cr CD H from the studies do suggest that the lower number of hospital admissions of black men for AMI may at least partially, be explained by the greater number of out-of-hospital deaths. Incidence of Sudden Death Sudden death comprises a large segment of total IHD death in U.S. black and white populations; however, as with other IHD end points, few reliable population-based data are available with regard to blacks. It is generally acknowledged that sudden death results from some type of arrhythmia--either asystole or ventricular fibrillation--and may have different underlying pathophysiological mechanisms from those of demonstrable AMI. The Charleston study (7) revealed sudden death in black men (32.2 per 1000) to be three times the rate in white men (10.2 per 1000). The rate in black women (13 per 1000), was 1.5 times the rate in white women (8.2 per 1000). In the 1967-68 Nashville study (14), investigators identified likely cases of sudden death from death certificates of people aged 35-74 years and then interviewed relatives or friends to confirm the diagnoses. The study showed black men to be at slightly greater risk of dying suddenly from apparent IHD than white men. The death rate was greater in black men (2.80 vs. 2.20 per 1000) than in white men. Black women had greater rates than white women (1.48 vs. 0.70 per 1000). The Baltimore study (15) measured total IHD rate and sudden death rate. Sudden death rates were similar for black and white women. The rate for black men was somewhat higher than for white men. No significant racial differences were found in the incidence of sudden death, with or without prior history of heart disease, but the number of sudden deaths was very small for black women. The incidence of instantaneous death (defined as death within 15 minutes of onset) was the same in black and white women, but was slightly lower in black men than in white men. In the Nashville study, more sudden deaths occurred within two hours of symptom onset than in the Baltimore study, but fewer unwitnessed deaths were recorded. Ratios of sudden death between white men and women were 5:1 for Baltimore and 3:1 for Nashville. Male-to-female ratios in blacks were identical (2:1) in both studies. The variation in statistics on incidence of sudden death may be traceable at least partially to the following: * varying definitions of terms, " varying population samples, and •'" inaccurate death certificates. The Charleston study defined sudden death as death within one hour. The Baltimore and Nashville studies defined it as death within 24 hours of symptom onset. The HIP study (12) defined sudden death as death occurring within 48 hours of symptom onset. Such conflicts in definition limit possible inferences from the studies. Population samples, too, varied from study to study. Baltimore studied a slightly younger population than Nashville, and the HIP study included only first MI, while the Nashville study included subsequent Mis. In addition, the HIP cohort consisted mostly of employed (and therefore relatively healthy) individuals, in contrast to the free-living population 353 in the Nashville study. These variations may account for the rate of sudden death in Nashville being higher for all age groups than in the HIP study. Inaccuracy of death certificates may also be a significant problem in available data. This is suggested by the comparison of the Baltimore study results with the 1975 findings of Kuller et. al. who reviewed government vital statistics published for Baltimore (18). Nonwhite death rates from IHD were found to be higher here than those reported in the 1970-72 Baltimore study, with a higher sudden death rate for black women than white women; little racial difference was found among men. In a representative sample, death certificate diagnosis of IHD was confirmed by other data for only 67% of black men and 52% of black women. The black male-female rate ratios was higher in the Baltimore study than in the official vital statistics, suggesting that vital statistics may not be a valid source of information on IHD trends in urban black populations. Since death certificates were also a primary source in the Nashville study, these findings call into question its conclusions. Incidence of IHD as Determined by Post-mortem Examination Another method for assessing the incidence of death from IHD is postmortem examination, particularly those in which autopsied cases are representative of either all deaths or a definable sample of deaths. Investigators in the Baltimore study performed detailed autopsies of 87% of all sudden deaths occurring in a section of Baltimore and certified by the medical examiner (15,18). In a total of 169 sudden deaths, 58% of black men and women had severe three or four-vessel coronary artery stenoses compared with 70% of white men and 34% of white women (18). In the 118 cases without prior history of coronary heart disease, black men had acute coronary thrombi less frequently than white men (16% vs. 26%). However, black and white men had approximately the same frequency of hemorrhage and plaque. As noted by Gillum (2) these investigators found greater heart weights in blacks, but they found fewer recent (6% vs. 13%) and old (26% vs. 43%) myocardial infarctions in black men. Assessment of Trends in Morbidity and Mortality in Blacks and Whites Moriyama et. al. (19) and Cooper (20) have summarized cardiovascular mortality in the U.S. in the last four decades. Gillum (2) points out that prior to the 1940s, mortality trends were difficult to assess due to inadequacies in reporting, diagnosis, and certification. Data from national statistics have consistently shown rates for white men to exceed those of nonwhite men, while rates for nonwhite women have exceeded those of white women since 1950 (21). It should be noted that in these early data, national statistics compared whites and nonwhites, with no direct comparison between whites and blacks (who comprise 85% of U.S. nonwhites) (19). The data show that mortality due to IHD seems to be greater in urban than suburban areas, but that mortality has declined nationwide (2,20). The only exceptions are the Pacific and Mountain states, where the number of deaths attributed to IHD increased in nonwhites aged 35-74 years in both 354 CD P- P cr CD rt H 09 Hi o H rt CD O P4 P P- O Pi cr CD CC pl CL CL- CD P <-H Pl Hi X O P H CD CD Cfl h-> CA 13 CD hh P* P C 3 P' CD Pi P CA 09 I P H H 09 Co P 09 rt H CD CD CA P rt p P- P 0 p p vi Cfl 7? CD Pi P Pi P' vi H P- O Pi Pi o ?r P o p1 CD a 09 CL CD cr CD pJ 13 •3 p- rt cr rt cr CD pi P a C-H Co o 7? P- CD > CL Pl 3 CA Hi o H Cfl CD O O h{ CL P CD Pl • rt rt pi Pi H - C P- CD pi O p< pi Pi Q rt p- cr P1 CD p* P o 3 Pi * H CD 3 Hi CC <-H > • z m > ca p a W CD n h O Cfl H CD Co CL P- P 09 3 T) Pi P4 P • P C3 CA • O - H P- C-.13 P rt > O 7t P 0 •5 p-1 CD a 09 CD 3 CD P rt Cfl 3 0 H cr p- CL P- rt vi i 3 0 H rt Pi rt cr < Vi 13 >9 CD pi H d rt H cr o CD 13 Hi Pl O P O Pi p P" co vi CD Cfl CL CD O Pi Hi 13 13 p- £ vj cr p cr o p' cr pi < p- p 09 Cfl P CD s-^ 13 P» 13 po Ln p-> rt - v; rt CD P rt CA hrj p, P- 3 P 13 pi p-1 P-1 CD P" CA cr v! po O 13 w* o CA rt CD P rt P- Pi Pi P a cr p- rt CD Cfl 3 Pi t_i. O H o o p rt h p- cr P rt O H rt O rt cr CD o P P CD CD CL Cfl rt O a4 CD 09 p- < CD p rt O 00 pi <~s CA 13 pl CD 0 Hi P CL H CD O CD P rt CD O cr o o rt CA rt P O CD Hi CD hH 13 CC a O Pi Cfl CL rt CD X O P Hi Pl rt P- P CT rt p p' CD 09 PO CD O 3 H JT CD P Pl 3 09 CD a4 CD P P Cfl hh • rt - rt CD pi cr h h rt p pl pl O O 09 p■ ?r cd rt Pi pi O < 09 pi CD •5 P- cr p^ pi p- pi P rt cr cl £ o cr 3 P- CD rt rt P Pi CD CD Cfl CD CL X Pl • Cfl Cfl P- cr cr 3 O O 13 P P H P1 p> O CL CL < CD cr p- cl CD Hi cr a P- CD Cfl H P- Pl 3 on CD -P- P vi Cfl 09 P' CD CD Pl Pl 3 S P CD P- 09 rt 09 P4 CD Cfl CD cr po po < 13 CA p) rt p^ P- P P Cfl O CL rt p(t C P CL CL Pl P- CD CL CD CL CD Cfl X P- P rt P Pi O rt CD P- P O P- CL <-3 cd cr P CD o CD 3 O CL Cfl po rt rt pl CD < CD 13 Pl CL H P Pl Cfl Cfl P cr < h CD Cfl P* CA pl 09 Pl CD O CA p. P Pl cr cr cd rt CD CD x a o - o o cr p o P H rt O cr rt Pi 3 P- cr P CD CL po - rt Pi CD ti- er Cfl CD P 09 CL09 CD CD CL CD 3 p- O p1 O 09 P- o Cfl rt P CL P- CD Cfl H CD < P- CD I g cfl CD 13 CL CD O Pl P- cr pi o P- < o CD cr pi < CD O •5 rt cr CD P Cfl CD O Hi PO 09 CD H CD hH CO CD rt CL CD - CL Cfl O CD Cfl O P po 13 Cfl p« rt cr H P- O P- H P rt pi cr CD H CD x Vi P H Pl CC pi p• !> rt Z CD 13 W H Cfl CD < Cfl CD Cfl O < rt Pl CD P" p CD P O CD H CD P CD rt X CD O Cfl pl Cfl 13 cfl H CL P p CD X S 13 cr cd p» p- p pi rt o P- CD CD P 3 H 09 Cfl CD pl CD O P rt Cfl a Pi Pl rt Pi Cfl p 09 O H P- rt O ft CD Hi CD H hH P- O CC Pl p a I-" Cfl vj p. cr P o Co p hi cr i—■ rt h-> cl CD po h o Cfl rt cfl cr p CD CL O CA P- Hi CD CD Cfl CA CL P Pl rt CT rt O <-■• CD CD CL O W cd rt cr rt CA O CD rt hH 3 CD CD Cfl X rt o o o cr Hi CD Hi Pl CD rt rt CL hH cr cfl ec pi cd a rt rt cr rt po p- rt pi p pi 09 CC CD cr 7? CD p. > p cr CD Cfl CD p P- Hi 13 P hh o O P- 13 P- O P a h- p CD CD pl P P rt O rt P- 0 3 On CD H> P o O P hH 3 P P 13 CD pl hi CD rt CL o p p H P I pi 3 cr rt cr pi CD CD Cfl Cfl H CD a O o Hi Hi Pl p ph cr cl O P- p a CD hH P ec h- a p- P- 3 P 13 p' cr p. p* o pl pi O rt 7? P- Cfl O p po H O CD Hi P- rt p cr pl p- CL Cfl CD X H P CD pl < rt P- CD CD O P- o cfl p Pi rt P cr Cfl pi P- rt p 09 13 - H CD pl < p pl CL P- CD 3 P P o Cfl CD rt pl cr p CD CL CD Pi o cr 5^ Vi Cfl n o p o P1 p Cfl p- o p Cfl pl p CL 3 13 h-' p- o pl rt p- o p Cfl £ rt cr cr P- CD rt CD P Cfl Pl • rt p- o p C3 CD o h-1 p- p CD Cfl p- p 3 o H rt pl p' p- rt vi •3 CD H CD p* Cfl po p < H CD < CD rt vi cr cd cd a cr " H- 3 09 CD cr rt CD H Cfl O rt 13 O O P- pl p- H rt CL Pl p- p CA Hi rt CL Pl p- p CA CL CD pl H CA p CD H 3 p-1 O hi TJ rt CD pl CD o p- a cr rt cd CA Vi CD CD H P- < p " CD Tl CL rt CD P- CD Pi CD o a H Cfl CD O CD cr p O rt p- rt cr ca cr pi ti- er p cr I— CL O pi p O 13 09 7? o cr CA CA rt Cfl Pi P- rt p cr o CL t-i V3 . p hi CL vi cr h-1 p- pl p o j»r cfl cfl O p O rt o cr 3 13 n po po H H CD O CL (-■ p- rt P O Pi cr ro P- N3 rt >--' CD Cfl Cfl • cr O 2 8. O o Cfl O CD P h-- Cfl CD P- O Cfl rt rt CD CD CL P rt pi (—' H v- CD po p- Cfl p- 09 «: cr CD CD H H CD rt Hi er h CD O 3 Cfl O H P CD rt 09 p4 p- € O CD P Cfl rt rt • O H CD 09 p- O p VD o pl p CL Hi p- p CL p- p 09 p- rt 09 H CD pl rt CD Cfl rt p- CL p- CA o p CA Cfl CD a rt cr CD H pl rt CD O Hi CL CD o P1 p- p CD Table 1 TOTAL NUMBER OF SUBJECTS WITH ECG READINGS IN THE NATIONAL HEALTH AND NUTRITION SURVEY 1971 - 1975 (Preliminary-1/28/85) Age Group ALL 25-44 45-54 55-64 65-74 Black men 379 120 97 75 87 White men 2688 1037 636 528 487 Black women 465 191 105 84 85 White women 3152 1329 739 556 528 Table 2 TOTAL NUMBER OF SUBJECTS WITH ECG READINGS IN THE NATIONAL HEALTH AND NUTRITION SURVEY 1976 - 1980 (Preliminary-1/28/85) Age Group ALL 25-44 45-54 55-64 65-74 Black men 528 209 62 129 128 White men 4302 1554 617 1086 1045 Black women 635 248 100 135 152 White women 4794 1726 647 1176 1245 356 Table 3A PREVALENCE OF CORONARY HEART DISEASE IN BLACKS AND WHITES AGE 15-75+ YEARS PER ONE THOUSAND SUBJECTS, (1960) EVANS COUNTY, GA N n Rates* Black men 537 12 22.4 White men 947 54 59.0 Black women 646 9 13.6 White women 972 18 18.2 N= number of persons in each race/sex group n=number of cases * Age-adjusted rates by indirect method using total population as standard Source: JC Cornoni et al, Ref 6a Table 3B PREVALENCE OF CORONARY HEART DISEASE IN BLACKS AND WHITES AGE 35-75+ YEARS PER ONE THOUSAND SUBJECTS, (1960-61) CHARLESTON, SC N n Rates* Black men 333 11 White men 650 49 Black women 453 13 White women 738 20 32.7 75.8 26.6 26.6 * Age adjusted by indirect method Source: Keil JE et al, Ref 7 357 Table 4 INCIDENCE OF ANTERIOR MYOCARDIAL INFARCTION IN BLACKS AND WHITES AGE 35-74 YEARS PER ONE THOUSAND SUBJECTS, (1967-68) NASHVILLE, TN N n Rates* Black men 14,628 31 White men 64,478 379 Black women 16,236 18 White women 71,469 137 2, .12 5. .88 1. .11 1. .92 N= number of persons in each race/sex group n= number of cases of AMI * Age group specific data showed similar trends Source: Hagstrom RH et al, Ref 14 Table 5 INCIDENCE OF ACUTE MYOCARDIAL INFARCTION IN BLACKS AND WHITES PER ONE THOUSAND SUBJECTS BY AGE DECADE, (1973) NEWARK, NJ Age Decade Total Cases 20-29 30-39 40-49 50-59 60-69 70-79 80+ Black men „ _ 1.16 2.41 4.61 4.04 2.27 - - (ID (15) (17) (6) (D 50 White men 0.10 0.78 1.09 3.38 5.37 6.28 2.41 (D (7) (11) (34) (41) (27) (4) 125 Black women 0.10 0.19 0.85 1.64 2.18 4.04 5.58 (2) (3) (4) (12) (10) (8) (4) 43 White women 0 0.11 0.28 0.88 1.31 3.43 3.28 - (D (3) (10) (12) (21) (8) 55 Numbers in parentheses represent numbers of cases Source: AB Weisse et al, Ref 16 358 Table 6 INCIDENCE OF IHD MANIFESTATIONS BY RACE AND SEX, 1960-1961 TO 1974-1975, CHARLESTON,SC Black men (N=322) White men Black men, high SES (N=601) (N=101) Manifestation Rate Rate n Rate All IHD 43 131.7 All nonfatal IHD 16 50.7 AMI 9 28.5 Angina pectoris 7 22.2 Black women (N= =440) Manifestation n Rate All IHD 72 161.0 All nonfatal IHD 42 100.9 AMI 15 36.3 Angina pectoris 27 63.6 ___ 114 188.4 60 95.2 51 80.5 8 12.8 Whit e women (N= 718) n Rate 84 113.8 50 67.4 27 36.3 22 29.8 "" 4 61.2 2 26.0 2 26.4 0 0 N= Total number of persons in each race/sex group n= Total number of cases in each IHD category SES=socioeconomic status * Rates per one thousand, age adjusted by indirect method Source: JE Keil et al, Ref 7 359 Table 7 INCIDENCE OF HOSPITALIZATION RATES FOR AMI IN U.S . BLACK AND WHITE POPULATIONS PER ONE THOUSAND SUBJECTS Age/Sex Evans County,GA 1960-1967 Nashville Baltimore Newark Columbia,SC 1967-1968 1970-1972 1973 1968 35-44 Black men 1.79 1.02 - 0.84 White men 8.28 1.84 0.783 1.25 Black women 3.72 0.19 0.193 0.38 White women 0.69 0.20 0.113 0.06 45-54 Black men 5.79 1.55 0.6 1.16 3.97 White men 10.48 5.50 1.9 1.09 5.76 Black women 5.10 1.41 0.2 0.85 0.68 White women 4.14 1.05 0.4 0.28 1.33 55-64 Black men 1.52 3.47 0.8 2.41 7.61 White men 17.79 9.82 2.9 3.38 10.75 Black women 8.97 1.64 1.0 1.64 2.91 White women 5.52 2.77 1.1 0.88 2.56 65-74 Black men 8.69 3.87 4.61 6.27 White men 32.14 12.75 5.37 18.82 Black women 6.21 1.77 2.18 5.75 White women 19.72 6.49 1.31 8.47 Source: RF Gillum, Ref 2 360 References 1. Adams L, Africano E, Doswell W, Frate D , Gillum R, Havlik R, Langford H, Mebane I, Neser W, Potts J, Saunders E, Savage D, Schachter J, Stamler J, Tillotson L, Watkins L, Williams R: Report of Epidemiology Working Group on Coronary Heart Disease in Black Populations. Am Heart Journal 1984;108:699 2. Gillum RF: Coronary heart disease in black populations I. Mortality and Morbidity. Am Heart Journal 1982;104:839. 3. National Center for Health Statistics: Coronary heart disease in adults. Washington, D.C, U.S. Government Printing Office. NCHS publication No. 1000, Series II, No. 101, 1965. 4. Prineas RJ, Crow RS, Blackburn H: The Minnesota Code Manual of Electrocardiographic Findings. Boston, 1982. John Wright- PSG Inc. 4a. Watkins LO: Epidemiology of coronary heart disease in black populations: methodologic proposals. Am Heart J 1983;108: 635. 5. National Center for Health Statistics: Prevalence of chronic circulatory conditions, U.S. 1972. Washington, D.C, U.S. Dept. of Health, Education, and Welfare. Publication No. (HRA), 75-1521. 1974. 6. McDonough JR, Hames CG, Stulb SC, et al: Coronary heart disease among Negroes and Whites in Evans County, Georgia. J Chronic Dis 1965;18:433. 6a. Cornoni JC,Waller LE,Cassel JC,Tyroler HA,Hames CG:The Incidence Study-Study Design and Methods. Arch Intern Med 1971;128:896. 7. Keil JE, Loadholt CB, Weinrich MC, Sandifer SH, Boyle E, Jr: Incidence of coronary heart disease in blacks in Charleston, South Carolina. Am Heart Journal 1984;108:779. 8. Langford HG, Oberman A, Borhani NO, Entwisle G, Tung BS: Black-white comparisons of indices of coronary heart disease and myocardial infarction in the stepped care cohort of the Hypertension Detection and Follow-Up Program. Am Heart Journal 1984;108:797. 9. Mihaley JP, Whiteman NC: Myocardial infarction in the negro: Historical survey as it relates to negroes. Am J Cardiol 1958;21:464. 361 10. Feldman, J: National morbidity studies. In Havlik RJ, Feinleib M, editors: Proceedings of the conference on the decline in coronary heart disease mortality. NIH publication No. 79-610, Washington, DC 1979, U.S. DHEW. 11. An Epidemiological investigation of cardiovascular disease: The Framingham Study: Section 11. Washington, D.C, U.S. Government Printing Office, September, 1968. 12. Shapiro S, Weinblatt E, Frank CW, et al: Incidence of coronary heart disease in a population insured for medical care. (HIP). Am J Public Health 1969; 59 (supp 2):1. 13. Cassel J, Heyden S, Bartel AG, Kaplan BH, Tyroler HA, Coroni JC, Hames CG: Incidence of coronary heart disease by ethnic group, social class and sex. Arch Intern Med 1971;128:901. 14. Hagstrom RH, Federspeil CF, Ho Y: Incidence of Myocardial infarction and sudden death from coronary heart disease in Nashville, Tennessee. Circulation 1971;44:884. 15. Kuller H, Cooper M, Perper J, et al: Myocardial Infarction and sudden death in an urban community: Bull NY Acad Med 1973; 49:532. 16. Weisse AB, Abiuso PD, Thind IS: Acute myocardial infarction in Newark, NJ: A study of racial incidence. Arch Intern Med 1977;137:1402. 17. Zmyslinski R, Lackland DT, Keil JE, and Higgins JE: Increased fatality and difficult diagnosis of in-hospital acute myocardial infarction: Comparison to lower mortality and more easily recognized pre-hospital infarction. Am Heart Journal 1981; 101:586. 17a. Keil JE, Saunders DE, Jr, Lackland DT, Weinrich MC, et.al: Acute myocardial infarction: Period prevalence, case fatality, and comparison of black and white cases in urban and rural area of South Carolina. Am Heart J 1985;109:776. 18. Kuller L, Perper J, Cooper M: Demographic characteristics and trends in arteriosclerotic heart disease mortality: Sudden death and myocardial infarction. Circulation, 1975; 51 and 52 (suppl III):III-1. 19. Moriyama I, Krueger DE, Stamler J: Cardiovascular Diseases in the United States. Cambridge, 1971. Harvard University Press, p. 49. 20. Cooper R, Stamler J, Dyer A, Garside D: The decline in mortality from coronary heart disease, U.S.A., 1968-1975. J Chronic Disease 1978;31:709. 362 21. Rosenberg, HM and Klebb, AJ: Trends in cardiovascular mortality with a focus on ischemic heart disease: US, 1950-76 in Proceedings of the Conference on the Decline in Coronary Heart Disease Mortality. U.S. DHEW, Public Health Services, NIH Pub. No. 79-1610, May 1979. 22. Keil JE, Lackland DT, Hudson MB, Saunders DE Jr, et al: Coronary heart disease and stroke mortality in South Carolina geographical and temporal trends. J SC Med Assn 1983; 79:65. 23. Savage DD: Echocardiographic assessment of cardiac anatomy and function in black and white hypertensive subjects. In Williams R, ed. Textbook of Ethnic Medicine (in press). 24. Haywood LJ: Coronary heart disease mortality/morbidity and risk in blacks. Clinical manifestations and diagnostic criteria: The experience with the Beta BLocker Heart Attack Trial. Am Heart Journal 1984;108:787. Supplemental References Ahmed SS, Razefort R, Franciato R: Incidence of acute myocardial infarction among blacks in an urban community. J Med Soc NJ 1977;74:1058. Five-year findings of the hypertension detection and follow up-program II. Mortality by race, sex and age. JAMA 1979; 242:2572. Kleinbaum DG, Kupper LL, Cassel JC, Tyroler NA: Multivariable analysis of risk of coronary heart disease in Evans County, Georgia. Arch Intern Med 1971;28:943. Kleinman JC, DeGruttola VG, Cohen BB, Madans JH: Regional and urban-suburban differentials in coronary heart disease mortality and risk factor prevalence. J Chronic Dis 1981;34:11. Kuller L: Sudden and unexpected nontraumatic deaths in adults. A review of epidemiological and clinical studies. Journal of Chronic Diseases 1966;19:11652. National Center for Health Statistics: Mortality, United States, 1977. By Number of Deaths. Washington, D.C, 1981, U.S. Government Printing Office. Stamler J, Berkson DM, Lindberg HA, Miller OW, Hall Y: Racial patterns of coronary heart disease: blood pressure, body weight, and serum cholesterol in whites and negroes. Geriatrics 1961;16:382. 363 Differences in Socioeconomic Status and Acculturation among Mexican Americans and Risk of Cardiovascular Disease Helen P. Hazuda, Ph.D. Division of Clinical Epidemiology Department of Medicine University of Texas Health Science Center at San Antonio San Antonio, Texas Introduction The purpose of this paper is to review the available literature to assess the extent to which increasing differences in socioeco- nomic status (SES) and acculturation among Mexican Americans are associated with differences in cardiovascular risk factors and to make public health recommendations based on the findings. It should be noted initially that there is no commonly agreed upon definition of Mexican American (MA) ethnicity and that the several indicators most commonly used to classify persons as Mexican American identify populations which differ significantly in composition and size (1-5). The studies reviewed in this paper relied on a variety of indicators for Mexican American eth- nicity, ranging from simple self-definition to a complex algorithm which considers surname and birthplace of subject's parents, ethnic origin of subject's grandparents, and subject's preferred ethnic identity. The study populations were referred to variously as Hispanics, Spanish-Americans, persons of Spanish- surname, Mexican Americans, or Mexican origin Hispanics. I have, nonetheless, restricted this review to studies which explicitly identified the target population as Mexican origin Hispanics or were carried out in a geographic region where the overwhelming majority of Hispanics are known to be Mexican origin, (e.g., over 80% of Hispanics in California and over 92% in Texas are Mexican origin (6)). To place the review in context I will begin by pre- senting a brief overview of changes in socioeconomic status and acculturation which have occurred among Mexican Americans during the last several decades, and, then, document the role of car- diovascular disease as a cause of mortality in this ethnic group before moving to a review of cardiovascular risk factors. Socioeconomic Status and Acculturation of_ Mexican Americans1 The three major indicators of socioeconomic status are education, occupation, and income. Of these, education is perhaps the most important since it is closely linked to upward mobility and the acculturation process. At the time of the 1970 census, Mexican Americans still lagged about a generation behind the total U.S. population in educational attainment and, together with the Puerto Ricans, ranked as the least schooled Hispanic subgroup. Nevertheless, Mexican Americans had made substantial gains in schooling over the last several decades. Around the time of World War I only about 10 percent of high-school-completion age "^Statistics reported in this section, unless otherwise indicated, were abstracted from Reference 7: Jaffe AJ, Cullen RM, Boswell TD: The Changing Demography of Spanish Americans. New York: Academic Press, 1980. 367 Mexican Americans had actually completed high school. By the end of World War II about 20 percent were high school graduates and, ty 1970, over 50 percent of the school-age cohort of U.S.-born Mexican Americans and 32 percent of the Mexico-born cohort were high school graduates. Gains were made in higher education as well, but in 1970 only about 4 percent of Mexican Americans aged 25-29 who entered college had actually completed a 4-year degree. Occupations held by Mexican American men in 1970 were the lowest among the Hispanic subgroups and fell substantially below those of non-Hispanic Whites. Nonetheless, considerable upward move- ment occurred among Mexican Americans during the past several decades. In fact, the movement of Mexican Americans up the occu- pational ladder between 1930 and 1970 was relatively greater than that for the total U.S. White male population. A major reason for this was the movement of Mexican American men out of agri- culture. In 1970 only 10 percent of Mexican American men were employed in agriculture compared to 40 percent in 1930. In addi- tion, in 1970 less than 25 percent of Mexican American men were on the bottom rungs of the occupational ladder (nonfarm and farm laborers) compared to 66 percent in 1930. While 61 percent of Mexican American men in 1970 were employed in lower level blue collar jobs compared to only 45 percent of non-Hispanic Whites, this figure represented a substantial decline from 1930, when 87 percent of Mexican American men were employed in such jobs. In addition, the number of Mexican American men employed in white collar jobs at both upper levels (professionals, nonfarm mana- gers) and lower levels (sales and clerical workers) tripled bet- ween 1930 and 1970. Within equal levels of schooling the occupational positions of Mexican American men closely approached those of non-Hispanic Whites; among college graduates, there were few differences between the two ethnic groups. Overall, about 37 percent of Mexican American women aged 16 and over were in the labor force in 1970, compared to 41 percent of non-Hispanic White women. The occupational position held by Mexican American women in 1970 were the lowest for all the Hispanic subgroups and fell substantially below those of non-Hispanic Whites. Between 1950 and 1970, however, Mexican American women had experienced considerable upward occupational mobility comparable to that for Mexican American men. The number of Mexican American women in upper level white collar jobs increased from 7 percent in 1950 to 8.5 percent in 1970. In lower level white collar occupations, the number of Mexican American women increased from 21 percent in 1950 to 31 percent in 1970. At comparable levels of schooling, Mexican American women closely approximated the occupational levels achieved by non-Hispanic Whites. 368 The income of Mexican Americans still lags substantially behind that of non-Hispanic Whites. In 1969 the median family income of Mexican Americans was $6,960, ranging from a median of $1,800 in families with no earners to a median of $9,080 in familes with two or more earners. These figures compared to a median family income among non-Hispanic Whites of $10,100, ranging from a median of $2,970 in familes with no earners to a median of $12,350 in familes with two or more earners. In general, fami lies headed by men had higher median incomes than families headed by women, and families headed by U.S. born MAs had higher incomes than families headed by Mexican born MAs. There is some indica- tion, however, that MAs are gaining ground on non-Hispanic Whites. In 1969, MA families headed by males averaged about 65% of the income of non-Hispanic White male-headed families. By 1976-1978, this percentage had increased to about 70 percent (MA median income, $12,600). Similarly, in 1969 MA families headed by females averaged about 50 percent of the income of non-Hispanic White female-headed families. By 1976-1978, this percentage had also increased to about 70 percent (MA median income, $6,200). Acculturation can be understood as a complex process whereby individuals whose primary learning has been in one culture take over characteristic ways of living from another culture (8-10). This process is complex in that it usually proceeds somewhat un- evenly in two major areas of life: the social and the cultural. The social area involves interaction with members of the host society which results in the adoption of the public behaviors, language, and material culture of the host society, while pre- serving the primary attitudes, values, language and behaviors of the culture of origin for participation in the private spheres of life. The cultural area involves psychological integration with the host society which results in the adoption of the central attitudes and values of the host society for participation in both public and private spheres of life. On a national level, two statistical measures can be examined as indicators of acculturation among Mexican Americans: the proportion who are first, second and third generation Americans; and the proportion who marry outside their own ethnic group (outmarriage), par- ticularly those who marry non-Hispanic Whites. In 1910 about 58 percent of all MAs in the U.S. were first generation Americans (born in Mexico); by 1970 that number had dropped to about 32 percent (7). In the five Southwestern states, 2 which contain about 90 percent of the total Mexican American California, Texas, Colorado, New Mexico and Arizona. 369 population, only 12 percent of MAs were first generation Americans in 1970 (11). Another 30 percent were second genera- tion (U.S. born of Mexico-born parents), while 58 percent, the vast majority, were third generation Americans (U.S.-born of U.S.-born parents) (11). The rate of outmarriage among U.S.-born MA women in 1970 was the lowest among U.S.-born Hispanic women.3 Only 16 percent of U.S.-born MA women had non-Mexican American husbands. However, 84 percent of these outmarriages were to non-Hispanic White males; this was the highest rate for such marriages among the Hispanic subgroups examined. Younger women were somewhat more likely to marry outside their own ethnic group than older women. The rate of outmarriage among U.S.-born MA women under 35 years old was 19 percent compared to only 13 percent for those aged 35 and over. The percentage of outmarriages increased substantially with increased schooling. For women younger than 35, the rate of outmarriage among high school graduates was 30 percent compared to only 11 percent among those who had not completed high school. For women aged 35 and over, the corresponding rates of out- marriage were 37 percent among high school graduates and 8 per- cent among those who had not completed twelfth grade. In summary, Mexican Americans are becoming an increasingly heterogeneous population, both socioeconomically and culturally. The majority of Mexican Americans, however, are still located in the lower SES strata and have experienced low or intermediate levels of acculturation into mainstream (non-Hispanic White) American society. Cardiovascular Disease As A Cause of Mortality Among Mexican Americans Two studies have examined the relative importance of car- diovascular disease (CVD) as a cause of death among Mexican Americans (12,13). One was conducted in California, the other in Texas, the two states where over 86 percent of the nation's Mexican American population resides (6). Both studies examined mortality data for the period 1969 to 1971. Persons were classified as Mexican American or non-Hispanic White using the U.S. Census surname identifier. Methodologies differed in that the California study compared cause-specific mortality using life table deaths computed for 5-year intervals ages 0 to 85 and stan- 3 Rates of outmarriage were based on the Public Use Sample tabula- tions for husbands and wives living together at the time of the 1970 Census enumeration. Rates could be determined for Mexican Americans, Puerto Ricans, Cubans, and Hispanos, but not for Central and South Americans(7). 370 dardized death rates directly standardized using the total popu- lation of California in 1960; while the Texas study compared only standardized death rates directly standardized on the Texas non-Spanish surname White female age distribution (13). The two studies yielded strikingly similar results. Both found cardiovascular disease to be the leading cause of death among Mexican Americans, just as it is among non-Hispanic Whites. CVD mortality rates were lower for Mexican American males, however, than for non-Hispanic White males (in California, 82% of the non-Hispanic White rate; in Texas, 85% of the non-Hispanic White rate). CVD mortality rates for Mexican American females, on the other hand, were virtually identical to those for non-Hispanic White females (in California, 99.8% of the non-Hispanic White rate; in Texas, 7% higher than the non-Hispanic White rate). The sex differential in CVD mortality was less favorable for Mexican American females than for non-Hispanic Whites (in California: a male-female sex ratio of 1.43 for MAs vs. 1.72 for non-Hispanic Whites; in Texas, a male-female sex ratio of 1.46 for MAs vs. 1.78 for non-Hispanic Whites). Two other studies have examined secular trends in cardiovascular mortality during the first half of the 1970s to determine whether Mexican Americans shared in the nation-wide decline in CVD mor- tality which occurred during that period (14,15). The first study, using data only from Bexar County (San Antonio) Texas, found that standardized CVD mortality rates4 for the period 1970 to 1976 declined significantly in both MA men and women and in non-Hispanic White males (14). Furthermore, the CVD mortality decline in Mexican Americans was equal to that in non-Hispanic whites. The second study, using state-wide data from Texas, found that standardized proportional CVD mortality rates for the period 1970 to 1976 declined in Mexican Americans, but that, par- ticularly for MA males, the decline was much less steep than in non-Hispanic Whites (15). Each of these studies suffered from the limitation that 1975-76 population denominators were derived from population estimates rather than actual census data. A new study, not yet published, which uses denominators from the 1980 census for the state of Texas, indicates an interesting sex- differential in the rate of CVD mortality decline in Mexican Americans during the period 1970 to 1980 (16). The rate of decline in Mexican American men was about only half as steep as in non-Hispanic Whites, while the rate of decline in Mexican 4Directly standardized to the 1970 U.S. population. 371 American women was one-third to two-thirds steeper than in non-Hispanic Whites. Socioeconomic Status, Acculturation, and Cardiovascular Risk Factors in Mexican Americans Evidence concerning the level of cardiovascular risk factors in Mexican Americans is growing, but research in this area has been quite limited over the past 15 to 20 years. This is particularly true for research on the association between differences in socioeconomic status and acculturation among Mexican Americans and the risk of cardiovascular disease. Table 1 provides a list of 21 research reports found in a search of the literature for the period 1966 to October 1984. These 21 reports are based on eight separate studies: 4 carried out on populations in California, 3 in Texas, and 1 in the four states along the U.S.-Mexico Border. Over half of the reports are from the San Antonio Heart Study (SAHS), a large population-based investiga- tion which compared diabetes and cardiovascular risk factors in Mexican Americans and non-Hispanic Whites residing in three socioculturally distinct neighborhoods of San Antonio, Texas. Almost one-fourth of the 21 reports, including all of those from the Starr County study, deal exclusively with diabetes or with the relationship between diabetes and obesity. Only about half of the reports, primarily those from the San Antonio Heart Study, examine the relationship between either SES or acculturation and cardiovascular risk factors in Mexican Americans. The evidence contained in these reports deals with six risk factors: lipids and lipoproteins, blood pressure and hypertension, cigarette smoking, exercise, obesity, and diabetes. 1. Lipids and lipoproteins. Cholesterol levels in Mexican Americans have been examined in five separate studies (17-21). All of these compared cholesterol levels in Mexican Americans with those in non-Hispanic Whites. One study looked only at very low SES Mexican Americans and found that the age-adjusted serum cholesterol levels in Mexican Americans of both sexes were higher than those recorded for non-Hispanic Whites in a national comparison sample (HANES I), which was undoubtedly of higher SES (17). Two studies used mixed samples which included both lower and middle SES Mexican Americans and non-Hispanic Whites, but did not carry out stratified analyses or adjust for differences in SES between the two ethnic groups (18,19). In one study Mexican Americans were predominantly of lower SES, while non-Hispanic Whites were predominantly of middle SES; cholesterol levels were no higher in Mexican Americans than in non-Hispanic Whites (18). In the other study lower and middle SES subjects were nearly equally distributed in the two ethnic groups; cholesterol levels 372 in Mexican American men, however, were higher than those in non-Hispanic Whites, while cholesterol levels in Mexican American women were similar to those in non-Hispanic Whites (19). These three studies revealed no consistent pattern of ethnic differen- ces in cholesterol levels nor do they permit inferences about the relationship between socioeconomic status and cholesterol levels in Mexican Americans. Two of the five studies do provide this this information; one of these was done in central California (20), the other in South Texas (21). The California study compared the proportion of hypercholestero- lemic males (cholesterol > 260 mg/dl) in six different levels of socioeconomic status. The overall prevalence of hypercholestero- lemia in Mexican Americans was higher than in non-Hispanic Whites (16.7 vs. 13.9%). There was some evidence of an inverse rela- tionship between SES and cholesterol levels in non-Hispanic Whites, but no evidence of a similar SES gradient in Mexican Americans (20). The South Texas study (SAHS) examined cho- lesterol as a continuous variable. SES diffrences were examined by comparing cholesterol levels for a random sample of Mexican Americans and Non-Hispanic Whites who resided in three socioculturally distinct neighborhoods of San Antonio, Texas. These neighborhoods had been purposively selected by the researchers to represent low, middle and high levels of SES and acculturation of Mexican Americans. Non-Hispanic Whites were sampled only in the two higher socioeconomic neighborhoods.5 Overall, cholesterol levels tended to be somewhat higher in Mexican American than in non-Hispanic White men, but similar in the two ethnic groups for women. Among Mexican Americans, cho- lesterol levels were nearly identical in the two lower SES groups, but increased markedly and significantly in the high SES group. For women, on the other hand, cholesterol levels remained similar in all three SES groups. For non-Hispanic Whites, similar sex difference was observed in the relationship between socioeconomic status and cholesterol. 5Strictly speaking, these neighborhood analyses do not permit a clear distinction between any purely socioeconomic or cultural factors which might be operating. Neighborhood differences pro- bably reflect the combined effect of socioeconomic status and acculturation. Nonetheless, since the three neighborhoods repre- sent significantly different levels of socioeconomic status (21), the investigators feel justified in referring to the three neigh- borhoods as "SES groups." 373 In men, cholesterol levels were significantly higher in the upper SES group than in the middle group; in women, cholesterol levels were similar for both SES groups. Four studies also examined triglyceride levels (17-19,21). All of them found triglyceride levels in Mexican Americans to be higher than those in non-Hispanic Whites, although the differences tended not to be statistically significant. Ethnic differences in triglyceride levels persisted in both sexes even when comparisons were made within the same SES strata6 (21). Among Mexican Americans triglyceride levels in men were similar in the two lower SES groups, and increased markedly, though not statisti- cally significantly, in the highest SES group. In Mexican American women, there was a clear tendency for triglyceride to decrease in the higher SES group. Two lipoproteins were also examined in the San Antonio Heart Study: low density lipoprotein cholesterol, associated with increased risk of coronary heart disease (CHD); and high density lipoprotein cholesterol, a protective factor against CHD (21). For low density lipoprotein cholesterol (LDL-C)7, the rela- tionship with socioeconomic status closely paralleled that for total cholesterol. In Mexican American men, LDL-C levels were similar in the two lower SES groups, but increased markedly and statistically significantly in the highest SES group. In non-Hispanic men LDL-C also increased significantly from the middle to upper SES group. For women in both ethnic groups, on the other hand, LDL-C levels were similar in all SES groups. For high density lipoprotein cholesterol (HDL-C), there was also a sex difference in the relationship with socioeconomic status. In men of both ethnic groups HDL-C remained constant across levels of SES. On the other hand, in women of both ethnic groups, HDL-C Strictly speaking, these neighborhood analyses do not permit a clear distinction between any purely socioeconomic or cultural factors which might be operating. Neighborhood differences pro- bably reflect the combined effect of socioeconomic status and acculturation. Nonetheless, since the three neighborhoods repre- sent significantly different levels of socioeconomic status (21), the investigators feel justified in referring to the three neigh- borhoods as "SES groups." 7A1though SES was not considered, a related report suggests that ethnic differences in triglyceride levels persist even when adjustments are made for a number of relevant behavioral variables (i.e., age, obesity, cigarette smoking, coffee and alcohol consumption and, in women, estrogen and oral contracep- tive usage) (22). 374 rose significantly with increasing socioeconomic status. In terms of lipids and lipoproteins, then, it appears that increased socioeconomic status may be associated with greater protection against CHD in women (i.e., lower triglyceride levels and higher levels of HDL-C), but with increased risk of CHD in men (higher cholesterol levels and higher levels of LDL-C). Two other reports from the San Antonio Heart Study have examined SES differences in dietary practices which may influence serum cholesterol (23,24). One report used data from a diet frequency questionnaire to devise a scale measuring avoidance of saturated fat and cholesterol (23). In Mexican Americans of both sexes, fat and cholesterol avoidance increased significantly with increasing socioeconomic status. Although non-Hispanic Whites in the middle SES group still scored significantly higher on the fat avoidance scale than Mexican Americans, no ethnic differences were present in the highest SES group. No difference in fat avoidance was observed between Mexican Americans born in the U.S. and those born in Mexico. The other SAHS report looked at how well-informed individuals were about the role of dietary cholesterol and saturated fat as risk factors for cardiovascular disease and at the extent to which individuals modified their diet to reduce heart disease risk (24). Overall, Mexican Americans were less well-informed than non-Hispanic Whites and were less likely to have attempted any dietary modifications to reduce risk of heart disease. Although increased socioeconomic status was associated with increased knowledge and preventive behavior in both ethnic groups, the gap in knowledge and behavior between Mexican Americans and non-Hispanic Whites began to close at the higher SES levels. 2. Blood pressure and hypertension. Three studies have compared blood pressure levels in Mexican Americans and non-Hispanic Whites (17-19). All three found diastolic blood pressure, in men, and systolic blood pressure, in both men and women, to be similar in the two ethnic groups. Two of the stu- dies (17,18) found diastolic blood pressure in Mexican American women to be lower than in non-Hispanic Whites, while the third found no ethnic differences (19). The Laredo Project (17) also examined prevalence of hypertension in Mexican Americans, using two separate definitions utilized in the Hypertension Detection and Follow-up Study (HDFP). One definition of hypertension ("elevated diastolic blood pressure") included only individuals with diastolic blood pressures (DBP) greater than or equal to 95mm Hg. The second definition ("actual" hypertension) included individuals with currently elevated diastolic blood pressures 375 plus those whose diastolic blood pressures were below 95 mm Hg, provided they gave a history of hypertension and were also currently taking anti-hypertensive medications. The second defi- nition provides for differences in treatment levels which may occur across ethnic groups or geographic regions. The prevalence of elevated diastolic blood pressure in Laredo Mexican Americans (very low socioeconomic status) was compared to that of HDFP Whites and Blacks. In Mexican American males the prevalence of hypertension was intermediate between those of HDFP Whites and Blacks. In Mexican American females, on the other hand, hypertension prevalence was lower than in either HDFP Whites or Blacks. Mexican American males had substantially higher rates of elevated diastolic blood pressure in each of three age groups (40-49 year olds: 19.4 vs. 7.9%; 50-59 year olds: 23.3 vs. 7.1%; 60-69 year olds: 23.8 vs. 5.7%) (17). The rate of "actual" hypertension in Laredo Mexican Americans was also intermediate between the rates for HDFP Whites and Blacks until age 59, after which it became essentially similar to the HDFP Black rate. For Mexican American women, however, the rate of actual hypertension was slightly higher than that in HDFP Whites except in the 60-69 year-old age category where the rate was as high as that in HDFP Blacks. The large sex differences observed for rates of elevated diastolic blood pressure in Laredo Mexican Americans were greatly reduced for actual hypertension and were, in fact, reversed in the oldest age category (40-49 year olds: 29.0 vs. 21.1%; 50-59 year olds: 37.2 vs. 26.2%; and 60-69 year olds: 38.1 vs. 44.3%). The latter finding suggests that among Laredo hypertensives, women were more likely than men to have their hypertension controlled (17). SES differences in hypertension prevalence were examined in a California study (20) and in the San Antonio Heart Study (21), but with two different definitions of hypertension. The California study looked at elevated diastolic blood pressure, while the SAHS examined actual hypertension. The California study found a clear inverse relationship between SES and preva- lence of elevated diastolic blood pressure in Mexican American males, ranging from a rate of 8.2 percent in the highest SES category to 19.6 percent in the lowest SES category (20). A similar SES gradient was present in non-Hispanic Whites. The overall prevalence of elevated diastolic blood pressure in Mexican American males was 15.6 percent; intermediate between the rates found for non-Hispanic Whites and Blacks. The San Antonio Heart Study found no SES gradient in prevalence of actual hypertension in either Mexican American or non-Hispanic White males (26). In Mexican American males, rates of actual 376 hypertension were 10.7 percent in the lowest SES group, 7.3 per- cent in the middle SES group, and 11.5 percent in the highest SES group. The corresponding rates for non-Hispanic White males were 10.0 percent in the middle SES group and 9.5 percent in the high SES group. In contrast to the finding for men, there was a strong inverse relationship between socioeconomic status and pre- valence of actual hypertension in women of both ethnic groups. For Mexican Americans, the rates declined from 9.2 percent in the low SES group to 8.4 percent in the middle SES group, and 4.7 percent in the high SES group; for non-Hispanic Whites, the rate declined from 10.7 percent in the middle SES group to 8.7 percent in the high SES group. Rates of hypertension appeared to be similar in males of both ethnic groups, but lower in Mexican American women than in non-Hispanic Whites, although this ethnic difference was not statistically significant. When adjustments were made for obesity (higher in MAs), Mexican Americans tended to have a lower rate of hypertension than non-Hispanic Whites at the same SES level. Data on the proportion of hypertensives previously diagnosed and under treatment with antihypertensive drugs were also reported from the San Antonio Heart Study. The proportion of hyperten- sives previously diagnosed and under treatment was lower in Mexican Americans of both sexes than in non-Hispanic Whites. Interestingly, the proportion of previously diagnosed hyperten- sives under treatment was as great among low SES Mexican American men as high SES men, and actually decreased from the low to high SES group in MA women. Among non-Hispanic Whites, on the other hand, there was either no SES difference in the proportion of previously diagnosed hypertensives under treatment (for women) or the proportion increased with increased socioeconomic status (for men). These data suggest that health care deliverers may be par- ticularly sensitive to the potential presence of hypertension among low SES minorities. Nonetheless, the proportion of Mexican American hypertensives both under treatment and under control (DBP < 95 mm Hg) was lower in both sexes at the lower SES levels. In addition, lower SES Mexican Americans of both sexes were also somewhat less likely than those at higher SES levels to report that they had attempted to control hypertension as a means of preventing heart attacks (24). The overall proportion of SAHS Mexican Americans under adequate control was greater in women than in men (87 vs. 64 percent). Comparable data from the Laredo Study showed 77 percent of MA hypertensive women and only 37 percent of hypertensive men under adequate control (17). While the figures for Mexican American women are similar to national figures for hypertension control in women, Mexican American men lag far behind national figures for hypertension control in men. 377 3. Cigarette smoking. Six studies have compared smoking behavior in Mexican Americans and non-Hispanic Whites (18-20,26-28). Overall rates of current smoking appear to be the same or higher in Mexican Americans than in non-Hispanic Whites. Overall rates, however, obscure an important sex-ethnic interac- tion: for males, the proportion of Mexican Americans who are current smokers is the same or higher than for non-Hispanic whi- tes; for women, on the other hand, the proportion of current smo- kers is much lower among Mexican Americans than among non-Hispanic Whites (19,27,28). Regardless of sex, however, Mexican Americans smoke significantly fewer cigarettes per day than non-Hispanic Whites (18,27,28). SES differences in smoking behavior were examined in three stu- dies (20,27,28). One study which included only women found that in both Mexican Americans and non-Hispanic Whites, the proportion of current smokers decreased with increasing socioeconomic status (27). At comparable levels of SES, though, the proportion of non-Hispanic White smokers was double that of Mexican Americans. Among Mexican Americans, similar proportions of U.S.-born and Mexico-born women were current smokers. The central California study (20) examined the proportion of heavy cigarette smokers (> 20 cigarettes per day) among males only. In both Mexican Americans and non-Hispanic Whites, the proportion of heavy smo- kers decreased with increasing socioeconomic status. For Mexican Americans, the proportion ranged from 6.1 percent in the highest SES category to 23.5 percent in the lowest category. For non-Hispanic Whites, rates were somewhat higher, ranging from 16.3 percent in the highest SES category to 48.3 percent in the lowest. Two reports from the San Antonio Heart Study compared SES dif- ferences in smoking behavior in both men and women. In the first report cited, subjects were grouped into two SES categories: low and high. Since no low SES non-Hispanic Whites were available in the SAHS sample, SES comparisons within ethnic groups could be made only for Mexican Americans. For men, the proportion of current smokers was greater among Mexican Americans of both SES levels than among high SES non-Hispanic Whites. For women, on the other hand, the proportion of current smokers was lower among Mexican Americans at both SES levels than among high SES non-Hispanic Whites. In Mexican Americans of both sexes the pro- portion of current smokers decreased markedly in the higher SES group. Within SES levels, the proportion of current smokers was twice as high among Mexican American men than among women. Among high SES non-Hispanic Whites, on the other hand, a slightly higher proportion of women were current smokers than men. High SES non-Hispanic Whites of both sexes smoked almost twice as many 378 cigarettes per day as Mexican Americans in either SES group. Among Mexican Americans, high SES men smoked slightly more cigarettes per day than low SES men (16.8 vs. 13.3), while Mexican American women in the two SES groups smoked about the same number (9.2 cigarettes per day). A second SAHS report focussed on how well-informed individuals are about the role of smoking as a risk factor for heart disease and whether they had modified their smoking behavior (i.e., quit or never started) to reduce heart disease risk (24). Mexican Americans of both sexes tended to be less well-informed about the risks of cigarette smoking than non-Hispanic Whites. High SES Mexican Americans, however, were much better informed on this issue than low SES Mexican Americans; ethnic differences in knowledge decreased at the higher SES levels. Within the same SES level, Mexican American men tended to be better informed about the risk of cigarette smoking than women (in the low SES group: 10% of men vs. 6% of women; in the middle SES group, 16 vs. 18%; and in the high SES group, 31 vs. 26% of women). For men, about equal proportions of Mexican Americans and non-Hispanic whites in the middle SES group reported modifying their smoking behavior to reduce CHD risk (11% of MAs vs. 12% of non-Hispanic Whites); but more non-Hispanic Whites than Mexican Americans were likely to report prevention-oriented behavior in the high SES group (18% of MAs vs. 24% of non-Hispanic Whites). A similar pattern was found for women; in the middle SES group, 10% of MAs vs. 11% of non-Hispanic Whites reported modifying their smoking behavior compared to 16% of MAs and 20% of non-Hispanic Whites in the high SES group. Low SES Mexican Americans of both sexes lagged far behind in prevention-oriented behavior only 6% in both sex groups reported modifying their behavior to reduce CHD risk. 4. Exercise. Three studies have compared the level of phy- sical exercise in Mexican Americans and non-Hispanic Whites (19,24,27). All three reported a lower level of physical exer- cise outside of work in Mexican Americans than in non-Hispanic Whites. Only one study (24) examined differences in physical exercise by level of SES. Although non-Hispanic Whites were generally better informed about the role of exercise in reducing risk of CHD and reported more prevention-oriented behaviors than Mexican Americans, both knowledge and preventive-behavior increased among Mexican Americans with increasing SES. For women, while only 21 percent of low SES Mexican Americans were informed about the role of exercise in reducing CHD risk, 37 per- cent of middle SES women and 52 percent of high SES women were informed on this issue. Similarly, the proportion of Mexican 379 American women reporting that they engaged in physical exercise to reduce risk of heart disease increased from 13 percent in the low SES group to 21 percent in the middle group, and 33 percent in the high SES group. Even in the highest SES group, though, MA women lagged substantially behind non-Hispanic Whites in both knowledge and preventive behaviors. For men, the proportion of Mexican Americans informed about exercise and CHD rose from 22 percent in the low SES group to 40 percent in the middle group, and 66 percent in the high SES group. This proportion in high SES Mexican American men equalled that observed in high SES non-Hispanic Whites. The proportion of Mexican American men reporting that they exercised to reduce CHD risk increased from 18 percent in the low SES group to 28 percent in the middle SES group, and 50 percent in the high SES group. Again the propor- tion of high SES Mexican American men reporting this prevention- oriented behavior equalled the proportion among high SES non-Hispanic Whites. On the four risk factors examined in this SAHS report (dietary fat, blood pressure, cigarette smoking, and exercise), Mexican American men ranked highest on exercise in their level of knowledge and prevention-oriented behavior. 5. Obesity. The available evidence clearly indicates that obesity is a major health problem in Mexican Americans, par- ticularly among women and those of low socioeconomic status (17,18,21,27,30-32). In the Laredo study of very low SES Mexican Americans, the age-adjusted prevalence of obesity (defined as being 20 percent or more over desirable weight) was 25.8 percent in men and 44.8 percent in women (17). Corresponding figures derived from the Health and Nutrition Examination Survey (HANES I) for the general U.S. population were only 15.6 percent for men and 29.0 percent for women (33). Another study of very low SES Mexican Americans in Starr County, Texas, found that the preva- lence of obesity (defined as body mass index > 30) was typically 30 percent or higher in adults of both sexes (30). Data from the San Antonio Heart Study found that body mass index decreased only slightly in Mexican American men with increasing SES, but decreased markedly and significantly in women at each higher SES level. Ethnic comparisons within SES-matched strata, however, still indicated greater adiposity among Mexican Americans of both sexes than among non-Hispanic Whites (21). Comparisons of attitudes and behavior related to obesity in middle and high SES subjects revealed that a greater percentage of Mexican Americans than non-Hispanic Whites at each SES level were likely to express the belief that Americans are too con- cerned about losing weight (34). Mexican Americans in both SES groups also scored lower than non-Hispanic Whites on scales measuring sugar avoidance and dieting behavior, but the ethnic 380 differences were statistically significant only for women in the middle SES group. Mexican Americans in the high SES group scored significantly higher on these scales than those in the middle SES group. In non-Hispanic Whites SES differences were statistically significant only for men and only for the dieting behavior scale. Only one report, also from the San Antonio Heart Study, examined the relationship between acculturation and obesity in Mexican Americans (35). Acculturation was measured using a multi- dimensional scale which reflected adoption during adulthood of non-Hispanic White behaviors, attitudes, and values. Even after adjusting for socioeconomic status, increased levels of acculturation were found to be associated with significant decreases in body mass index in Mexican Americans of both sexes. This finding suggests that SES and acculturation may play inde- pendent and distinct roles in reducing obesity in Mexican Americans. Mechanisms (e.g., diet, exercise, stress, attitudes toward obesity) which may mediate the SES-obesity and acculturation-obesity relationships are in need of further investigation. In addition to the role played by overall adiposity, there is evidence that the distribution, or patterning, of body fat may be an important determinant of some metabolic disorders (such as diabetes) which may be related to heart disease. Limited evi- dence, based on subscapular and triceps skinfolds, suggests that Mexican Americans have a relatively more central distribution of body fat compared to non-Hispanic Whites (32,36). There is also some data to suggest that the centrality of the fat distribution in Mexican American men increases with increasing SES, while for Mexican American women centrality decreases (32). Interest has been expressed recently as well in the hypothesis that persons with upper body obesity have a greater propensity toward metabo- lic disorders (e.g. diabetes) than persons with lower body obe- sity. While no data are currently available on ethnic differences in upper and lower body obesity between Mexican Americans and non-Hispanic Whites, the Starr County study has reported that Mexican American diabetics have relatively more upper body fat and less lower body fat than non-diabetics (30,31). Data needed to examine ethnic difference in upper vs. lower body fat distribution across SES levels are currently being collected in the San Antonio Heart Study II. 6. Diabetes. A recent review of current research on diabe- tes mellitus in Hispanics indicated that the prevalence of insulin-dependent diabetes mellitus (IDDM) is less than 5 percent and found no evidence for increased prevalence of IDDM in Mexican Americans relative to non-Hispanic Whites (37). Non-insulin 381 dependent diabetes mellitus (NIDDM), on the other hand, appears to be a major health problem in Mexican Americans, particularly those of low socioeconomic status (17,21,37-41). In Laredo, Texas, the age-adjusted prevalence of NIDDM in low SES Mexican Americans was 15.7 percent in males and 16.1 percent in females (17,37); corresponding rates for low SES Mexican Americans in Starr County, Texas were 8.3 percent in both sexes (38). Comparisons to the general U.S. population made in both studies suggest that low socioeconomic status Mexican Americans are at substantially greater risk of NIDDM. The San Antonio Heart Study compared the prevalence of NIDDM (defined according to the National Diabetes Data Group (NDDG) criteria (42)) in Mexican Americans and non-Hispanic Whites across three SES levels (21). In the lowest SES group the age- adjusted prevalences for Mexican Americans were comparable to those reported in the Laredo study: 13.7 percent in males and 14.8 percent in females. NIDDM prevalence declined in both sexes and in both ethnic groups with increasing levels of socioeconomic status. In the middle SES group the age-adjusted prevalence for males was 14.6 percent in Mexican Americans and 7.0 percent in non-Hispanic Whites; the corresponding rates for females were 8.2 percent in Mexican Americans and 7.3 percent in non-Hispanic Whites. In the highest SES group the rates for males dropped to 6.1 percent in Mexican Americans and 3.6 percent in non-Hispanic Whites; for females the corresponding rates fell to 3.7 percent in Mexican Americans and 2.2 percent in non-Hispanic Whites. Further analyses of the data for Mexican Americans indicates that the inverse relationship between socioeconomic status and NIDDM may be largely mediated through obesity in MA women, but not in men (41). In addition, when SES and acculturation were examined separately in relation to NIDDM prevalence, acculturation was consistently associated with a reduction in NIDDM prevalence for both MA men and women, even after linear adjustments for socioeconomic status. The effect of acculturation on NIDDM appeared to be largely mediated through obesity in women, but was largely independent of obesity in men (42). Since Mexican Americans tend to be much more obese than non-Hispanic Whites, efforts were made to determine whether obe- sity could explain the excess prevalence of diabetes. Even when Mexican Americans were closely matched to non-Hispanic Whites on degree of obesity, the prevalence of NIDDM was significantly greater in Mexicans than in non-Hispanic Whites (40). The sum- mary prevalence ratio, after matching for obesity, was 2.54 for men and 1.70 for women. While obesity contributes to NIDDM in Mexican Americans, particularly Mexican American women, it does 382 not entirely explain the excess prevalence rate that has been observed relative to non-Hispanic Whites. There is limited evidence to suggest that the excess NIDDM preva- lence in Mexicans which cannot be attributed to obesity may have a genetic component related to degree of native American admix- ture. In the San Antonio Heart Study NIDDM prevalence in Mexican Americans was intermediate between the rates observed for the predominantly non-Hispanic White HANES II population (3.1%), which presumably has few if any native American genes, and for Pima Indians (49.9%), who are believed to have close to 100 per- cent native American genes (39). NIDDM prevalence for non-Hispanic Whites in the SAHS sample was nearly identical to that observed for the HANES II population (3.0%). In addition, the stepwise decline in NIDDM prevalence from the low to high Mexican American SES groups observed in the San Antonio Heart Study was paralleled by a similar stepwise decline in percent native American admixture (39,43). Low SES Mexican Americans had about 46 percent native American admixture, MAs middle SES Mexican Americans about 26 percent native American admixture, and high SES Mexican Americans about 18 percent native American admixture. It should be kept in mind, however, that this evi- dence is currently ecological in nature, although more definitive studies are now in progress. Summary and Recommendations This review confirms that the profile of cardiovascular risk fac- tors is less favorable in Mexican Americans than in non-Hispanic Whites. Within the Mexican American population, those in the lowest socioeconomic group and those who are the least acculturated have a significantly worse profile of cardiovascular risks than those in higher socioeconomic groups and those who are more acculturated. Obesity and non-insulin dependent diabetes mellitus rank as major health problems in Mexican Americans of both sexes. For Mexican American males, hypertension and cigarette smoking rank as significant health problems, as well. Finally, the less favorable profile of lipids and lipoproteins in Mexican Americans (greater hypertriglyceridemia and lower levels of high density lipoprotein cholesterol in both sexes; and greater hypercholesterolemia and higher levels of low density lipoprotein cholesterol in males) relative to non-Hispanic Whites should also be recognized as a public health concern. In light of these findings, the following recommendations seem warranted: 1 Major priority should be given to public health ini- tiatives directed toward reducing the high rates of obesity and non-insulin dependent diabetes mellitus in the Mexican American 383 population. Toward this end, research monies should be appropriated to determine whether genetic as well as lifestyle factors are related to excess NIDDM prevalence in Mexican Americans. 2. Educational programs and improvements in health care delivery should be directed toward increasing compliance with antihypertensive regimens and reducing the rate of current smoking in Mexican American males. 3. Health care deliverers should be alerted to the need for careful monitoring of lipids and lipoproteins in Mexican Americans associated with increased risk of coronary heart disease. This seems particularly important in low SES Mexican Americans of both sexes and in upwardly mobile Mexican American males. 4. Public health initiatives to lower the profile of coro- nary risk in Mexican Americans should be directed primarily toward those at lower levels of socioeconomic status. However, since Mexican Americans at all SES levels tend to lag behind non-Hispanic Whites in levels of knowledge and behavior related to the prevention of coronary heart disease, educational programs designed to reduce CHD risk should be directed toward Mexican Americans at all socioeconomic levels. 384 Table 1. Research Reports on Cardiovascular Risk Factors in Mexican Americans Researchers Locale ex 1. Stern et al (18) 2. Roberts and Lee (27) 3. Kraus et al (20) 4. Friia et al (19) 5. Stern et al (17, 25) Three semi-rural communities in northern California Alameda County, CA Sacramento, CA Orange County, CA Year Rink Factors 1972 Lipids and Lipoproteins Dietary Fat Blood Pressure Cigarette Smoking 1974 Cigarette Smoking and Exercise 1975 Obesity _5*5Pi.? Characteri^stics_ Examined Differences in Acculturation or SES: Yes or No? Laredo, TX 1975- Lipids and Lipoproteins 1976 Blood Pressure Cigarette Smoking 1978 Lipids and Lipoproteins Blood Pressure Cigarette Smoking Exercise Obesity Glucose Levels 1979 Lipids and lipoproteins Hypertension Overweight 299 Mexican American and 1367 non-Hispanic White men and women, aged 35-59, residing in households randomly sampled in each community MAs: predominantly lower SES; non-Hispanic Whites: predominantly middle SES 417 MA, 382 Black, and 2638 non-Hispanic White men and women, aged 20-65+, residing in house- holds randomly sampled from the county; Mixed lower to upper SES; SES structure for MAs significatnly lower than that for non-Hispanic Whites 19,141 employed male volunteera, aged 35-57 738 MAs, 923 Blacks, 818 Asians, and 16,412 non-Hispanic Whites; Five levels of SES, ranging from lower to upper; SES structure for MAs substantially lower than for non-Hispanic Whites 450 MA and non-Hispanic men and women who were incoming government employees of Orange County County in 1978; age ranges not reported; Mixed lower and middle SES 127 MA men and 262 MA women, aged 40-47, residing in household* randomly sampled from two low SES cenaua tracts No No; SES used as a control variable in ethnic comparisons Yes—SES only No 6. Stern et al (21) 7. Hasuda et al (24) 8. Haffner et al (23) 9. Haffner et al (22) 10. Franco at al (26) San Antonio, TX* San Antonio, TX* San Antonio, TX* San Antonio, TX* San Antonio, TX* 1979- 1982 1979- 1981 1979- 1981 1979- 1982 1979- 1982 Lipids and Lipoproteins Obesity Diabetes Dietary Fat and Cholesterol Blood Pressure Cigarette Smoking Exercise Dietary Fat and Cholesterol Dietary Factors Related to Lipida and Lipoproteins 1288 MA and 929 non-Hispanic Whit* men and women, aged 25-64, residing in households randomly sampled from three neighborhooda; Neighborhoods were purposively selected to represent three distinct levels of SES and acculturation of MAs; low, middle, and high 932 MA and 382 non-Hispanic White men and women, aged 25-64, residing in households randomly sampled from three neighborhoods; Neighborhoods were purposively selected to represent three distinct levels of SES and acculturation of MAs: low, middle, and high Same as Research Report 7; sample sise varies slightly (MAs: 915; non-Hispanic Whites: 388) Same as Research Study 6; sample site variea slightly (MAs: 1241; non-Hispanic Whites: 923) Hypertension Same as Research Report 6 "••—SIS and Acculturation Combined Yea—SES and Acculturation Combined Yes—SES and Acculturat ion Combined Yes—SES and Acculturat ion Combined and Separately Yea—SES and Acculturation Combined Table I. Continued Researchers Locale 11. Haffner et al San Antonio, TX* (29) 12. Hoick et al (28) 51 U.S. counties in four states along the U.S.- Mexico border 13. Malina et al San Antonio, TX* (32) 1979- 1982 1979 1979- 1981 Risk Factors Cigarette Smoking Cigarette Smoking Obesity and Fat Patterning SamplejCharscteristic«__ 1605 MA and non-Hispanic White men and women, aged 25-64, residing in households randomly selected from three neighborhoods; Neighborhoods were purposively selected to represent three distinct levels of SES end scculturation of MAs: low, middle, and high 1255 MA, 798 non-Hispanic White, and 82 other ethnic group women, aged 15-44, residing in a sample of households; Three SES levels and two levels of acculturation represented among MAs Same aa Research Report 7; Sample sice variea slightly (MAs: 934; non-Hispanic Whites: 394) Examined Differen in Acculturation SES: Yes or No? Yes--SES snd Acculturstion Combined Yes —SES and Acculturation Separately Yes—SES and Acculturation Combined 14. Mueller et al Starr County, TX* (30) 15. Stern et al San Antonio, TX* (34) 16. Hasuda et al San Antonio, TX* (35) 1981- Obesity and Fat 1983 Patterning 1979- Obesity-related Attitudes 1980 Knowledge, and Behavior 1979- Obesity 1982 Glucose Intolerance 1155 MA men and women, ages 15-70*. residing in households randomly selected from three densely populated towns in the County; SES is extremely low, 52X of familiea were classified below the poverty level in the 1970 Census 454 MA and 395 non-Hiapanic White men and women, ages 25-64, residing in a random aample of households in two neighborhoods; Neighborhoods were purposively selected to represent two distinct levels of SES and acculturation of MAs Same as Research Report 6, MAs only No Yes—SES and Acculturation Combined Ye*—SES and Acculturatioa Separately 17. Stern et al San Antonio, TX* (40) 1979- 1981 Obesity and Diabetes Same aa Research Report 7 Yea—SES and Acculturation Combined 18. Jooa et al Starr County, TX* (31) 19. Gardner et al San Antonio, TX* (39) 1981- 1983 Obeaity and Diabetea 20. Hasuda and Haffner (41) 21. Hani* et al (38) San Antonio, TX* Starr County, TX* 1979- Diabetes 1981 1979- Diabetes 1981 1981- Diabetea 1983 59 male and 109 female diabetica, agea 30-69, and the same number of randomly aelected matched non-diabetic controla; a subset of the sample in Research Report 14 Same as Research Report 7 Same aa Reaearch Report 6; MA ethnic group only 2498 MA Men and women ages 15-70+, residing in households randomly sampled from three towns in the county; SES is extremely low Yes—SES and Acculturation Combined Yea—SES and Acculturation Combined Yea—SES and Acculturation Separately No *Data preaented in theae studies were all collected as part of the San Antonio Heart Study, a population-based study of diabetea and cardiovascular risk factors in Mexican Americana and non-Hispanic Whites. References 1. Hernandez J, Estrada L, Alvirez D: Census data and the problem of conceptually defining the Mexican American popu- lation. Social Science Quarterly 1973; 53:671-687. 2. Siegel JS, Passel JS: Coverage of the Hispanic population of the United States in the 1970 census: a methodological analysis. Current Population Reports, Special Reports, Series P-23, 1979, No. 82. 3. Hayes-Bautista DE: Identifying "Hispanic" populations: the influences of research methodology upon public policy. Am J Public Health 1980; 70:353-356. 4. Haub C: The U.S. Hispanic population: a question of defi- nition. Intercom 1981; 9:8-11. 5. Hayes-Bautista DE: On comparing studies of different Raza populations. Am J Public Health 1983; 73:274-276. 6. U.S. Bureau of the Census. 1980 Census of Population. Volume 1. General Population Characteristics: United States Summary, PC80-1-B1. Washington, DC: U.S. Govt. Printing Office. Issued May 1983. 7. Jaffe AJ, Cullen RM, Boswell TD: The Changing Demography of Spanish Americans. New York: Academic Press, 1980. 8. Marden CF, Meyer G: Minorities in American Society (2nd ed.), New York: McGraw-Hill, 1962. 9. Krech D, Crutchfield RS, Ballachey EL: Individual in Society. New York: McGraw-Hill, 1962. 10. Prosser MH: The Cultural Dialogue: An Introduction to Intercultural Communication. Boston: Houghton Mifflin, 1978. 11. Schreiber JM, Homiak JP: Mexican Americans. Pp. 264-336 in Harwood A (ed.): Ethnicity and Medical Care. Cambridge, Mass: Harvard University Press, 1981. 12. Schoen R, Nelson W: Mortality by cause among Spanish- surnamed Californians, 1969-1971. Social Science Quarterly 1981; 62:259-74. 13. Bradshaw BS, Fonner E Jr.: The mortality of Spanish- surnamed persons in Texas, 1969-71. In: Bean FD, Frisbie WP, eds. The Demography of Racial and Ethnic Groups. New York: Academic Press, 1978; 261-82. 14. Stern MP, Gaskill SP: Secular trends in ischemic heart disease and stroke mortality from 1970 to 1976 in Spanish- surnamed and other white individuals in Bexar County, Texas. Circulation 1978; 58:537-43. 15. Kautz JA, Bradshaw BS, Fonner E: Trends in cardiovascular mortality in Spanish-surnamed, other white, and black per- sons in Texas, 1970-1975. Circulation 1981; 64:730-735. 16. Rosenthal M, Hazuda HP, Stern MP: Trends in ischemic heart disease (IHD) mortality among Mexican Americans (MAs) in 387 Texas. Abstract submitted to the 25th Conference on Cardiovascular Disease Epidemiology of the American Heart Association. March 7-9, 1985, Tucson, Arizona. 17. Stern M, Gaskill SP, Allen CR, Garza V, Gonzales JL, Waldrop RH: Cardiovascular risk factors in Mexican Americans in Laredo, Texas. I. Prevalence of overweight and diabetes and distribution of serum lipids. Am J Epidemiol 1981; 113:546-55. 18. Stern MP, Haskell WL, Wood PD, Osann KE, King AB, Farquhar W: Affluence and cardiovascular risk factors in Mexican-Americans and other whites in three northern California communities. J Chronic Dis 1975; 28:623-36. 19. Friis R, Nanjundappa G, Prendergast TJ Jr, Welsh M: Coronary heart disease mortality and risk among Hispanics and non-Hispanics in Orange County, California. Public Health Rep. 1981; 96:418-22. 20. Kraus JF, Borhani NO, Franti CE: Socioeconomic status, eth- nicity, and risk of coronary heart disease. Am J Epidemiol 1980; 111:407-414. 21. Stern MP, Rosenthal M, Haffner SM, Hazuda HP, Franco LJ: Sex difference in the effects of sociocultural status on diabetes and cardiovascular risk factors in Mexican Americans: The San Antonio Heart Study. Am J Epidemiol (in press). 22. Haffner SM, Knapp JA, Stern MP, Hazuda HP, Rosenthal M, Franco L: Coffee consumption, diet, and lipids. Am J Epidemiol (forthcoming: July 1985). 23. Haffner SM, Gaskill SP, Hazuda HP, Gardner LI, Stern MP: Saturated fat and cholesterol avoidance by Mexican Americans and Anglos: The San Antonio Heart Study. Clinical Research 1982; 30:237A. 24. Hazuda HP, Stern MP, Gaskill SP, Haffner SM, Gardner LI: Ethnic differences in health knowledge and behaviors related to the prevention and treatment of coronary heart disease. The San Antonio Heart Study. Am J Epidemiol 1983; 117:717-28. 25. Stern MP, Gaskill SP, Allen CR Jr, Garza V, Gonzales JL, Waldrop RH: Cardiovascular risk factors in Mexican Americans in Laredo, Texas. II. Prevalence and control of hypertension. Am J Epidemiol 1981; 113:556-62. 26. Franco U, Stern MP, Rosenthal M, Haffner SM, Hazuda HP, Comeaux PJ: Prevalence, detection and control of hyperten- sion in a bi-ethnic community: The San Antonio Heart Study. Am J Epidemiol (in press). 27. Roberts RE, Lee ES: Health practices among Mexican Americans: further evidence from the human population laboratory studies. Prev Med 1980; 9:675-88. 28. Hoick SE, Warren CW, Rochat RW, Smith JC: Lung cancer mor- 388 tality and smoking habits: Mexican-American women. Am J Public Health 1982; 72:38-42. 29. Haffner S, Rosenthal M, Hazuda H, Stern M: Cigarette smoking (SMK) in Mexican Americans (MA) and Anglo Americans (AA): The San Antonio Heart Study. Paper presented at the 111th annual meeting of the American Public Health Association, Dallas, TX, November 13-17, 1983. 30. Mueller WH, Joos SK, Hanis CL, Zavaleta AN, Eichner J, Schull WJ: The diabetes alert study—growth, fatness, and fat patterning, adolescence through adulthood in Mexican-Americans. Am J Phys Anthropol 1984; 64:389-399. 31. Joos SK, Mueller WH, Hanis CL, Schull WJ: Diabetes alert study: weight history and upper body obesity in diabetic and non-diabetic Mexican American adults. Ann Hum Biol 1984; 11:167-71. 32. Malina RM, Little BB, Stern MP, Gaskill SP, Hazuda HP: Ethnic and social class differences in selected anthropo- metric characteristics of Mexican American and Anglo American adults: The San Antonio Heart Study. Human Biology 1983; 55:867-883. 33. National Center for Health Statistics. Overweight adults in the United States. Advance data. Vital and Health Statistics. 1979; 51:1-9. 34. Stern MP, Pugh JA, Gaskill SP, Hazuda HP: Knowledge, atti- tudes, and behavior related to obesity and dieting in Mexican Americans and Anglos: The San Antonio Heart Study. Am J Epidemiol 1982; 115:917-28. 35. Hazuda H, Haffner S, Stern M, Rosenthal M, Franco L: Effects of acculturation and socioeconomic status on obesity and glucose intolerance in Mexican American men and women. Am J Epidemiol 1984; 120:494. 36. Garn SM: Ten-state nutrition survey, subscapular and tri- ceps fatfold data, total U.S.A. Center for Human Growth and Development, University of Michigan, Ann Arbor. No date. 37. Stern MP: Diabetes in Hispanic Americans. In Harris MI, Hamman R (eds). Diabetes Data—Compiled 1983. Washington, D.C: U.S. Govt. Printing Office (in press). 38. Hanis CL, Ferrell RE, Barton SA, Aguilar L, Garza-Ibarra A, Tulloch BR, Garcia CA, Schull WJ: Diabetes among Mexican Americans in Starr County, Texas. Am J Epidemiol 1983; 118:659-72. 39. Gardner LI Jr, Stern MP, Haffner SM, Gaskill SP, Hazuda HP, Relethford JH, Eifler CW: Prevalence of diabetes in Mexican Americans. Relationship to percent of gene pool derived from native American sources. Diabetes 1984; 33:86-92. 40. Stern MP, Gaskill SP, Hazuda HP, Gardner LI, Haffner SM: Does obesity explain excess prevalence of diabetes among Mexican Americans? Results of the San Antonio Heart Study. 389 Diabetologia 1983; 24:272-7. 41. Hazuda H, Haffner S: Acculturation as a protective factor against diabetes in Mexican Americans: San Antonio Heart Study. Diabetes 1984; 33(Suppl.l):30A. 42. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose tolerance. Diabetes 1979; 28:1039-57. 43. Relethford JH, Stern MP, Gaskill SP, Hazuda HP: Social class, admixture, and skin color variation in Mexican Americans and Anglo Americans living in San Antonio, Texas. Amer J Phys Anthropol 1983; 61:97-102. 390 Ischemic Heart Disease Risk Factors in Hispanic Americans Shiriki K. Kumanyika, Ph.D. M.P.H. Department of Epidemiology Johns Hopkins School of Hygiene and Public Health Baltimore, Maryland Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland TABLE OF CONTENTS 1.0 INTRODUCTION.................394 2.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG MEXICAN AMERICANS AND HISPANIC-HERITAGE POPULATIONS IN THE SOUTHWEST .... 395 2.1 Elevated Blood Pressure and Hypertension 2.2 Cholesterol 2.3 Cigarette Smoking 3.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG PUERTO RICANS . . .405 3.1 Data Issue 3.2 Ischemic Heart Disease Risk Patterns 4.0 SUMMARY...................407 5.0 REFERENCES..................408 6.0 SUPPLEMENTARY REFERENCES ............. 410 393 1.0 INTRODUCTION This paper presents selected data on three of the standard ischemic heart disease (IHD) risk factors (hypertension, elevated cholesterol, and cigarette smoking) among Hispanic Americans. In addition to the material presented here, the Cardiovascular Diseases Subcommittee report analysis of IHD risk among Hispanics draws upon information compiled by Hector F. Meyers, Ph.D., one of the subcommittee's consultant writers, and upon the commissioned paper by Hazuda on socioeconomic factors and CVD (1). However, in their entirety, the data available at the time of the Task Force deliberations do not provide an adequate picture of IHD risk patterns among Hispanics--either in the aggregate, according to ethnic-regional subgroups, or according to sociocultural and socioeconomic distinctions which may have a direct bearing on IHD risk. Some mortality, morbidity and risk factor data which relate to Hispanics are inaccesible in that they are grouped within data for "non-whites". "Spanish-surname" tabulations are ambiguous as to the homogeneity of the population so classified, and the ethnicity of "Spanish-surname" populations must be inferred on a geographical basis. For example, according to Aday et al. (2), 83% of a southwest Spanish surname sample drawn for a 1975-76 national survey were Mexican American persons (Colorado, Texas, New Mexico, Arizona, and California were the sample states). In the New York City area, most Spanish-surname persons are Puerto Rican. In Florida, most are Cuban Americans. Other Spanish-surname persons have origins in Central or Latin America or other Spanish-speaking areas of the world. The numbers of Hispanic respondents in health surveys of the overall United States population (e.g., the National Health and Nutrition Examination Survey (NHANES) and the National Health Interview Survey (NHIS)) have been small, proportional to the relatively small numbers of Hispanics in the population. However, the Hispanic Health and Nutrition Examination Survey (HHANES), to be completed in 1985, has been specially designed to provide data on a range of health status variables for sizeable, representative samples of Mexican Americans, Puerto Ricans, and Cuban Americans (3). HHANES will become the appropriate reference point for many IHD risk factor considerations related to Hispanic populations after late 1985 or early 1986. Most of the information in this paper relates to Mexican Americans and Hispanics in the Southwest who are presumed to be primarily Mexican Americans. This information is presented, by risk factor, in section 2. 394 Section 3 includes relatively brief comments related to the assessment of IHD risk among Puerto Ricans. Essentially no relevant data on Cuban Americans were identified. 2.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG MEXICAN AMERICANS AND HISPANIC-HERITAGE POPULATIONS IN THE SOUTHWEST 2.1 Elevated Blood Pressure and Hypertension Kraus et al. (4) reported the prevalence of elevated blood pressure (DBP >=95 mm Hg) for Hispanic men screened at the University of California Multiple Risk Factor Intervention Trial (MRFIT) center during 1975-76. The approximately 19,000 screenees included 738 Spanish American men: 340 and 398 in the 35-44 and 45-57 year age-groups, respectively. In the white comparison group, a higher proportion of white men were in the 45-57 year age-group compared to the Spanish American men. The prevalences of elevated blood pressure among Spanish American men and white men are compared in Table 1, overall and according to categories of a 77 point education-occupation scale of socioeconomic status. Table 1: Prevalence of elevated blood pressure among Spanish American and white men in the U.C. Davis MRFIT screenee population, overall and according to socioeconomic status (SES). Elevated blood pressure is defined as diastolic blood pressure > or = 95 mm Hg) (from 4; Table 3) SES Spanish White American % % all SES 15.6 11.7 11-21 high SES 8.2 9.0 22-32 12.3 10.8 33-43 11.5 12.0 44-54 17.4 13.1 55-65 19.1 14.7 66+ low SES 19.6 16.1 The data in Table 1 indicate an overall higher prevalence of elevated blood pressure among the Spanish American men. High blood pressure prevalence increases with decreasing socioeconomic status among the men in both ethnic groups. The excess prevalence among the Spanish American men is most evident in the lower SES strata. The black men in this population had the highest prevalence of elevated blood pressure of all the ethnic groups reported (18.8% prevalence), but did not show the SES gradient. 395 Data for Hispanic men and women in California relating to several hypertension variables are included in the report of the California Hypertension Survey. This survey, conducted in 1979, used a multistage, random, probability sample of households. The Hispanic sample included 1,647 persons. The survey results on hypertension prevalence, awareness, treatment and control for Hispanic men and women in two age-groups are shown in Table 2. Table 2. Hypertension variables for Hispanic men and women- California Hypertension Survey (from reference 5; Tables 4.3 and 4.4.) Comparison data for white respondents in the same survey are shown in brackets. H: Lspanic Hispanic Hispan ic Hispanic Men 18- ■49 Men 50+ Women 18-49 Women 50+ % hyperten 14.1 39.4 7.6 44.7 (140/90+ or [ 16.3 45.6 6.4 46.7 ] on meds) %140/90+ 13.9 35.7 6.9 29.4 [ 15.0 38.5 4.8 36.4 ] % hyperten 28 64 45 68 aware [ 42 60 61 70 ] % hyperten 6 38 20 55 on drug Rx [ 15 40 41 52 ] % hyperten 1 7 8 27 controlled [ 7 15 26 21 ] Among men, the prevalence of hypertension by either of the criteria shown is somewhat lower among Hispanic vs. whites in both age-groups. Awareness, treatment, and control are lower among Hispanic men compared to white men in the younger age-group. Among the older men (aged 50 or over), hypertension awareness is slightly higher among Hispanics but treatment and control are somewhat lower among Hispanics, compared to whites. Among women, the prevalence of hypertension is higher in Hispanic women than in white women in the younger age-group but somehwat lower in the older age-group. Younger Hispanic women are less aware and less likely to be treated and controlled than their white counterparts. For older women, the treatment and control appear to be somewhat better among Hispanic than among white women. 396 In comparisons of overall hypertension rates among the ethnic groups surveyed, adjusting for differences in the age, sex, and body weight distributions of the populations being compared, the prevalence of hypertension (systolic blood pressure >140 mm Hg or diastolic blood pressure >90 mm Hg or on medication) was 21.5% among Hispanics, compared to 23.4% among whites and 35.2% among blacks (5; Table 4.2). The adjusted mean levels of systolic and diastolic blood pressure among Hispanics and whites were not markedly different (e.g., Hispanics mean DBP was 74.9 compared to 76.4 for whites). Additional probability sample data which permit comparisons of blood pressure levels among Mexican Americans and whites have been reported from cycle 1 of the NHANES. These comparison data are shown in Table 3, for men, and Table 4 for women. Table 3. Systolic and diastolic blood pressure levels of Spanish"/ Mexican American men, with comparison values for white men, U.S., 1971-1974, NHANES I. (from reference 6; Tables 25 and 27). SYSTOLIC(mm Hg) DIASTOLIC (mm Hg) AGE- Spanish/ Spanish/ GROUP Mex American White Mex American White Men Men Men Men 18-24 119.1 123.7 73.4 76.4 25-34 121.6 125.2 78.4 80.8 35-44 124.7 127.0 81.8 84.2 45-54 140.1 134.7 86.5 87.5 55-64 139.9 139.6 86.8 86.4 65-74 146.0 146.0 82.4 84.9 * examinees who identified themselves as Spanish Americans or Mexican Americans 397 Table 4. Systolic and diastolic blood pressure levels of Spanish"/ Mexican American women, with comparison values for white women, U.S., 1971-1974, NHANES I. (from reference 6; Tables 25 and 27). SYSTOLIC(mm Hg) DIASTOLIC (mm Hg) AGE Spanish/ Spanish/ GROUP Mex American White Mex American White Women Women Women Women 18-24 109.6 115.1 67.1 71.3 25-34 117.2 116.2 74.4 74.6 35-44 122.4 122.6 78.3 79.3 45-54 130.0 131.1 83.7 82.6 55-64 144.8 143.0 85.8 86.2 65-74 150.1 151.6 81.6 85.4 " examinees who identified themselves as Spanish American or Mexican American. Among men in the NHANES I sample (Table 3), systolic blood pressures of the Spanish/Mexican American men were lower than those of white men between the ages of 18 and 44 but comparable to or higher than those of white men in the 45-74 year age-groups. Diastolic blood pressures of the Hispanic men were comparable to or somewhat lower than those of their white counterparts. Among women (Table 4), systolic and diastolic blood presures of the Spanish/Mexican American women were generally similar to those of white women, slightly lower in several of the age groups. Christensen et al. (7) reported blood pressure levels of Mexican American and Anglo students (mean age 16 years) and parents (mean ages 46 and 43 years for fathers and mothers) from the Houston-Baylor Prevalence Study. Data are for the 1972-75 period. The number of Mexican American parents who participated was small. Data for Mexican Americans were reported for 165 male and 169 female students but only 44 fathers and 80 mothers. Among Anglos there were 500 to 600 in each of the four categories (i.e., male and female students and parents). Mean systolic and diastolic blood pressure levels for Mexican American and Anglo fathers were almost identical (e.g.,mean systolic blood pressures were 126.6 mm Hg and 126.9 mm Hg for Mexican American and Anglo fathers, respectively. Mean systolic and diastolic blood pressures of Mexican American and Anglo male students were similar also--120.2 mm Hg systolic for Mexican American boys vs. 118.7 mm Hg for white boys. Blood pressure levels of the Mexican American mothers were higher than those of the Anglo mothers--125.1 mm Hg systolic for Mexican American mothers compared to 118.2 mm Hg systolic for Anglo mothers; 80.5 mm Hg diastolic vs. 76.7 mm Hg diastolic for Mexican American mothers vs. Anglo mothers. Blood pressure 398 levels of the Mexican American girls were approximately 2 mm Hg lower than those of Anglo girls (109.6 vs. 111.3 mm Hg systolic and 68.6 vs. 70.4 mm Hg diastolic). The large standard deviations and the lack of adjustment for age or body weight limit the ability to interpret these small differences. Stern et al. (8) reported blood pressure levels from a 1979 cardiovascular survey of Mexican Americans in Laredo, Texas. Eighty-eight percent of 462 age-eligible Mexican Americans participated (the age range was 40 to 74 years of age). Mean systolic and diastolic blood pressure levels of the "Laredo Project" men and women were consistently lower than the NHANES I levels for black men and women in all sex-age-groups (men and women between 40 to 74 years). Compared to NHANES I data for whites and Spanish/Mexican Americans (shown in Tables 3 and 4 above), mean systolic blood pressure levels for Laredo Project men and women were lower in some age-sex groups and higher in other groups (8; Tables 1 and 2). The prevalence of hypertension (i.e., medication for hypertension and/or uncontrolled high blood pressure taken together) among Mexican Americans was estimated to be higher than among whites but lower than among blacks, using estimates for the Hypertension Detection and Follow-up Program (HDFP) as the comparison data—but only in the 40 to 59 year olds. Hypertension prevalence rates among Mexican Americans in the 60-69 year age-group were equal to the high levels reported for blacks. Hypertension prevalences for Laredo Mexican American men ages 40-49 and 60-69 years were 29 and 38% respectively, compared to estimates of 20.6% and 18.5% respectively for HDFP white men and 39% and 36.3%, respectively for HDFP black men. The same data for women--Laredo: hypertension prevalences of 21.1% and 44.3% for 40-49 and 60-69 year old Mexican American women, compared to 18.0% and 23.6% estimated for HDFP white women and 40.8% and 42.1% for HDFP black women (8, Table 4). Stern et al. (8) note that this intermediate blood pressure prevalence status of Mexican Americans in Laredo (between levels for blacks and whites) is similar to the findings of Kraus et al. (4) among the MRFIT screenees. Hypertension awareness, treatment and control levels were greater among Laredo women than men in every age category. In addition, Laredo project women appeared to have better awareness, treatment, and control than white or black women in a comparison group (Impact of Hypertension Information (IHI) Program; reference 9). Laredo Project men were somewhat less likely than white or black men in the IHI population to be aware, treated or controlled. Overall, these data on blood pressure levels and hypertension prevalences among Hispanics are very difficult to interpret. Variances around mean levels reported are often large and relevant tests of statistical significance are not reported. Comparison data for some reports are taken from external sources; thus situational or methodologic differences across studies may affect results. The SES and weight effects on blood pressure may contribute to the discrepancies in the results. With these considerations in mind, it seems safe to infer that the problem of hypertension among Mexican Americans and Hispanics in the southwestern United States is generally not of the same magnitude reported for U.S. black populations overall, although it may be substantial, and greater than for whites, among certain subgroups. 399 2.2 Cholesterol Kraus et al. (4) reported a 16.7% prevalence of elevated serum cholesterol (defined as > or = 260 mg/dl) for the Spanish American men in the MRFIT screenee population described above, compared to 13.9% prevalence among the white men. Prevalences of elevated cholesterol by socioeconomic status subgroup are shown in Table 5. The prevalence of elevated cholesterol among Spanish American men is higher than for white men in all SES classes. The direction or nature of the SES-cholesterol relationship, if any, within these two ethnic groups is not clear, although a U-shaped relationship is suggested. Table 5: Prevalence of elevated serum cholesterol among Spanish American and white men in the U.C. Davis MRFIT screenee population, overall and according to socioeconomic status (SES) (from 4; Table 3) SES Spanish American White % % all SES 16.7 13.9 11-21 high SES 20.4 12.5 22-32 18.5 14.1 33-43 17.2 14.4 44-54 14.0 13.9 55-65 17.8 15.4 66+ low SES 19.6 16.1 The Houston-Baylor Prevalence Study data reported by Christensen et al. (7) also compared plasma cholesterol levels of Mexican American and Anglo parents and children. There were no differences between Mexican American vs. white students of either sex. Among the adults, the Mexican American men had cholesterol levels higher than either white or black fathers (Mexican American fathers: 218 mg/dl; white fathers: 210 mg/dl; black fathers: 209 mg/dl). Mexican American mothers had the lowest cholesterol levels of the three groups (Mexican American mothers: 192 mg/dl; white mothers: 198 mg/dl; black mothers: 200 mg/dl). As noted earlier when reporting the blood pressure findings from this study, the differences reported cannot be interpreted in light of the large standard deviations and lack of age or relative weight adjustments. It should also be noted that the overall cholesterol values in Houston were notably lower than values reported in national surveys. Even with consideration of methodological differences in cholesterol determinations, Houston values were probably 15mg% lower than in NHANES I. 400 Fortmann et al. (10) report plasma cholesterol levels at baseline in the Stanford Three Community Study and report changes in cholesterol levels after the 1972-1975 intervention period. The three communities were semi-rural agricultural communities. The Hispanic persons surveyed were classified as either bilingual or Spanish-speaking. Five classes of the Hollingshead SES index are reported for stratification of results; however, the Hispanic persons are almost all confined to the two lowest categories (70 out of 80 bilingual persons and 51 out of 52 Spanish-speaking, with 45 of the Spanish-speaking in the lowest category). The only useful comparison seems to be of baseline plasma cholesterol levels across ethnic groups within the lowest category (class 5). Within SES class 5, plasma cholesterol levels were lowest in the Spanish-speaking group (199 mg/dl), intermediate in the bilingual group (210 mg/dl), and highest among Anglo persons (220 mg/dl) (10; Table 1). Stern et al. (11) compared levels of serum cholesterol in the Laredo, Texas population (described earlier) to cholesterol levels of the Lipid Research Clinics' population (comparison data were from reference 12). The values reported are shown in Table 6. Table 6: Serum cholesterol levels of Laredo Project Mexican American men and women with comparison data for white men and women in the Lipid Research Clinics (LRC) population (from 11; Table 5). Mexican Mexican American American Age Men White Men Women White Women Mean cholesterol levels (mg/100 ml) 40-44 241.8 206.5 215.5 194.5 45-54 220.4 212.4 218.5 210.9 55-64 224.9 213.6 234.6 227.2 65-74 207.1 210.9 238.6 228.5 The prevalence of obesity in this Mexican American population was intermediate between levels reported for U.S. whites and Pima Indians. Prevalence of hyperglycemia was also intermediate between whites and Pima Indians and this was apparently not due to differences in the % of diabetics controlled. With the exception of 65-74 year old men, the serum cholesterol levels of Mexican Americans were somewhat higher than those of the LRC whites. Stern et al. (13) reported total and HDL cholesterol and related variables in Mexican and Anglo men and women in San Antonio Texas (San Antonio Heart Study). The major purpose of this report was to look at effects of cultural and socioeconomic status on these variables. However, ethnic comparisons were briefly discussed. The prevalence of diabetes 401 among Mexican American men was twice that of Anglo men. It was also somewhat higher in Mexican American women vs. Anglo women. The prevalence of obesity was consistently higher in Mexican Americans vs. Anglos--in both sexes. HDL cholesterol levels were consistently lower in Mexican Americans in both sexes. No consistent patterns of ethnic differences in total cholesterol were observed. Overall, no consistent pattern of Hispanic-white cholesterol differences can be inferred from these data. As noted earlier regarding findings on blood pressure, some of the inconsistencies may be due to weight and SES differences across comparison groups. 2.3 Cigarette Smoking Samet et al. (14) reported smoking data from a survey in Bernillo County, New Mexico. The survey was conducted around 1980. Data were collected through mailed questionnaires supplemented by telephone and home interviews to increase response rates. Response rates were 75-78% for Anglo men and women; 60 to 69% for Hispanic men and women. The Hispanic sample was of lower SES than the Anglo sample. Thus, it is not possible to separate ethnic and socioeonomic effects in the smoking data. Comparable smoking patterns (current, former, never smoked) were observed in Hispanic and Anglo samples except that among 40-59 year old Anglo men the proportion of ex-smokers was higher than the proportion of ex-smokers among Hispanic men. The cumulative cigarette exposure was lower among Hispanics due to fewer average cigarettes smoked, not shorter smoking duration. Smoking data for Hispanics (unspecified) and whites are given in an Advance Data report from the National Center for Health Statistics (15). Preventive health practices were surveyed in a 1977 National Health Interview Survey. Age-, sex-, and ethnic-specific tables are not given. Data are shown in Table 7. A larger proportion of Hispanics than whites had never smoked, although the percentages of current smokers were similar. A substantially larger percentage of Hispanic than white smokers reported smoking less than 15 cigarettes per day. 402 Table 7. Reported smoking status of Hispanics and whites in a 1977 National Health Interview Survey (from reference 15) Never Smoked % White 43.1 % Hispanic 54.1 Former Current Smoker Smoker 21.7 35.2 12.3 33.5 Average # cigs/day for smokers <15 >25 25.2 29.8 59.1 13.6 Roberts and Lee (16) reported smoking data from the Human Population Laboratory Studies (1974 and 1975). Households in Alameda County, California census tracts were sampled (Table 8). In the 1975 sample, which was much smaller than the 1974 sample, there is an impression that Chicanos smoke less than whites. Smoking patterns of the two groups appear comparable in the 1974 data. Multivariate analyses controlling for age and education or income clarified the ethnic differences such that the authors concluded that Chicano rates of "never smoking" were higher than for Anglos, Table 8: Smoking behavior of Chicanos and Anglos in Alameda County in two survey samples (from 16; Table 2) n current former never 1974 White 2292 34.3 22.5 43.2 Chicanos 162 36.0 16.8 47.2 1975 White 346 45.1 17.7 37.2 Chicanos 225 36.2 16.5 47.3 Smoking patterns among Spanish American and white MRFIT screenees in California are shown in Table 9. Spanish American men smoke less than white men in all SES strata. The prevalence of smoking 20 or more cigarettes per day decreases with increasing SES (4). Table 9: Percentages of Spanish American and white U.C. Davis MRFIT screenees smoking 20 or more cigarettes per day, overall and by socioeconomic status (from 4; Table 3). SES Spanish American White all SES 20. .7 11-21 high SES 6. ,1 22-32 10, ,8 35-43 21, ,7 44-54 22, .0 55-65 25, .7 66+ low SES 23 .5 30.2 16.3 25.6 32.6 36.7 42.2 48.3 404 Smoking status data for Hispanic men and women in the California Hypertension Survey population are shown in Table 10. Rates of current and former smoking are similar among Hispanic and white men, although Hispanic men smoke fewer cigarettes per day. Among women, all smoking variables are lower among the Hispanic women. Hispanic men who smoke may be less likely to want to quit than white men, in both age-groups. Younger Hispanic women smokers appear to be more likely to want to quit than white women in this age-group. Table 10: Cigarette Smoking Variables among Hispanic American men and women in the California Hypertension Survey, 1979 (from reference 5; Table 7.5). Comparison data for whites are shown in brackets Men Men Women Women 18-49 50+ 18-49 50+ % current regular 38.8 34.0 21.3 20.2 smokers [ 34.6 30.5 36.5 25.9 % current or 58.1 73.6 37.1 37.2 former smokers [ 58.0 72.6 52.2 48.7 average cigarettes 17.7 16.7 9.7 16.4 per day among [ 26.5 28.3 23.4 23.2 current smokers % of current smokers 66.3 50.8 72.3 57.9 who would like to quit [ 71.0 59.0 63.8 60.8 Overall, the smoking data are relatively consistent in showing lesser smoking prevalences and less heavy smoking among Hispanics when compared to whites, particularly when SES factors are controlled. 3.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG PUERTO RICANS 3.1 Data Issue A recently published paper by Rosenwaike and Preston (17) questioning the validity of age-reporting in some data for Puerto Ricans may be of interest regarding the interpretation of mortality patterns for this group. The authors note that recent official statistics show a longer life expectancy among both males and females in Puerto Rico compared to that for males and females in the United States (and all but a few other countries). They suggest, however, that there appears to be strong evidence of widespread overstatement of age among Puerto Ricans at the older ages, resulting in an underestimation of mortality rates. 405 Table 1 in the Rosenwaike and Preston paper (17) notes life expectancies after age 45 (1969-71) of 30.3 for Puerto Rican men vs. 27.5 and 24.6 for mainland white and non-white men. For Puerto Rican women the life expectancy after age 45 years is 34.2 years vs. 33.5 and 29.8 for mainland whites and non-whites. The authors present several graphs and tables which demonstrate an excess of "age-heaping" (tendency to report ages with terminal digits of zero and five) in the distribution of ages among Puerto Ricans over age 40. Also, there is some inconsistency in the data from successive censuses regarding the numbers of persons at different ages. In discussing this analysis, the authors note that an overstatement of one year would introduce a 10% error and two years a 20% error. 3.2 Ischemic Heart Disease Risk Nearly all of the papers on IHD in Puerto Ricans are for the San Juan area. Gordon et al. have published two reports (18,19) comparing men in Framingham, Honolulu, and Puerto Rico. Some baseline data from this comparison are shown in Table 11. The Puerto Rican men have lower levels in all IHD risk factor and morbidity categories except diabetes. Table 11: Baseline data for Puerto Rican and Framingham men on IHD risk variables (from reference 18; Table 2) FRAMINGHAM PUERTO RI Relative weight (%) 120.04 112.50 SBP (mm Hg) 136.41 130.35 DBP (mm Hg) 85.07 82.19 Serum cholesterol (mg%) 233.96 202.45 # cigarettes/day 12.26 percent of group 7.63 diabetes by Hx 3.7 5.3 LVH-ECG 2.3 1.5 Smokers 57.7 44.4 Angina Pectoris 3.8 1.3 Coronary Insuffici ency 0.5 0.5 Myocardial Infarct ion 1.2 0.1 - difference is 25mg% for Puerto Ricans after adjustment for laboratory differences) At two-year follow up (18), Framingham men were reported to have a two-fold greater risk of coronary heart disease (CHD) than men in Puerto Rico, even after adjustments for blood pressure and cholesterol. The differences in rates between Puerto Rican men and Framingham rates were small or nonexistent in younger men and larger in the older men. The apparent absence of a smoking effect on CHD among Puerto Rican men was noted as an important exception to the overall pattern of similar risk factor/outcome associations across ethnic groups (19). Host risk 406 factor/outcome associations were generally similar in all three cohorts in spite of lower overall risk in the Honolulu Japanese and Puerto Rican cohorts compared to Framingham. However, a later analysis of the Puerto Rican Heart Program data using 8 year incidence (which gave a larger number of incident cases) established an association of smoking with increased incidence of myocardial infarction in both urban and rural Puerto Rican men (20). This association was not found with other CHD endpoints. 4.0 SUMMARY Available data on patterns of IHD risk among Mexican Americans and Puerto Ricans are limited. Data relating to Cuban Americans appear to be lacking. Data from the HHANES, which will describe all three groups, are not yet available. Concurrent data for whites will not be available for comparison with HHANES; however, older data from NHANES II are available for comparison. A comparison of age-adjusted heart disease mortality in predominantly Hispanic census tracts in Los Angeles with rates of Los Angeles whites, gives the impression that Hispanics are at lower risk of dying of heart disease than whites (21). Levels of smoking appear to be lower in Hispanic populations. Available data on other IHD risk factors do not clearly establish whether levels for Hispanics are lower, higher, or variable in relation to those of whites. Analyses are needed which examine risk factor patterns within cultural and SES subgroups of Hispanics and also adjust for levels of overweight and diabetes. ACKNOWLEDGEMENTS We gratefully acknowledge the editorial assistance of Sandra J. Anderson and Elisabeth Pitt. 407 5.0 REFERENCES 1. Hazuda HP. Differences in socioeconomic status and acculturation among Mexican Americans and Risk of Cardiovascular Disease. Background paper prepared for the Secretary's Task Force on Black and Minority Health. 1984. 2. Aday LA, Chiu GY, Andersen R. Methodological issues in health care surveys of the Spanish heritage population. Am J Pub Health 1980;70:367-374. 3. Transcript of remarks by Fernando M. Trevino, Ph.D., M.P.H. to the Secretary's Task Force on Black and Minority Health. National Institutes of Health. Bethesda, MD. August 8, 1984. 4. Kraus JF, Borhani NO, Franti CE. Socioeconomic status, ethnicity, and risk of coronary heart disease. Am J Epidemiol 1980;111:407-414. 5. Igra A, Stavig GR, Leonard AR, with special contributions by Gamber W, Hawthorne A, Jang VL, Ransom B. Hypertension and Related Health Problems in California. Results from the 1979 California Hypertension Survey. Hyertension Clearinghouse. Department of Health Services. Hypertension Control Program. 714 P. Street. Sacramento, California 95814. 6. U.S. DHEW. California Conference on High Blood Pressure Control in the Spanish-Speaking Community. Summary Report. Conference date April 1-2, 1978. NIH Publication No. 79-1959. August 1979. 7. Christensen BL, Stallones RA, Insull W, Gotto AM, Taunton D. Cardiovascular risk factors in a tri-ethnic population. Houston, Texas 1972-1975. J Chron Dis 1981;34:105-118. 8. Stern MP, Gaskill SP, Allen CR, Garza V, Gonzales JL, Waldrop RH. Cardiovascular risk factors in Mexican Americans in Laredo, Texas. II. Prevalence and Control of Hypertension. Am J Epidemiol 1981;113:556-562. 9. Apostolides AY, Cutter G, Kraus JF, Oberman A, Blaszkowski T, Borhani NO, Entwisle G. Impact of hypertension information on high blood pressure control between 1973 and 1978. Hypertension 1980;2:708-713. 10. Fortmann SP, Williams PT, Hulley SB, Maccoby N, Farquhar JW. Does dietary health education reach only the privileged. The Stanford Three Community Study. Circ 1982;66:77-82. 11. Stern MP, Gaskill SP, Allen CR, Garza V, Gonzales JL, Waldrop RH. Cardiovascular risk factors in Mexican Americans in Laredo, Texas. I. Prevalence of overweight and diabetes and distribution of serum lipids. Am J Epidemiol 1981;113:546-555. 408 12. The Lipid Research Clinics Program Epidemiology Committee. Plasma lipid distribution in selected North American Populations: Lipid Research Clinics program prevalence study. Circulation 1979;60:427-439. 13. Stern MP, Rosenthal M, Haffner SM, Hazuda HP, Franco LJ. Sex differences in the effects of sociocultural status on diabetes and cardiovascular risk factors in Mexican Americans. Am J Epidemiol 1984;120:834-851. 14. Samet JM, Schrag SD, Howard CA, Key CR, Pathak DR. Respiratory disease in a New Mexico population sample of Hispanic and Non-Hispanic whites. Am Rev Respir Dis 1982;125:152-157. 15. National Center for Health Statistics. Health Practices Among Adults. United States, 1977, by Schoenborn CA, Danchik KM. Advancedata. No 64, November 4, 1980. 16. Roberts RE, Lee ES. Health Practices among Mexican Americans. Further evidence from the human population laboratory studies. Prev Med 1980;9:675-688. 17. Rosenwaike I, Preston SH. Age overstatement and Puerto Rican longevity. Human Biology 1984;56:503-525. 18. Gordon T, Garcia-Palmieri MR, Kagan A, Kannel WB, Schiffman J. Differences in coronary heart disease in Framingham, Honolulu, and Puerto Rico. J Chron Dis 1974;17:329-344. 19. Gordon T, Kagan A, Garcia-Palmieri M, Kannel WB, Zukel WJ, Tillotson J, Sorlie P, Hjortland M. Diet and its relation to coronary heart disease and death in three populations. Circ 1981;63:500-515. 20. Sorlie PD, Garcia-Palmieri MR, Costas R, Cruz-Vidal M, Havlik R. Cigarette Smoking and Coronary Heart Diseaes in Puerto Rico. Prev Med 1982;11:304-316. 21. Frerichs RR, Chapman JM, Maes EF. Mortality due to all causes and to cardiovascular disease among seven race-ethnic populations in Los Angeles, County 1980. Int J Epidemiol 1984;13:291-298. 409 6.0 SUPPLEMENTARY REFERENCES Ramirez EA, Garcia-Pont PH. Relationship of serum cholesterol to socioeconomic and dietary factors in Puerto Rican veterans. Dis Chest 1969;55:197-201. Garcia-Palmieri MR, Tillotson J, Cordero E, Costas R, Sorlie P, Gordon T, Kannel WB, Colon AA. Nutrient intake and serum lipids in urban and rural Puerto Rican men. Am J Clin Nutr 1977;30:2092-2100. Garcia-Palmieri MR, Sorlie P, Tillotson J, Costas R, Cordero E, Rodriguez M. Relationship of dietary intake to subsequent coronary heart disease incidence. The Puerto Rico Heart Health Program. Am J Clin Nutr 1980;33:1818-1827. Costas R, Garcia-Palmieri MR, Sorlie P, Hertzmark E. Coronary heart disease risk factors in men with light and dark skin in Puerto Rico. Am J Pub Health 1981;71:614-619. Health of Minorities and Women. Chartbook. American Public Health Association. 1015 Fifteenth Street, N.W. Washington, D.C. August 1982. Stock No. 072. Health of the Disadvantaged. Chartbook II. DHHS Publication No. (HRA)80-633. September 1980. Frerichs RR, Chapman JM, Nourjah P, Maes M. Cardiovascular Diseases in Los Angeles. 1979-1981. Hansen VE, Ed. American Heart Association. Greater Los Angeles Affiliate, Inc. 1984 edition. Chapman JM, Frerichs RR, Maes EF. Cardiovascular Diseases in Los Angeles. 1980. Hansen VE, Ed. American Heart Association. Greater Los Angeles Affiliate, Inc. 1983 edition. U.S. DHHS. Hispanic Health Services Research. NCHSR Research Proceedings Series. Conference held September 5-7, 1979. DHHS Publication No. (PHS) 80-3288. Hayes-Bautista DE. Identifying "Hispanic" populations. The influence of research methdology upon public policy. Am J Pub Health 1980;70:353-356. Roberts RE, Lee ES. The health of Mexican Americans. Evidence from human population laboratory studies. Am J Pub Health 1980;70:375-384. Benson H, Costas R, Garcia-Palmieri MR, Feliberti M, Aixala R, Blanton JH, Colon AA. Coronary heart disease risk factors. A comparison of two Puerto Rican populations. Am J Pub Health 1966;56:1057-1060. 410 Costas R, Garcia-Palmieri MR, Nazario E, Sorlie PD. Relation of lipids, weight, and physical activity to incidence of coronary heart disease. The Puerto Rico Heart Study. Am J Cardiol 1978;42:653-658. Garcia-Palmieri MR, Costas R, Schiffman J, Colon AA, Torres R, Nazario E. Interrelationships of serum lipids with relative weight, blood glucose, and physical activity. Circ 1972;45:829-836. Garcia-Palmieri MR. Precursors of coronary artery disease in Puerto Rico. Am J Clin Nutr 1973;26:1133-1137. Garcia-Palmieri MR, Sorlie P, Tillotson J, Costas R, Cordero E, Rodriguez M. Relationship of dietary intake to subsequent coronary heart disease incidence. The Puerto Rico Heart Health Program. Am J Clin Nutr 1980;33:1818-1827. Caetano R. Drinking patterns and alcohol problems among Hispanics in the U.S. A Review. Drug and Alcohol Dependence 1983;12:37-59. National Center for Health Statistics. Blood Pressure Levels of Persons 6-74 years, United States 1971-1974, By Roberts J and K Maurer. Vital and Health Statistics, Series 11, No. 203. DHEW Publication No. (HRA) 78-1648. Caetano R. Self-reported intoxication among Hispanics in Northern California. J Stud Alcohol 1984;45:349-354. Gordon AJ. The cultural context of drinking and indigenous therapy for alcohol problems in three migrant Hispanic cultures. J Stud Alcohol 1981;(Supp)9:217-240. Ross CE, Mirowsky J. Social epidemiology of overweight. A substantive and methodological investigation. J Health Soc Behavior 1983;24:288-298. Stern MP, Gaskill SP, Hazuda H, Gardner LI, Haffner SM. Does obesity explain excess prevalence of diabetes among Mexian Americans? Results of the San Antonio Heart Study. Diabetologia 1983;24:272-277. Ailinger RL. Hypertension knowledge in a Hispanic community. Nurs Res 1982;31:207-210. Hazuda HP, Stern MP, Gaskill SP, Haffner SM, Gardner LI. Ethnic differences in health knowledge and behaviors related to the prevention and treatment of coronary heart disease. Am J Epidemiol 1983;117:717-728. 411 Nader PR, Baranowski T, Vanderpool NA, Dunn K, Dworkin E, Ray L. The Family Health Project. Cardiovascular risk reduction education for children and parents. Develop and Behavioral Pediatrics 1983;4:3-10. Stern MP, Pugh JA, Gaskill SP, Hazuda HP. Knowledge, attitudes, and behavior related to obesity and dieting in Mexican Americans and Anglos. The San Antonio Heart Study. Am J Epidemiol 1982;115:917-928. Ramirez AG, Herrick AG, Herrick KL, Weaver FJ. El asesino silencioso. A methodology for alerting the Spanish-Speaking community. Urban Health 1981;10:44-48. Weaver FR, Herrick KL, Ramirez AG, Deatrick DA. Establishing a community data base for cardiovascular health education programs. Health Values 1978;2:249-256. U.S. DHHS. Proceedings of the Conference on Communicating with Mexican Americans. Por Su Buena Salud. (Comunicando Con Mexico Americanos. For Their Good Health. Conference dates September 13-14, 1979. Moore TJ, Ramirez AG, Slayton PL, eds. NIH Publication No. 81-1961. June 1981. 412 Ischemic Heart Disease Risk Factors in Asian/ Pacific Islander Americans Shiriki K. Kumanyika, Ph.D., M.P.H. Department of Epidemiology Johns Hopkins School of Hygiene and Public Health Baltimore, Maryland Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland TABLE OF CONTENTS 1.0 INTRODUCTION..................416 1.1 Data Sources 1.2 Approach 2.0 ISCHEMIC HEART DISEASE MORTALITY AMONG ASIAN/PACIFIC ISLANDERS: Trends and Geographic Comparisons ......... 417 3.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG ASIAN/PACIFIC ISLANDERS . 419 3.1 Japanese Americans 3.1.1 Blood Pressure 3.1.2 Blood Cholesterol 3.1.3 Smoking 3.1.4 Overall Impact of Risk Factors on Ischemic Heart Disease in Japanese Americans 3.1.5 Summary 3.2 Chinese Americans 3.3 Filipino Americans 3.3.1 Blood Pressure 3.3.2 Smoking 3.4 Other Asian/Pacific Islanders 4.0 CONCLUSION...................436 5.0 REFERENCES ...............437 415 1.0 INTRODUCTION 1.1 Data Sources The census definition of Asian/Pacific Islanders refers to a set of ethnically and culturally diverse U.S. population subgroups whose origins were in the "Far East, Southeast Asia, or the Pacific Islands' (1). Relatively recent (1980 or later publication date) ischemic heart disease (IHD) or IHD risk studies on Asian/Pacific Islanders identified through the Medline search related almost exclusively to Japanese Americans--drawing primarily on the Ni-Hon-San Study comparisons of Japanese in Hawaii, California, and Japan (2-7) and on the prospective study of the Honolulu cohort of Japanese (Honolulu Heart Study) (8-20). Papers on heart disease mortality in Hawaii and Los Angeles compare mortality rates of several groups (the Hawaii paper includes data for Hawaiian, Filipino, Japanese, and Caucasian men and women (8); the Los Angeles County paper includes data for Japanese, whites, Filipinos, Chinese, and Koreans (21)). A review of disease trends in Pacific Basin countries (22) includes an overview of CHD mortality trends in Japan, China, Hawaii, and the Philippines. A California statewide hypertension survey presents cardiovascular risk factor data (except lipids) for white, Japanese, Filipino, Chinese, and other Asian/Pacific Islander men and women (23,24). Even recognizing the relatively small amount of data available on Asian/Pacific Islander subgroups other than Japanese Americans, the low yield of studies of other Asian/Pacific Islander subgroups from the computerized literature search suggests that the NLM searching protocol for minorities (i.e. MEDLINE heading: "minorities") may need expanding to ensure that Filipinos, Hawaiians, Koreans, and other Asian sub-groups (e.g., Vietnamese) are included. Several informative papers on Chinese, Filipino-, and Hawaiian-Americans (25-36) were identified through ad hoc methods, using the bibliograpy of a recent paper on diabetes mortality among New York City Chinese (Current Contents citation) as the index paper. 1.2 Approach An attempt is made to summarize available data on levels and trends of IHD risk factors among Asian/Pacific Islander subgroups in the United States. Patterns of IHD or heart disease mortality and incidence are presented to provide a perspective in considering the related risk factors (and to some extent as a substitute for risk factor data to give at least a general impression of patterns in various sub-groups). Available information on risk factor levels, trends, and impact in the Asian/Pacific Islander population is then presented by risk factor separately for Japanese, Chinese, and Filipinos. Where possible, comparisons with levels in the U.S. total population or the white population are made. Comparisons between or among Asian/Pacific Islander subgroups are noted where they appear to give insight into differential effects of migration and adaptation among Asian/Pacific Islanders of different origins. 416 2.0 ISCHEMIC HEART DISEASE MORTALITY AMONG ASIAN/PACIFIC ISLANDERS: Trends and Geographic Comparisons A comparison of age-adjusted coronary heart disease (CHD) mortality rates among males in the state of Hawaii between 1940 and 1978 (8) indicates that rates for Caucasians followed very closely the pattern observed for the total population of white men in the U.S. Hawaiian men had the highest rates, Caucasian men second highest, Filipino men third, and Japanese men lowest. Over the 38 year period between 1940 and 1978, rates for Hawaiian and Caucasian men showed almost no net increase--a slight upwards slope changed to a downward trend around 1970. Rates for Hawaiian men were in the range of 3 per 1000 and for Caucasian men in the range of 2 per 1000. The pattern for Filipino men in Hawaii was quite different from that seen in Hawaiian and Caucasian men. CHD mortality among Filipino men increased from less than 0.6 per 1000 in 1948 to slightly more than 2 per 1000 in 1978 with no evidence of the downturn seen in the other three ethnic categories. In a detailed analysis of CHD mortality data for Filipinos in Hawaii, Gerber noted that CHD mortality rates more than doubled during the 20 year period between 1950 to 1970 (1970 rates 2.6 times 1950 rates), and were significantly related to degree of urbanization and to marital status (lowest rates among married males)(27). CHD mortality for Chinese in Hawaii was not reported in the paper by Reed et al. (8). King reported that 1949-1956 CHD mortality rates for Chinese in Hawaii were lower than for Hawaiians or Caucasians but higher than for Japanese (28). More specific data for this period were reported by Bennett et al. (29). Age-adjusted death rates for arteriosclerotic heart disease including coronary disease (ICD 420) for Polynesian (Hawaiian), Chinese, Caucasian, Filipino and Japanese men (ages 35-74 years of age, 1949-1956) were 550, 357, 488, 140 and 161/100,000 respectively (29). Chinese men had rates lower than Caucasians but considerably higher rates than Filipinos and Japanese. Among women, equivalent rates for Polynesians, Chinese, Caucasians, Filipinos, and Japanese were 303, 251, 190, 130, and 94 per 100,000. Thus, Chinese women were at higher risk of mortality than women of all ethnicities other than Hawaiian. Data on the slope of CHD mortality among Chinese men and women in Hawaii since 1956 have not been identified at this writing. Gerber and Madhavan reported 277 CHD deaths among Hawaii Chinese males age 25 or over in 1968-72, but no denominator is given for conversion to a rate (30, Table 1). The remaining data are presented as proportional mortality. CHD mortality rates also increased substantially among Japanese men in Hawaii between 1940 and 1978 (less than 0.3 in 1940 to about 1.5 in 1970) but declined somewhat between 1970 and 1978. Patterns for women in Hawaii were similar to those of the men in the respective ethnic groups, except that rates were lower and downward trends in rates were observed among women in all four groups and began in 1960 (rather than 1970) for Hawaiian and Filipino women. 417 Although the above findings, which were based on Hawaii state death certificate data, indicate a decline in CHD mortality for Japanese men from the late 1960's on, no such decline was indicated in data for the Honolulu Heart Study cohort (8). In noting this discrepancy, Reed and coworkers suggest that the apparent declines indicated in the state mortality records may be artifacts of coding changes or misclassification of other cardiovascular diseases. Los Angeles 1980 age-sex adjusted mortality due to heart disease is shown in Table 1, below; cerebrovascular disease rates are also shown, for comparison. Heart disease mortality is substantially less among Asian/Pacific Islanders vs. the white or overall U.S. populations. Among Asian/Pacific Islanders, heart disease death rates were highest for Japanese, intermediate for Chinese and Korean and lowest for Filipinos. The relative mortality ranks for Japanese, Chinese and Filipinos were the same for both heart and cerebrovascular disease, although Japanese American cerebrovascular disease rates were higher than those for whites or for the general U.S. population. Table 1. Deaths due to heart and cerebrovascular diseases in Los Angeles County, 1980 (from ref 21, Tables 3 and 4) MORTALITY: LOS ANGELES COUNTY, 1980 US White Japanese Filipino Chinese Korean Heart Diseases 309 331 162 58 99 82 Cerebrovascular Dis 69 76 80 20 49 48 " rate per 100,000 age adjusted for age and sex by direct method, Los Angeles County population 1980 as standard. # adjusted for age by direct method, LA county 1980 population as standard. Comparison of the adjusted Hawaii and Los Angeles County death rates is not appropriate. However, the difference in the relative positions of Filipinos and Japanese in Hawaii in 1978 vs. Los Angeles in 1980 is noted. In discussing the Los Angeles data, Frerichs et al. (21) suggest that the recency of migration of the Chinese, Filipino and Korean populations to the Los Angeles area may be responsible for a "healthy migrant" effect on mortality rates in these groups. In the same vein, rates for the more established Japanese population are closer to those of the Los Angeles general population. Gerber and Madhavan have compared proportional mortality due to CHD among Chinese in Hawaii, native and foreign-born Chinese in New York City, and whites in New York City between 1968 and 1972 (30). CHD deaths are proportionately higher in Hawaii Chinese vs. New York City Chinese in every age-group (for population ages 35 and over) with the size of the difference 418 narrowing with increasing age and disappearing in the 75+ age-group. Among the Chinese in New York City, CHD deaths were proportionately greater in U.S. born vs. foreign-born Chinese at all ages (ages 25 and over). CHD mortality represented a lower proportion of deaths among Chinese in both Hawaii and New York City vs. whites in New York City, except in the 25-44 year age-group where proportionate mortality was higher in Hawaiian Chinese. Age at CHD death was later in Chinese populations than in New York City whites and later in foreign-born New York City Chinese than in the other two Chinese subgroups. These findings are consistent with an increasing and earlier CHD risk with increasing U.S. exposure. These limited data on IHD mortality data for Asian/Pacific Islanders indicate generally lower IHD mortality among Asian/Pacific Islander subgroups than among United States whites. Asian/Pacific Islander women appear to have lower risk than men. The subgroup variations in IHD mortality rates and trends among Asian/Pacific Islanders are noteworthy and suggest the need to go beyond generalities to analyses for each specific subgroup. Subgroup differences are related to biocultural factors which are inherently different among groups classified as Asian/Pacific Islanders and also to factors such as place of birth (native-born or immigrant), age of migration, length of U.S. residence, region of residence (e.g., Hawaii, California, New York City), socioeconomic status, and level of acculturation in diet and lifestyle. IHD risk appears to increase with increasing length of U.S. residence and increasing acculturation. 3.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG ASIAN/PACIFIC ISLANDERS 3.1 Japanese Americans 3.1.1 Blood Pressure [The high rate of stroke and potential effect of stroke mortality among Japanese men with elevated blood pressure should be kept in mind when considering the data on blood pressure distributions] A report by Yano et al. gives mean baseline blood pressure levels in 45-54 year old men in the Honolulu Heart Study Cohort (1965) as 130.7/82.0 (systolic/diastolic, mmHg) and the prevalence of elevated blood pressure (>160/95) as 16.7% (37). Gordon reported that mean blood pressure levels of Honolulu Japanese men 45 to 64 years old (note that age span is 10 years longer than that cited from Yano et al.; i.e., more older men in the Gordon data) were similar to those of men in Framingham (mean SBP 136 in both groups). Mean diastolic blood pressure levels were somewhat lower in the Honolulu men compared to Framingham men (82.68 vs. 85.07) (38). A substantially higher prevalence of ECG-LVH in Framingham vs. Honolulu men was noted (2.3 vs. 0.7%) in spite of the similarity of blood pressure levels. The authors note that the readings were not standardized. 419 Comparisons of (1967-1970) systolic blood pressure levels of Japanese men living in Japan, Honolulu and California, reported by Winkelstein et al. (6), resulted in an overall impression that blood pressure levels of home-island Japanese men were intermediate between those of Northern California Japanese men, who had the highest levels, and those of Hawaii Japanese men. The concluding point of their abstract is that . . . the blood pressure distributions of the Japanese populations residing in Japan, Hawaii, and Northern California cannot of themselves account for the observed differences in coronary heart disease and stroke occurrence among these populations in which there is a gradient from high stroke rates in Japan to low rates in California and a reverse trend for coronary heart disease" (6). The specific analyses indicated that Issei Japanese men (men born in Japan who migrated to Hawaii or California) under age 55 in all three areas have similar blood pressures, but that over the age of 55 California Issei men had higher readings than their counterparts in the other two areas. Blood pressures of Issei men over age 55 in California were somewhat higher than those of the Issei Hawaiian men. Diastolic blood pressure levels among Nisei men in California were higher than those of the other cohorts at all ages studied. Systolic blood pressure levels of both Hawaiian and California Nisei men under age 55 were substantially higher than those of the Japanese men in Japan or the Issei men in Hawaii and California. Diastolic blood pressures of the men in Japan and those of first and second generation Japanese men in Hawaii were similar. Systolic blood pressure levels increased with age for all cohorts; diastolic blood pressure levels were unrelated to age in the Japan and Hawaii cohorts and in the Nisei men in California. Appendix A in the Winkelstein et al. paper gives mean blood pressure levels as shown in Table 2, below (however, most of the analyses presented by Winkelstein et al. are based on comparisons of selected percentiles--means are presented here as a convention; standard deviations are in the source Table: SDs range from 16 to 27 for SBP and from 11 to 15 for DBP). 420 Table 2. Baseline systolic and diastolic blood pressure levels of cohorts of 45-64 year old Japanese men in three geographic regions ref.6; Appendix A age 45-49 50-54 55-59 60-64 65-69 JAPAN SBP 126 130 136 140 143 DBP 81 82 85 83 83 HAWAII Issei SBP 131 134 135 137 141 DBP 80 86 82 80 81 Nisei SBP 128 132 134 139 141 DBP 82 82 83 83 81 CALIFORNIA Issei SBP 123 132 139 146 148 DBP 79 84 88 89 89 Nisei SBP 133 137 142 142 146 DBP 88 89 90 89 90 The differences in blood pressure levels among these cohorts were primarily explained by weight. Weight-blood pressure relationships were unstable in the men in Japan; weight and blood pressure were postively associated in the other cohorts and adjustment for relative weight decreased the blood pressure differences between cohorts. Blood pressure levels of the California cohort remained somewhat higher than those of the other two cohorts in most of the comparisons. Mean blood pressure levels among Japanese men in California in 1979 were estimated from the California Hypertension Survey data, based on a multistage probability sample of the state population, with oversampling in primarily Asian American communities (24). The authors of this paper noted that in the 1980 census, the Asian/Pacific Islander population of California comprised 35.8% of the Asians and Pacific Islanders in the United States. The California Hypertension Survey (24) estimate of elevated blood pressure prevalence (>140/90) among California Japanese men ages 18-49 was slightly higher than for white men (19.2 vs. 15.0 %). For men over age 50, the prevalence of elevated blood pressure among Japanese men by this criterion was lower than for white men (29.1 vs. 38.5 %). Hypertension prevalence (140/90) for Japanese women was much lower than for white women 421 at all ages: 0.4% vs. 4.8% among Japanese vs. white women ages 18-49 and 13.9 vs. 36.4% for Japanese vs. white women over age 50. (See Table 9 in section 3.4 for 160/95+ prevalence comparisons with sexes pooled.) The sex differences in hypertension prevalence among Asian/Pacific Islander men and women were reduced 5% by adjustment for age and relative weight and further by adjustment for alcohol intake and other environmental variables. Rates for the men were only 2.3% higher than those of women after adjustment for all significant predictors available (these included psychological, social support, nutrition, smoking, and acculturation factors). In a more detailed report of the survey results, the prevalence of hypertension (i.e., rather than elevated bp) (>140/90 or taking medication) among Japanese men and women was as follows (from Table 4.3 in Reference 23): Japanese American men 18-49: 19.8% (vs. 19.2% with bp > or =140/90) 50+: 32.3% (vs. 29.1% with bp > or =140/90) Japanese American women 18-49: 0.9% (vs. 0.4 % with bp > or =140/90) 50+: 17.5% (vs. 13.9% with bp > or =140/90) The parenthetical data are repeated here to indicate the relative percents of the populations who are hypertensive under control vs. those whose blood pressures were elevated at time surveyed, using the 140/90+ definition. In the California Hypertension Survey (23), levels of hypertension awareness, treatment and control differed depending on the definition of hypertension used. In almost all instances higher percentages of awareness, treatment and control were reported for the groups defined by the more severe criteria (i.e., 160/95+ or 95+). (This also applies to the whites, blacks, and the other Asian/Pacific Islander subgroups surveyed.) Awareness among hypertensive Japanese adults ages 50 and older was 72% (using 140/90 mm Hg) and 94% (using a diastolic blood pressure of 95 or greater mm Hg) in men (60 to 87% in white men) and 60 to 87% in women (white women, 70 and 93%). Awareness levels among Japanese males 18-49 were 35 to 30% (i.e., 35% aware using 140/90 and 30% aware using 95+ mm Hg, a reversed trend); (white males 42 to 62%). Percentage of hypertensives under drug treatment were as follows: 12 to 29% for Japanese men 18-49 (white men 15-38%); (no stable estimates for Japanese women ages 18-49);37 to 61% for males ages 50 and over (white males 40-75%) and 32 to 79% for women ages 50 and over (white women 52-87%). Percentages of hypertensives controlled were 3 to 29% for Japanese men 18-49 (7-32% for white men) and 10-56% for men 50 and over (15-58% for white men). Eight to 38% of Japanese women ages 50 and over were controlled (21-78% of white women). 422 3.1.2 Blood Cholesterol Yano et al. report a mean level of baseline serum cholesterol in the younger segment of the Honolulu cohort (ages 45-54) as 219.4 mg% and a 13.4% prevalence of hypercholesterolemia (>=260 mg%) (37). Comparisons of total cholesterol levels in 45-64 year old Japanese men in the Honolulu cohort with levels of Framingham men were reported by Gordon et al. in 1974 (38). Methodologies for cholesterol determinations were sufficiently similar for Honolulu and Framingham to support direct comparisons. Serum cholesterol levels were approximately 15% lower in Honolulu vs. Framingham (218.55 vs. 233.96 mg%). Baseline comparisons of serum cholesterol levels in Honolulu, Japan, and California were reported by Nichaman et al. (4). The essence of the overall findings was that "in every age-group, mean, median and percentiles for each of the biochemical variables are considerably lower for men in Japan than in California or Hawaii. . . at all ages the cholesterol values in Hawaii are somewhat lower than in California" (4, p. 494) (the other biochemical variables examined were glucose, uric acid, and triglycerides). Mean levels for the three cohorts are shown in Table 3, below. Table 3. Mean serum cholesterol levels at baseline (1967-1970) for Japanese men in Japan, Hawaii, and California (from ref 4, Table 3; mg%)* AGE JAPAN 45-49 179.8 50-54 182.5 55-59 181.5 60-64 182.2 65-69 180.9 AWAII CALIFORNIA 219.4 223.4 219.4 228.2 218.7 226.8 216.7 223.6 211.1 224.0 a footnote to the source table notes that the values from Japan were taken from the 1967 cycle and that diabetics were excluded from this analysis--therefore other published values may differ slightly. Serum cholesterol levels of a sample of men from the Honolulu and San Francisco cohorts were among those analyzed in the Cooperative Lipoprotein Phenotyping Study (9). White comparison groups were from Albany NY, Evans County GA and Framingham. Mean HDL levels reported in Table 1 of that paper are shown in Table 4, below. 423 Table 4. Mean HDL-cholesterol level in samples of white, Japanese, and black men over age 50 by age-group (ref 9, Table 1) GROUP AGE 50-59 60-69 70+ Albany (white) 48.7 49.6 51.4 Framingham ((white) 44.8 45.5 45.5 Evans County (white) 48.3 48.1 50.9 Evans County (black) 59.0 56.3 51.0 San Francisco (Japanese) 45.9 48.8 49.7 Honolulu (Japanese) 44.0 46.0 46.6 The focus of the Cooperative Lipoprotein Phenotyping analysis was the relationship of lipoproteins to CHD. Total cholesterol levels and HDL-C to total-C ratios were not reported as such. This limits the ability to directly evaluate the data in Table 4. However, considering the earlier-cited data indicating similar total cholesterol levels among Framingham and Honolulu men and higher total cholesterol levels in San Francisco vs. Honolulu men, the data in Table 4, above, do not suggest any large difference in proportionate HDL-levels of Japanese American men compared to white men in the same age range. 3.1.3 Smoking The prevalence of cigarette smoking in the Honolulu cohort men ages 45-54 was reported by Yano et al. (37) as 46.4%. The percent of men smoking a pack of cigarettes per day or less was 27.2; 19.2 smoked more than one pack per day (37). In the comparison of 45-64 year old Framingham and Honolulu men reported by Gordon et al. (38), 44.1% of Honolulu men vs. 57.7% of Framingham men were smokers; Honolulu men smoked 10.52 cigarettes on average, vs. a 12.26 average among Framingham men. Robertson et al. (5) reported the following frequency distributions of baseline smoking status for home-island and Honolulu Japanese men ages 45-68 years (ref. 5; Fig. 1): none'" <11 11-20 21+ Japan 24.4 7.9 56.1 11.6 Honolulu 55.8 2.5 25.1 16.6 ( * includes former smokers ) Taken together, these data indicate that Japanese men in Honolulu were less often smokers than Japanese men in Japan and white men in Framingham. Marmot and Syme discuss smoking habits in relation to acculturation of Japanese men in San Francisco (3). The data are based on a survey conducted in 1969-70 (response rate 66% of the men enumerated in San 424 Francisco and Oakland plus additional men in Santa Clara County). The study included 14.6% Issei men. The rest were Nisei (including some Sansei (third generation)) men ages 30 and over. Issei men were the older men in the cohort. Defining traditional vs. nontraditional upbringing according to years spent in Japan, age left parents' home, ever lived on a farm, where schooling took place (Japan or U.S.), years spent in Japanese language school, religion while growing up, friends while growing up, wife s cultural background (if married) (3; pp.230-231)--the following data were reported (3; p.232): % of smokers who smoke: 1-14 15-24 25+ cigarettes per day traditional 20.7 50.5 28.8 nontraditional 20.0 50.6 29.4 Thus, although there may be differences in the percentages of traditional and nontraditional men who smoke, there are no differences in level of smoking among those who do smoke. Also, the authors note a possible interaction between Japanese vs. Western diet preference and smoking among the men surveyed--there was a somewhat higher percent of smokers among men who preferred a Japanese diet. More recent estimates of cigarette use among Japanese Americans in California are available from the 1979 survey results. Overall, 50.6% of the Japanese American men were classified as "ever smoked" (24). A more detailed breakdown of smoking habits in this population is shown in Table 5, below. Table 5. Cigarette Smoking Variables among Japanese American men and women in the California Hypertension Survey, 1979 (from reference 23, Table 7.5) [percentages for comparable white sex-age-group] ^current regular smokers ^current or former smokers average cigarettes smoked per day among current smokers %of current smokers who would like to quit Men Men 18-49 50+ 34.8 17.7 [34.6 30.5 55.0 81.1 [ 58.0 72.6 19.3 19.9 [ 26.5 28.3 80.4 40.2 [ 71.0 59.0 Women Women 18-49 50+ 19.0 18.1 36.5 25.9 ] 47.4 29.5 52.2 48.7 ] 17.6 11.7 23.4 23.2 ] 56.1 5.9* 63.8 60.8 ] *=based on fewer than 20 cases 425 3.1.4 Overall Impact of Risk Factors on CHD in Japanese Americans Earlier, the discrepancy between Hawaii state mortality data and Honolulu Heart Study findings regarding CHD mortality was noted. The Honolulu data did not show declines in CHD mortality in analysis of data for 60 to 69 year old men between 1968 and 1978 using causes coded from medical records review, whereas the data coded according to state mortality codes did show a decline (see Reed et al. (8, Figures 3 and 6)). Reed et al. present total and fatal myocardial infarction (MI) incidence data for 50-59, 60-63, and 64-67 year old men in this cohort in Figure 4 of their report. Incidence rates for total MI increased overall between 1967-70 and 1975-78 for men ages 60-67 and appear to have remained constant in the 56-59 year old men. Fatal MI rates showed a slight increase in the 60-67 year old men with evidence of tapering off after 1971-74; fatal MI rates were constant in the 56-59 year old men. Case-fatality rates for the three age groups are presented in Table 1 of the report: 37, 39, and 36% in men 56-59 during the three time periods; 36, 42, and 33% in 60-63 year old men; and 46, 45, 45 percent in men ages 64-67. Analyses by birth cohort indicate upward slopes of total and fatal MI incidence during this time period in all cohorts of men born between 1900 and 1919. Gordon et al. (38) reported that although the standard risk factor associations were observed among men in the Honolulu cohort, the 2 year CHD incidence (by ECG) or mortality in Honolulu was less than in Framingham by one-half to one fourth. Incidence was 2 times more in Framingham and mortality was 4 times more i.e., the case-fatality in Honolulu was much lower than in Framingham. CHD mortality in Framingham was greater by a factor of 2.1 even after adjustment for differences in levels of blood pressure, smoking, and cholesterol (this finding was not substantially changed by exclusion of men with prior CHD from the analysis). Castelli et al. point out that Honolulu men have CHD incidence half that of Framingham men in spite of the fact that HDL cholesterol levels of the two groups of men are similar (9). Framingham CHD rates were higher than Honolulu rates even in the group with low risk factor levels. Differences in Framingham and Honolulu rates increased in magnitude with increasing age. In a later analysis of six year follow-up data (10) Gordon et al. attempted to identify factors other than blood pressure, smoking and cholesterol which could explain the differences between Framingham and Honolulu CHD rates. Alcohol intake was found to be a strong inverse predictor of MI and CHD death although the authors noted that total mortality was increased with increased alcohol consumption (possibly positive effect of alcohol on blood pressure and/or stroke). Higher starch intake in the Honolulu cohort was also noted as a significant protective factor for CHD. Elsewhere, the possible differential factor of later age of acquisition of risk factors among Japanese migrants vs. U.S.-born men has been noted as a possible explanation for CHD rates less than those predicted with Framingham logistic functions (5). 426 Robertson compared risk factor-CHD associations in Japanese men in Honolulu and Japan (5). Gradients for blood pressure and cholesterol on CHD incidence were similar in both cohorts of men. Smoking was the most significant risk factor in Honolulu but was not related to CHD incidence in Japan (suggesting to the authors that cigarette smoking may be a less important CHD risk factor in men with low serum cholesterol levels). An influence of relative weight on CHD was observed in Honolulu Japanese men but not among the men in Japan (this finding may have been due to the relatively small number of obese men in the Japan cohort (authors' note)). In spite of higher blood pressure levels in the California cohort, as reported by Winkelstein et al. (6), the authors note that a separate study of CHD prevalence in the three cohorts did not find a higher CHD prevalence among the California men (there are several possible explanations for this other than the lack of a true relationship). However, the Honolulu Heart Study prospective data found a substantial impact of blood pressure on CHD at ten-year follow-up (and throughout)(37). In the ten-year Honolulu Study data, the significance of blood pressure levels for total, fatal, and nonfatal CHD, nonfatal MI and acute coronary insufficiency compared with nonCHD was great. In fact, systolic (or diastolic) blood pressure was one of the strongest and most consistently related risk factors of those measured, for all CHD measures except angina pectoris (see Table 6, below). Table 6. Impact of standard risk factors, body mass index and alcohol intake on CHD among Japanese American men in Honolulu at ten-year follow-up (from 37, Table 2 (age-adjusted means for baseline characteristics) SBP DBP CHOL* CIGS# BMI** ALCH! nonCHD 133.2 81.9 217.0 10.2 23.8 14.3 Total CHD 144.7 86.6 229.9 13.0 24.7 9.7 Fatal CHD 151.6 89.0 231.3 14.2 24.8 9.1 nonfatal MI 143.2 86.0 223.0 14.6 24.8 8.3 Acute CI 147.1 87.4 222.7 14.2 24.8 15.6 uncomp AP 137.2 84.0 225.4 7.3 24.5 10.5 * serum chol mg%;#no. per day;** kg/m squared;!ml/day Levels of these variables were significantly different from the nonCHD group for the total CHD cases, for fatal CHD cases (except alcohol consumption), and for nonfatal MI. Acute coronary insufficiency cases differed from the nonCHD only for blood pressure (the authors note that less than valid ascertainment of angina pectoris may be involved in the apparent lack of associations with this CHD manifestation). 427 Multivariate analyses (see 31, Table 4) indicated that, in addition to blood pressure, cigarette smoking was next in line as a strong predictor of CHD in all categories except angina. Alcohol consumption was a strong, independently associated protective factor for both fatal CHD and for nonfatal MI. Serum cholesterol was strongly associated with total CHD and nonfatal MI and significantly, but less strongly related to fatal CHD. Relative weight was not independently associated with any of the CHD manifestations in the multivariate analysis. 3.1.5 Summary These data, although far from adequate in describing the overall IHD risk patterns among Japanese American men and women, can be summarized as follows. Systolic blood pressure levels of Japanese men in the Honolulu Heart Study cohort were similar to those of Framingham men; diastolic blood pressures were somewhat lower. The EKG-LVH prevalence among the Japanese men was lower than expected based on the Framingham model, possibly related to methodology. A consistent gradient in blood pressure and blood pressure related risk is not evident when comparing Japanese men in Japan, Hawaii, and California. U.S. born Japanese men in California have higher blood pressure levels than men in Japan or Japanese-born men in California; these differences appear to be weight-related. When compared to white men in California, Japanese men under age 50 had a hypertension prevalence slightly higher than that of white men whereas Japanese men above 50 had a lower hypertension prevalence than white men. Hypertension awareness was somewhat higher among Japanese men over age 50 compared to white men in that age-group but was lower among Japanese men under age 50 compared to white men. Percentages of Japanese men under drug treatment were somewhat lower than for white men. Percentages of those treated who were controlled were similar among Japanese men and white men. Japanese women in California had substantially lower hypertension prevalence than white women in both the under 50 and 50 and over age groups. Weight and other environmental factors would appear to explain most of the gender differences in hypertension prevalence. Small numbers precluded estimates of hypertension awareness, treatment, and control among Japanese women under age 50. Data for hypertensive women over age 50 indicated lower levels of awareness, treatment, and control compared to white women in this age-group. The limited available data on serum cholesterol levels of Japanese men indicate a consistent gradient in serum cholesterol levels from low levels among Japanese men in Japan to intermediate levels among Japanese men in Hawaii and the highest levels among Japanese men in California. Serum cholesterol levels of Japanese men in all three areas were lower than for men in Framingham. No data on cholesterol levels of Japanese American women were identified. 428 Regarding smoking, the prospective study baseline data suggest that Japanese men in Hawaii were less likely to be smokers than men in Japan or Framingham. The more recent California survey data suggest similar smoking rates among Japanese and white men under age 50, but lower smoking rates among the older Japanese men compared to older white men. Smoking rates among Japanese women respondents in the California survey were lower than for white women, in both the younger and older age-groups. The average number of cigarettes smoked by Japanese men and women was consistently less than for the comparable white age-sex group. Somewhat more of the younger Japanese men wanted to quit smoking compared to the white men in the same age-group; otherwise, the Japanese surveyed were less likely than their white counterparts to state that they wanted to stop smoking. The impact of the standard risk factors on IHD morbidity and mortality, although less than predicted at short-term prospective follow-up, has been in line with expectations upon longer (10-year) follow-up. Thus, at least tentatively, these data indicate that Japanese Americans have IHD risks comparable to those of white Americans to the extent that they acquire comparable risk factor profiles. The lower IHD mortality rates among Japanese Americans may reflect somewhat lower levels of certain IHD risk factors, particularly among the older Japanese men who are primarily Japan-born. 3.2 Chinese Americans Primary findings from the available data sources on risk factors among Chinese Americans are summarized in this section. Refer to the earlier description of CHD mortality data for Chinese Americans (section 2.0). A report by Gerber of higher proportional diabetes mortality among Chinese in New York City compared to whites and other nonwhites in New York City should be noted as well (31). Among 45-64 year old foreign-born Chinese, the proportion of diabetes deaths increased with increased length of residence in New York City. Proportional mortality from diabetes was slightly higher in foreign born vs. U.S. born Chinese above age 65, but slightly lower among foreign-born Chinese ages 45-64. A component of the Hawaii Cardiovascular Study compared cardiovascular risk factors in 30 Chinese and 68 Japanese males who had Mis with those of CHD-free controls in 1966 and 1967 (25). The comparison was prompted by the higher risk of CHD in Chinese men--2.2 times that of the Japanese men. Cases were identified through hospital discharge records. Examinations were at least 3 months post-discharge. Controls were age- and race-matched men identified through hospital records as well as an additional group of population controls randomly selected from a health department survey sample. Absence of CHD in controls was electrocardiographically confirmed. Racially-mixed Chinese and Japanese men were not included in the study. Chinese cases were older than Japanese cases by about 5 years (58.4 vs. 53.2 years of age). Racial comparisons were therefore age-adjusted. Recognizing the inherent possibilities for sample biases among the cases, the emphasis of the analyses was in racial rather than case-control comparisons. Significant differences between the Chinese and Japanese 429 cases which were also seen in the population controls were considered to have greater probable validity than differences which were apparent between Japanese and Chinese controls only. The Chinese men were more obese than the Japanese (skinfolds) and appeared to have gained more weight since age 25, but were not more overweight by a weight/height index (Quetelet). In a follow-up report on the Chinese MI survivors and controls, these authors confirmed that the moderate, "nonoverweight" adiposity among the Chinese men was of pathological significance, i.e., it was related to higher levels of CHD and CHD risk factors (32). Serum cholesterol levels of the Chinese population controls were approximately 20 mg% higher than those of the Japanese controls (242.7 vs. 220.7); higher cholesterol among the Chinese vs. Japanese men was generally consistent in age-specific comparisons although not always statistically significant. Blood pressure levels and LVH patterns were not consistently or significantly different between the two racial groups. A slightly higher proportion of Japanese vs. Chinese cases were smokers (75 vs. 70%) in the immediate pre-MI period; higher proportions of Chinese patients and both groups of controls had never smoked and fewer were current smokers than the respective groups of Japanese men. Physical activity was lower in the Chinese than Japanese men. According to Stavig et al., the Chinese population of California increased by a factor of 2.5 between the 1970 and 1980 censuses (24). Estimates of elevated blood pressure prevalence (> or =140/90) among California Chinese men ages 18-49 was slightly lower than for comparable white men (11.8 vs. 15.0%). For Chinese men over age 50 the prevalence of elevated blood pressure by this criterion was higher, 45.0 vs. 38.5% for white men. Elevated blood pressure prevalence (140/90+) for Chinese women vs. white women was similar: slightly higher in the 18-49 year age-group (6.4 vs. 4.8%) and essentially the same (34.3 vs. 36.4%) for women ages 50 and over (also see Table 9, section 3.4). The overall influence of age, relative weight and other predictive variables on the prevalence of elevated blood pressure among Asian/Pacific Islanders in the 1979 California survey was described earlier (section 3.1.1). Chinese in California were intermediate (between Japanese and Filipinos) in reported alcohol intake, the least likely of the three sub-groups to be foreign-born, and varied in comparability to the other two groups on other blood pressure predictors. In the more detailed report of the survey results, the prevalence of hypertension (rather than elevated blood pressure) (> or =140/90 or taking medication) among Chinese men and women was as follows (from Table 4.3 in reference 23): Chinese American men 18-49: 13.3 % (vs. 11.87% with bp > or =140/90) 50+: 50.3% (vs. 45.0% with bp > or =140/90) Chinese American women 18-49: 7.1% (vs. 6.4 % with bp > or =140/90) 50+: 42.6% (vs. 34.3% with bp > or =140/90) 430 The parenthetical data are repeated here to indicate the relative percents of the populations who are hypertensive under control vs. those whose blood pressures were elevated at time surveyed, using the 140/90+ definition. Levels of hypertension awareness among hypertensive Chinese ages 50 and older in California in 1979 were 58 to 83% in men (white men, 60-87%) and 55 to 98% in women (white women 70-93%), depending on the definition of hypertension used. Awareness levels among Chinese males 18-49 were 32 to 79% (white men 42-62%)(23, Table 4.4). Percentages (of hypertensives) under drug treatment were 21 to 63% for Chinese men 18-49 (white men 15-38%); no stable estimates for Chinese women aged 18-49; 40 to 68% for males ages 50 and over (white males 40-75%); and 52 to 97% for women ages 50 and over (white women, 52-87%). Percentages of hypertensives under control were 11 to 48% (white men 7-32%) for Chinese men 18-49 and 11-59% for men 50 and over (white men 15-58%). Nineteen to 94% of Chinese women ages 50 and over (white women 21-78%) were controlled. Smoking data for Chinese men and women from the detailed California Hypertension report are summarized in Table 7. Table 7. Cigarette Smoking Variables among Chinese American men and women in the California Hypertension Survey, 1979 (from reference 23, Table 7.5) [comparison data for whites] Men Men 18-49 50+ % current regular 26.6 35.2 smokers [34.6 30.5 % current or 42.0 69.8 former smokers [58.0 72.6 average cigarettes 12.5 16.4 per day among [26.5 28.3 current smokers % of current smokers 54.3 59.0 who would like to quit [71.0 59.0 Women Women 18-49 50+ 3.6 9.7 36.5 25.9 ] 7.7 19.2 52.2 48.7 ] 12.0 17.1 23.4 23.2 ] 48.0* 66.5* 63.8 60.8 ] based on fewer than 20 cases 431 In summary, pertinent data which permit interpretation of IHD risk factor levels among Chinese American men and women in the context of levels of risk typical for the U.S. white population are essentially limited to the data from the California survey. The California data indicate higher hypertension prevalence among Chinese men over age 50 compared to white men, but somewhat lower rates among Chinese men compared to whites in the under 50 age-group. Among Chinese and white women in both age-groups the rates of hypertension are similar and are lower than for men. No serum cholesterol data are available from the California survey. The smoking data indicate that more of the older Chinese men, but fewer of the younger men, smoke compared to their white age peers. Substantially fewer Chinese women smoke than white women in both age-groups. At all ages, the Chinese who smoke smoke fewer cigarettes per day. Younger Chinese men and women are less likely than their white counterparts to report wanting to quit smoking. Older Chinese men and women were equally or more likely to report wanting to quit compared to whites. Only proportionate mortality data were identified for the Chinese population in New York City. These data indicate lower IHD risk among Chinese relative to whites, although the proportion of diabetes related deaths is higher among Chinese than among whites. These data are too limited to support any overall conclusions about patterns of IHD risk among Chinese Americans. There does not appear to be any striking evidence of excess IHD risk on the basis of these data; most of the available data suggest less than average risk among Chinese Americans compared to whites. 3.3 Filipino Americans Other than the mortality data for Filipinos reported in section 2.0 above, the only relevant data identified specific to Filipino Americans was from the California Hypertension Survey. Data from that survey for blood pressure and smoking are noted below. No serum cholesterol data for Filipino Americans have been identified at this writing. 3.3.1 Blood Pressure Mean blood pressure levels among Filipino men in California in 1979 were estimated from the California Hypertension Survey data (24). Stavig et al. note that the Filipino population of California tripled between 1970 and 1980. Estimates of elevated blood pressure prevalence (> or =140/90) among California Filipino men and women were higher at all ages and substantially higher than for whites in the age-sex groups with relatively high prevalences: Filipino men 18-49, 29.5% vs. 15.0% for whites; Filpino men ages 50+, 50.8% vs. 38.5% for whites; Filipino women 18-49, 6.5% vs. 4.8% for whites; Filipino women ages 50+, 61.3% vs. 36.4% for whites (24, Table 1) (also see Table 9, section 3.4). 432 The influence of age, relative weight, and other predictors on blood pressure levels of California Asian/Pacific Islanders was noted earlier in the discussion of blood pressure levels of Japanese Americans. Filipinos and other Asian-Pacific Islanders in California reported a greater prevalence of excess weight (higher body mass index), higher average number of alcoholic drinks per sitting, and fewer close friends, were less likely to live in Asian neighborhoods and more likely to be foreign born than Japanese and Chinese-_consistent with the higher prevalence of uncontrolled hypertension in Filipinos vs. Chinese and Japanese. Educational attainment was lower among Filipinos, particularly among Filipino men, vs. the general population or Japanese American males. In the more detailed report of the survey results, the prevalence of hypertension (>140/90 or taking medication) among California Filipino American men and women was as follows (from Table 4.3, ref. 23): Filipino American men 18-49: 30.5% (vs. 29.5% with bp > or =140/90) 50+: 60.0% (vs. 50.8% with bp > or =140/90) Filipino American women 18-49: 6.7% (vs. 6.5 % with bp > or =140/90) 50+: 65.2% (vs. 61.3% with bp > or =140/90) The parenthetical data are repeated here to indicate the relative percents of the populations who are hypertensive under control vs. those whose blood pressures were elevated at the time surveyed, using the 140/90+ definition. Levels of hypertension awareness among hypertensive Filipino men and women ages 50 and older in California in 1979 were 67 to 91% in men (white men 60-87%) and 49 to 83% in women (white women, 70-93%), depending on the definition of hypertension used. Awareness levels among Filipino males 18-49 were 62 to 89% (white males 42-62%) (23, Table 4.4) (stable estimates for women 18-49 were not possible). Percentages under drug treatment were: 52 to 82% for Filipino men 18-49 (white men 15-38%); (no stable estimates for Filipino women aged 18-49); 43 to 76% for males ages 50 and over (white males 40-75%) and 39 to 82% for women ages 50 and over (white women 52-87%). Percentages of hypertensives who were controlled were 3 to 38% for Filipino men ages 18-49 (white men 7-32%) and 15 to 60% for men ages 50 and over (white men 15-58%). Six to 54% of Filipino women ages 50 and over (white women, 21-78%) were controlled. In summary, an excess prevalence of hypertension among Filipinos in California is evident, a noteworthy departure from the picture seen in the other Asian/Pacific Islander populations discussed in this review. Awareness of hypertension was relatively high, particularly among the younger Filipino men compared to white men in the same age-group. Percentages of Filipino men under drug treatment and controlled were comparable to or better than for white men. Among hypertensive Filipino women over age 50, awareness, treatment, and control levels were lower than for white women. 433 3.3.2 Smoking Fewer Filipino than white men and women in the California survey were smokers, in both the under 50 and 50 and over age-groups (as shown in Table 8). Among men and women who smoked, Filipino men and women smoked fewer cigarettes per day. The estimates of percentages of Filipino smokers wanting.to quit are not stable. Table 8. Cigarette Smoking Variables among Filipino American men and women in the California Hypertension Survey, 1979 (from reference 23, Table 7.5) [comparison data for whites] Men Men Women Women 18-49 50+ 18-49 50+ % current regular 30.7 21.2 12.4 16.1 smokers [ 34.6 30.5 36.5 25.9 ] % current or 58.3 67.2 32.9 25.9 former smokers [ 58.0 72.6 52.2 48.7 ] average cigarettes 18.2 17.2 15.0 1.1 per day among [ 26.5 28.3 23.4 23.2 ] current smokers % of current smokers 75.6 90.8* 67.6* 13.2* who would like to quit [ 71.0 59.0 63.8 60.8 ] * based on fewer than 20 cases 3.4 Other Asian/Pacific Islanders CHD mortality data for Koreans in California are included in Table 1 (section 2.0). Blood pressure data for "other Asians" in the California survey are shown in Table 9, below, with data for Caucasian, Japanese, Chinese and Filipinos included for summary and comparison purposes. 434 (from reference 24; Table 1) White Japanese Filipino Chinese Other Asian/PI MALES 18-49 15.0 19.2 29.5 11.8 28.5 weighted % hypert 140/90+ MALES 50+ 38.5 29.1 50.8 45.0 45.2 140/90+ FEMALES 18-49 4.8 0.4 6.5 6.4 3.2 140/90+ FEMALES 50+ 140/90+ 36.4 13.9 61.3 34.3 42.2 BOTH SEXES 20.2 12.5 24.5 15.7 20.1 18+ 140/90+ BOTH SEXES 18+ 160/95+ 6.6 4.3 9.9 5.8 8.2 The overall rate of hypertension for Asian/Pacific Islanders is 18.9% compared to an overall rate of 19.3 for whites. However, this similarity in rates is not a true reflection of the heterogeneous picture evident in Table 9. The high rates of hypertension among the Filipino men and women, the older Chinese men, and the younger Chinese women relative to whites in the same sex-age-groups have been noted. Also, hypertension rates are higher than those of whites among three of the four sex-age-groups of "other" Asian/Pacific Islanders. Additional relevant information is sparse but includes more detailed data on "other Asians" from the California Survey (23), the Kraus et al. analyses of CHD risk factors in several ethnic groups (data are for Asian Americans, not further specified) (39), and data for Hawaiians (Polynesians) from the Hawaii Cardiovascuar Study (33-36). The paucity of data reflects the relatively small numbers of other Asians rather than a lack of need for concern about significant levels of CHD risk. 435 4.0 CONCLUSION In conclusion, the data on IHD risk among Asian/Pacific Islanders as a group and especially for specific subgroups are extremely limited. What data there are indicate that IHD risk levels are generally lower than for whites, but with some noteworthy exceptions. The extent of IHD risk appears to parallel the degree of acculturation. Risks associated with diet and lifestyle adaptations to the U.S. environment may be important contributors to excess risks as they occur. ACKNOWLEDGEMENTS We gratefully acknowledge the editorial assistance of Sandra J. Anderson and Elisabeth Pitt. 436 5.0 REFERENCES 1. Health of Minorities and Women. Chartbook. American Public Health Association. 1015 Fifteenth Street, N.W. Washington, D.C. August 1982. Stock No. 072. 2. Kato H, Tillotson J, Nichaman MZ, Rhoads GC, Hamilton HB. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California. Serum lipids and diet. Am J Epidemiol 1973;97:372-385. 3. Marmot MG, Syme SL. Acculturation and coronary heart disease in Japanese Americans. Am J Epidemiol 1976;104:225-247. 4. Nichaman MZ, Hamilton HB, Kagan A, Grier T, Sacks ST, Syme SL. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California. Distribution of biochemical risk factors. Am J Epidemiol 1975;102:491-501 5. Robertson TL, Kato H. Gordon T, Kagan A, Rhoads GG, Land CE, Worth RM, Belsky J, Dock DS, Miyanishi M, Kawamoto S. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Coronary heart disease risk factors in Japan and Hawaii. Am J Cardiol 1977;39:244-249. 6. Winklestein W, Kagan A, Kato H, Sacks S. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California. Blood pressure distributions. Am J Epidemiol 1975;102:502-513. 7. Tillotson JL, Kato H, Nichaman MZ, Miller DC, Gay ML, Johnson KG, Rhoads GC. Epidemiology of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California. Methodology for comparison of diet. Am J Clin Nutr 1973;26: 177-184. 8. Reed D, McGee D., Yano K. Trends of coronary heart disease among men of Japanese ancestry in Hawaii. J Comm Health 1983;8:149-159. 9. Castelli T, Doyle JT, Gordon T, Hames CG, Hjortland MC, Hulley SB, Kagan A, Zukel WJ. HDL-cholesterol and other lipids in coronary heart disease. The Cooperative Lipoprotein Phenotyping Study. Circ 1977;55;767-772. 10. Gordon T, Kagan A, Garcia-Palmieri M, Kannel WB, Zukel WJ, Tillotson J, Sorlie P, Hjortland M. Diet and its relation to coronary heart disease and death in three populations. Circ 1981;63:500-515. 437 11. Kagan A, Gordon T, Rhoads GG, Schiffman JC. Some factors related to coronary heart disease incidence in Honolulu Japanese men: The Honolulu Heart Study. Int J Epidemiol 1975;4:271-279. 12. Kagan A, McGee D, Yano K, Rhoads GG, Nomura A. Serum cholesterol and mortality in a Japanese American population. The Honolulu Heart Program. Am J Epidemiol 1981;114:11-20. 13. Reed D, McGee D, Yano K, Feinleib M. Social networks and coronary heart disease among Japanese men in Hawaii. Am J Epidemiol 1983;117:384-396. 14. Rhoads GG, Gulbrandsen CL, Kagan A. Serum lipoproteins and coronary heart disease in a population study of Hawaii Japanese men. New Eng J Med 1976;294:293-298. 15. Yamamoto L, Yano K, Rhoads GG. Characteristics of joggers among Japanese men in Hawaii. Am J Pub Health 1983;73:147-152. 16. Blackwelder WC., Kagan A, Gordon T, Rhoads GG. Comparison of methods for diagnosing angina pectoris. The Honolulu Heart Study. Int J Epidemiol 1981;10:211-215. 17. Worth RM, Kagan A. Ascertainment of men of Japanese ancestry in Hawaii through world war II selective service registration. J Chron Dis 1970;23:389-397. 18. McGee DL, Reed DM, Yano K, Kagan A, Tillotson J. Ten-year incidence of coronary heart disease in the Honolulu Heart Program. Relationship to nutrient intake. Am J Epidemiol 1984;119:667-676. 19. Yano K, Rhoads GC, Kagan A. Coffee, alcohol and risk of coronary heart disease among Japanese men living in Hawaii. New Eng J Med 1977;297:405-409. 20. Yano K, Rhoads GC, Kagan A, Tillotson J. Dietary intake and the risk of coronary heart disease in Japanese men living in Hawaii. Am J Clin Nutr 1978;31:1270-1279. 21. Frerichs RR, Chapman JM, Maes EF. Mortality due to all causes and to cardiovascular diseases among seven race-ethnic populations in Los Angeles, County, 1980. Int J Epidemiol 1984;13:291-298. 22. Reed DM, Feinleib M. Changing patterns of disease in the Pacific Basin. Report of an international workshop. J Comm Health 1983;8:182-205. 438 23. Igra A, Stavig GR, Leonard AR., with special contributions by Gamber W, Hawthorne A , Jang VL, Ransom B. Hypertension and Related Problems in California. Results from the 1979 California Hypertension Survey. Hypertension Clearing -house. Department of Health Services. Hypertension Control Program. 714 P. Street. Sacramento, California 95814. 24. Stavig GR, Igra A, Leonard AR. Hypertension among Asians and Pacific Islanders in California. Am J Epidemiol 1984;119: 677-691. 25. Basset DR, Schroffner WG. Coronary heart disease in Chinese men in Hawaii. Arch Int Med 1970;125:478-487. 26. Sloan NR. Ethnic distribution of diabetes mellitus in Hawaii. JAMA 1963;183:419-424. 27. Gerber LM. The influence of environmental factors on Mortality from coronary heart diseaes among Filipinos in Hawaii. Human Biol 1980;52:269-278. 28. King H. Selected epidemiologic aspects of major disease and causes of death among Chinese in the United States and Asia. In: Medicine in Chinese cultures. Comparative Studies of Health Care in Chinese and other Societies. A Kleinman, P Kunstadter, ER Alexander, and JL Gale (eds). DHEW Pub No. (NIH)75-653, pp. 487-550. 29. Bennett CG, Tokuyama GH, McBride TC. Cardiovascular-renal mortality in Hawaii. Am J Pub Health 1962;52:1418-1431. 30. Gerber LM, Madhavan S. Epidemiology of coronary heart disease in migrant Chinese populations. Med Anthrop 1980; 4:307-320. 31. Gerber LM. Diabetes mortality among Chinese migrants to New York City. Human Biol 1984;56:449-458. 32. Bassett DR, Schroffner WG. Blood lipids and lipoproteins, glucose tolerance, and plasma insulin response in Chinese men with and without coronary heart disease in Hawaii. Isr J Med Sci 1969;5:666-670. 33. Chung CS, Bassett DR, Moellering RC, Rosenblatt G, Stokes J, Yoskizaki H. Risk Factors for Coronary Heart Disease in Hawaiian and Japanese Males in Hawaii. J Med Genet 1969;6:59-66. 439 Bassett DR, Abel M, Moellering RC, Rosenblatt G, Stokes J. Dietary intake, smoking history, energy balance, and stress" in relation to age, and to coronary heart disease risk in Hawaiian and Japanese Men in Hawaii. Am J Clin Nutr. 1969;22:1504-1520. Bassett DR, Abel M, Moellering RC, Rosenblatt G, Stokes J. Coronary heart disease in Hawaii, Dietary intake, depot fat, "stress," smoking, and energy balance in Hawaiian and Japanese men. Am J Clin Nutr 1969;22:1483-1503. Bassett DR, Moellering RC, Rosenblatt G, Greenberg D, Stokes J. Coronary heart disease in Hawaii. Serum lipids and cardiovascular, anthropometric and related findings in Japanese and Hawaiian men. J Chron Dis 1969;21: 565-583. Yano K, Reed DM, McGee DL. Ten-year incidence of coronary heart disease in the Honolulu Heart Program. Relationship to biological and lifestyle characteristics. Am J Epidemiol 1984;119:653-666. Gordon T, Garcia-Palmieri MR, Kagan A, Kannel WB, Schiffman J. Differences in coronary heart disease in Framingham, Honolulu, and Puerto Rico. J Chron Dis 1974;17:329-344. Kraus JF, Borhani NO, Franti CE. Socioeconomic status, ethnicity, and risk of coronary heart disease. Am J Epidemiol 1980;111:407-414. SUPPLEMENTAL REFERENCES Chapman JM, Frerichs RR, Maes EF. Cardiovascular Diseases in Los Angeles. 1980. Hansen VE, Ed. American Heart Association. Greater Los Angeles Affiliate, Inc. 1983 edition. Dock DS, Fukushima K. A longitudinal study of blood pressure in the Japanese 1958-1972. J Chron Dis 1978;31:669-689. Frerichs RR, Chapman JM, Nourjah P, Maes M. Cardiovascular Diseases in Los Angeles. 1979-1981. Hansen VE, Ed. American Heart Association. Greater Los Angeles Affiliate, Inc. 1984 edition. Health of the Disadvantaged. Chartbook II. DHHS Publication No. (HRA) 80-633. September 1980. 440 S-5. Ludman EK, Newman JM. Yin and yang in the health-related food practices of three Chinese groups. J Nutr Ed 1984; 16:3-5. S-6. Szatrowski TP, Peterson AV, Shimizu Y, Prentice RL, Mason MW, Fukunaga Y, Kato H. Serum cholesterol, other risk factors, and cardiovascular disease in a Japanese cohort. J Chron Dis 1984;37:569-584. S-7. Tanaka H, Ueda Y, Hayashi M, Date C, Baba T, Yamashita H, Shoji H, Tanaka Y, Owada K, Detels R. Risk factors for cerebral hemorrhage and cerebral infarction in a Japanese rural community Stroke, 1982;13:62-73. S-8. Ting RY. Diseases of the Chinese. in Rothschild H, op.cit., Chapter 11, pp. 327-357. S-9. Ueshima H, Shimamoto T, Iida M, Konishi M, Tanigaki M, Doi M, Tsujioka K, Nagano E, Tsuda C, Ozawa H, Kojima S, Komachi Y. Alcohol intake and hypertension among urban and rural Japanese populations. J Chron Dis 1984;37:585-592. S-10. Ueshima H, Iida M, Shimamoto T, Konishi M, Tanigaki M, Doi M, Nakanishi N, Takayama Y, Ozawa H, Komachi Y. Dietary intake and serum total cholesterol level: their relationship to different lifestyles in several Japanese populations. Circ 1982;66:519-526. S-ll. Ueshima H, Iida M, Shimamoto T, Konishi M, Tanigaki M, Nakanishi N, Takayama Y,Ozawa H, Kojima S, Komachi Y. High- density-lipoprotein cholesterol levels in Japan. JAMA 1982;247: 1985-1987 S-12. USDHEW. National Symposium on High blood Pressure Control in U.S. Asian/Pacific Populations. DHEW Publication No. (NIH) 79-1609. (conference date October 1-2, 1977) S-13. Wenkham NS, Wolff RJ. A half century of changing food habits among Japanese in Hawaii. J Am Dietet A 1970;57:29-32. 441 Ischemic Heart Disease Risk Factors in American Indians and Alaska Natives Shiriki K. Kumanyika, Ph.D., M.P.H. Department of Epidemiology Johns Hopkins School of Hygiene and Public Health Baltimore, Maryland Daniel D. Savage, M.D., Ph.D. Medical Advisor National Center for Health Statistics Hyattsville, Maryland TABLE OF CONTENTS 1.0 APPROACH 446 2.0 OVERALL PERSPECTIVE ON PREDISPOSITION TO ISCHEMIC HEART DISEASE IN NATIVE AMERICANS ............. 447 2.1 Background Comments on the Native American Population 2.2 Heart Disease in the Mortality Profile of the Native American Population 3.0 ISCHEMIC HEART DISEASE IN SPECIFIC INDIAN POPULATIONS . . . .453 3.1 Southwestern Indians 3.2 Indians outside of the Southwest 3.3 Summary 4.0 ISCHEMIC HEART DISEASE RISK FACTORS IN NATIVE AMERICANS . . .457 4.1 General Comments 4.2 Hypertension 4.3 Cholesterol 4.4 Smoking and Alcohol Use 5.0 MULTIVARIATE PERSPECTIVE .............. 466 .....467 6.0 IMPLICATIONS AND RECOMMENDATIONS ....... 7.0 REFERENCES..................469 445 1.0 APPROACH In general, ischemic heart disease (IHD) risk factor data sources for the Native American population are in no sense comparable to those for the other minority groups, due to the relatively greater emphasis on health problems other than IHD in this population and to the general paucity of data specific to the Native American subset of the U.S. non-white population. The size of the Native American population is small compared to the larger minority of black Americans. Numbers in research reports and vital statistics categories relevant to cardiovascular risk are extremely small due to the young age of the Native American population and to their excess mortality from noncardiovascular causes (1,2). Rates and percentages in this report should be interpreted with the small numbers in mind. The anthropological, socio-political, and cultural history of the Native American population is unique and varies greatly among Indian tribes. These factors have had a large influence on the amount and nature of health and disease data available for this population. Related to this, there is considerable redundancy in the authorship of cardiovascular studies in Native American populations, a greater than desired reliance on inherently biased data sources and anecdotal material, and relatively little written about Native American health by Native American people. Socioeconomic status of Native Americans is generally at the low end of the continuum. Thus, it is difficult to separate genetic and cultural factors from socioeconomic factors. Varying degrees of racial admixture and inconsistencies in racial classification of Native Americans of mixed ancestry (i.e., as Indian vs. white) also complicate the picture of Native American-white differences in disease patterns. A summary of selected background information on Native American population and mortality characteristics has been included, to indicate our perspective on IHD risk in this population group (section 2). Patterns of IHD and IHD risk in specific Indian populations are then discussed (sections 3 to 6). A large portion of the literature pertinent to cardiovascular risk factors relates to Indians in the Southwest and, in particular, to Pima Indians, who are reported to have the "highest recorded prevalence of diabetes in the world" (50% prevalence in Pima over age 35; almost exclusively type II) (3, page 199). Pima or the closely-related Papago Indians are a focus or sub-focus of a third of the papers identified as relevant to this report 3,7,12-25). There are also several references which provide information on Navajo or Apache Indians (3 7,11,22-27,29-31,38). Papers on southwestern Indians include the above groups as well as Hopi Indians and other tribes in the Arizona and New Mexico area (3,7,22-24,29,30) Papers from Minnesota (33,51) provide data primarily on Ojibwa (Chippewa) Indians; small percentages of Sioux and Winnebago Indians are also represented in the Minnesota Native American population. Other 446 relevant studies provide data on aspects of CHD risk among Crow and Northern Cheyenne Indians in southeast Montana (34), Arapaho and Shoshone Indians in Wyoming (35), Seminole Indians in Oklahoma and Florida (36), Alaska Natives (37,50); Ojibwa Indians in Canada (38), Seneca Indians in upstate New York (39,40), American Indian university students in Georgia (41), and the Tarahumara Indians in Mexico (32,42). The literature identified gives a less than complete picture of IHD risk for the sub-groups listed above and leaves a substantial portion of the Indian population uncovered (4,5). Published tables for the U.S. ^ population usually present Native American data in an aggregate "non-white category along with data for black- and/or Hispanic- and/or Asian-Americans (9,10). One tabulation of blood pressure levels of nonreservation Indians was identified in an NCHS report from the National Health and Nutrition Examination Survey, 1971-74 (43). Mortality data for some years were obtained from Indian Health Service reports (2,6,8,44) for Indians and Alaska Natives in reservation states. 2.0 OVERALL PERSPECTIVE ON PREDISPOSITION TO ISCHEMIC HEART DISEASE IN NATIVE AMERICANS 2.1 Background Comments on the Native American Population The census designation "Native American" includes American Indians, Eskimos, and Aleuts. Native Americans were 0.6% of the United States population in the 1980 census--l,418,000 of the 226,505 million Americans counted (9, Table 1). In the 1970 census, Native Americans were 0.4% of the population--793,000 of 203,212,000 Americans counted (10, Table 1). The Native American population is disproportionately poor, has a lower life expectancy and is younger than the U.S. population as a whole (median age 18.4 vs. U.S. population median of 28.1 years in the 1970 census (3,8) The Navajo, the largest of several hundred American Indian tribes residing in the U.S., numbered approximately 150,000 in the mid-70's (3). States with the largest Native American populations are Arizona, Oklahoma, California, New Mexico (3), and North Carolina (45) but the federally recognized Indian tribes are spread throughout more than 25 states (3,5). Ninety percent of Indians lived on reservations prior to 1940. In 1970 and 1977, 40% and 51% of Indians lived off of reservations, primarily in urban settings (45). Sievers and Fisher (3) point out that southwestern Indians have remained more isolated and less racially mixed than Indians in other regions. In recent decades however, social and environment changes have been occurring, leading to reduced ethnicity. For example, whereas 60% or more of White Mountain Apaches lived in traditional tribal dwellings in 1959, these "wickiups" have become rare. At the time the Sievers and Fisher review was written, (1981 publication date) 6% of southwestern Indians over age 15 vs. 20% of the population under age 15 were reported to have some non-Indian admixture. 447 2.2 Heart Disease in the Mortality Profile of the Native American Population The mortality profile of Native Americans differs from that of the overall U.S. population (Tables 1 and 2) and to some extent from that of other racial minorities (Table 2). Age-adjusted total mortality comparisons are not given in the 1979 Indian Health Service tables available at this writing. Age-adjusted total mortality among Native Americans was 1.3 times that for U.S., all races in 1975 (see Table 3 in this report). There were 19,474 deaths among Indians and Alaska Natives in 1979 (1,913,814 deaths among all races, U.S.). Table 1 indicates that, although the majority of deaths are classified as due to heart disease in both the general population and among Native Americans, the proportionate mortality from heart disease in Native Americans is half that of the general population (Note however, that proportionate mortality from heart disease varies significantly by region. See Table 5 and accompanying text.) The proportion of accidental deaths among Native Americans is nearly as high as that due to heart disease and is more than three times as high as the proportion of accidental deaths in the general population. The proportion of Native American deaths associated with chronic liver disease or cirrhosis is 3.8 times higher than in the general population. Table 1. Leading causes of death among the U.S., all races 1979 and comparable data for Indians and Alaska Natives, 1978-1980 [taken from Table 4.1 in reference 2]. Percent Distribution Causes of Death Diseases of the heart Malignant neoplasms Cerebrovascular diseases Accidents COPD* Pneumonia and influenza Diabetes mellitus Chronic liver disease & cirrhosis Atherosclerosis Suicide All other ALL CAUSES * chronic obstructive pulmonary disease and associated conditions ** cirrhosis of the liver, 1978 # among Indians and Alaska Natives, "all other" includes 3.3% homicide, and 2.7 percent deaths attributed to "certain causes of mortality in early infancy (1978)" U.S. , Ind ians and All Races Alas ka Natives 38.3 20.8 21.1 10.1 8.9 4.8 5.5 19.5 2.6 0.9 2.4 3.8 1.7 2.9 s** 1.6 6.0 1.5 0.7 1.4 2.6 15.1 28. 0# 100.0 100.0 448 The distinction between proportionate mortality from heart disease, crude, and age-adjusted heart disease rates among Native Americans vs. U.S. all races is critical due to the younger age and excess noncardiovascular mortality in the Native American population. Disparities in age-adjusted rates between Native Americans and whites are most pertinent for this report. Unfortunately, of the data suitable for comparison purposes, proportionate mortality data were more readily available than age-adjusted rates [the Indian Health Service may have more specific recent tabulations than those provided to date (2)]. As shown in Table 2, Native American rates of infant mortality and mortality due to tuberculosis, gastrointestinal disease, accidents, and alcoholism have decreased markedly since 1955. However, in 1980 Native American mortality from tuberculosis, accidents, and alcoholism was still disproportionately high when compared to the general population. Native American mortality rates were more similar to mortality rates of other non-whites in the population than to rates for whites. Native Americans were at somewhat higher risk than other non-whites in 1979/1980 for all mortality classes shown in Table 2 except infant mortality. Table 2. Comparison of mortality rates for Indians and Alaska Natives with those of the general population and other U.S. non white populations in two time periods for selected causes @ CAUSE OF DEATH CLASSIFICATION Infant Mortality Tuberculosis Gastrointestinal Disease Accidents Alcoholism RATE 1955 62.7* 1979 14.6* 1955 57. 9// 1980 3.6// 1955 15.4** 1980 4. 0** 1955 184.0//// 1980 107.3//// 1969 56.6! 1980 41.3! RATIO TO RATIO TO U.S. S. ALL RACES OTHER NON-WHITES 2.4 1.5 1.1 0.7 6.9 2.4 6.0 1.5 4.3 2.3 1.3 1.3 3.3 2.6 2.5 2.1 7.4 not given 5.5 @ data for Indians and Alaska Natives are based on reservation states. * deaths per 1000 live births; from reference 2, Table 3.3 // age-adjusted deaths per 100,000 population; from reference 2, Table 4.11 ** age-adjusted deaths per 100,000 population; from reference 2, Table 4.13 //// age-adjusted deaths per 100,000 population; from reference 2, Table 4.7 ! age-adjusted deaths per 100,000 population; from reference 2, Table 4.10 449 Data on heart disease mortality in the available Indian Health Service tabulations covering the time period up to 1979 and 1980 (2) are limited in detail. More detail is available in published IHS reports for the period up to 1975 (6,8). Using these reports, Tables 3, 4 and 5 have been constructed to show time trends in heart disease mortality among Native Americans and their relationship to certain trends in other classes of death, variations in mortality patterns by region, and age-specific patterns vs. those of the total U.S. population. Table 3. Percent change in age-adjusted mortality (rate per 100,000 population) between 1970 and 1975 for Indians and Alaska Natives and the U.S., all races (from ref. 6) Cause of Death Indians and U.S., ratio* Alaska Natives^ All Races All Causes - 7.3 -10.6 1.3 Disease of the Heart -12.7 -13.1 0.7 Cerebrovascular -20.9 -17.8 0.8 Arteriosclerosis -14.4 -21.4 1.5 Hypertension - 5.9 -34.5 0.8 Diabetes Mellitus -12.2 -17.7 2.1 Homicide + 19.4 +15.4 2.5 Suicide +45.3 + 6.8 2.1 Cirrhosis of the liver + 7.9 - 6.1 4.4 @ based on reservation states * 1975 ratio of rate for Indians and Alaska Natives to rate for U.S., all races. Table 3 indicates a decrease in heart, cerebrovascular, and arteriosclerosis mortality among Native Americans to an extent roughly comparable to that for the general population during the period between 1970-75 but a substantially lesser decrease in deaths due to hypertension. The disproportionate increase in deaths attributed to suicide and liver cirrhosis among Native Americans influences interpretation of the decreases in the other categories. Similar cardiovascular mortality may not represent similar incidence of cardiovascular disease, (e.g.,if heart disease mortality is pre-empted by competing causes of death). 450 Table 4. Age-specific death rates for cardiovascular diseases and diabetes for Indians and Alaska Natives (1973-75 average) and U.S., all races (1974) in 25 reservation states (deaths per 100,000 population) (from ref. 6) Cause of Death 25-34 35-44 Age Group 45-54 55-64 65-74 Heart Ind US 19.6 9.4 65.0 55.7 192.4 215.5 451.0 590.8 1426.6 2758.4 Hyper-tension Ind US 0.3 1.3 0.9 2.8 2.0 9.2 4.3 9.8 26.3 Cerebro-vascular Ind US 4.8 3.6 17.2 13.3 47.4 35.3 123.8 99.6 492.7 869.0 Arterio-sclerosis Ind US 1.0 0 1.8 0.2 7.8 1.0 23.7 5.9 83.3 153.3 Diabetes Mellitus Ind US 3.8 1.9 15.5 4.3 35.1 11.4 86.4 31.7 171.6 132.6 Forty-three percent of the heart disease deaths among Native Americans are due to myocardial infarction; 32% are due to chronic ischemic heart disease (6). At all ages below age 35 (ages 1-24 not shown), the heart disease death rate for Native Americans was approximately twice as high as for U.S., all races. Above the age of 44, heart disease mortality increases less steeply with age in Native Americans than in the general population, and Native American rates are lower than U.S. rates for all age groups over 45. This cross-over in heart disease mortality is hidden in the overall 0.7 ratio of Native American to U.S., all races rates (Table 3), since the numbers of deaths at the younger ages are quite small. A later cross-over in arteriosclerosis and cerebrovascular disease death rates is observed. Native American rates are similar or moderately higher than U.S. rates at younger ages (under the age of 65), and substantially lower at older ages (over age 65 years). It should be noted that heart disease other than ischemic disease (i.e., congenital heart disease) makes up a larger proportion of the heart disease under age 35 compared to the proportion in older age-groups. Diabetes deaths are higher among Native Americans at all ages over 25. The differences in the proportion of mortality due to heart disease among Native Americans in different regions of the country are shown in Table 5. Since the rates for Indian Health Service service areas are not age-adjusted, the percent of the population under one year and over 30 years of age is shown for reference in making comparisons. In 1977, 88% of Indians lived in the states where the IHS had responsibility for providing 451 health services; 67.5% of the Indians living in these states lived within the service unit boundaries (8). Among the Indians represented in the IHS data for 1975-77, heart disease mortality is lowest in the Southwest. Table 5. All cause and heart disease mortality and proportionate mortality from heart disease in Indian Health Service service areas, 1975-1977 (constructed from reference 8) Service Area % < lyr % > 30 All Causes Heart % heart old years old /100,000 /100,000 i All areas 2.57 31.0 751.2 127.6 17.0 Tuscon 2.68 30.5 837.7 82.8 9.9 Aberdeen* 2.95 27.9 1,135.6 214.6 18. 9@ Bemidji** 2.31 30.0 918.2 252.7 27.5 Albuquerque// 2.67 30.3 648.1 56.3 8.7 Alaska 2.39 29.1 696.5 84.4 12.1 Billings//// 2.77 28.6 869.5 178.6 20.5 Oklahoma City 1.98 40.3 683.6 190.4 27.9 USET! 2.68 30.6 662.7 111.3 16.8 Phoenix!! 2.61 29.2 751.1 90.1 12.0 Portland+ 2.57 33.1 889.2 174.6 19.6 Navajo++ 3.07 25.1 613.8 49.8 8.1 * North and South Dakota, Nebraska, Iowa ** Minnesota, Wisconsin, Michigan // Colorado, New Mexico //// Montana, Wyoming ! North Carolina, Florida, Louisiana, Mississippi !! Arizona,Nevada, Utah, California, Oregon, Idaho + Washington, Oregon, Idaho ++ Utah, Arizona, New Mexico @ rate given in Table 18 (ref 8) is in error, but correct data are shown elsewhere in Table 18 and in Table 20. When the tabulations of mortality data are taken together, heart disease appears to be a significant contributor to mortality in Native Americans outside of the Southwest, except Alaska, but proportionately less so than for the general population—apparently related to the mortality contribution of noncardiovascular causes. The impression gained from overall rates of heart disease mortality for the total Native American population could be misleading if these two caveats are overlooked: 1) In 1975-77, Native American rates were lower than those of whites in older but not younger age-groups; and 2) the proportion of deaths due to heart disease was still relatively low among a large segment of the Indian population. 452 3.0 ISCHEMIC HEART DISEASE IN SPECIFIC NATIVE AMERICAN POPULATIONS 3.1 Southwestern Indians Review of older papers on coronary heart disease (CHD) in southwestern Indians aids examination of the baseline from which increases or decreases are judged. Although the data bases used have limitations (e.g., dependence on hospital cases, autopsy records), the methodologies and their limitations are generally well-documented and data-based inferences separated from impressions and opinions. Moreover, in many instances, the magnitude of differences between Indian and white comparison populations is often much larger than the probable error. The clinical impression in the 1960's was of an absence of CHD among the Navajo and Apache, based on the failure to observe CHD in Indian Health Service hospitals for periods of several years (26,28,31). This impression was supported by a prospective study conducted by Cornell Medical School investigators in a rural, isolated, traditional Navajo community (Many Farms) between 1956 and 1962 (31). With a very high probability that all symptomatic CHD in Many Farms was identified during the study period, the six year incidence was four in 508 persons. In a comparison with data from Framingham (taking age and sex into account), the CHD incidence was significantly less than expected in the Navajo men but not the women. Prevalence of hypertension, ECG abnormalities and cigarette smoking were also very low in this Navajo population. ECG measurements on 70 male and 77 female White Mountain Apache Indians ages 30 and over in 1957-58 found no records characteristic of coronary disease (26). However, a 1957-1966 study of myocardial infarction in a large southwestern Indian Health Service population suggests that CHD, although of low occurrence, was definitely present to some extent in all tribes (23). Rates were approximately one-fourth of Framingham-based expected rates. One hundred and thirty-eight myocardial infarctions (Mis) were identified in an estimated population of 15,905 Indian adults (over age 30) at risk vs. 507 expected; only 56 of the 138 Mis were recent. The Indian rate was 86 vs. 347.8 (per 100,000) estimated for Framingham (estimate based on age and sex, not risk factors). The sex ratio for MI was lower in the Indian populations (2.3:1 vs. 4.6:1 in Framingham), possibly related to the high prevalence of diabetes among the Indian women. MI rates were higher in the desert tribes (Pima, Papago, Colorado River) than in the mountain or canyon tribes (Apache, Navajo, and others). Cigarette smoking patterns paralleled the MI patterns. In reporting these data Sievers emphasizes that no tribal group is immune from CHD, as had been implied in some earlier reports. More recent studies among the southwestern Indians suggest that CHD is increasingly less rare, particularly when diabetes is present (3). However, a 1976 report points out that CHD incidence in both diabetic and nondiabetic Pima Indians was lower than among white comparison populations (ECG data (Q wave changes) on 85% of half to full-blooded Pimas on the Gila River Reservation (46)). Sievers (47), as noted below, describes increasing rates of ischemic heart disease in Indians. In contrast, Kunitz 453 H pi rt rt cl P- Cl 0 H- pi hh 0 /---N o G 3 pj pi Pl > O 3d rt 0 h( N3 /~~\ 3 hh /^ P- h| p- Cfl ►a X! 3* rt p- cr 3 3- 3d CD 3 H- cr a Cfl hi p- w 3d 0 0 OQ 3 3 0 P- 3d 0 Pl P1 z 0 0 X 3 CD 3 rt 3 Pi 0 cr 3 CD CL CD CD X hh pi n Cfl CD OQ a 0 m 3 z CD CL CD 3 OQ CD 3 rt ON 3d h( h( hi *d t—' hh Pi Pi h) Cfl CD Cl M CD Cl hh 3 h( 0 ►o Pi f p* i CD 1 h| a cr h-1 CD H- rt CD M CD Pi rt 3 rt X p- pi Cfl CD rt 3 P- CD OQ CD o 3 hi CD PH h( Cfl 3 P- p- Ui CC CL Cfl ^J 0 Pi rt Cfl CD < P CD rt p- p- z n 3 3 £ Cl hh h| m CD pi p- CD CD n Cfl /—N 3 CD 0 o rt 0 p- o rr )—• cr 3 Cfl Pi O Cfl p- O rt o> pi rt 00 Pl 0 hh CD p- Cfl Cfl pi 3 rt ^—/ rt ON CL h| X a pl P- h| CD X cr Pl X p- CD 3 Cfl 3 rt rt 3 pi < rt CT OQ H 3 CD 3 CD CD rt O rt CD Vw/ p- Pl CD 3 0 < p- CD hi CD p- rt Pl P- rt P- rt P- >—■ »—• Pi CD 0 CD CD rt X hi O < P- Cfl CD "< CD rt h) pi rt pl hi 3 h-1 CD CD h| 3 "^ £ cr 0 0 cr CD Pi - (_^. Cfl h-t Cfl p- 0 CD CD CD 3 CL 0 0 3 CD CL Pl PH 0 hi CL p- O 3d h-- 3 3 3 pi Cfl hi 0 Cfl rt a 3 3 3 Cfl hi P- 0 Cfl h-' Cfl L_i. rt 3 £ ?: Cfl pl 3 Pl P1 pj CL P- ^ - 3 Cfl rt Pl Cfl rt 0 CD rt Cfl p-1 3 K hh 3 h-' hd Pi 3 CD CL cr rt rt a O 3 3 rt 3 0 P- 3 3d 3 rt rt rt P- p- P- 0 0 h-' 0 CL Pi Cfl * P- p- hh Pl Pi CD t_i. o CL CD CD pi cr cr P' — hh 3 CL O Pi 0 3d X 3* rt 3 Pl hh n rt cr 7? € CD Cfl 3 3 rt Pi 0 hi 3 h| 3 V CD OQ 3 CD Pi rt rt 0 rt rt pi 3 CD CD hi 3 CD a 3d h-1 P- < Cfl p- 0 CD P- 3 cr CD CL p< X Cfl o Cfl rt X CD rt h| Cfl cr Cfl CD O 3 CD rt Pi M P- h( Cl Cfl a 0 3 Cl CD p- X 3 a 3 Cfl ►a 0 hi rt o 0 0 1 Cfl CD Pl hi 0 *d P- C_i. P- CD cr Pl rt hi 0 OQ CD l—' cr rt OQ 0 z p- 3 •n c CD CD o rt >d pi cr p- h| rt CD 3 Pi Cfl rt n 0 0 ►O pi rt ►*J ^ p- CD CL 3 Cfl pj 3 Cfl G G pi hi X Cfl CD p- CD rt CL 3d CL 3 CL CD 0 3 CD 3 Cfl 3 h) O 3- CC H Cl 3 pj CD H- h-1 3 Cl Cfl X CD hi Cfl 0 • o CD 3 Cfl h| CD hh CD P- h-1 i_i. hh O h| rt 3 CD cr pi a Cfl p- Cfl Cfl CD 3 OQ P- p- CD 0 3 in C/3 3 Pl p- M o 3d Cfl CL rt 0 Cfl Pl 3 OQ h( P- Pl 3 a CD rt X < OQ 3d < n h| 0 rt 3- CL 3 rt rt Cfl 3 CD ^ CD 0 h( CC CD rt Cfl p- cr CD Cfl rt rt p- 3 s: M Pi Pi hi 0 CD Pl pi CD 0 hhv; P- CD Pl rt cr i 1 CD P- h-1 hi hi Pi 0 Pl p- rt < 0 Pl CD 0 rt rt p- = P- 0 3 pj 0 hh hi 3 hi Cl 3 hi OQ rt 3 cr S pi Pl 3 Cfl Pl pl * Pl Cfl rt rt X Cfl 0 CD CD Cfl Cfl 3 CL *< O O P- (-■ cr a CD 3 a P- X CD pj cr CD CL £ CD pj pJ 3 OQ X rt 3 rt cr 3d p- CD 3 Cl Cfl X CD 0 Cfl 3 i- p-- 3 X P- ^-\ 0 n >a P- CD hi CD Cfl pj pJ ^ p- P- 0 P- CD Cfl CD CD P1 P* P* Cfl p- CL hh 0 Cfl rt cr 0 CD p- O Cfl 3 LO p- 3 w O 3 X hi 3 Pl Cfl rt 0 hi rt < 3 3 Cfl 3d h) VD ^D \D "5 pi 3 rt CD Pl 3 rt cr CD a cr 3 3 CD 0 * 3 Cfl a CD 0 P- o rt CD PH hi o Pl CD CD OQ 0 G G CD »-J ON Ln 0 Pi 3 0 Pi CL < rt CD 3 CD pi CD CD Cfl hi ■P- O Pi 3 ^d hi 3 0 CD hi 3 Pl pl rt P1 Pi hh £ Cfl LO Ln ■P- hh Cfl Cl p-1 3 P- CL P 3 rt Pl CD »-J h| PH hi 3 O M Pl Pi 3 hi 3 CL o M CD 0 ^< z 0 p- m m CD 1 1 1 CL P- h| CL p- P- p- G rt p> p. p- 3d 3 h^ Pl \s CD 3 p- 0 ^t 0 rt 3 pj 3 p- CD CL Cfl hh 0 Cfl h| 3 hi ^J ON Ln rt O CL CD N P- CD 3 3 CD • p-1 3 3 CD a 3d Cfl Pl C^ Cfl 3 rt P- rt rt hH pi Cfl P- z hi 3 OQ - < Ln ^J ON 3d p- 3 3 Pi 3 3 < cfl m CL o CD CD 0 Cfl p- ?^0Q 0 CD 0 Cfl 3 PH 3 3 Pl 0 cr 3 rt 3 rt pi Pl CD rt CD 0 rt Oq O rt CD • P- 3 H rt rt pi hh CD Pi hh P- Cfl hi 3 CL Cfl P* PJ hh p- 3 er 3 cr Pi OQ pi rt H> 4> M CD 3 P- CD 3d Cfl rt Cfl h-1 CD CT £T n < 0 hh a 3 Cfl 3 0 Pl CL P- Ln OQ 0 Pi CD 3 CD h-< P- ON ho VD cr Cl rt p. 0 rt CD rt 3^0Q CD O pi CD H CD Pl hh CD 0 P- 0 hh rt P- Pl p> M P- cr cr 0 > hi h-> 0 1 1 1 CD 0 3 3 cr rt P- P CD Pl CD Cfl CD CD h-1 h| Cl rt H 3 3 O hh P- Pl 3 LO 3 ^ 3 3 ^ CD p- CL 3 Ln ■P- LO pi Pi Cfl CD 0 rt OQ O 3 Cfl hi H- a CD CL P- Pl p- hi hT rt rt 3d rt 0 3 Cfl Ln -(> hhT3 h| 3 CD Pi 3 P- h| LO VC ■f> h| S rt P Pi 3 CD CD Cfl 3 pi o Cl O rt Pi P- CD 3" CD 0 3d 3d 3 Cfl Pi CD Pl 0 hi 0 hi o Cfl Pl hh rt W PH O Pi p> P- Pi rt OQ hi " OQ rt rt 3 P- 3 3* cr cr £ M pj P- P- p> OX hi h| rt 3 P- hh h-' h| rt n 0 ^d X X Pi w hj 3 VD 3 CL 0 cr Pl 3* PJ pl h| rt rt CD CD CD P- CD ^d CD OQ Cfl 0 p- H> CD 0 rt CD n P- CD hi CD hi CD CD CD P- hi Cfl CL ^-J CL P- hi h-' Pl Pl *< P- hh v- CD rt Cfl 3 Cfl CD Cfl 3^ 0 3 X h( rt CD Cfl Pi pi CL CD Pl p- Cfl hi hi h( Cfl h-1 0 • 0 P- rt Cfl p* l—1 < cr hi Cl CD X) rt h( rt i 1—' 3 >^J CD P- 3 3 3 P- hi Cfl cr rt P- hh CL - O P- vo X rt CD Cfl rt pi 0 0 3 Pi Cfl Pl hi CD Cfl CD Cl 0 X ^d h| h-1 0 Cfl 0 0 Cfl CD 3 P- cr P1 h-■ P1 OQ CD CD Cfl Pl 0 CL ON 0 3d CD 3d p-< 3 hh v; rt M CD hi 0 M CD £ Pi pi X o 3 p- hh 3 PH CD rt l rt 3 CD O O O P- hi P> >• Pl z rt 3 3 ^1 x 0 CD CD CD 0 Pi pl CD Cfl 3 3 0 3 CD hi hi CD P1 o 0 OQ 3 Pl P- l P- O O O < N3 3 rt CD Pi CD 3 3 x s OQ Cfl rt o 3 Cfl Pi h-' Cfl h( CD rt hi O w 3 3 CL Cl Cfl 3 X 0 P1 hh >. - >. CD P- cr ^~N p-" /—•s t—' OQ cr rt H pi 3d pl P- Cl 0 P- PH i—1 p- CL o o 0 p- hh p- CD CD pl 3 VO p- O O O 3 3 rt pi Cfl p> rt rt h-l P) 3d p- pi H 0 3 CD hi 3 0 0 3 ^-\ rt n rt P1 " o hi ^-\ pi 3 Pi Cfl hi ^J hi O O O O < P1 3d Z 0 p* rt cr g CD 3 X cl 3d 3 3 CL O fvej w rt 0 O o o rt LO cr h-1 3 >d rt p- ON Cfl O O O pi rt CD p- w. Pi Pi N—" P- 3 P- 0 3 X pj cl p- H ^ 3 P- 0 a a 0 O o Pi Vw' CD p-1 Cfl -5 CD P- Cfl rt Cfl cr h-1 M 3 rt < rt ' O 0 rt 3 a P- CD p- pi Pi X CD 0 Pl 3 tr1 h-> Pi <* o h-1 rt i ■* cr hi O /"> hh CD -P- Pl X Pi cr 3 rt CD OQ Cfl 3 hh p-1 cr CD 3 hh 3 X i P- pi PH o P- CD 3d p- P- 3 3 P1 rt 0 hh OQ Cfl rt C_i. CD 1 CL oq cr hh pj h( Cfl Cfl CD nx h( 3 h-' o rt Cfl CD p- o 0 pj 0 p> « h| O Cfl hh CD 3 3d 0 H 1 3 Cfl p- Cfl p- CD p- CD rt 3 rt i rt v_/ 0 CD CL CD o "< H hi 3 cr CL Pi rt ^—^ 0 h-1 pi 0 >d VD CD 3 3d O 3 pi 0 3 pi hi 3 CD hi cr i P- - X P- Cfl cr 3 P- O Cfl hi rt h-' 3 hi pl >d 00 hh Cfl P- P- 7? 3 CL Cfl CD hh ON 3 Pl CD Pi rt hi h-1 Pl Cfl pl hi CD CD rt h-1 0 h-1 0 /—\ rt cr 0 CD 3 0 N—/ H 0 CD Cfl rt 3 cfl rt P- 3 Pl Cfl rt Cfl pj P- I—1 0 0 3 CD 3 Pl Cfl h-1 Pi 3 o ^3 pj ?: P- CD z Z 3 hi ■ Pl 0 P- 0 H- 3d o > p- o h( rt P- CD CD pj < 0 rt •5 Cfl h( CL rt CD PJ OQ CL 0 CD cr 3 Cfl O Pi 3 rt rt CL 0 hi 3 -s 3 s pi Cfl 3 CD 0 0 0 CD s CD CD CD P- CD CD 3 P- Pl CD CD Z cr < Pi CD Cfl hh Pi OQ 3 P- CD rt Cfl Cfl rt 3 £ Pi 0 3 CD P- h( CD CL Pl Cfl pi rt hi hh h| rt pi Pi pi 1—' 0 CC ■ hi P-" Cfl CD 3 rt P- P- H- C hh CD M 3 pj CD rt 3 3 Cfl n v_/ crt3 < >d i_i. hh CD 0 Cfl p. pj pi rt CL 3 3" Cfl 0 P-1 /-> rt ti- Cfl ^3 Cfl OQ rt Pl rt Cfl 0 CC P- Cfl CD CD pi rt O 3 hh 3 rt 3 P- hi CD p- CD rr 3 P-1 w 3d er £ Cfl X O h-1 CD hi hi Cfl 3d - 3 CJ 3 3 h( h( i_i. CD Cfl 0 rt 0 « crvs rt Cfl P. Pl 3 0 3 ^^/ 0 cr X h( cr Pl Z CD P- 0 rt cr P- 0 hi M 3d p- 3 3 Cfl 0 rt 3 P, OP Cl £ 0 Cfl rt hi p- CD 0 Cfl cr CD cr hh hi 0 3d pi 3 rt H rt 0 - < rt P- rt v< 3 rt 3d X hi h{ CD Pi 3 < rt hi Cfl o pi < pj 3 CD 3 0 Cfl cr £ Cfl 3 h| cr cr CL pi Cfl pi Cfl CD 0 p- Pi pi Cfl Pl p. PH rt rt CD hi P- cr 3 Pl CD Pi hi 0 rt h-1 CD CD 3 pl CD CD Cfl hi CD h-1 Cfl O rt 3 Cfl v<- hi cr 3 Cfl 3* Pi £ hi Pi cr P- CL h-1 Cfl CD hh CD P o Cfl Cfl pl Cfl O Cfl ^-/ CD Cfl P- Pi N rt » CD Pl CL rt Pi 0 rt Pi Cfl pi CD rt 0 p- p- rt 3 rt CD 2; Pl rt p> rt h( p- cr p-> p- hi rt hh 3 rt CD p-1 hi 3d 3 CD p- rt £ rt rt CD Pi P- pi < P- Ln w< CL Pi Pi CD Pl pi P- 0 h-1 H- rt CD P- 0 hi 3 cr Cfl CD ^ h| p^ 0 h) < Cfl 3 00 P- p-1 h-1 CD 3 3 ?r h( a 3 3d OQ CD 0 CD 1 - 3 p-1 3 Pl pi ■ Pi • M Pl Cfl 3 0 Cfl p-3 CD Pi rt cr CD h| 0 hh h-1 rt 3" CD 1 *< 0 hh rt CD Cfl 0 rt CD CL vj pj rt CD h-1 0 0 hh 3 OQ Table 6. Comparison of rates of acute myocardial infarction during 1957-66 vs. 1975-78 in three sub-groups of southwestern Indians (taken from 3,47) rate per 100,000 per year 1957-66 1975-78 % increase Piman 50 74 48% (Pima and Papago) Athapascan 18 53 194% (Navajo and Apache) Other 33 97 194% Southwestern* In the text accompanying these data, the authors note that 35% percent of the infarctions in this category were among Hopi Indians, who are only 19% of this group 3.2 Indians Outside the Southwest Mayberry and Lindeman examined cardiovascular disease death rates among Seminole Indians integrated into the Seminole County, Oklahoma population vs. those of whites in the same county (36). The authors note that in addition to other limitations of death certificate data, mixed-heritage Indians may have been recorded as either Indian or white on death certificates. The percent of mortality attributed to coronary artery disease (using data for persons age 25 and over) was less among the Seminoles than among whites (15.5 vs. 31.0% for men; 11.1 vs. 20.7% for women) (36, table III). Coronary artery disease (CAD) death rates calculated from Table III are as follows: Seminole males 15.5/1000; white males 37.5/1000; Seminole women 8.4/1000; white women 14.1/1000. Tabulations of death rates for male Indians aged 25 and over in other Oklahoma counties showed proportionate CAD mortality ranging from 12 to 33% (36, table IV), varying with the overall mortality pattern for the Seminole men in the different counties. These data demonstrate higher CAD mortality among Seminoles than among southwestern Indians at approximately the same time period. Pinkerton and Badke (34) retrospectively determined the incidence of myocardial infarction (MI) among Crow and Cheyenne Indians admitted to Crow Hospital (Crow Agency, Montana) during the 14-year period from 1956 to 1970. The study methodology was designed for comparability with the 14-year Framingham data, to the extent possible. Of the CHD categories ascertained, the authors were most confident of the completeness of the MI data (compared to angina pectoris, for example, which may have been treated primarily on an outpatient basis). Therefore, the authors base their 455 conclusions on the MI data. In calculating denominators for MI rates, off-reservation Indians who lived too far away to be in the hospital service population were excluded. Otherwise, maximum estimates of populations of Crow and Cheyenne men and women at risk were used so that the rates calculated would be conservative (i.e., if in error the rates would underestimate rather than overestimate the incidence of MI. The racial criterion in the tribal census was one-quarter or more Crow (this may also have applied to the Cheyenne). Equivalent or higher MI rates for Crow men and both Crow and Cheyenne women vs. rates for Framingham age-sex counterparts (10-year age-groups between 30 and 59) were reported. Indian women had MI rates at least twice as high as Framingham wo-nen at all ages, and had rates six times higher in the 45-54 year age-group. Rates for Cheyenne men were substantially lower than those of the Crow and Framingham men. When maximum Framingham rates were calculated assuming that all persons excluded at the first exam due to known CHD had had infarctions during the subsequent 14 years, Framingham MI rates for men were significantly higher than for both Crow and Cheyenne men. Rates among women were not significantly different using the upper limit Framingham rates. As noted above for the Seminole Indians, rates for Crow and Cheyenne Indians were higher than those reported for southwestern Indians. Pinkerton and Badke point out that their findings were contrary to the clinical impression of the Crow hospital staff that infarction was rare among Crow and Cheyenne Indians. Gillum et al. (33) report that cardiovascular disease was the leading cause of death among Minnesota Indians (primarily Chippewa) between 1968 and 1973 with rates comparable to those for Minnesota whites (age-adjusted rates for heart disease and stroke combined: 448.7/100,000 for Indians 455.2/100,000 for all races; proportionate mortality 26% among Indians). These authors draw attention to the similarity in heart disease and stroke mortality rates between Indians and whites vs. the substantially higher age-adjusted total mortality for the Indians (due to excesses in other categories. Thus, this is an example of similar cause-specific rates but lower proportionate mortality. Fifty-six percent of Minnesota Indians lived in urban areas in 1970. Rates in reservation and urban areas were similar. 3.3 Summary These studies of CHD in specific populations, although neither comprehensive nor up-to-date, confirm the general impression that CHD is lower among Indians in the Southwest than among Indians elsewhere and that patterns for Indians elsewhere are similar to those of whites in the same areas. Also, the higher heart disease death rates among Indians in IHS health service areas in 1973-75 vs. the U.S. general population through age 45 may suggest higher risk in younger cohorts (Table 4). However, the proportions of ischemic vs. other forms of heart disease at the younger ages are not clear. 456 Due to marked differences in the risk of death from other causes among Native Americans and whites, proportionate mortality figures, which are relative, tend to minimize the potential importance of heart disease risk. The overall impression of lower heart disease mortality among Native Americans (1975 ratio to U.S., all races = 0.7) may be influenced downward by the lower rates in the Southwest where the majority of Indians live. The CHD picture for Indians outside of the Southwest may be less favorable. There are many as yet unresolved issues related to CHD in Indians. Those encountered most often are issues of measurement (e.g., appropriate interpretation of ECG findings (46)); risk factor-disease associations (e.g., obesity and diabetes on CHD (13,20); consequences of hypertension (45)); and heredity-environment interactions (20,25). 4.0 ISCHEMIC HEART DISEASE RISK FACTORS AMONG NATIVE AMERICANS 4.1 General Comments Lifestyle-related risk factors appear to have increased substantially in Native American communities in the last 50 years with increasing urbanization and acculturation (3,33,45). However, in the absence of systematic risk factor surveys among most Indian populations, quantification of these risk factor changes is difficult. Some apparent increases are possibly due to increased screening efforts (45). In addition to personal and environmental factors which affect IHD risk factor trends in the general population, risk factor changes of Native Americans vary with tribe (both hereditary and cultural factors), extent of admixture for Indians who are not full-heritage, extent and duration of acculturation to white society, and regional factors (e.g., origins in the Southwest vs. in Seattle, Minnesota, or upstate New York (3,45)). The picture is further complicated by the differential rates of change of the admixture and acculturation variables in different Indian communities and by the different socioeconomic and quality of health care patterns for on and off-reservation Indians (45). The variation is both qualitative and quantitative; summary statements about IHD risk in Native Americans can only be made within groups with similar origins and experiences. 4.2 Hypertension Mortality attributed to hypertension and stroke was lower than that for heart disease in the IHS health service areas in 1973-75 (Table 4). However, between ages 25 and 64 years, rates among Indians were similar to or higher than those for the general population. Thus, hypertension is not inconsequential in the Native American population and it may increase in importance as other causes of mortality decline. At the 1977 conference on hypertension control in Native American communities, Maurice Sievers, M.D., Director of Research for the Phoenix Indian Health Service Area presented an historical overview (and critique) of reports of hypertension among American Indians and Alaskan Natives (45, 457 pp.77-82). Dr. Sievers notes the inadequacy of the "numerous bits of information" on this subject. Several speakers at this 1977 conference noted that the extent of the problem of hypertension among Native Americans was not at all well-documented. Dr. Sievers points out that the data which are available are often difficult to identify; they are buried within studies of diabetes or other aspects of Indian disease. He also points out that there are major tribal differences in cardiovascular disease, including a higher prevalence of congenital cardiovascular abnormalities in the Athapascans (Navajo and Apache) than in other tribes. Highlights of Dr. Sievers' summary follow--grouped by tribe and region to the extent possible. The talk by Sievers (45) is reported without citations. Reference 48 is a documented version of this text. Where the study described was recognized as one included in the bibliography for this report, a notation to this effect has been made in parentheses and the impression conveyed in Sievers' report verified by examination of the primary data. Regarding southwestern Indians: - a rate of hypertension among Navajo Indians less than one-fourth the white rate was reported in 1948-52; - 6.2% of the men and 3.7% of the women ages 30 and over in the Cornell-Navajo 1956-62 study at Many Farms Arizona (reference 31) were hypertensive; - a 1970 report indicated that migration to an urban environment was associated with higher blood pressures among Navajo men (49); - 1.4% of admissions to the Phoenix Indian Medical Center in 1953 had primary hypertension (based on one, unstandardized casual blood pressure reading); - hypertension-related morbidity among Phoenix-area outpatients in 1952 was 7 per 1,000 for Papagos, 4 per 1,000 for Pimas, and 2 per 1,000 for Apaches vs. 25 per 1,000 for whites. Two-fifths of the MI cases identified in a 10-year study (1957-66) (23) were hypertensive; - an analysis of computerized clinic records of more than 4,000 Papago Indians for the period 1969-1972 yielded a hypertension prevalence rate of 20% (Sievers notes, however, that the methodology used has rather serious limitations); the prevalence of hypertension and complications was greater in hypertensives who also had diabetes; - blood pressures of Pima Indians studied by Ingelfinger (46) were lower compared to whites but were higher in diabetic vs. nondiabetic Pimas; 458 - Leo observed a 34% prevalence of diastolic blood pressure greater than 95 mmHg among White Mountain Apache men and a 29% prevalence among the women (28) in a 1958 study of 110 men and 110 women volunteers; - Clifford et al. also noted a high prevalence of hypertension among 327 White Mountain Apaches studied in 1963 (26% in men, 23% in women) (26); examination of the Clifford et al. data indicates particularly high hypertension prevalence (BP >160/95) in Apache men ages 30-49 (43-47% vs. rates of 12-23% among older men). Highest rates among Apache women were in the 50-59 year old group (Figure 1 in reference 26)). Regarding Indians outside of the Southwest: - a systematic study of blood pressure found a very low rate of hypertension (1 in 200 Indians over age 20) in a northwestern coastal village; - A blood pressure survey of Seminole Indians in Seminole County, Oklahoma, Seminoles on a reservation in Florida, and Seminole County whites indicated similar blood pressure levels among the three groups (36); - no differences in blood pressure levels were observed among Sioux and white admissions to a South Dakota hospital (50 consecutive admissions of each race) in spite of the fact that 50% of the Sioux vs. 10% of the whites were diabetic (and most of the Indians were overweight). Regarding Eskimos and Alaska Natives: - 2.6% of 842 Eskimo men surveyed by Scott et al. (50) in 1958 had a mean blood pressure of 125; blood pressures were similar to those in a large group of U.S. men. Sievers' account of hypertension studies gives an adequate impression of the patterns of hypertension among Native Americans to the limited degree that these had been documented up to the mid 1970's. Hypertension has previously been relatively infrequent in southwestern Indians-- White Mountain Apaches excepted. No striking excess prevalence of hypertension in Indians outside of the Southwest (vs. whites in the same areas) has been reported. Sievers notes that the Indian Health Service age-adjusted death rate for hypertension for Native Americans in 1971 was lower than the rate for white Americans by only 10% (45). NCHS estimates of mean blood pressure levels for American Indians (off-reservation only) in 1971-74 do not show a consistent pattern of differences from blood pressure levels of whites (see Tables 7 and 8). However, given earlier statements about the intertribal differences and the substantial proportion (60%) of Indians still on reservations in 1970 (45) the value of the NHANES data may be limited to showing that there were no large differences in off-reservation Native American vs. white blood pressure distributions in 1971-74. 459 Table 7. Systolic and diastolic blood pressure levels of American Indian men, with comparison values for white men, U.S., 1971-1974. (from reference 43) Systolic Indian* White Diastolic Indian* White Age (mill Group 18-24 121.2 123.7 25-34 128.0 125.2 35-44 127.7 127.0 45-54 130.9 134.7 55-64 139.4 139.6 65-74 139.9 146.0 (millimeters of Hg) 77.7 76.4 76.5 80.8 84.9 84.2 89.8 87.5 85.4 86.4 84.9 84.9 examinees who identified themselves as American Indians and were not living on a reservation or in an institution at the time of the survey Table 8. Systolic and diastolic blood pressure levels of American Indian women, with comparison values for white women, U.S., 1971-1974. (from reference 43) Systol ic Indian* White Women Women Age (mill Group 18-24 114.9 115.1 25-34 119.0 116.2 35-44 117.4 122.6 45-54 138.9 131.1 55-64 144.3 143.0 65-74 145.0 151.6 Diastolic Indian* White Women Women 73.8 71.3 75.0 74.6 77.6 79.3 80.1 82.6 88.6 86.2 81.5 85.4 examinees who identified themselves as American Indians and were not living on a reservation or in an institution at the time of the survey 460 Only three recent reports of hypertension surveys in Indian populations were identified--two from Minnesota (33,51) and one among Navajo (27). The Navajo survey reported blood pressure data from 1977 clinic-based screenings of Navajo people ages 20 years and older at Fort Defiance, Arizona and Crownpoint, New Mexico—two demographically similar Navajo communities (n=640) (27). The populations were relatively diverse but were not randomly selected. Blood pressure levels of the Navajo women showed a linear increase with age. There was no age trend in either systolic or diastolic blood pressure among the Navajo men, but their blood pressure levels and hypertension prevalence were greater than among Navajo women. The prevalence of DBP >=90 was higher among the Navajo men ages 20 to 59 (24 to 28%) than among the small number of men aged 60 years and older (16%). The pattern was similar among the women. Prevalence of DBP >=95 was 9,16,8,and 11% in the 10-year age-groups of Navajo men between 20 and 59 years of age; 0,8,3, and 10% among the women. One hundred and ten of the Navajo screened had DBP >=90, of whom 57 reported a prior knowledge of having high blood pressure. Eighty and 72 percent of the Navajo men and women on medication had DBP less than 95 mmHg. The authors note that the estimates of hypertension are based on one casual reading and are thus probably overestimates. In an analysis of determinants of blood pressure levels in these Navajo subjects, male sex, obesity and alcohol were found to be important variables (alcohol use was measured only as yes or no). Overweight was a statistically significant hypertension risk factor for both men and women. Relative risks were estimated as 3.0, 3.6, and 9.0 for overweight females, normal weight males, and overweight males if normal weight females were assigned a value of 1.0. The highest reported use of alcohol was among the younger men (43% of men under age 40 vs. 22% of those over 40). The prevalence of DBP >=90 was significantly higher among the men who used alcohol. Very few Navajo smoke and smoking was not a determinant of hypertension. An index of acculturation was not significantly related to blood pressure levels (this index included ratings on alcohol use and smoking). In their discussion, DeStefano et al. (27) compare the Navajo blood pressure levels in 1977 with levels in the 1956-62 Many Farms Study (31) for an impression of changes in blood pressure over time. Blood pressures of Navajo females were essentially the same at both points in time. Blood pressures of the younger Navajo men measured in 1977 were higher than those of the men measured in 1956-62 at Many Farms. The authors suggest that a trend towards increasing hypertension may be present in the younger cohort of men. The 1977 data were also compared with blood pressure levels of black and white men in a 1973-75 study of one million Americans. Again, the trend towards hypertension in the young Navajo men was apparent. Systolic and diastolic blood pressure levels and the prevalence of DBP >=90 were equal to or greater than levels of young black and white males in the mid 70's. Although this trend in the younger cohort of Navajo men is presumably associated with some aspect of acculturation, the nature of such an acculturation factor was not evident from the measures used in this 461 study. Alcohol use may be an important factor. Blood pressures of the Navajo women are lower than those of women in the comparison populations. Gillum et al. reported on a blood pressure survey of essentially all first, second, and third grade children in the Minneapolis Public Schools (who were not absent from school on the day concerned) during the spring of 1978 (51). The population of 10,640 6- to 9-year old children surveyed included 307 Native American children. Analyses of the blood pressure data indicated higher systolic (by 1-3 mmHg) and lower 4th and 5th phase diastolic (by 2-5 mmHg) blood pressures (and higher pulse pressures) among the Native American vs. the white children. The systolic blood pressure differences were statistically explained by higher body mass index of the Native American children. No good explanation for the lower diastolic blood pressure among Native American children could be uncovered. Several methodological explanations were tested and ruled out (51). Gillum et al. have also reported on a blood pressure survey among adult Indians in Minneapolis in 1980-81. Two populations of Indians were reached—one in an Indian housing project (n=173) and the other through community screenings (n=295) during American Indian Week (33). Comparison data for whites were taken from the Minnesota Heart Survey (probability sample) (clinic measurements; n=1950). Self-reported prevalence of hypertension was somewhat higher among Indians than whites (33% vs. 27%), but unlike whites who reported a positive history, the Indians with a positive history were more likely than whites with such a history to have blood presssures in the normal range. This discrepancy was not fully accounted for by the proportions of Indians and whites on medications. The prevalence of hypertension (defined as diastolic blood pressure >=90 mm Hg or diastolic blood pressure <90 but taking blood pressure medication) was 13-17% among the Indians and 14% in whites. Levels of control among Indians and whites on medication were similar. A smaller proportion of Indians than whites with a history of high blood pressure were on medication; a larger proportion were on special diets (33). Ten percent of Indians reporting a history of high blood pressure also reported a history of diabetes. Diabetes, smoking, and obesity were more prevalent among Indians than whites. Systolic and diastolic blood pressure were significantly correlated with the body mass index (coefficients were 0.33 and 0.36 (33) in the Indians). The overall impression was that blood pressure levels of Indians and whites were similar, but that associated risks were higher in Indians due to higher prevalences of obesity, smoking and diabetes. These later studies add to the evidence that obesity- and diabetes-related hypertension is an important health problem in Indians and suggest the need for more systematic and careful studies of blood presssure levels and trends in different Indian communities. Attention to trends in younger cohorts seems particularly important (26,27). The workshop summaries in the 1977 conference report (45) do not contain data on hypertension prevalence (except in references to various 462 studies). However, this report is a useful point of reference for issues of Native American health services in general and hypertension control issues in particular. Many of the relevant research and data needs have already been identified and attention to some of these needs may already be underway. However, the picture obtained from literature identified in conjunction with the present report does not appear to have changed much from that reported at the 1977 conference. 4.3 Cholesterol Comparisons of cholesterol levels of Native Americans vs. whites indicate lower or similar levels in Native Americans. In a 1968 report of cholesterol levels of 746 southwestern Indians compared with 70 nonsouthwestern Indians and 163 whites, Sievers observed lower levels in both groups of Indians compared to whites, no increase in cholesterol levels with age, and no sex differences (peak levels between ages 30-44) (29). Cholesterol determinations were for Phoenix Public Health Service hospital patients over the age 15 during the period July, 1963 through December, 1965. The lowest levels were observed among Pima and Papago Indians. Pima and Papago Indians had lower cholesterol levels than Apache and Navajo, in spite of greater obesity among the Pimans. In a separate report, Sievers also comments on the paradox of lower cholesterol levels among the Pimans vs. Navajo and Apache in light of the higher frequency of myocardial infarction among Pimas (23). Cholesterol levels for nonsouthwestern Indians were slightly higher than for southwestern Indians but still lower than those of whites (29). For example, among 45-59 year old men in the study, mean cholesterol level was 193.7 for southwestern Indians; 214.7 for nonsouthwestern Indians; 302.8 for whites (levels for whites were duplicated in outside laboratories because they seemed unusually high; however, the levels of the Indians would have been significantly lower even if the levels for whites had been substantially overestimated). Hamman et al., in a 1975 report (52) noted similar cholesterol levels in 148 pre-menopausal (ages 35-54) and 75 age-matched post-menopausal Pima women. Similar pre- and post-menopausal cholesterol levels in women were also observed in 65 women followed longitudinally. The findings were not affected by the presence of diabetes. Cholesterol levels of males were also measured in the cross-sectional survey. The levels of Pima males and females were 20-30% lower than those of Caucasians and did not rise with age. A 1976 report of serum cholesterol levels in Pima children and adults indicates that levels of Pima and white children are similar at birth but that levels among the Pima remain low and do not rise with age in adulthood (17). Ingelfinger et al. (46) reported mean cholesterol levels of 189 and 188 mg/dl for nondiabetic Pima men and women over age 40 and slightly higher, but still low, levels for diabetic Pimas (197 and 198 mg/dl for men and women), with no age-related changes in either group. The most recent report of cholesterol levels in Pimas (from the same laboratories) presents data collected between 1979 and 1982 (14). Although levels in Pimas are 463 lower than those of the Lipid Research Clinic whites, the finding of an increase in plasma cholesterol with age in Pima women is a departure from the previous findings that no such rise occurred. It is not clear whether this is due to sample factors or to a "westernization" of cholesterol levels among the Pima women. Levels in men peaked in the 45-54 year age group and then tapered off. HDL cholesterol levels were similar in Pima males and women. The lack of a gender difference was not explained by obesity although obesity and HDL levels were negatively correlated. The fat content of the traditional Pima diet is primarily animal fat (from beef and pork). Meat consumption (and thus the percentage of fat calories) in the traditional Pima diet was relatively low compared to typical U.S. diets (29). However, Savage notes that the cholesterol levels of Pima remained lower than those of whites at a point when the usual Pima diet had become quite similar to that of the general population (17). Metabolic studies in Pima volunteers suggest that there are significant differences in apoprotein and lipoprotein metabolism which may account for the different lipid profiles of Pima Indians (15,24). The high prevalence of gallstones among southwestern Indians, including the Pima, has been suggested as a possible explanation for the lower blood cholesterol levels (3,11); however, this view is not widely supported. The 1956-62 Many Farms Study among the Navajo indicated relatively low cholesterol values in 89 and 113 randomly-selected men and women for whom cholesterol determinations were made (31). The authors were reluctant to generalize to the Navajo as a whole, partly because of their belief that the Navajo diet was high in saturated fat. In fact, the study was motivated by what appeared to be a paradox of high fat intake and low CHD in this population. Although fat from mutton or lard is an important part of traditional Navajo food, other authors, including Keys (25,26) have pointed out that the total percent fat in the Navajo diet was around 24%, much lower than the 40% typical of Americans in general. Also, dairy fat is not common in the Navajo diet (26). Two studies among the White Mountain Apache (26,28) indicated low cholesterol levels in this group (in the range of 185 to 200 mg/dl) compared to whites. However, Clifford et al. noted a lower percentage of cholesterol levels under 225 mg/dl among younger Apache men ages 25-44 vs. 45-64 year old men (72% vs. 86%) and commented that cholesterol patterns in the younger Apache men might be changing (26). In the study of Seminole Indians in Oklahoma (36), the prevalences of serum cholesterol >260 mg/dl were 9.1% and 5.4% in Seminole men and women vs. 15.6 and 9.1% in white men and women. The authors attribute most of this difference to the younger ages of the Seminoles in their sample and state that the overall serum cholesterol distributions were similar. The Minnesota study (33) is the only recent report of cholesterol levels among Indians other than Pimas which could be identified. Total cholesterol levels of the Minneapolis Indians were similar to those of whites under age 55 and lower after age 55. HDL cholesterol levels of Indian and white men were simliar; lower HDL levels among the Indian vs. 464 ■p- ON O t-tX X o hh 3 CD CD hh H O rt rt TJ pi cr cr s: p-» m CD CD CD CD n g Cfl CD pl o cd p-« 7? p- 3 sc pi • 3 CL p. Cfl rt rt O 3 3 CD CD 3 OQ O H Z h{ JC O < Cfl pi CD /-\ 3 P- Pi 3 CL p« CD CD 3 O vd £ TJ 3 H ^J O CD t—> OQ CD ^-J hh CL CD Cfl w rt Cfl rt £ O 3d Cfl CT CD hh CD rt rt CD h( 3" 3 CD Ln 2 pi CLTJ 3 ^ CD O O P)tJ 3 O 3 CD Ln pa 3 hh *< Pi 3 M p. pi h) t3 CW w 7? p- g ' CD CD Cfl h{ Cfl Z pi 3 O rt t-i JC 3d < en CD CD > pj pi CL Cl < P- O CL pi Cfl pi 3 pi i_i. g c* Cl ^-n ^< O O O P- p> ^ rt < vd ^ CD 3d P- <-n N3 CL CD CL On hr) ^ h{ 3 I O n—' 3 Pl ON hi * 0 to h-' to rt hi ON Cfl v^-> ^ P1 CD w LO Cfl s tj ^s rt g s~\ pl CD 3- 0 LO Cfl h S pi JC P1 0 CD 3 CD w Pi CD hi CL ■ Cfl 3 CD 0 Cfl rt 3 N5 CD Cfl CD hrj Cfl 0 3 rt 0 Cfl hh 0 X 0 hi CD 7? Pi rt CL rt CD o Ln v< 3d M 7? i cr CD Cfl 0 0 << X P- 3 Z Pl CD OQ CD p- Pi 3 h( Pi 3 < CL h( 0 rt pi CL CD hh CD i_i. VD Pl rt h( 0 4>v<- rt ^J < ^ CD Cfl -P> P-CD hd p- g pi pi 3 Cl CC O X p- rt rt O 3d P- cr c/i CD hi cr cd CD x pi o < M KJ rt pl s-\ h-« p- o • O CD 3d • o - h- 3 X CD hi O CD 3 rt P- 3 rt 3d CD 0 rt 3d CD h( cfl o CD 3 Cfl CL - P- 3 ZOQ Pi < Cfl pi o <-*• 3 O CD o P- rt OQ £ Pi O hi CD OQ rt p. h-to >>3 rt p. p Co $ TJ 3 O X 3 Cl g P- 3 pi 0 O 3 CTOQ 3 CD CD rt hj pl pl 3* CD g 3 CD O CL 3 3 Cfl O CD OQ O 0 O 3 3 rt Cfl pi 3 CD JC cd cr S 3 O CD rt hh 3d hi O OQ 3 H O Pi 3 TJ Cfl Cfl p^ rt 3 cr o cr hh rt O CD hi £ hi c o cr Pl CD CD CL hh p-> pi O rt 3 O < CD CD 3 rt SI Pi Sd M CD C_i. p. hj hi O h| Pl 3 OQ CD 0 hh 3 3 CL O rt pi cr 3 pi Cl ph p- 3 3 3 Pi OQ T) Cl 3 hh Pi P- Cfl 3 OQ O TJ Pi - CD 0 3 hh CL p- CL pl Pl h{ CJ "" p. Cfl 3 7? rt P- 3d 3 Pi CT 3d 0 CL Cfl p« TJ P- O rt 3 Co P (D P 3 3 „ Pi Cfl rt £ rt CD O P- 3 O pi CD 3 (_i 3 v.. O O rt ph ET O 3 O CL PCT H- CD Pl Pi 3 3 hi o cr ht CD CD Pl O < 3 rt v< CD cr 3 pl Cfl 3 g O o 3 CL P- o CC hH 0 0 3 • pj Cfl CL ■F> h-' TJ p- 0 P- pl Ui z rt 3 3 Cfl Pi Cfl 0 .. pj Ui 7? p- P- p- cr 3 CD 3 cr CD < OQ CD rt Pi CD pi s hi Pi < CD p1 Cfl 3 ^ CD VD CL 3 ON 00 p» VD hi ON CD h-» TJ o CL o o 3 g CD 3 rt CD CL 7? rt CD Cfl pi Cfl TJ 3 g pi Cl pi rt h-" rt n p ic O CD M !-■ hi 0 o ca i-t rt • Pl h( CL P- o cr o CD hh PH Cfl 3 n-> CL- P- Pl > 3 TJ cfl pi O 3 cr tr CD CD CD X Cfl Cfl O H 3d rt CD 0 rt on hi ^5 P- cr p. Pi 3 pj rt X 3d Pi CD rt rt Ui CD 0 hi pi 3 OQ Cfl rt 0 - 3d CD 3d CD Pl 3 < O VJ OQ CC h( CD O CD Cfl pl S 3 3 o O 3 g 3 CD CD hi < pl CD h-> h| 3 P- P- 3 3 rt 3d hH £ pi o cr 3 £ P- pi rt —' CD ■ Cfl 0 3 o cr cd ca cl £ • CD CD TJ Cfl CD 0 rt M CC hh hH 3 3 Cfl CL Pl pl P- 3 OQ pi CL CD 3 Cfl p» r-■. VD CT Pl O Cfl P' ON CD OQ Ln CL Pl hi ^ O ro 3 LO CD hi S 3 CD CD "-< CD hH JC jc 3 p. OQ W rt Cfl CD rt 3 3 rt Cfl 3d O CD hh Cfl rt cr CD < Pl Cfl < • pl O OQ 3 O 3d rt CD 3 CL P- £ Pl Pl 3 ca Cfl pj pi pi cd cr 3d 3 o hi 3 Cfl 3 pj OQ CD 3 •-< Pl Cfl h-- O O 3 O rt O £ h-1 CD Cfl 3 rt CD ""d Cfl pl o cfl p- rt hh g pi pl pl 3 g CL O Pi 3 ' O O OQ O O g OQ TJ pl pl hi h( g pi o 3 Cfl Cl rt Cfl pi £ CO O OQ CD P- O CD hi CD hi TJ TJ CD CD Pl hi O O 7? CD 3 TJ rt CD P. P' 3 CD rt rt P- CD 3 hh 3 P- i-hOQ Cl O 3 O p. Cfl CD 7? rt a CD p-> h-* P- ^ p. 3 - JC CD O 3 TJ CD co CC w O O Pi 3 TJ 3 w rt p. O rt 3 CD V 3 p* O • H ON CD rt O rt 3d O Cfl CD CL Pl 3 V! O hh hh CD hh P- 3 CL O P- hh 3 OQ TJ CA Pl pi OQ 3 H CL O 3 p> OQ CL pi TJ 3 P- h-> CL rt CD O 3 3d CL O Pi 3 CD • CT Cfl 0 O Cfl h-1 rt 3 Oh-« rt Cfl 3 P- p- O 3 7? vj 3 O CD rt (D P oq cr cfl vj hi % CD CD O rt pi Cfl hh 3d rt rt CD CD h{ h( 3 Pi pi g h-1 o CD 3 Cfl OQ Cfl 3 o 7? P- 3 OQ rt CD P- CD TJ 3* Pl CD 3 O 13 3 3d OQ O > 3 3 P- CD X h| P- hd O CC Pi Ui 3 o 0 cr o a Vi CD 3 O ht cr cr o pi (—• 3 CD p. Cfl N rt pl Pl CD p^ rt h( h-1 P- O O h-- CD 3 hh P^ hh K CD O CD < hi rt CD 3 hh P- O CD rt M P- rt Cfl CT pi O 3 h- CD p. pj 3 H hh P- rt 0 3 3 h-' CD h-' CL P' O • 3 £ O h| hH H CD CC cr pi a cd w Cfl P hi CD p- Cfl p- fC Cfl pl Cfl O hh 3 O pl CD 3 O Cfl P P- CD 3 O OQ CD CD 3 ►-! O Pl CD h-' 3 OQ O Cfl 3 hi cr M o Pi rt 3 rt h{ h-« P- CD CL O 3 3 cl cr - Cfl CD h| p- pi pl 3 Cl O CL P- rt h( pi C (t p p. Cfl Cfl Cfl pl CD CL cl g p- rt 3d Pi rt Cfl CD hi 3 g o 3d O \-> CD Cfl rt CD P- H O O TJ O TJ pi 3 Cfl pl Pl h| rt (D P- TJ O Pi Pi 3 H 3 • rt Cl 0 hh o 3 Cfl 3d P- 3 TJ CL Cfl O hh P- 3 rt 3d CD Cfl CD Pl 3 0 3 pi OQ rt Cfl 0 g CD 3 CL p- pi 3 Cfl Pi 3 rt CD 3* P- CD M O Cfl M O p-- rt cd cr hi CD 0 Cfl 0 p. < Cfl CD hi Pl Pl 3 I- 0 h- 3 OQ hh Pl M rt Cfl PC o P- o 3 3 0 rt Cfl CD 3 3^ rt Pi Cfl O Hi cr CD rt CD 3d 3 CD CL W O Cfl O 3 3 CD 3 rt CD CL CD rt ro 3 Ln O Vw/ h{ rt cr CD pi cr Cfl CD 3 O CD 0 hh cr cl pi pl CD 3 < hh rt cd cd cr 3 O Pl CL hf p. Cfl cr 3 P- OQ CD OQ 3 CT rt TJ cr cr hh CD pi pi Cfl rt O P- CT N CL O CD P- h-» CD CD rt rt Cfl cr rt pi CT CD rt 3 M rt O rt i- cr h-1 I CD POM rt CC CD rt a i- S cd cr Pi «-< Cfl pi TJ rt O O cr rt o cd cr 3 hi CD Cfl O Cfl P- Cfl p- CL O Cfl CD h-> M CD pl Pl ht OQ CT O pi Cfl p. P- 3 Cfl Cfl h| " rt CD OQ WCo H> CD Cfl O VJ Cfl 3 Cfl CD pl ht h- TS rt O P- P- o 3 3 rt Cfl Cfl CD < Pl hi P- Pl rt rt i o cr p- 3 Pi o rt rt 3 rt W CT Cfl Pi 5C ph rt p- 3 3 O Cfl rt cr CD Ui > o h-- o pi rt Cfl rt fC pi CD rt 3 pi Pi rt (-■ P- • 0 3 ^ pi Ln h-> O OQ 3 H pl CD ht TJ CL 0 • hi rt CD s a CD pi o 3 3 o o tr cr o o pj pj CD CD Cfl Cfl rt rt CD CD ht hj o o CD CD < < CD CD h-» P-1 Cfl pi o Cfl hh ro 00 P* H> 4^ to W Cfl fC p- 3 O 3 CD 3 P- 3 O £ cr cr o P- h-> rt CD CD Cfl rt £ CD O h g O CD h-> 3 V £ to CD LO h( O CD 3 3 OQ O \ rt CL CD H* CL C " pi TJ Cfl pi h( rt P- O 3 h-' Pl h( ro pi P» hh rt CD h| LO 3 CL P- Pl 3 Cfl £ p- rt 3d O O 3 TJ p-1 CD rt CD pl OQ CD Ln Ln P TJ P- CL Cl Pi rt Pi h-3 3d CD TJ h( CD < Pl h-' CD 3 O CD O hh Cfl CD ht 3 3 men under age 40 compared to 22% over age 40 reported using alcohol in the 1977 study. The authors noted that this may have indicated a trend towards higher alcohol use in the younger cohort. Ingelfinger reported that only 2.5% of Pima males and no Pima females smoked more than one pack of cigarettes per day (46). In contrast, smoking was highly prevalent among the Minneapolis Indians surveyed by Gillum et al. in 1980 (33). However, this probably does not represent a trend towards increased smoking among the Chippewa. According to the authors, smoking has been highly valued in traditional Chippewa culture. Seventy-seven and sixty-seven percent of the Indian men and women surveyed were current smokers; 45% smoked more than 20 cigarettes per day (there was no sex difference in number smoked per day). Of those who smoked, only 25% reported wanting to stop smoking, although most had tried to quit. Reported alcohol intake in the Minnesota population was low. Forty-eight percent reported abstinence; those who drank reported consumption of alcohol at least two days per week. Cigarette and alcohol use of 76 Native American students in the University System of Georgia were ascertained in a 1972 study of 20,547 students (41). Reported cigarette use among the Native American students was highest of all the ethnic groups (white, black, Asian, other) and reported alcohol use was second highest. A 1973 study of drug use patterns among American Indian and white high school students in Wyoming reported more favorable attitudes towards drug use among the Indian students, but alcohol was not among the drugs specifically queried (35). In a survey of students between 10 and 20 years of age in Anchorage Alaska, Native American students (5.7% of 15,634 respondents) were less likely than white students to be nonusers of any drugs and more likely than any of the other ethnic groups to have tried drugs in addition to alcohol or tobacco (37). The association of cigarette and alcohol use with IHD risk among Native Americans has apparently not been explored, although the social and other health consequences of heavy alcohol use in Indian populations have received great attention in the Native American literature (3,53). Insufficient data are available to draw any conclusions about trends in these variables or their contributions to IHD patterns. 5.0 MULTIVARIATE PERSPECTIVE A discussion of obesity and diabetes among Native Americans is beyond the scope of this paper. However, it is appropriate to draw attention to the apparently large public health importance of these two conditions in most Native American populations and to their enhancement of what might be (for many tribes) an otherwise relatively benign IHD risk profile. A greater prevalence of obesity in American Indians vs. whites was mentioned in the majority of papers reviewed (7,11,12,20,23,26,27,30,31,33,36,39,40, 45,46,48,49,51) and in others not noted above (54-57). Two papers (39,40) note the excess prevalence of obesity among Seneca Indians--a group for 466 whom other papers related to IHD risk were not identified. The mortality risk associated with obesity among Pima Indians does not appear to be the same as for the general population (20). The excess risk of diabetes among Native Americans is evident in the mortality data. The Pima Indians have an unusually high prevalence of diabetes, but an excess of glucose intolerance prevalence is apparently typical of many adult Indian populations (3,12,13,20,33,36,39,40,46). The absence of diabetes among the leading causes of death for the Navajo (11) is consistent with a low or moderate prevalence of diabetes among some Athapascan tribes (Navajo and Apache)(3,23). Diabetes is reportedly common in Apache tribes in Oklahoma (3). From a multivariate perspective, Native Americans are at high potential IHD risk. With the exception of elevated cholesterol and smoking, standard risk factors and predisposing co-morbid conditions are equivalent or higher in American Indian populations where comparisons have been made. This suggests that the future of IHD risk may become worse for American Indians at a time when the picture is improving for large segments of the general population. 6.0 IMPLICATIONS AND RECOMMENDATIONS To draw conclusions about current IHD risk perspectives among Native Americans on the basis of the limited data available would be inappropriate. Probably there is no one conclusion that will apply to all Native American people; there are at least two distinct patterns implied. Indians in the Southwest have lower overall IHD risk than Indians in other parts of the country. However, Native Americans in all parts of the country may have excess IHD risk associated with excess obesity and diabetes. In some areas of the country Native Americans are at excess IHD risk due to lifestyle factors (cigarette use). Although the risk of hypercholesterolemia among Native Americans may be lower than in the general population, in no sense does this result in "immunity" to IHD. In spite of an impression from available literature that IHD is not an important contributor to overall Native American-white mortality disparities (due to extreme disparities in certain other mortality classes), the fact that heart disease is the leading cause of death among Native Americans must be kept in mind. Moreover, there is evidence of excess heart disease mortality among Native Americans at younger ages. In conclusion, the following recommendations are put forth: --systematic data collection is needed, if not tribe-specific then at least specific to major subgroups which show different risk patterns (e.g., Southwest vs non-Southwest); 467 --more of the Indian Health Service data should be included in frequently used publications of health statistics; --intervention on those risk factors already identified is indicated, particularly prevention of obesity and diabetes; the disparities in this area are huge and are bound to result in increasing health problems for Native Americans as their life span increases and deaths from other causes are reduced. ACKNOWLEDGEMENTS We gratefully acknowledge the editorial assistance of Sandra J. Anderson and Elisabeth Pitt. 468 7.0 REFERENCES 1. DHHS. Task Force on Black and Minority Health. Black-White Mortality Cross Over. TFBMH/CHPatrick 7/12/84 2. DHHS. Indian Health Service. Indian Health Service Chart Book Series. June 1984. (excerpts only) 3. Sievers ML, Fisher JR. Disease of North American Indians. IN Rothschild HR, ed., Biocultural Aspects of Disease. New York: Academic Press, 1981, Chapter 8, pp. 191-252. 4. Gonzalez NL. Changing dietary patterns of North American Indians. In Moore WM et al., eds. Nutrition, Growth, and Development of North American Indian Children. Washington, D.C. U.S. Governmnent Printing Office. DHEW Publication No. (NIH) 72-26, 1972), pp. 15-33. 5. Taylor TW. The States and Their Indian Citizens. United States Department of the Interior. Bureau of Indian Affairs. Washington, D.C. 1972. 6. U.S. Department of Health, Education, and Welfare. The Indian Health Trends and Services. U.S. Government Printing Office. Washington, D.C. 1978. HSA 78-12009. 7. Sievers ML. Disease patterns among Southwestern Indians. Public Health Reports 1966;81:1075-1083. 8. U.S. DHEW. Selected vital statistics for Indian Health Service Areas and Service Units, 1972-1977. DHEW Publication (HSA)79-1005. 9. Health of Minorities and Women. Chartbook. American Public Health Association. Washington, D.C, 1982 10. Health of the Disadvantaged. Chartbook II. September 1980 DHHS Publication No. (HRA) 80-633. 11. Kunitz SJ. Disease Change and the Role of Medicine. The Navajo Experience. University of California Press. Berkeley CA. 1983. 12. Strotz CR, Shorr GI. Hypertension in the Papago Indians. Circ 1973;48:1299-1303. 13. Howard BV, Lisse JR, Knowler WC, Davis MP, Pettitt DJ, Bennett PH. Diabetes and atherosclerosis in the Pima Indians. Mount Sinai J Med (NY) 1982;49:169-175. 469 14. Howard BV, Davis MP, Pettitt DJ, Knowler WC, Bennett PH. Plasma and lipoprotein cholesterol and triglceride concen- trations in the Pima Indians. Distributions differing from those of Caucasians. Circ 1983;68:714-724. 15. Howard BV, Zech L, Davis M, Bennion LJ, Savage PJ, Nagulesparan M, Bilheimer D, Bennett PH, Grundy SM. Studies of very low density lipoprotein triglyceride metabolism in an obese population with low plasma lipids. Lack of influence of body weight or plasma insulin. J Lipid Res 1980;21:1032-1041. 16. Reitman JS, Kosmakos FC, Howard BV, Taskinen MR, Kuusi T, Nikkila EA. Characterization of lipase activities in obese Pima Indians. J Clin Invest 1982;70:791-797. 17. Savage PJ, Hamman RF, Bartha G, Dippe SE, Miller M, Bennett PH. Serum cholesterol levels in American (Pima) Indian children and adolescents. Pediatrics 1976;58:274-282. 18. Klimes I, Nagulesparan M, Unger RH, Aronoff SL, Mott DM. Reduced Na+, K+ -ATPase activity in intact red cells and isolated membranes from obese man. J Clin Endocrinol Met 1982;54:721-724. 19. Nagulesparan M, Savage PJ, Mott DM, Johnson GJ, Unger RH, Bennett PH. Increased insulin resistance in obese, glucose- intolerant Southwestern American Indians. Evidence for a defect not explained by obesity . J Clin Endocrinol Met 1980;51:739-742. 20. Pettitt DJ, Lisse JR, Knowler WC, Bennett PH. Mortality as a function of obesity and diabetes mellitus. Am J Epidemiol 1982;115:359-366. 21. Sasaki H, Nagulesparan M, Dubois A, Samloff M, Straus E, Sievers ML, Unger RH. Gastric function and obesity. Gastric emptying, gastric acid secretion, and plasma pepsinogen Int J Obesity 1984;8:183-190. 22. Sievers ML. Cigarette and alchohol usage by Southwestern American Indians. Am J Pub Health 1968;58:77-82. 23. Sievers ML. Myocardial infarction among Southwestern American Indians. Ann Int Med 1967;67:800-807. 24. Garnick MB, Bennett PH, Langer T. Low density lipoprotein metabolism and lipoprotein cholesterol content in southwestern American Indians. J Lipid Res 1979;20:31-39. 25. Keys A. Coronary heart disease--the global picture. Atherosclerosis 1975;22:149-192. 470 26. Clifford NJ, Kelly JJ, Leo TF, Eder HA. Coronary heart disease and hypertension in the White Mountain Apache tribe. Circ 1963;28:926-931. 27. DeStefano F, Coulehan JL, Wiant MK. Blood pressure survey on the Navajo Indian reservation. Am J Epidemiol 1979;109: 335-345. 28. Leo TF, Kelly JJ, Eder HA. Cardiovascular survey in population of Arizona Indians. Circ 1958;18:748 (abstract). 29. Sievers ML. Serum cholesterol levels in Southwestern American Indians. J Chron Dis 1968;21:107-115. 30. Chase-the-Bear R, Bonnell M. Morse HG, Rate RG. Hopis and Navajos not lean. N Eng J Med 1979;301:1348 (letter). 31. Fulmer HS, Roberts RW. Coronary heart disease among the Navajo Indians. Ann Int Med 1963;59:740-764. 32. Connor WE, Cerqueira MT, Connor RW, Wallace RB, Malinow R, Casdorph HR. The plasma lipids, lipoproteins, and diet of the Tarahumara Indians of Mexico. Am J CLin Nutr 1978; 31:1131-1142. 33. Gillum RF, Gillum BS, Smith N. Cardiovascular risk factors among urban American Indians. Blood pressure, serum lipids, smoking, diabetes, health knowledge, and behavior. Am Heart J 1984;107:756-776 34. Pinkerton RE, Badke FR. Coronary heart disease. An epidemi- ologic study of Crow and Northern Cheyenne Indians. Rocky Mount Med J 1974;71:577-583. 35. Cockerham WC, Forslund MA, Raboin RM. Drug use among white and American Indian high school youth. Int J Addict 1976; 11:209-220. 36. Mayberry RH, Lindeman RD. A survey of chronic disease and diet in Seminole Indians in Oklahoma. Am J Clin Nutr 1963;13:127-134. 37. Porter MR, Vieira TA, Kaplan GJ, Heesch JR, Colyar AB. Drug use in Anchorage, Alaska. A survey of 15,634 students in grades 6 through 12-1971. JAMA 1973;223:657-664. 38. Longclaws L, Barnes GE, Grieve L, Dumoff R. Alcohol and drug use among the Brokenhead Ojibwa. J Stud Alcohol 1980;41:21-36. 471 39. Judkins RA. American Indian medicine and contemporary health problems. IV. Diabetes and perception of diabetes among Seneca Indians. NY State J Med 1978;78:1320-1323. 40. Doeblin TD, Evans K, Ingall GB, Dowling K, Chilcote ME, Elsea W, Bannerman RM. Diabetes and hyperglycemia in Seneca Indians. Hum Hered 1969;19:613-627. 41. Strimbu JL, Sims OS Jr. A university system drug profile. Int J Addict 1974;9:569-583. 42. McMurry MP, Connor WE, Cerqueira MT. Dietary cholesterol and the plasma lipids and lipoproteins in the Tarahumara Indians--a People habituated to a low cholesterol diet after weaning. Am J Clin Nutr 1982;35:741-744. 43. NCHS. Blood pressure levels of persons 6-74 years, United States, 1971-74, by Roberts J, Maurer K. Vital and Health Statistics. Series 11. No 203. DHEW Publication No. (HRA) 78-1648. September 1977. 44. U.S.DHEW. Indian Health Service. Sample Tribal-specific Comprehensive Health Plan. San Carlos Reservation. 1978. 45. DHHS. Proceedings of the National Conference on High Blood Pressure Control in Native American Communities. NIH Publication No. 79-1960. April 1977 (reprinted August 1979) 46. Ingelfinger JA, Bennett PH, Liebow IM, Miller M. Coronary heart disease in Pima Indians. Electrocardiographic findings and postmortem evidence of myocardial infarction in a population with a high prevalence of diabetes mellitus. Diabetes 1976;25:561-565. 47. Sievers ML, Fisher JR. Increasing rate of acute myocardial infarction in southwestern American Indians. Ariz Med 1979;36;739-742. (Cited in 3) 48. Sievers ML. Historical overview of hypertension among American Indians and Alaskan Natives. Ariz. Med 1977; 34:607-610. (cited in 3) 49. Alfred BM. Blood pressure changes among male Navajo migrants to an urban environment. Canad Rev Soc Anthropol 1970;7:189-200. (cited in 33). 50. Scott Em, Griffith IV, Hoskins DD, Whaley RD. Serum cholesterol levels and blood pressures of Alaskan Eskimo men. Lancet 1958;2:667-668. 472 51. Gillum RF, Prineas RJ, Palta M., Horibe H. Blood pressures of urban Native American school children. Hypertension 1980;2:744-749. 52. Hamman RF, Bennett PH, Miller M. The effect of menopause on serum cholesterol in American (Pima) Indian women. Am J Epidemiol 1975;102:164-169. 53. Thomas RK. The history of North American Indian alchohol use as a community-based phenomenon. J Studies Alcohol 1981;suppl)9:29-39. 54. Savage PJ, Dippe SE, Bennett PH, Gorden P, Roth J, Rushforth NB, Miller M. Hyperinsulinemia and hypoinsulinemia. Insulin responses to oral carbohydrate over a wide spectrum of glucose tolerance. Diabetes 1975;24:362-368. 55. Sievers ML, Hendrix ME. Two weight-reduction programs among southwestern Indians. Health Serv Rep 1972;87:530-536. 56. Westfall DN, Rosenblood AL. Diabetes mellitus among the Florida Seminoles. HSMHA Health Rep 1971;86:1037-1041. 57. Reid JM, Fullmer SD, Pettigrew KD, Burch TA, Bennett PH, Miller M, Whedon GD. Nutrient intake of Pima Indian women. Relationships to diabetes mellitus and gallbladder disease. Am J Clin Nutr 1971;24:1281-1289. 473 Stroke Report Lewis H. Kuller, M.D., Dr.P.H. Chairman, Department of Epidemiology Graduate School of Public Health University of Pittsburgh Pittsburgh, Pennsylvania MORTALITY Stroke is a major cause of death and disability in the United States. In 1981 there were 163,504 stroke deaths (Table 1), including 9,429 among non-white men and 11,310 non-white women. Most of these deaths, 8,760 in men, and 10,656 were among black men and women. Age adjusted stroke death rates are approximately two times as high in the all other or black men and 1.8 times as high in black women than for white v. mien who have the lowest death rates (Table 1). Age specific stroke death rates increase substantially with age (Table 2). Approximately doubling with each increase in 5 year age group (Table 2). Stroke death rates are much higher in the younger and middle aged blacks than whites (Table 2). The mortality rate is generally over 3-fold higher through age 65 (Table 3). Stroke is estimated to reduce life expectancy at age 45 for non-white men by 12.3 years and for non- white women by 14.7 years (Manton and Baum, STROKE 1984;15:451-457). Stroke accounts for about seven percent of all deaths in 1981, including about 10% among black women (Table 4). Stroke is also an important contributing cause of death on the death certificate, that is not listed as the underlying cause (Table 5). In 1976-1978, the last year in which current data is available, stroke was the underlying cause for 31,368 deaths among black men and contributed to another 19,216 based on the death certificate listings and for women it was the underlying cause in 35,864 and in 21,313 a contri- buting cause. Previous studies have shown that when stroke is the under- lying cause of death, the individual has usually died from the acute complications of stroke usually within the first several months after the stroke. However, many other individuals who survive the initial stroke event may subsequently die of a stroke or related cause without any mention of this on the death certificate. There is a very substantial variation in stroke mortality by geographic area. These differences have persisted for many years. In 1969-1971 (Table 6), stroke death rates were 10 times higher for black men in Savannah, Georgia than for white men in Baltimore, Maryland aged 45~54. In 1978 stroke death rates aged 45~54 were still four times higher in Georgia for black men than for white men and 3.5 times higher for black than white women. Rates were still substantially higher in Georgia and other areas of the southeast than in the Great Plains or Rocky Mountain areas (Tables 7 end 8). Previous studies have clearly demonstrated a marked socioeconomic gradient in stroke mortality for both black men and women (Table 9). There are also marked variations in stroke mortality within a state such as North Carolina (Table 10). A recent study in Allegheny County, Pennsylvania, for white men continues to demonstrate a considerable socio- economic gradient in stroke mortality. These observations are consistent with a greater prevalence of risk factors especially hypertension in lower socioeconomic and less educated populations and also poor control of these risk factors. 477 High stroke death rates among blacks are not limited to the United States. Stroke is reported to be increasing and common in Africa. A recent WHO report noted stroke rates in Cape Verde Islands (Table 11), similar to those of blacks in the United States. Unfortunately mortality statistics for other countries in Africa with predominantly black populations are not available. However, clinical studies suggest an increase stroke mortality associated with hypertension especially in urban areas. In the March, 1983 report of the East Africa Medical Journal, Daresi, et al., noted that cerebral vascular disease was becoming a major cause of mortality and morbidity. One hundred eighty stroke cases and 180 age and sex matched controls were studies over a two year period in Lagos, Nigeria. The stroke patients had a much higher prevalence of systolic and diastolic hypertension, diabetes mellitus and obesity. In Lagos University Teaching Hospital, cerebrovascular accidents were the second most common of all neurological admissions between 1962 and 1967, the most common being tetanns. Evidence from the Stroke Registry in Ibadan, Nigeria, has shown that the incidence of stroke was 26/100,000. Daresi, et. al., East Africa Medical Journal, March 1983. The prevalence of cerebro atherosclerosis increases in Nigeria with age and with the presence of a history of hypertension (Table 12). The prevalence is not as high as in U.S. populations (Table 12). Therefore, the current evidence from Africa suggest that stroke is becoming increasingly prevalent, that it is related to hypertension and perhaps diabetes, and that cerebral atherosclerotic disease is increasing especially among hypertensives. Coronary artery disease and myocardial infarction is still very uncommon. The picture in Africa especially in urban west Africa is probably a replication of changes in risk factors and stroke that occurred many years ago in the United States. Blacks in the United States are also reported to have a higher prevalence of both intracranial and extracranial pathology as compared to whites (Tables 13-14). Stroke death rates are not elevated among the Chinese, Japanese, Koreans and Philippino's compared with U.S. whites in the United States (Tables 15-17). There also does not appear to be an excess stroke mortality among Hispanics in Southern Texas (Table 18) as compared to whites. There may be a slight excess of stroke mortality among Puerto Ricans in New York especially in the younger age groups as compared to whites (Table 19). The apparent absence of an increase in stroke death rates among Mexican-American populations may be of particular importance. This population is generally of lower socioeconomic status, has a very high prevalence of obesity especially among women, and diabetes, but little increase in hypertension. The failure to note an apparent increase in hypertension with acculturation may be due to genetic factors or diet. Further investigation of this population should be done especially in search of possible protective constituents in the diet or selected genetic factors. Stroke death rates among American Indians appear to be slightly higher than the rest of the United States population, but lower than those for U.S. blacks. Stroke death rates also appear to be declining among American Indians (Tables 20-21). 478 VO X M O 00 M § Cl rr CD O P* M CD CD Pl M n> cr p co «*rj p X P4 CD 0 to <-t PJ < rt p CD CD •>< P a oi o CD P« M < CD X rt M P4 o CD Oq H O rt p4 P (X CD CD P- CO Cl CO CD CD P- Pl O P CO O a 3 P- rt CD cl P« O 3 O l-h p O rt M M 0 O pi pJ rt P-TJ O M P O OQ h- M CO pi „ 0 a- co CD pl p* I-1 H ^O 03 -.J P* O I CD H-» VO U3 00 o o m CD H P4 co a co 13 g H- 0 rt CD CD Cl P Pi Ui M rt H oi CD TJ • CD CD O CD CO CD rr rt X P* P4 O CD CD CD M TJ CD Pi rt O- p. p- co O Pi CD P CO p O 0 Pl v<< co O O hh CT cd cr Pi CD P H CD Pl P CD P X s rt Pi CD p P) co p«0Q p. p O < - * CD £ O P4 c rt p- co M O CD P4 01 0i rt rt rt rt 0 P- CD CD CD 0 P X rt rt POP CD 0 P CD TJ P p- oi oq m CD Mi P- 01 P- CD p rt rt CD CL O M CD o CD Cl CD Pl rt p- M pi rt CD co Pi X X CD Pl M CO rt O ^ cr P4 CD P- 0 tj p> o rt O Ml P4 CD f» CD CO rt O Ml rt P4 CD n o p p rt M CD M CD H« CO P O n p M P* H Pl CD O P 7? rt co a. oi p> P rt Cl 0i rrl pi rt rt CD co Ui P o p4 cr CD TJ IX - CD 0 Pl o P P cl CO rt Ml h o pi M rt CD Pl CO 0Q CD 3 P*0Q CD M It o P4 p CD X M CO rt OO P4 Ui CD 1 00 CO ■"> rt H o 7? CD H O Pl Ml cr i-1 rt CD p4 CD to ^4 CW v_x {D • O 0Q M H Pi p*TJ CD P4 P« Cl O CD O < P-* Oi P« M P H- CD 01 rt p. p. P o P CO CO rt •1 H* O P K* CD 01 rt Cl M CD O 01 7? rt CD *B H O Pl M rt rt CD pi co p-1 P« cr rt CD VCJ rt CD p CD P ft p4 CD If?. CD CO i-» tj p« pi o p pi p* P O CO H P P- D. P P- Pl Ui P 01 co p ui h a it cr cd 01 CD O It h-> CD O- p. CO CD P O CD P- M> O 3 CO CO CO fc CD r CD C X ? "1 C CD >* < P«TJ O 1 P O co a* p> rt a4 01 P4 i-1 CD Pl co 0 w O • P CM ^ n P- P4 CD CD Pu C_i CD 0i O X p- p P CD CD CO CD P* Oi M" CO p n X CD Oi c pi P-1 M- rt p> CD ex. pi p p- p- p pi 0 O p P« P CO It M O !^ It CD P* •"» M O 0 7? O P* CD H O rt O 0 Pi C O M1 0 M P- CD rt rt P P> "- 3* CD CO H- P- 0 P. P'OQ P- P- Pi M Pi P P- O Ml O Ml CD It rt C It p* H p4 rt c B O Cl Oi rt CD 0 P4 Ml M gg TJ 01 M It O O rt P4 CL CD CL M CL p- It P4 H pi cr M1 CD ro CD Cl CD O 01 Ml to rt H O 7? CD 0 O M It 03 P« (t "^ cr Pi CD CL CD n M CD Pl CO CD cr pi CO a4 CD CD p p O P It O CD P Pl P-1 P. P-CO CD Pl CO CD CD CD CO T3 CD O P« Mi P-O M c 0Q rt P4 CD n cr O CD 7? * CD O • C P-1 O CD CO P. pi CD p Pl P. CO O M» P o CD M CD TJ CD P* M CD It O CD 1 P* CD Cfl Pl P« CO < CD CD Ml P-O M M- C P'OQ O p. rr P« CO 3 Pi cr CD It p4 CD CO pl 0 CD a4 CD M o c P1 O CO p« CO O o o c M H CD CL CD P1 TJ CD < P-O P CO pj p O rt It O CD CD P H0Q o-B ^ H' p. 3 pj P O 0Q Pi 0 0 P CD O ft p CD CO CD cr co O CD ^ MiOQ P- 01 ft p c P4 P 0Q CD P«0Q P CD O 0 pd O o p4 CD 01 rt CD M S H« P P ft CD CO P« 01 fX o CD O C O M Ml 01 O ft *^ P4 CD O Ml P-O ft co p4 TJ CD P- 01 C o a It VJ P'~ o CD /-s CO H 0> p. cr CD Ml p- ro h-» OJ p* >-* p--P 0Q 01 CL p-p OQ O Ml ft p4 CD Pu CD 01 It P4 CD P« 3 O M CD 01 CO CD P-P P4 CD Pl M O H CD 03 CO CD P-P O 0) M PL P-O < 01 P 0Q it M CD Pi It 1 P rt Ml O M P4 H P4 CD CL CD a M1 p-p CD p« p pi 0Q P O CD H« CO O p rt P4 CD CL CD 01 Oi P rt P4 •i O & O CO p-CO B3 O Ml CD O It 01 Pl P* P-Pl 0Q CD OQ M O P Ui rt M O 7? CD CL CD Pl ft P4 M Pl It P4 CD CO O P rt P4 CD Pl CO rt Pi P CL rr CD H TJ P-M CD CD CL 01 9 C 0 cr p. CD P rt Pl 03 p p rt rt M1 P-v- 03 i-1 o O p M Pl Ti P Ml ft 0Q O P4 CD »-( 01 o pl X CD P. Cl 0Q CO CD ft o pi CD CO O It P-Pl P- O CL P n CD It CL CD n P TJ CD 01 It M It P4 3 CD < M"T3 P« 01 P • cr CD rt CD CL It CD 01 TJ M O cr pl cr M1 P- CD 0 TJ !_.. p. It CL CD CD CL 0 P. Ve-en M CD It TJ P4 « CD CD CD P o p. CO Cu O CD CT Ml P- CD Pl M 1 CD P CD M rt ** cr pi 3 Pl P-P o M CD CD p. p- It 01 P* 3 p- Pi 01 O CO >i CD 0 P> ft 01 CD n o M1 P* rt p 01 p H rt It P p« CD CD p- p. CO rt cr 01 CD CD P« O Ml P4 o CD cr It O o p P ft rt P4 p4 o a4 P ft O rt p p4 p- It P O M P« CD Pl CL CD h-' O O 7? p O 0 P CD p, CD CD M1 P4 0 •*< CD CL 01 p it O pj CD P rt pl P' p. 0 P, M rt M CD P« P- it pi o CO CD h-1 • Pi 01 01 CO O it P4 o P rt Ph cr p 0 1 P P. CD CD p4 o rt p p P P"*4 CD M Pl P^ It O O P1 P CD ?^ Ml CO It c OQ CD p i-l CD 0i CD M P 3 rt 0 P« CD O P' 03 CO It P4 CD 01 CD H ►o n £, 01 pj O It M M It M" Oi cr O 3 it P cr P It P* rt o CD O it Pl P, P- o 7? & *• CD O CD P» CD CD 1-1 "^ P« M CD CD CD VO pi CD cp 1 rt cr O Mi p> CD P- P CO CD CD p Ml P^ 0 Mi 01 M rt CD CO Oi CD CL to p P o P- 7? p. c P- O It P" P CL 01 P- CL Pl 0i a4 o p P4 o Cl I—* 0 rt Mi P4 p It It H CD O CD It P-1 CD P- O O o It p-1 CL CD CO >-<$ Pl P« — O It CD p p.V< P4 O O . < 01 P4 M P 01 P p CD P- CD 0Q p 0Q 01 P CD £ CD CL 0 0 Pi O X O CD CL CD M CD CD rt rt CD CO it O Pi CD O H P4 CD CD Ml p CO c •I M> < Oi co CD "^ P-1 p TJ Ml CD cr 1 <*-s P- 1 O "-< CL M It o 3 v< p. 01 0Q C 01 01 P- CD P- CD Pl CD CD H rt Ui 1 CD It CO Pi p, rt c: rt P M1 CD CD P P CO 0> P TJ O CO rt CD O 03 CD CD rt rt rj4 O CD TJ O p* o P- P* O P TJ Pi ^ TJ P CD M ft TJ CD p CD cr CD O 03 Pl CD P' P CD CD ?^ O p4 0 VO P« Oi OQ CD 01 It CD CL P« • CD 1 P i-1 o It h-1 P> CO O CD CD p ro -P> a rt M O o P4 O TJ O O 0Q It CD Pi P" CD H- M p p- P- O O O CD P- O ft ft p4 CD P« *3 Pl M P« Mi l-l M O co rt CD CD O Ml O 01 0 Ml 0 - P O c cr oi 01 01 M Ml O CD ^ to CD CD ^< M P P- p P Ml oi co o P TJ CD CD P cd P* 01 Pl It P- M Pl CD to o cr Pl X O p- P* it «< CO CO 0Q 01 P-1 P- O O It 03 1 Cl O l_l pj It p, 01 pi CD P4 ^> O • TJ M CD P-1^ p- I-1 n CD M to CD p Pi 0 CD H« P It M it t: CD a4 h o Ml CD CD CO rt p CD ft ft CD •i 01 LO n O O O P- O P- P4"^ p4 CD CD Pl Mi O 01 p- Pi < M1 P4 •^ M v-^ •1 0 7? P P, CD CD n CL a4 It £ H- 01 P- 0 0 CD CD CD TJ CD Ml ✓—N • CD P- 01 0Q It X M Pl p« CD CO p-1 01 P4 O 01 O P O O M CL CD rr O H Pl o *• o rt M Pl O P> Ml O p* CD rt CD * h-* p fi P CD P- M 11 0i 01 0 O CD TJ O P P- h-1 0 1 CD M 01 O CD 0Q CD Pl P O pi cr H P- 0Q CD P- Cu < "^ CD ft O P> H-M O 01 ^ CD P1 It Mi p* rr h-1 P4 P M CD P P> It CD P M P I-1 CT\ i*4 O CL Mi M Pi 0 ^ CD hi p-1 CD CD OQ O TJ O CD P* P cr It O Pl CD 01 ^^ CD O CD Mi O ft P CD H« O CO p4 Pl • P CD 1 It CD It O c P. ^ M • P It O n h« CL P M 3 Cl ft CD o to CL TJ O P- P- it pi C c P- CD M Oi M Ml • CD P- P 1 0 7? ■p* CD 01 P- It O P- O p4 it o 0Q P It 01 01 P« It p. s P1 M O O o ■^s O H • pl P« P, P P P' • It O H TJ O M> CL P4 o CD 01 Pl rt :* M K M P4 h-1 CD it Ml O P' • rt P4 CD "^ CD CD p. pj 0 > It M P« CD M o Pi CD P- M1 CO n ^ O O p O P" It CD O Cl 01 01 it TJ CD O P4 0 p Pl 01 ^ CD O Ml M p CD P4 H- ■ O rt P< ■ P4 M Cl O P Pl CD Cl CO 01 P. p. O M CD Pl Ml i CO P4 CL O P' ■ CD < 0Q P CD It TJ c c M i O 01 M1 rt co CD P CO 1 rt CD O CD P*TJ o o MiTJ M i P O O M Ml P' O CD It rt ^< ^ CL P-1 Pl P 01 Pl P' ■ p« • P- P- M 1 CD M O O CD P P- 1 CO p p 01 CD MORBIDITY Information pertaining to stroke morbidity as compared to mortality is more difficult to identify especially as related to specific minority groups. The 1972 National Health Interview Survey estimated the prevalence of stroke as 7.2/1,000 among whites and 9.1/1,000 among all others. The prevalence of stroke reported in the Evans County, Georgia study in 1967-1969 was 53.2/1,000 for white males, 15.0 for white women, 58.6 for black men and 43.4 for black women age adjusted (Table 32). Among participants eligible for randomization in the Hypertension Detection and Prevention Trial, the prevalence of stroke was much higher in blacks than whites. The data on the prevalence of stroke among whites has been provided by previous national stroke surveys, but no data for blacks was included in that report or for other minority groups. The need for including larger numbers of minorities in future prevalence surveys is obvious if good prevalence data is to be obtained. The prevalence of stroke among Japanese in Hawaii and California was much lower than for Japanese in Japan (Table 33). There were 806,000 stroke hospital discharges in which stroke was the first listed diagnosis in 1981 in the United States including 100,000 among non-whites (Table 34). The average length of stay was slightly longer for other racial groups than for whites. Based on these data, there were approximately 1.4 million hospital days directly related to stroke care among non-whites. The estimated annual medical care cost of stroke in 1980 was 5.1 billion dollars, half of which was for hospital care. Non-whites accounted for about 1/8 of the hospital days or a minimum of 300 million dollars for their hospital care alone. The total cost of stroke care for both whites and non-whites in the United States is very substantial (Table 35). Data on the incidence of stroke among several minority groups is available. The southern Alabama study is the most recent and most complete (Table 36). The age adjusted incidence is higher in blacks than whites especially for black women. The risk of stroke was also substantially higher among black men and women in the Hypertension Detection and Follow-up Program clinical trial for both the stepped and referred care populations (Table 37). Stroke incidence has also been measured among blacks in Ibadan, Nigeria (Table 38). Stroke rates in Ibadan appear to be in the same range as those noted in the southern Alabama study. For example, age specific incidence per 100,000 in Ibadan, ages 50-59, is 440/100,000 as compared to a rate of 642/100,000 aged 55-64 in southern Alabama. It is unclear however, whether the Ibadan rates are from one or three years. It is also unclear what percentage of stroke cases are actually ascertained 480 in Ibadan as compared to the southern Alabama study. The high stroke incidence rates may not be limited to the U.S. black population. Stroke incidence is comparable among Japanese living in Hawaii and the U.S. white population. Stroke incidence is much lower among the Japanese in Hawaii than for the Japanese in Japan (Table 39). There is no evidence for increased stroke death rates among Chinese in the United States. Chinese in Taiwan have the highest stroke death rates reported (Table 31). Stroke death rates are declining in Taiwan even though there is no apparent well-organized hypertension control program (Table 31). Prophylactic approaches to hypertensive diseases (Y. Yamori, Raven Press, New York 1979). A stroke registry in part of Taipei in 1975 reported a stroke incidence rate aged 55~64 substantially higher than in the United States whites, 11.6/100,000 in men, as compared to 2.7/100,000 in Southern Alabama (Table 40). The remarkable apparent decrease in stroke among both Chinese and Japanese migrants to the United States certainly requires further evaluation especially as it does not appear to be directly related to changes in blood pressure levels. Stroke is not a single disease but rather a group of related diagnoses. Many investigators believe that blacks and Japanese have more intracranial and more hemorrhagic disease than whites. Few studies have documented the specific types of stroke based on good clinical and laboratory-diagnostic evaluations. The stroke study in Southern Alabama included a detailed review of the specific types of stroke based on careful clinical evaluations. Most strokes in both blacks and whites were due to infarction rather than hemorrhage (Table 41). Thromboembolic stroke also predominants among Japanese in Japan as well as among Japanese in Hawaii (Tables 42-43). Further evaluation of the specific types of strokes will be possible with the development of new technology for evaluating stroke diagnosis such as CT scan and NMR, etc. The availability of improved diagnostic methods may provide a better understanding of the interrelationship of some of the key risk factors for stroke especially the relationships if any between the types of hypertensive disease and specific clinical- pathology of stroke. RISK FACTORS Hypertension is the major risk factor for stroke. Diabetes, alcohol intake, high hematocrit or hemoglobin and lipoprotein abnormalities may also be important risk factors. Other cardiovascular disease, specific electrocardiographic abnormalities especially signs of left ventricular hypertrophy, as well as other manifestations of increased left ventricular mass such as an increased left ventricular wall mass by echocardiography and cardiac arrhythmias including atrial fibrillation are major risk factors for stroke. Both systolic and diastolic blood pressures are higher in blacks than whites (Table 44). 481 The prevalence of elevated blood pressure increases substantially with increasing age among both blacks and whites (Table 45). The decline in the prevalence of elevated blood pressure, as noted in Table 45, is apparently primarily due to the treatment of hypertension which has substantially improved over time. Blood pressure increased with age among Mexican Americans. Blood pressure levels appear to be similar among Mexican American and U.S. whites (Table 46). Blood pressure among American Indians in Minnesota were similar to those of U.S. whites (Table 47). No data could be obtained on the blood pressure distribution of Indians living on the reservations in the south- western United States. Blood pressures are higher among Japanese in the United States than among Japanese in Japan (Table 48). This is surprising because of the much higher stroke death rates among Japanese in Japan than those in the United States. The blood pressures of Japanese in Hawaii and California are similar to those of the U.S. whites. Part of the difference in blood pressure between Japanese in Japan and Hawaii may be related to a greater prevalence of obesity among U.S. Japanese. However, obesity adjusted blood pressures are either similar or slightly higher among U.S. Japanese than those in Japan (Table 49). Similar studies of blood pressure distribution of Chinese in Taiwan fail to demonstrate any higher levels as compared to other populations inspite of the much higher stroke death rates (Table 50). The prevalence of hypertension is also higher among Japanese in the United States (Table 51). The prevalence of electrocardiographic evidence of left ventricular hypertrophy and hypertensive heart disease is however much higher among Japanese in Japan than in Hawaii or California (Table 52). This is also true even when comparing electrocardiographic abnormalities among Japanese in Japan working for the NTT Telephone Company versus U.S. telephone workers (Table 53). There is also a relationship between the high R-waves on the electrocardiograph and elevated blood pressure in both of these populations. There is evidence from studies both in Japan and in Hawaii that these electrocardiographic abnormalities may be an independent predictor of stroke incidence and mortality. The higher prevalence of left ventricular hypertrophy in relationship to hypertension has been noted in Chinese in Taiwan (see ref. Fig. 50). The prevalence of electrocardiographic abnormalities was also noted to be much higher among blacks even after controlling for the degree of hypertension. This has been noted most recently in both the Multiple Risk Factor Intervention Trial (MRFIT) (Table 54) and the Hypertension Detection and Follow-up Program (HDFP). There have been numerous surveys of selected population groups in Africa. Few of the studies represent true population samples or comparable methods of measurement. The studies were reviewed by Adenounmu (Clinical and Experimental Hypertension, 1981;3(4):597-621). He noted that blood pressure increases with age and is higher in urban than rural men in 482 Africa (Table 55). The blood pressure levels are similar to those in the Carribean Islands but lower than in the U.S. or south Africian blacks. Stroke, congestive heart failure and renal diseases are the major complications of hypertension. Coronary artery disease remains very rare. Earlier studies such as from Evans County, Georgia (Table 56) suggested relatively little difference in the risk of stroke in relationship to blood pressure level among blacks and whites. More recent studies, however, suggest that black hypertensives may be at greater risk of stroke than white hypertensives. The estimated incidence of stroke in the referred care group for blacks in the Hypertension Detection and Follow-up Program was 35.3/1,000 as compared to 24.2/1,000 for whites. There was a substantial decrease in stroke morbidity and mortality in the special care as compared to the referred care for all four race, sex groups (Table 57). In the Multiple Risk Factor Intervention Trial the ineligible screenees were followed initially for five years. Among the screenees there were 23,490, black men between the ages of 35~57. The incidence of stroke, was 1.3/1,000 among black men and 0.5/1,000 for white men (Table 58). The risk of stroke death increased with level of diastolic blood pressure (Table 59). At most, levels of diastolic blood pressure the stroke rates were higher in blacks than whites. The logistic regression coefficient for the risk of stroke death was greater for black men than white men (Table 60) suggesting that at a given level of blood pressure the risk of stroke death is greater for blacks than whites. However, part of this difference could be due to differential subsequent treatment for hypertension. The risk of stroke was increased among individuals with electro- cardiographic abnormalities in the Evans County Study, this was true for all four race/sex groups but especially for men (Table 61). In the HDFP trial the mortality rates including stroke were increased among men with electrocardiographic abnormalities based on the "Minnesota Code" (Table 62-63). Stroke incidence is also directly related to blood pressure levels among Japanese men in Hawaii (Table 64). The increased risk appears to be related to both intracranial hemorrhage and thromboembolic disease (Table 64). It appears that at higher levels of systolic blood pressure the Japanese in Japan have a greater risk of stroke than Japanese in Hawaii (Table 65). Surveys in Taipai also demonstrate a striking association between hypertension and stroke, as well as the high prevalence of left ventricular hypertrophy (Table 66). The relationship between lipoprotein levels and risk of stroke is equivocal. In the Evans County Study there was no consistent relationship between the serum cholesterol level and risk of stroke among blacks or whites (Table 67). 483 In the MRFIT screenees there was also no consistent relationship between the serum cholesterol level and risk of stroke death for either black or white men (Table 68). None of the regression coefficients related to serum cholesterol and risk of stroke were significant (Table 69). Prior studies in Japan had suggested a possible inverse relationship between stroke and serum cholesterol levels. Studies of Japanese in Hawaii are consistent in demonstrating a possible inverse relationship between serum cholesterol levels and cerebral hemorrhage but not for cerebral thrombosis or infarction (Table 70). More recent studies from Japan (Table 71) do not demonstrate a consistent relationship between the serum cholesterol level and stroke except for a lower cholesterol levels among the small number of individuals with cerebral hemorrhage. These initial studies however do demonstrate that lower animal protein and very low saturated fat intake may be a risk factor for stroke (Tables 72-73). A further analysis of the Japanese in Hawaii based on post-mortem measurement of cerebrovascular disease again demonstrated a slightly lower cholesterol level among deaths with cerebral hemorrhage and higher cholesterol for those found with a brain infarction. Animal protein intake was decreased only for deaths with cerebral hemorrhage. This data set further demonstrated a relationship between cerebral and aortic atherosclerosis and brain infarction and also a high prevalence of prior myocardial infarction among those also found to have brain infarction by post-mortem examination (Tables 73~74). The results of these studies therefore suggest that the relationship between lipoproteins, dietary factors and stroke may be different for those who have cerebral thromboembolic disease as compared to cerebral hemorrhage. Cerebral thromboembolic disease secondary to atherosclerosis is by far the most common stroke disorder in the U.S. black and white population as well as among the Japanese in Hawaii and California. The primary emphasis should obviously be placed on the risk factors for this disorder. On the other hand, the relationships between low protein intake and cerebral hemorrhage especially animal protein may be a clue to the apparent sharp decline in cerebral hemorrhage over the past 20 and 30 years associated with improved diet and lifestyle including a probable increase in animal protein intake. The prevalence of diabetes is also substantially higher among blacks as compared to whites (Table 75). Diabetes is also very prevalent among several American Indian tribes and among Mexican Americans in the southwestern United States. Diabetes is an important risk factor for stroke. Blacks have a higher prevalence of diabetes than whites. In the Three-Area Study of Stroke Risk Factors, blacks had higher blood sugars and history of diabetes than whites especially in the southeastern United States (Table 76). Diabetes is clearly associated with an increase in cerebro atherosclerosis (Table 77). Prospective studies such as the American Cancer Society Followup Study, (Table 78) or the Framingham Study (Table 79) have noted the increase risk of stroke among diabetics. 484 A study in the 1960's demonstrated that diabetes was a frequent contributing cause of death on a death certificate in which stroke was the underlying cause especially among black women (Table 80). Studies in the late 1960's (Table 81) showed that diabetes was listed on about 16% of all stroke hospital discharges. A more recent, unpublished study, from Pennsylvania showed that stroke was listed as a secondary diagnosis on 2.7% of hospital discharges for diabetes among blacks and 4.8% among whites, and in 5.2% of the records among blacks when diabetes was listed as a secondary diagnosis on the hospital discharge record. A followup study of a diabetic cohort in Rochester, Minnesota suggested that the increased risk of stroke among diabetics may be limited to those with concurrent hypertension (Table 82). This observation could explain the lower rates of stroke inspite of the high prevalence of diabetes among Hispanics and Mexican Americans due to their apparent lower prevalence of hypertension as compared to U.S. blacks. The prospective Honolulu Heart Study has further documented an increased risk of stroke especially thromboembolic associated with increasing levels of blood sugar (Table 83). In multivariate analysis, the blood sugar levels remained an independent predictor of thromboembolic stroke but not hemorrhage. OTHER RISK FACTORS Alcohol Consumption Alcohol consumption has been identified as an independent predictor of stroke as well as a possible risk factor for elevated blood pressure. Among Japanese in Hawaii, there was a direct relationship between alcohol consumption and cerebral hemorrhage but not cerebral thrombosis (Table 84, 85). Alcohol consumption and hypertension may be an especially lethal combination for cerebral hemorrhage. A major effort must be made to treat hypertension among alcohol consumers, as well as reducing prevalence of alcoholism. Cigarette Smoking The relationship between cigarette smoking and stroke is uncertain. Among Japanese in Hawaii, there was a statistically non-significant greater number of cigarettes smoked among subsequent cerebral hemorrhage or thrombosis cases than non-strokes. In the MRFIT trial screenees, the risk of stroke appeared to increase up to 26-35 cigarettes per day and then decrease among black men (Table 86). Cigarette smoking has an important effect on platelets and thrombogenesis. It is possible that smoking could be an important precipitant of stroke among individuals with a high risk of cerebro thromboembolism especially those with carotid artery stenosis, heart disease and cardiac arrhythmias. 485 Hematocrit and Hemoglobin Many studies such as the "Framingham Heart Study", have shown a direct relationship between hematocrit or hemoglobin levels and risk of stroke. The relationship may be due to an association of elevated hemo- globin or hematocrit with blood pressure to greater viscosity of the blood or to other unknown pathophysiological changes associated with the level of hematocrit or hemoglobin. In the Evans County Study, there appeared to be an increased risk of stroke among both blacks and whites with increasing hematocrit levels (Table 87). These studies may have important implications for defining ideal hematocrit or hemoglobin levels especially among women. It is possible that the lower levels of hemoglobin or hematocrit in women secondary to recurrent menstrual bleeding offer some protection against thromboembolic disease. Attempts to raise these levels could increase the risk of stroke or other thromboembolic diseases. Sickle cell anemia is an important cause of stroke especially in the younger age groups. There is no evidence however that sickle cell trait or other hemoglobin disorders are associated with a substantial increased risk of stroke. COMBINATION OF RISK FACTORS A combination of these risk factors such as in the Hawaii Heart Study of Japanese men (Table 88) can identify a high risk of stroke. The combination of several of these risk factors plus hypertension (Table 88) defines a much higher risk group. Many of these risk factors are clearly ameanable to preventive efforts. DISCUSSION AND IMPLICATIONS The death rates due to stroke have continued their long-term decline. The slope of the decline accentuated in the raid 1970's most likely due to an improvement in the treatment and control of hypertension. The changes in awareness, treatment and control of hypertension as noted in the "Impact Study" (Table 89) have resulted in a decrease in the prevalence of elevated diastolic blood pressure. Blood pressure remains substantially higher among blacks than whites. Improvement in the control of hypertension has also occurred among Mexican Americans (Table 90) and probably among Japanese and other minority groups in Hawaii and California. Inspite of these good efforts the stroke death rates and incidence still remain much higher among blacks than whites. There is little evidence that other minority groups such as Hispanics, Chinese, Japanese, Koreans or native Americans are at substantially higher risk of stroke than whites. The risk may be elevated among Puerto Ricans in New York. No data could be obtained for Cubans in the southeast or for Vietnamese or other Asian groups. The risks of stroke are related primarily to hypertension. Prevention and treatment of hypertension remains the cornerstone of any stroke control program. The increased prevalence of hypertension among blacks is probably 486 a combination of genetic and environmental factors. Clinical and epidemiological studies have identified differences in renin, sodium, lithium counter transport, the ability to excrete a salt load and urinary kallikrein excretion between blacks and whites. Elevated blood pressure is clearly not due to any one single factor. The type of blood pressure elevation and relationship to outcome may vary in relationship to the key risk factors. Obesity is the most important risk factor for elevated blood pressure in the United States. However, obesity related hypertension may be a different type of disease than non-obesity hypertension especially in relation to specific hypertensive outcomes, such as, cerebral hemorrhage, intracranial vascular disease, renal disease, and retinopathy. Obesity hypertension may be a more important factor for atherosclerotic disease including myocardial infarction. The decline in stroke that occurred from the 1920's and 1930's may be related to this non-obesity related hypertension. A recent report from the Hypertension Detection and Followup Program noted a substantially higher mortality among those with non-obese, as compared to obesity hypertension (Table 91). Other hypertensive studies have shown similar results. The origins of non-obesity hypertension is unknown but low protein intake, high intake of salt and low consumption of potassium and calcium may be playing important roles. There is also most likely an important genetic component. It is important to reiterate that both obese and non-obese hypertension must be treated in order to reduce the specific hypertensive complications especially stroke. The lower the blood pressure the greater the decrease in hypertensive complications especially stroke and congestive heart failure. The only issue perhaps that still remains is the interaction between the treatment of hypertension and atherosclerotic complications especially myocardial infarction. The highest rates of hypertension and stroke among blacks are found in the southeast, in lower socioeconomic groups and among less educated individuals. Key environmental factors appear to play an important role in these geographic and socioeconomic variations including obesity, sodium, potassium, and calcium intake, possibly stress and "hostility", alcohol intake and health care especially as related to the treatment of hypertension and its complications. The control of diabetes and hypertension especially the combination of diabetes and hypertension among black women may be associated with an exceptionally high risk of stroke. The joint prevalence of hypertension and diabetes may both be primarily related to nutritional factors especially the development of obesity. The future accentuation of the rapid reduction of stroke mortality and morbidity especially among blacks will require efforts focused at identifying the high risk individuals and communities in the United States, and aggressive control of hypertension and other risk factors in these high risk communities. Such communities can be easily identified by monitoring the high stroke death rates. Such communities should be candidates for aggressive hypertension treatment programs and careful monitoring to determine a consistent reduction in stroke mortality. The stroke death 487 rates in general reflect the acute stroke case fatality which is generally at least 20-30%. Mortality rates reflect the trends in the incidence of stroke and are the easiest and best endpoint to monitor in hypertension control programs. The reduction in stroke mortality especially among younger blacks, i.e., 35~64, is probably the most important medical factor in decreasing the black-white difference in mortality rates among adults. Major efforts should be continued to reduce the prevalence of elevated blood pressure by control of risk factors for hypertension. The ultimate control of this epidemic of hypertension and stroke will depend on primary prevention. Much of the health habits which lead to the elevated risk factors are learned early in life and therefore may be ameanable to modifi- cation. Nutrition and genetic factors are probably the key elements. The role of diabetes and hyperlipoproteinemia in the etiology of stroke especially among blacks needs further investigation. It is unclear whether elevated levels of HDL cholesterol or its subfractions are associated with a reduced risk of stroke especially thromboembolic stroke as well as heart attack. There is little evidence that treatment of non-insulin dependent diabetes with either oral hypoglycemic agents or insulin reduces the risk of stroke. There is a strong genetic component to diabetes, as well as a clear relationship to dietary factors especially obesity and perhaps a low fiber intake. Prevention of diabetes and a better understanding of its pathophysiology and improved methods of treatment offer the best potential for the prevention of the complications of this disease including stroke. Treatment of stroke may have improved in recent years. There is some evidence for reduction in the case fatality. Diagnostic methods have certainly improved substantially. Early detection and treatment of potential strokes may be of substantial value especially the identification of transient ischemic attacks and the possible treatment with antiplatelet agents or surgical correction of underlying carotid artery disease or identification of carotid artery bruit and asymptomatic stenosis and occlusion. There is also a strong association between vascular disease at other sites especially coronary artery disease and subsequent risk of stroke. Patients who have had a myocardial infarctions or other manifestations of coronary artery disease are ideal candidates for preventive efforts to reduce the subsequent risk of stroke. Furthermore, there appears to be a strong association between certain electrocardiographic and echocardiographic abnormalities, and subsequent risk of stroke. The identification of such high risk individuals and effective treatment may also be important in reducing the risk of stroke. There is strong evidence that blacks, as well as Japanese may have more intracranial and less extracranial disease than whites (Tables 92-93) and therefore such approaches as carotid artery surgery, antiplatelet drugs may be less effective in black populations. The major effort to reduce stroke mortality should continue to emphasize identification, treatment and control of hypertension. The risk factors for stroke are identifiable and can be controlled effectively. 488 TABLE 1 Age Adjusted Death Rates Due to Cerebrovascular Disease (430-438) 1981—United States by Race and Sex Number of Deaths Rate/100,000 Total 163,504 38.1 Men 66,429 41.7 Women 97,075 35.4 White Men 57,000 38.9 White Women 85,765 33.1 All Other Men 9,429 65.6 All Other Women 11,310 53.2 Black Men 8,760 72.7 Black Women 10,656 58.1 Monthly Vital Statistics Reports. National Center for Health Statistics, Vol. 33, No. 3, Supplement, June, 1984. Ratio to White Women 1, .1 1, ,3 1. ,1 1. ,2 1 2. ,0 1. 6 2. 2 1. 8 489 TABLE 2 Death rates for cerebrovascular diseases, according to race, sex, and age: United States, selected years 1950-82 (Data are based on the National Vital Statistics System) Race, sex, and age 1950J Total3 All ages, age adjusted ......... 88.8 All ages, crude................. 104.0 Under 1 year.......................... 5.1 1 -4 years............................. 0.9 5-14 years............................ 0-5 15-24 years........................... 1-6 25-34 years........................... 4-2 35-44 years........................... 18.7 45-54 years........................... 70«! 55-64 years........................... 195-3 65-74 years........................... 549-7 75-84 years........................... 1,499.6 85 years and over..................... 2,990.1 Year 1960J 1970 1975 1979 1980 1981' Number of deaths per 100,000 resident population 1982' 79.7 66.3 53.7 41.6 40.8 38.3 36.1 108.0 101.9 90.1 75.5 75.1 71.3 68.9 4.1 5.0 5.0 4.6 4.4 3.6 3.8 0.8 1.0 0.8 0.3 Ml 0.2 0.2 0.7 0.7 0.5 0.3 0.3' 1.8 1.6 1.4 0.9 1.0 0.9 1.0 4.7 4.5 3.4 2.6 2.6 2.6 2.3 14.7 15.6 11.7 9.1 8.5 8.1 7.9 49.2 147.3 41.6 115.8 32.3 90.4 26.4 68.1 25.2 65.2 26.3 64.5 24.2 57.2 469.2 384.1 302.2 226.9 219.5 209.8 195.3 1,491.3 1,254.2 1,028.8 793.8 788.6 702.6 678.6 3,680.5 3,234.6 2,736.4 2,264.9 2,288.9 2,119.8 2,056.4 White male 4 All ages, age adjusted ........... All ages, crude................... Under 1 year.......................... 1-4 years............................. 5-14 years............................ 15-24 years........................... 25-34 years........................... 35-44 years.....................*..... 45-54 years........................... 55-64 years........................... 65-74 years........................... 75-84 years........................... * 85 years and over..................... 3 87.0 80.3 68.8 56.7 42.9 41.9 100.5 102.7 93.5 80.2 64.2 63.3 5.9 4.3 4.5 4.6 3.6 3.8 1.1 0.8 1.2 0.9 0.3 0.4 0.5 0.7 0.8 0.5 0.3 0.2 1.6 1.7 1.6 1.4 0.9 1.0 3.4 3.5 3.2 2.6 2.2 2.0 13.1 11.3 11.8 8.7 6.8 6.5 53.7 40.9 35.6 27.6 22.2 21.7 182.2 139.0 119.9 93.8 68.0 64.2 569.7 501.0 420.0 339.4 249.5 240.4 ,556.3 1,564.8 1,361.6 1,134.9 867.0 854.8 ,127.1 3,734.8 3,317.6 2,807.4 2,224.5 2,236.9 White female All ages, age adjusted ........... All ages, crude................... Under 1 year.......................... 1-4 years............................. 5-14 years............................ 15-24 years........................... 25-34 years........................... 35-44 years........................... 45-54 years........................... 55-64 years........................... 65-74 years........................... . 75-84 years........................... *■ 85 years and over..................... J See footnotes at end of table. 79.7 68.7 56.2 46.1 35.9 35.2 103.3 110.1 109.8 101.6 88.5 88.8 2.9 2.6 3.2 3.9 3.3 3.3 0.6 0.5 0.6 0.7 0.3 0.4 0.4 0.6 0.6 .0.5 0.3 0.3 1.2 1.4 1.1 1.1 0.7 0.7 2.9 3.4 3.4 3.0 2.0 2.0 13.6 10.1 11.5 9.3 7.0 6.7 55.0 33.8 30.5 25.0 20.1 18.7 156.9 103.0 78.1 64.5 50.6 48.7 498.1 383.3 303.2 233.6 179.2 172.8 ,471.3 1,444.7 1,176.8 966.2 739.3 730.3 ,017.9 3,795.7 3,316.1 2,794.9 2,335.7 2,367.8 Health-United States, 1983. 490 TABLE 2 (cont.) Death rates fqr cerebrovascular diseases, according to race, sex, and age: United States, selected years 1950-82--Continued (Data are based on the National Vital Statistics System) Race, sex, and age 1950' Black male 4 All ages, age adjusted ........... All ages, crude................... Under 1 year..........................> 1-4 years.............................. 5-14 years............................ 15-24 years........................... 25-34 years........................... 35-44 years........................... 45-54 years........................... 55-64 years........................... 65-74 years........................... 75-84 years...........................j 85 years and over.....................\ 146.2 122.0 2.5 0.7 3.3 12.0 59.3 211.9 522.8 783.6 1,504.9 Year 1960J 1970 1975 1979 1980 1981' 1982' Number of deaths per 100,000 resident population 141.2 122.9 8.5 1.9 *0.9 3.7 12.8 47.4 166.1 439.9 899.2 1,475.2 2,700.0 124.2 108.7 95.0 87.8 12.2 9.1 •1.4 1.1 0.8 0.7 3.0 2.5 14.6 9.4 52.7 36.9 136.2 98.0 343.4 253.7 780.0 626.9 1,442.6 1,123.0 2.315.4 2.115.2 77.9 73.8 12.7 0.5 0.3 1.4 7.3 34.4 88.8 204.0 470.9 963.9 1,840.4 77.5 73.1 11.2 0.6 0.5 2.1 7.7 29.2 82.1 189.8 472.8 1,067.6 1,873.2 Black female 4 All ages, age adjusted . All ages, crude......... Under 1 year............... 1-4 years.................. 5-14 years................. 15-24 years................ 25-34 years................ 35-44 years................ 45-54 years................ 55-64 years................ 65-74 years................ 75-84 years................ 85 years and over.......... 155.6 128.3 2.8 0.6 4.2 15.9 75.0 248.9 567.7 754.4 1,496.7 139 127 *6.7 *1.3 1.0 3.4 17.4 57.4 166.2 452.0 830.5 1,413.1 2,578.9 107.9 112.1 9.1 8. *1.4 0. 0.8 0. 3.0 1, 14.3 8, 49.1 31. 119.4 81. 272.5 187 673.4 502 1.337.8 998 2,504.8 2,126 78.6 91.4 60. 76. 9. 0, 0 1 6 21 62 133 373 865 1.881 61.7 77.9 6.4 0.5 0.3 1.7 7.0 21.6 61.9 138.7 362.2 918.6 1,896.3 includes deaths of nonresidents of the United States. ^Provisional data. WXTJll byC"/dfrect me^oV'to the total population of the United States as enumerated in 1940, using 11 age groups. N0TE: For the data years shown^ «-0«-?h.--SS SSll^TSr'^aSirf'SriJS. Vin^^^^1^r^K.,"3Sc^i*3S,;0?- SiJ£^!^?1^th to'"». «iS>» u2^^*«« States. „os.'430-438; and for .979-82. the Ninth Revision. Nos. 430-438! , , *„ uaaith statistics- Vital statistics of the United States, Vol. II, 1950-80. Public Health SOURCES: National Center for Health Statistici. mai >tat st f bWth deaths marriages, and divorces, United Service. Washington U.>,^erment■>"«£%*f™' TiVff DHHS Pub. No. (PHS) 83-1120. Public Health Service States, 1982. Monthly V •a\nSff ^"CuK^'by the Division of Analysis from data compiled by the Division of Vital Hyattsville, Md., Sept. 27, J^i. Data £3"? °yes™-mates and projections. Current Population Reports. Series P-25. No. Statistics; U.S. Bureau of the Census POP"1*1;00^^. 1950 Nonwhite Population by Race. Special Report P-E, No. 38. 310. Washington. U.S. Government Printingi Officej ^nyy* population characteristics, United States sugary. 1960 and ^n^V^^ Wl)-Bl. ««"1"9ton. U.S. Government Printing Office. 1961 and 1971: 491 TABLE 3 Ratio of Stroke Mortality By Age 1980: Black and White Black Men/ Black Women/ White Men White Women 25-34 3.5 3.5 35-44 4.5 3.2 45-54 3.8 3.3 55-64 3.0 2.8 65-74 2.0 2.1 75-84 1.1 1.3 85 + 0.8 0.8 Total 1.8 1.8 Monthly Vital Statistics Reports. National Center for Health Statistics, Vol. 33, No. 3, Supplement, June, 1984. 492 TABLE 4 Percentage of Deaths Due to White Men White Women Black Men Black Women Total Monthly Vital Statistics Health Statistics, Vol. : Stroke By Race and Sex 1981 Percent 5.3 8.2 7.2 9.7 6.7 Reports. National Center for 3, No. 3, Supplement, June, 1984. 493 TABLE 5 NUMBER OF DEATHS 1968 - 1970 Stroke as Stroke as Stroke Hot Underlying Associated Mentioned Cause Cause WM 246,313 145,496 2,447,487 WF 299,625 160,170 1,756,574 NM 38,042 20,653 348,661 NF 41,691 21,940 246,305 625,671 348,259 4,799,027 NUMBER OF DEATHS 1976 - 1978 Stroke as Underlying Cause Stroke as Associated Cause Stroke Hot Mentioned WM 200,429 138,515 2,416,466 WF 278,924 166,131 1,834,235 NM 31,368 19,216 351,681 NF 35,864 21,313 248,673 546,585 345,175 4,851,055 Personal communication from Herb Baum, 494 TABLE 6 Age-Race-Sex Specific Death Rates Due to Stroke Baltimore, Maryland and Savannah, Georgia (1969-71/100,000) 45-54 55-64 65-74 Baltimore, Maryland White Males White Females Black Males Black Females Savannah, Georgia White Males White Females Black Males Black Females 34 26 137 109 52 53 379 260 106 64 237 168 142 96 657 710 355 252 567 429 551 400 1266 1195 963 790 952 913 1600 1207 1599 1355 Computed from Vital Statistics of United States. TABLE 7 Age Specific Stroke Death Rates: 1978 for Selected States by Race and Sex/100,000 Age 45-54 Georgia Pennsylvania Colorado Mississippi All White Women 26 22 16 18 22 White Men 35 24 21 28 23 Black Women 97 53 — 65 58 Black Men 143 92 — 116 81 Data provided by the National Center for Health Statistics. 496 TABLE 8 Age Adusted Stroke Death Rates By Geographic Area 1978/100,000 Ages (35-74) Colorado Kansas Utah New York Maryland South Carolina Georgia Mississippi North Carolina White White Black Black Men Women Men Women 43 38 50 43 38 47 48 39 91 75 47 38 115 73 79 54 231 173 82 59 283 158 77 45 169 103 70 47 197 145 Data provided by National Center for Health Statistics 497 TABLE 9 Age-adjusted death rates per 100,000 {age 4(H)9) for cerebrovascular disease in Baltimore based on multiple-cause tabulation of 1960 death certificates by socioeconomic status: based on combination distribution of socioeconomic areas in whites and non-whites* Socioeconomic status Race & sex Upper half Lower half No. deaths Rate No. deaths Rate Unknown WM 85 129 78 189 1 WF 73 90 77 174 2 NM 26 280 108 422 2 NF 43 410 124 472 2 •Excludes 30 deaths in state hospitals out- side Baltimore City. Kuller L, Seltser R: Cerebrovascular Disease Mortality in Maryland. Am J Epidemiol 1967;86:442-450. 498 TABLE 10 560r Mountain I I PwJmonl ■I Plains HH Coastal tfnife Males White Femoies Block Moies Slock fero*> Age-adjusted mortality rates (per 100.000) for stroke (ICu 430 to 438) in North Carolina in 1969 according to region o' state and sex-race groups (all entries). The numbers within each bo' indicate the number of stroke deaths for that region. Heyman A, Tyroler HA, et al: Geographic Differences in Mortality From Stroke in North Carolina. Stroke 1976;7:41-45. 499 TABLE 11 Stroke Mortality By Age and Sex: Cape Verde Islands (1980) Stroke U.S. Blacks Sex Deaths Rate/100,000 1980 Age 45-54 M 10 128 82 F 12 131 62 55-64 M 13 158 190 F 9 94 139 65-74 M 23 424 472 F 21 338 362 WHO.Statistics Annual 1983. 500 TABLE 12 100 80 _ 60 c 0J U Vw QJ 40 20 0 O—O Hypertensives ■—•"■ Normotensives D Hypertensives Normotensives Percent of Nigerian and Minnesota subjects with cerebral atherosclerosis in the presence or absence of hypertension. Note the similarity in percent of Nigerian hypertensive subjects and nor- motensive Minnesota subjects. Williams 0, Loewenson RB, et al: Cerebral Atherosclerosis and Its Relationship to Selected Diseases in Nigerians: A Pathological Study. Stroke 1975;6:395-401. 501 TABLE 13 Carotid Arteries 100 90 80 ro 60 50H 40 30H 20 10 Itoqro-Main • si »i*^ «r* i* i*1 i—p-i—r»i Ntgro-Ftmaltt » ■ » * *■ | ,0° | *H ^ 80 § 70H * 60 50H 40 30 20 H iO Whitt-Moin *• • • "»"-»t—i—i**i—r»i—i—r—«t—i 20 30 40 50 60 70 WhilfFtmoiti 20 30 40 50 60 70 Age Percentage ot iutimal surface involved with raised atherosclerotic lesions in carotid arteries oi autopsied N'ew Orleans inon and women, basal group of cases. Each point represents a case; the triangle represents the mean of the age group. Solberg LA, McGarry PA: Cerebral Atherosclerosis in Negroes and Caucasians. Atherosclerosis 1972;16:141-145. 502 TABLE 14 Intracranial Arteries 100 90H BO 70 60 H 50 40 30 H 20 ioH 90 BO 70H GO 50 40 30H 20 10 Ncqro-Malt* • :7 • • • / rN^fO* rtflU i*n m S • ^%«^d*v: • •• V* Whitt-IKain ■■yt—i Or'+rfr ••• Whltt-Fwwlt* 20 30 40 50 60 70 20 30 40 50 60 70 Age Percentage of intimal surface involved with raised atherosclerotic lesions in intracranial arteries of autopsied New Orleans men and women, basal group of cases. Each point represents a case; the triangle represents the mean of the age group. Solberg LA, McGarry PA: Cerebral Atherosclerosis in Negroes and Caucasians. Atherosclerosis 1972;16:141-145. 503 TABLE 15 Age- and sex-adjusted mortality rates, by ethnicity, Los Angeles County, 1980 Age- and sex-adjustedf mortality rates per 100,000 population Asians and Pacific Islanders Los Angeles Cause of death County White Black Hispanicf Japanese Chinese Filipino Korean All causes 819.9 870.2 1038.3 814.8 482.5 362.8 137.2 421.8 Major cardiovascular diseases 409.4 429.5 472.0 390.6 255.3 157.0 84.2 143.8 Diseases of heart 313.4 331.2 353.4 307.8 161.7 99.2 57.8 82.1 Total IHD 194.0 207.7 192.5 177.7 106.9 47.4 31.4 63.9 MI and other acute IHD 90.9 97.9 88.3 82.3 55.4 28.1 18.8 13.1 Chronic IHD 103.1 109.8 104.3 95.4 51.4 19.3 12.6 50.9 Hypertensive disease 21.0 19.5 47.4 — — — — — Cerebrovascular diseases 74.3 75.8 94.2 63.5 79.6 48.7 19.7 48.3 IHD - Ischemic heart disease; MI «■ myocardial infarction. 'From Chapman J, Frerichs Ft, Maes R: Cardiovascular diseases in Los Angeles, 1980. Los Angeles, 1983, American Heart Association, Greater Los Angeles Affiliate, Inc. t Direct method of adjustment with Los Angeles County population, 1980, as standard. {Census tracts in which 75r< or more of the population are persons of Spanish-Hispanic origin or descent. Haywood LJ: -Issues in the natural history and treatment of coronary heart disease in black populations: Medical management. Am Heart J 1984;108:683-687. 504 TABLE 16 181.7 ■ MALE E3 FEMALE ( ) RATIO MALE:FEMALE JAPAN HAWAII JAPANESE JAPANESE (1.4) (1.3) HAWAII U.S. WHITES WHITES (.94) (1.0) Age-adjusted stroke mortality (1970). Kagan A, Popper JS, Rhoads GG: Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 505 TABLE 17 Ratio of age-adjusted Chinese-White death rates by sex, for leading causes of death: United States, 1969-71 8 Heart disease Cancer Stroke Accidents Influenza Et pneumonia Diabetes Cirrhosis Suicide Homicide Heart disease Cancer Stroke Accidents Influenza & pneumonia Diabetes Cirrhosis Suicide Homicide Male Female W////////////////////////A -50 V///////////////////////////////////////A -74 V//////////////////////////////////77777A .73 V7/////////////////////////A -so Y/////////////////////////////A .57 W////////////////////////////////A -64 W/////////////A .30 V////////////////////////////////////W^ 1 -42 1__________I__________L 0 .2 .4 .6 .8 1.0 1.2 1.4 Ratio of age-adjusted death rates SOURCE: National Center for Health Statistics Rice DP: Figures and Tables for "Health of the Chinese in America". Handout at Conference on Health Problems Related to the Chinese in America. Department of Health and Human Services. May 23, 1982, San Francisco, CA. TABLE 18 O _i Q. O 0. o o o o" o a: ui o. >- cr. O o IS (7) D -a O < I Ui < 400 -i 300- 200- 100 80 40- 20- 10- SScr" 0W9 0Wor» -------ss 1970 1971 1972 1974 1975 —i 1976 1973 YEARS Secular trends in age-adjusted cerebrovascular disease mortality (ICDA codes 430-8) in Bexar County, Texas, from 1970-1976 by sex and ethnic group. Stern MP, Gaskill SP: Secular Trends in Ischemic Heart Disease and Stroke Mortality from 1970 to 1976 in Spanish-surnamed and Other White Individuals in Bexar County, Texas. Circulation 1978;58:537-543. SS=Spanish Surname 507 TABLE 19 Average Annual Death Rate3 from Selected Leading Causes of f>ath: Puerto RIcan-Born and White Non-Puerto RIcan-Born by Age and Sex: New York City, 1969-71 (number of deaths per 100,000 population) Males Females Ratio of Rater; ol Puerto Rican Rnm to Nnn PijpMo Riran While Aqe (yenrs) Puerto nican Non-Puerto Rican White Puerlo Non Rican Rica -Puerto n While Mains F ornalns Heart Disease (ICDA* 390-98. 402. 404-29) 25-34 19 10 9 6 1 77 1 A') 35-44 05 94 30 27 0 90 1 to 45-54 265 378 117 111 0 70 1 nri 55-64 616 995 393 347 0 0? i 1:1 65-74 1.610 2.166 1.01 T 1.102 0 74 0 92 75 and over 2.713 5,197 3.037 4.578 052 Ofifi Malignant Neoplasms (ICDA 140-209) 25-34 13 16 11 17 079 0 66 35-44 48 52 53 75 0.9? 0 7(1 45-54 143 184 148 210 0.78 0 70 55-64 414 510 276 383 0.81 0 72 65-74 726 1.102 570 588 O.fifi D.Q7 75 and over 1.529 1,768 886 1.062 0.86 0.03 Cerebrovnscul; lr Disease (ICDA 430-38) 25-34 6 3 4 3 2.21 1.10 35-44 21 10 16 10 2.11 1.65 45-54 44 29 36 25 1.52 1.45 55-64 97 80 92 59 1.21 1.r)H 65-74 216 267 302 196 0.81 1.5S 75 and over 495 918 771 974 0.54 0.79 Diabetes Mellitus (ICDA 250) 25-34 2 2 2 2 0 82 O.G0 35-44 7 6 4 4 1.12 n/m 45-54 19 15 17 10 1-31 1.73 55-64 52 35 72 71 1.47 2-60 Rosenwaike I: Mortality Among the Puerto Rican Born in New York City. Social Science Quarterly 1983;64:375-385. 508 Age-specific daath rates for selected causes, Indian and Alaska Native population in 25 reservation states, 3-year average (1973-1975) and U.S. All Races, 1974: rates per 100,000 population Under 1 year 1-4 years 5-14 years 15-24 years 25-34 years 35-44 years 45-54 years 55-64 years 65 Years and over All Causes Indian ..................... 2,342.3 1,755.7 127.1 48.3 32.2 7.5 94.9 40.9 49.2 28.4 42.4 22.0 6.8 5.1 0.0 1.7 3.8 239.0 86.9 0.5 1.7 1.1 15.3 2.8 150.2 73.9 80.7 29.2 35.8 9.9 44.9 19.3 5.4 2.9 4.6 2.0 0.0 0.8 0.8 5.4 5.9 11.7 4.7 0.1 0.4 0.1 0.4 0.4 63.6 38.2 40.4 18.4 19.2 8.7 21.2 9.7 2.9 1.7 1.7 0.9 0.0 1.1 0.6 0.2 0.0 4.0 5.2 2.2 1.3 0.2 0.1 0.2 376.2 121.7 230.5 61.6 160.5 40.5 70.0 21.1 6.9 4.5 4.5 2.8 0.0 1.5 1.4 0.9 0.0 7.1 6.9 2.6 1.7 0.6 0.5 3.4 0.4 0.4 0.4 578.0 146.8 245.1 45.4 155.5 25.3 89.6 20.0 25.4 14.0 19.6 9.4 0.3 4.8 3.6 1.0 0.0 12.7 15.0 10.0 2.5 3.8 1.9 60.8 4.2 0.7 0.4 773.1 278.6 202.0 40.0 105.6 19.0 96.4 21.0 84.4 72.0 65.0 55.7 1.3 0.9 17.2 13.3 1.8 0.2 37.1 55.1 26.5 5.7 15.5 4.3 139.7 18.6 1.8 1.7 1,150.7 675.0 208.1 43.5 97.6 18.3 110.4 25.2 251.6 259.8 192.4 215.5 2.8 2.0 47.4 35.3 7.8 1.0 101.5 183.8 49.6 11.1 35.1 11.4 179.0 38.7 2.8 7.0 1,729.7 1,549.2 188.0 51.2 86.4 19.5 101.7 31.7 607.6 720.9 451.0 590.8 9.2 4.3 123.8 99.6 23.7 5.9 269.0 437.0 64.2 26.6 86.4 31.7 123.8 50.9 14.5 27.3 4,6402 All Races ................... 6.172.9 Accidents Indian...................... All Races ................... 267.2 125.0 Motor Vehicle Accidents Indian ................... All Races ................ 90.7 27.9 All Other Accidents All Races ................ 174.0 97.1 Major Cardiovascular Diseases All Races .................... 2.0639 3.096.5 Diseases of the Heart Indian ................... 1.426.6 All RAces................ 2,758.4 Hypertension 9.8 26.3 Cerebrovascular Diseases 492.7 All Races ................ 869.0 Arteriosclerosis 83.3 153.3 Malignant Neoplasms 683.9 1.040.4 Influenza and Pneumonia 304.0 200.9 Diabetes Mellitus 171.6 132.6 Cirrhosis of Liver 68.6 42.1 Bronchitis, Emphysema and Asthma 31.9 94.7 Indian Health Trends and Services. U.S. Department of Health, Education and Welfare. Public Health Service, Health Services Administration. Publication No. 78-12009. 509 TABLE 21 Age adjusted mortality rates for selected causes among Indians and Alaska Natives in reservation states and the U.S. All Races population, 1970 to 1975 INDIAN AND ALASKA NATIVE Cause of death Number of deaths in: Rate per 100,000 Populat ion 1975 1970 1975 1974 1973 1972 1971 1970 All causes................. 5.774 1,335 965 291 70 9 1,256 704 552 508 355 281 145 64 27 185 180 5,268 1.340 933 326 71 10 1,116 595 521 421 276 314 143 63 33 125 105 824.8 202.5 147.4 43.5 10.1 1.6 170.5 94.1 76.4 79.8 61.4 36.1 23.8 9.9 3.7 26.5 26.0 872.6 235.6 169.5 50.1 12.9 3.0 163.2 86.7 76.5 88.4 66.7 30.7 24.4 9.8 4.3 31.7 21.8 968.8 237.0 167.7 53.0 13.4 2.9 202.7 117.1 85.6 81.6 66.0 49.9 28.4 9.1 6.9 31.7 22.9 881.6 229.3 165.4 490 6.8 2.7 185.1 107.1 78.0 81.3 60.5 39.0 29.3 10.7 2.8 25.7 20.6 935.5 251.1 182.6 32.8 12.0 2.6 183.0 96.5 86.5 84.4 66.8 41.6 31.7 10.6 6.6 26.4 21.8 889.6 Major Cardiovascular Diseases ................ 237.2 Diseases of the Heart....................... 168.8 Cerebrovascular diseases.................... 55.0 Arteriosclerosis.............................. 11.8 Hypertension................................ 1.7 Accidents.................................... 181.8 Motor Vehicle.............................. 98.5 All Other .•.................................. 83.3 Malignant Neoplasms.......................... 79.5 Cirrhosis of the Liver.......................... 56.9 Influenza and Pneumonia..................... 46.7 Diabetes Mellitus............................. 27.1 Tuberculosis, all forms........................ 11.4 Bronchitis, Emphysema and Asthma........... 5.7 Homicide..................................... 22.2 17.9 Indian Health Trends and Services. U.S. Department of Health, Education and Welfare, Public Health Service, Health Services Administration. Publication No. 78-12009. 510 TABLE 22 i0° t" to : 40 to iob 4 600 I»ii0 m>o txo —r—i iii i; 40 to DIPHTHERIA I HAT 6 UNCtA«QUAH»| 8. o S .0 -1 3 * X I ItOQ IfcQ littQ HM l»40 l>00 IttO 1920 ISM 1*40 IT , SCARLET FEVER ; i*AU UAOW 10YIAASI 4 400 too 100 to 40 too 400 it 00 . WHOOPING COUGH . IhAlC UhOfcft KOUAtttl ■»---1---1---1---1 t I---HH---h TUBERCULOSIS (ALL formsl J—I t t I t---1—HH---* 8 * I i—I I I I I, I t I » £ . - DIARRHEA AND ENTERITIS. (fCM UMOUVt ttfllHtl too 400 100 i* 100 =*—1---1 I I---►—»----•---r—♦; E PNEUMONIA lALL FORMS) 5 to 40 DISEASES OF THE HEART (ALL FORMS) -t-H—i—lit!—t ■■ t t «00t INTRACRANIAL LESIONS OF VASCULAR ORIGIN 1001- 100 to -I- I I I I___I I I, I 1 3 *o © to * too ►- too a x tO 3* H--> I «--1--H-*---1--HH- I SYPHILIS (ALLFORMS)- +-H—I—hH—till—*■ NEPHRITIS(ALL FORMS) l i l J___I---LI -L a—I—h—1—I i t t I—r I" 56 40 « M > 5 • rt 4 I I 400 COO : DIABETES MELLITUS : ^—i—I I t i—i—|—i—r-j ACUTE RHEUMATIC FEVER * i i—i i l i I i i . CANCER 100- to- J--1___I___L ttOO 1*0 l»«0 i»W i»*0 UA* rtOO UtO HiO imO i»40 ?»AA l»00 IttO t I I ■ ■ I x t»iO i»*0 UAH itio Course of mortality for selected causes of death in the expanding death registration States (logarithmic scale), 1900-1945. Moriyama IM, Gover M. Statistical studies of heart diseases U.S. Public Health Service, April 23, 1948. 511 TA&LE 23 AGE ADJUSTED DEATH RATES (AGE 45-74) FOR CEREBROVASCULAR DISEASE (UNDERLYING AND CONTRIBUTING) BY RACE AND SEX EACH DECADE BALTIMORE AND MEMPHIS* IOOO- 900- 800H 700 600 500- 400 O o o o" 2 300- UI < 200- 100 NM-NF NM — Baltimore —— Memphis 'Age adjusted to the estimated I960 Baltimore population. 1930 31 1940 1 41 —, , 1948 I960 1 1 49 61 Kuller L, Seltser R, Paffenbarger RS, Krueger DE. Trends in cerebrovascular disease mortality based on multiple cause tabulation of death certificates 1930-1960. Am J Epidemiol 1969;88:307-317. 512 TABLE 24 Age Adjusted Annual Decrease in Stroke Mortality 1968-1978 By Race and Sex in Percentage White White Black Black Women Men Women Men 1968-1978 5.0 4.6 5.5 6.8 1968-1973 2.0 2.3 3.1 3.6 1974=1978 7.8 6.8 6.7 8.8 Arteriosclerosis 1981, Vol. 2. U.S. Department of Health and Human Services, PHS NIH # 82-2035. 513 Table 25 Trends of Mortality Rates from Cerebrovascular Diseases (ICD 430-438), Age-Adjusted, Persons Age 35 to 74 by Sex-Color, United States, 1968-78 White White Nonwhite Nonwhite Year Men Women Men Women All 1968 155.4 110.8 340.2 292.1 148.1 1969 150.3 106.8 317.1 274.3 142.2 1970 146.8 105.8 304.0 261.9 139.3 1971 143.8 100.7 292.9 248.6 134.2 1972 144.8 99.7 280.4 249.8 133.6 1973 137.4 96.8 282.0 239.8 128.7 1974 129.4 91.3 260.8 216.6 120.7 1975 117.8 83.7 233.7 190.1 109.7 1976 108.9 78.5 216.6 176.9 102.2 1977 100.9 72.4 202.2 162.4 94.6 1978 93.9 68.7 194.7 148.6 88.7 Change 196fi i-78 -62.1 -42.1 -145.5 -143.5 -59.4 Percent Change -40.0 -38.0 -42.8 -49.1 -40.1 Slope 1968- •73 -0.0174 -0.0190 -0.0348 -0.0359 -0.0288 Standard Error 0.0035 0.0036 0.0049 0.0048 0.0032 Slope 1973- 78 -0.0789 -0.0707 -0.0769 -0.0951 -0.0761 Standard Error 0.0017 0.0021 0.0054 0.0041 0.0020 Report of the Working Group on Arteriosclerosis of the National Heart, Lung and Blood Institute. Arteriosclerosis 1981. Volume 2. 514 TABLE 26 Average annual incidence rates 300 200 1< 70 r 60 so 10 Men Women ^^ ^*"*** 1945-49 1955-59 1965-69 1975-79 1950-54 1960-64 1970-74 Calendar years Five-year average annual incidence rates per 100,000 population for all first episodes of stroke in men and women in Rochester. Minnesota, age adjusted to I960 US white population. Whisnant JP: The Decline of Stroke. Stroke 1984;15:160-168 515 TABLE 2 7 Trends in Age Specific Stroke Mortality in Georgia and Pennsylvania, and U.S. by Race and Sex 1968-1978 Age 45-54 Georgia Pennsylvania Year White Women White Men Black Women Black Men White Women White Men Black Women Black Men 1968 33 53 234 302 35 38 101 105 1969 36 50 216 271 28 38 95 127 1970 45 47 158 301 35 39 96 103 1971 34 49 227 259 29 36 101 111 1972 39 48 181 313 31 34 111 131 1973 41 31 182 229 30 36 79 123 1974 26 35 140 201 23 28 91 93 i 1975 29 42 140 148 25 32 56 81 1976 31 40 143 162 24 24 62 56 1977 22 30 113 137 22 27 60 74 1978 1 26 35 97 143 21 24 53 92 Data provided by National Center for Health Statistics. TABLE 28 Trends in Stroke Mortality 1974-1978 By State: In Percent, Annual Change, Age Adjusted (35-74) White White Black Black Women Men Women Men 6.6 6.7 9.2 6.0 10.0 11.0 18.2 -1.0 6.9 8.2 11.2 6.1 4.8 9.1 10.9 7.3 6.8 8.4 9.0 8.3 6.7 6.1 10.4 6.1 5.8 10.9 6.4 6.6 5.0 10.0 9.9 5.3 Center for Health Statistics. Colorado Connecticut Massachusetts Maryland North Carolina Pennsylvania Georgia South Carolina Mississippi National 517 TABLE 29 Changes in Cerebrovascular Disease Death Rates by Country for Men Ages 55-64, 1970 and 1980, per 100,000 Country 1970 1980 Change % Change Japan 444 204 240 54 U.S. Black 343 190 153 45 Australia 175 120 55 31 Canada 105 67 38 36 England and Wales 155 114 41 26 West Germany 160 111 49 31 U.S. Whites 120 64 56 47 New Zealand 137 112 25 18 Sweden 87 72 15 17 Poland 85 125 40 47 WHO Annual Report. TABLE 30 Percentage Decline in Stroke Mortality By Age, Race and Sex 1970-1980 WM WW BM BW 35-44 44.9 41.7 44.6 56.0 45-54 39.0 38.7 39.8 48.2 55-64 47.6 37.6 44.8 48.9 65-74 42.9 43.0 39.3 46.2 7 5-84 37.2 37.9 26.0 31.3 Health United States, 1983. 519 TABLE 31 L i ' « »_________i_________i________i_________i________i_________i_________i_________' * «.....■ ■ 1955 196C 196S 1970 1975 Mortality from cerebrovascular disease in population aged 55-64, (1955-1974). Tseng WP, Epidemiological study of hypertension and stroke in Taiwan. In: Prophylatic Approach to Hypertensive Diseases, Y. Yamori et al. (eds.), Raven Press, New York, 1979. 520 TABLE 32 Age-Specific Prevalence Rates of Stroke (per 1,000) Found in Examination, 1967 Through 1969 Age White Men White Women Black Men Black Women First Exam-ination, yr Population Examined No. of Cases Rata Population Examined No. of Cases Rate i-Population Examined No. of Cases Rate Population Examined No. of Cases Rate _i34 198 *» 196 1 5.1 82 110 35-44 152 3 19.7 175 63 2 31.7 93 2 21.5 45-54 223 9 40.4 221 2 9.0 133 9 67.6 147 7 47.6 55-64 138 7 50.7 161 5 31.0 61 5 82.0 109 2 18.3 >65' 60 9 150.0 102 3 29.4 47 2 42.6 59 7 118.6 Total _S35 yr 573 28 659 10 304 18 408 18 All ages 771 28 855 11 386 18 518 18 Age-adjusted rate <35 53.2 15.0 58.6 43.4 All ages 40.9 12.7 45.0 33.4 • Only 11 persons were greater than 75 years of age. two of whom had cerebrovascular disease. Heyman A, Karp HR, Heyden S, et al: Cerebrovascular Disease in the Biracial Population of Evans County, Georgia. Arch Intern Med 1971;128:949-955 521 TABLE 33 Prevalence of stroke in examined cohorts as determined by neurologists Japan" Hawaii* 1 California' Significance level" Age group N Rate/1.000 /V Rate/1,000 N Rate/1,000 Definite cases only 45-49 0 - 0 - 2 2.7 NS 50-54 5 20.3 14 4.9 4 7.6 <0.02 55-59 16 41.3 12 6.2 2 7.4 <0.001 60-64 19 49.5 25 19.2 3 18.1 <0.01 65-69 28 72.4 29 34.2 3 19.4 <0.01 Total 68 46.6 80 10.7 14 7.6 Age-adjusted 35.4 10.7 10.4 rate' Definite and possible cases 45-49 0 — 1 2.0 4 5.5 NS 50-54 7 28.5 21 7.3 5 9.6 <0.01 55-59 17 43.9 18 9.3 3 11.0 <0.001 60-64 23 59.9 31 23.8 3 18.1 <0.01 65-69 33 85.3 41 48.4 4 25.8 <0.01 Total 80 549 112 15.0 19 10.3 Age-adjusted 42.5 15.0 13.0 rate' "In Japan,-85% of definite and possible stroke cases were seen by the neurologist. Remaining cases diagnosed by neurologist from review of clinic records and occasionally from hospital records. * In Hawaii. 62% of definite and possible stroke cases were seen by the neurologist. Remaining cases diagnosed by neurologist from review of clinic and hospitalization records. The latter were usually available. r In California, 74% of definite and possible stroke cases were seen by the neurologist. Remaining cases were classified as definite stroke or no stroke on the basis of the screen- ing test results. "Chi square test with two degrees of freedom. ' Age-adjusted by direct method to age structure of Hawaii cohort. Kagan A, Popper J, Rhoads GG, et al: Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California: Prevalence of stroke. Cerebrovascular Diseases, edited by P. Scheinberg, Raven Press,-New York, 1976. 522 TABLE 34 Number of Hospital Discharges* for Stroke By Race and Average Length of Stay 1981 Discharges Average Length of Stay ______________________(in 1, OOP's)________________(DaVs) ______. White 707 All Others 100 Total 806 *First Level Diagnosis Only National Center for Health Statistics, 12.2 14.2 12.4 Hospital Discharge Survey. 523 TABLE 35 Stroke Economic Costs (1980) Total $5.1 Billion Hospital 52% Nursing Home 35% Professional Ser. 11% Drugs 2% Per Capita <65 Men $8 Women $7 _„ >65 Men $128 Women $147 Total Per Capita <65 Men $627 Women $791 >65 Men $2,278 Women $2,667 Health Care Financing, Summer, 1984. 524 TABLE 36 Swniur «'/ ("i.M. and liuidctuv" Kttln tor Slmkc. .S<»i.//» .\Uihmim Study /'cpuluiiim. l9ir«tkc*« Female number of Mr« »kes Total number of strokes As_c All l-irv| Incidence rates All Hirvt Incidence rales All First Incidence rales 211-54 5 4 4V 4 3 35 9 7 42 55- 64 5 5 274 6 s 233 II 10 252 65-74 K 7 541 6 6 365 l-l 13 442 75 + II It) 1.3X5 8 K 719 19 18 981 All uecs 29 2ft 24 •>■> 53 48 Crude rate 136 107 121 Ape-adjusted ratc+ 139 88 109 Black Male number of stroke* Female number of strokes Total number of strokes Age All Kirsl Incidence rates All First Incidence rates All First Incidence rates 20-54 7 3 32 9 8 64 16 II 50 55-64 14 13 566 19 18 712 33 31 642 65-74 16 14 1.106 19. 16 957 35 30 1.021 75 + 5* 3 482 IK 16 1.824 23 19 1.268 All aces 42 33 65 58 107 91 Crude rate 110 174 144 Ace-adjusted rate* 172 236 208 'Rate of initial strokes per 100.000 persons. tAgc-adjustcd rates based on the NINCDS National Survey of Stroke, indirect method. ?0nc case had unknown jxior stroke history in each of these categories. Gross CR, Kase CS, et al: Stroke in South Alabama: Incidence and Diagnostic Features—A Population Based Study. Stroke 1984- 15:249-255. 525 TABLE 37 Ftva-yaar incidanca of fatal and nonfatal atroke by race and aax for Slopped Care (SC) and Rata.rod Care (RC) participanta. Data are adtualad for *o« and entry dieatoilc blood proaaura. Shaded portion indicalaa fatal atroke; white portion, nonfatal atroke. a a 2.9 m 'RC i t a 7" Ksc 34 6% 37 i RC I Ifl2% ,---, y 33 3.1 'SC4 1 RQ< _■_•_■ '46 6% i.a F77 2.6 VA 1 1.6 7"" 40.0% 2.3 W> [ 3C 30.4% i.a sc All tiucfc Man Buck WomM Wlwia Man ww» Woman Five-Year Findings of the Hypertension Detection and Follow-up Program. III. Reduction in Stroke Incidence Among Persons with High Blood Pressure. JAMA 1982;247:633-638. 526 TABLE 38 Stroke Registry in Vxuian 197.1-76 Incidence Rales________________ Malta ____________Females No. of Incitlrnoe patients per l.(KX) 0 — 1 0.01 3 0.01 8 0.08 16 0.4 25 1.4 27 2*1 4 l.:i 5 2.1 RO 0 1 i No. of Incidence Age io yra patieola per 1.000 0-0 0 — 10-10 4 0.03 20 20 r, 0.01 ;io no 18 0.12 40-40 64 0.0 m-u\ 66 2.6 li»l *M 50 6.4 70-70 2H 7.8 WJ 6 220 2.0 0.26 Osuntokun BO, Bademosi 0, Akinkugbe 00, et al: Incidence of Stroke an African City: Results from the Stroke Registry at Ibadan, Nigeria, 1973-1975. Stroke 1979;10:205-207. TABLE 39 u .•.«.*•«• Animal Ituidnu c of Dt funic and Pa>.\ihh' Sin ike per HMK) by Aft' Japan I972-78 Haw aii l9o5-73 No. «i| No. of No. of No. of Age subjects cases Ratet subjects cases Kate Test+t 45-49 54 0 0.0 1825 II 1.0 NS 50-54 239 7 7.3 2766 39 2.3 * * 55-59 J67 11 7.5 1569 21 2.2 *** 60-64 357 17 11.9 1306 37 4.7 *** 65-69 349 27 19.3 429 18 7.0 . *« Total 1366 62 11.3 7895 126 2.7 Age adjusted raict+t 7.4 2.7 .*« ^Annual Incidence rate is calculated as follows: Japan: (No. of cases/No. of subjects)'4 (Years follow-up). Hawaii: (No. of cases,No. of subjects.,6 i Years follow-upl ■f*X2 lcsl of ,xvo ra,C;> between two cohorts NS: /> > 0.10. •: p < 0.05. *•: p <»0.0I. ***: p < 0.001 t+^Calculated by the indirect method wiih Hawaii as standard Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 528 TABLE 40 Incidence rate, death rate, and prevalence rate of stroke per 1,000 per annum, by age and sex, Taipei, 1975 Age and sex Male <54 55-64 65+ Total Female <54 55-64 65+ Total No. of subjects 23,632 1,467 892 25,991 21,832 1,268 1,069 24,169 Incidence No. % 13 17 23 53 2 8 20 30 0.6 11.6 25.8 2.0 0.1 6.3 18.7 1.2 4 11 12 27 2 7 15 24 Death No. % 0.2 7.5 13.5 1.0 0.1 5.5 14.0 1.0 Prevalence No. 28 31 66 125 6 17 46 69 % 1.2 21.1 74.0 4.8 0.3 13.4 43.0 2.9 Tseng WP. Epidemiological study of hypertension and stroke in Taiwan. In: Prophylactic Approach to Hypertensive Diseases, Y. Yamori et al. (eds.), Raven Press, New York, 1979. TABLE 41 0\ trail Incidence Rates* for Stroke by Diagnostic Category. Rate and Sex South Alabama Study Population, 19H0 Diacnosis Total"! Athcro- ihrombotic infarction Unspecified origin infarction Embolic infarction Lacunar infarction Parenchy- matous hemorrhage Subarachnoid hemorrhaec All cases 160(1001$ 9(6) 64(40) 42(26) 20(13) 13(8) 9(6) crude rate 282.4 15.9 113.0 74.1 35 J 22.9 15.9 White male number 29( 100) 0 9(31) 12(41) 5(17) 2(7) 0 rate 242.7 0 75.3 100.4 41.8 16.7 0 female number 24(100) 0 10(42) 6(25) 2(8) 1(4) 3(13) rate 177.3 0 73.9 44.3 14.8 7.4 22.2 Black male number 42(100) 5(12) 13(31) 15(36) 4(10) 5(12) 0 rate 307.5 36.6 95.2 109.8 29.3 36.6 0 female number 65(100) 4(6) 32(49) 9(14) 9(14) 5(8) 6(9) rate 371.2 22.8 182.8 51.4 51.4 28.6 34.3 •All strokes per 100.000 persons aged 20 and over. ^Includes 139 initial stroke cases. 19 recurrent stroke cases. 2 cases with unknown CVD history. Three cases which had other stroke diagnoses arc in the total but not listed by diagnosis. .Numbers in parentheses are the piuponions of cases in each diagnostic category. Gross CR, Kase CS, et al: Stroke in South Alabama and Diagnostic Features—A Population Based Study. 15:249-255. Incidence Stroke 1984; 530 TABLE 42 Xuinbcr of Stroke Cases by Subiypr and Certainty Diagnosis — Japan and Hawaii Japan Huwaii Defi- Pos- Defi- Pos- Subt\pc Total nite sible Total nite sible Tolal 62 38 24 126 71 55 ICH 18 12 6 34 26 S T-E 44 26 18 74 42 32 Unknown 0 0 0 18 3 15 Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 531 TABLE 43 Average Annual Stroke Incidence Per 1000 by Age and Subtype — Definite Cases Only__________________________ Total Intracranial hemorrhage Thromboembolic stroke Japan Hawaii Japan Hawaii Japan Age No. of cases Rate No. of cases Rate No. of cases Rate No. of cases Rate No. of cases Rate No. of cases Rale 45-49 0 0.0 6 0.55 0 0.0 2 0.18 0 0.0 3 0.27 50-54 6 6.3 22 1.3 4 4.2 10 0.60 •» 2.1 II 0.66 55-59 7 4.8 13 1.4 2 1.4 3 0.32 5 3.4 10 I.l 60-64 II 7.7 21 2.7 i 1.4 9 i.l 9 6.3 12 1.5 65-69 14 10.0 9 3.5 4 2.9 ■> 0.77 10 7.: 6 2.3 Total 38 7.0 71 1.5 12 2.2 26 0.55 26 4.S 42 0.S9 Age ad- juslcd rate 4.7 1.5 1.7 0.55 _____JJ) Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 532 Risk factor and r_)ce-sex group_______ TABLE 44 Mean systolic and diastolic blood pressure and mean serum cholesterol of adults 25-74 years by race, sex and survey period showing standard errors and age-adjusted* means: United States, 1971-75 and 1976-80. 1971-75 Mean Standard error 1976-80 Mean Standard error 1971-75 Age-adjusted* mean Standard error 1976-80 Age-adjusted* Standard mean error Systolic blood pressure (mmHg) White men 132.8 0.5 130.5 0.6 132.8 0.5 130.6 0.7 White women 130.1 0.5 125.6 0.7 129.4 0.4 125.1 0.7 Black men 140.0 1.6 133.2 1.1 140.4 1.7 133.7 1.1 Black women 137.8 1.4 130.6 1.2 138.6 1.4 131.7 1.1 Diastolic blood pressure (mm Hg) White men 83.7 0.3 82.3 0.5 83.6 0.3 82.2 0.5 White women 80.4 0.4 78.1 0.5 80.1 0.3 78.0 0.5 Black men 89.9 1.0 85.3 0.8 89.8 1.0 85.5 0.8 Black women 87.0 0.9 82.3 0.6 87.1 0.8 82.6 0.6 Serum cholesterol (mg/dl) White men 218.0 1.1 216.7 1.2 217.5 1.0 216.7 1.1 White women 222.8 1.1 220.7 1.2 221.4 1.0 219.9 1.2 Black men 225.5 4.5 214.7 3.1 225.7 4.9 214.9 2.8 Black women 219.4 2.9 216.9 3.0 221.2 2.6 219.0 2.6 *Age-adjusted by direct method to the total U.S. population as estimated at the midpoint of the 1976-80 National Health and Nutrition Examination Survey. SOtBCE: Division of Health Examination Statistics, National Center for Health Statistics: Data from the National Health ahd Nutrition Examination Survey. Rowland ML, Fulwood R: Coronary "Heart Disease Risk Factor Trends in Blacks Between the First and Second National Health and Nutrition Surveys, United States, 1971-1980. Presented at American Heart Association c.vmmm iimi on rnrmv.rv llrirl l.i,ci .r> in R1 1 r l- Pnnnlil ii.n«-- Mirrli S 1Q81 TABLE 45 Prevalence rates of elevated blood pressure levels* for black persons 25-74 years by sex and age: United States, 1971-75 and 1976-80. Sex and Age Men Age-adjusted* rate for 25-74 years, 25-34 years..........................■ 45-44 years..........................■ 45-54 years..........................- 55-64 years.......................... 65-74 years...............•.......... Women Age-adjusted* rate for 25-74 years 25-34 years..................... 35-44 years..................... 45-54 years..................... 55-64 years..................... 65-74 years..................... Elevated blood pressure* 1971-75 Standard 1976-80 Standard Rate error Rate error Per 100 population 35.7 3.7 23.1 1.8 16.4 5.7 11.7 2.4 37.7 8.8 22.3 4.4 34.7 6.8 23.0 5.7 59.9 6.5 39.2 4.5 43.7 8.0 27.5 3.0 30.5 2.7 24.4 2.4 ..... 12.4 5.1 5.7 5.4 7.4 6.7 4.3 17.6 37.3 36.4 43.4 l.S ..... 23.9 4.1 ..... 39.4 5.6 ..... 46.0 6.2 ..... 46.7 5.8 * Systolic blood pressure of at least 160 mm Hg and/or diastolic blood pressure of at least 95 mm Hg. ♦ Ase adiusted by direct method to the total U.S. population as estimated at the midpoint of the 1976-80 National Health and Nutrition Examination Survey. SOURCE- Division of Health Examination Statistics, National Center for Health Statistics: Data from the National Health and Nutrition Examination Survey, Rowland ML, Fulwood R: Coronary Heart Disease Risk Factor Trends in Blacks Between the Firsthand Second National Health and Nutrition Examination Surveys, United States, 1971-80. Presented at American Heart Association Symposium on Coronary Heart Disease in Black Populations, March 5, 1983. 534 TABLE 46 Mean systolic blood pressure ( zSD) in Laredo Project participants and HANES I subjects, by sex :e n Laredo HANES 1(1) Ag "Spa nish- US US (years) __to ject Mexican white black American" Men 35- ■44 18* 129.2 £ 19.1 124.7 £ 12.8 127.0 = 14.8 136.7 £ 18.8 45- -54 37 134.4 £ 18.1 140.1 £ 15.0 134.7 £ 19.7 141.7 £ 28.2 55- -64 42 132.1 r 25.7 139.9 r 19.9 139.6 £ 20.4 144.2 £ 23.0 65- -74 30 150.3 .t 29.2 146.0 £ 19.4 146.0 = 24.1 156.6 £ 28.3 Women 35- -44 34* 119.2 _c 20.6 122.4 £ 19.8 122.6 £ 18.7 130.5 £ 21.4 45- -54 93 126.9 £ 18.3 130.0 r 17.5 131.1 £ 22.2 150.8 £ 35.1 55- -64 70 134.8 r 21.6 144.8 ~ 28.0 143.0 £ 25.2 153.4 r 27.4 65- -74 65 155.7 £ 25.5 150.1 £ 21.2 151.6 £ 24.7 161.3 r 28.7 Mean diastolic blood pressure (±SD) in Laredo Project participants and HANES I subjects, by sex Men 35-44 18* 89.6 £ 13.3 81.8 £ 9.7 84.2 £ 11.3 91.2 £ 12.1 45-54 37 88.3 £ 10.6 86.5 £ 9.3 87.5 £ 12.7 91.9 £ 16.5 55-64 42 82.6 £ 12.6 86.8 £ 7.3 86.4 £ 12.0 93.4 r 14.1 65-74 30 84.5 £ 11.5 82.4 £ 10.6 Women 84.9 £ 13.0 90.9 _= 14.0 35-44 34* 79.2 £ 13.0 78.3 £ 11.8 79.3 £ 12.0 86.9 £ 13.7 45-54 93 80.5 £ 9.3 83.7 £ 10.5 82.6 £ 13.1 93.5 r lo.o 55-64 70 80.9 £ 10.3 85.8 £ 9.9 86.2 £ 12.4 90.6 £ 13.9 65-74 65 77.5 £ 12.4 81.6 £ 10.8 85.4 £ 12.5 90.4 = 15.9 ♦ Age range 40-44 years for Laredo Project participants. Stern MP, Gaskill SP, Allen CR, et al: Cardiovascular Risk Factors in Mexican Americans in Laredo, Texas. Am J Epidemiol 1981;113:556-562. 535 TABLE 47 Systolic blood pressure levels by age and sex among Minneapolis American Indians and Twin Cities whites Clinical screenings (whites) Little Earth Other site screenings Average Average second/ Average second/ second/ First SBP third SBP First SBP ' third SBP First SBP third SBP n X (SD) X (SD) a X (SD) X (SD) n X (SD) X (SD) Men 15-24 13 119.5 (15.8) 117.5 (14.3) 36 121.7 (11.6) 120.6 (13.1) 25-34 11 121.5 (13.3) 120.4 (13.3) 49 125.2 (13.8) 125.3 (13.2) 265 117.7 (11-5) 117.7 (U.l) 35-54 13 138.2 (17.8) 139.4 (18.9) 32 127.1 (11.7) 126.8 (11.7) 419 121.4 (13.9) 121.4 (13.5) 55-34 8 136.8 (19.0) L31.5 (15.8) 10 138.0 (19.1) 141.2 (17.3) 235t 131.1 (17.6) 131.0 (17.1) All ages 45* 128.5 (18.1) 127.2 (17.9) 127* 125.7 (13.7) 125.6 (14.0) 919 122.8 (15.3) 122.8 (14.9) Women 15-24 34 114.3 (10.3) 113.2 (10.2) 35 113.7 (12.6) 112.8 (12.2) — 25-34 33 112.9 (10.6) 113.8 (11.0) 42 115.4 ( 9.2) 117.0 (13.0) 343 108.5 (11.3) 103.5 (11.0) 35-54 25 122.9 (17.1) 123.4 (18.6) 46 122.9 (12.0) 122.9 (13.7) 451 116.1 (14.9) 116.0 (14.8) 55-34 13 129.1 (10.1) 129.7 ( 8.2) 15 132.1 (21.2) 133.3 (18.9) 237t 134.3 (19.5) 133.6 (19.0) AH a,;es 105* 117.7 (13.6) 117.8 (13.9) 138* 119.3 (13.8) 119.6 (15.0) 1031 117.8 (13.4) 117.6 (18.0) SBP = systolic blood pressure 'Excludes participants with missing age. sex, or blood pressure data. Twin Cities whites, 55 to 74 years old. Gillum RF, Gillum BS, et al: Cardiovascular risk factors among urban American Indians: Blood pressure, serum lipids, smoking, diabetes, health knowledge, and behavior. Am Heart J 1984;107:765-777. 536 TABLE 48 Systolic Blood Pressure Medians by Age, Area, and Generation mmHg 150 120 JAPAN'/ ... •' CALIF ISSEI 45-49 50-54 55-59 60-64 05-69 AGE Winkelstein W, Kagan A, Kato H, Sacks S: Epidemiological Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Blood Pressure Distributions. Presented at 101st Annual Meeting, American Public Health Association, San Francisco, California, November, 1973. 537 TABLE 48 (cont.) Diastolic Blood Pressure Medians by Age, A rea, and Generation mm Hg 100 CALIF NISEI 90 z < ~ 80 CALIF ISSEI PQ Q 70 45-49 50-54 55-59 60-64 65-69 AGE Winkelstein W, Kagan A, Kato H, Sacks S: Epidemiological Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Blood Pressure Distributions. Presented at 101st Annual Meeting, American Public Health Association, San Francisco, California, November, 1973. 538 TABLE 49 Systolic Blood Pressure Msdians* Adjusted to Relative Weight Differences by Age, Area and Generation: Japan, Hawaii and California (Unadjusted medians shown in parentheses) Age 45 - 49 50 - 54 55 - 59 GO - 64 65 - 69 Japan Hawaii Califo rnia Issei Nisei Issei Nisei 123 124** 122 119** 129 j (120) (122) (128) (118) (130) 127 125 124 127 132 (128) (128) (130) (128) (136) 131 130 126 132 141 (130) (132) (132) (140) (140) 138 130 132 140 134 (138) (134) (136) (114) (140) 142 134 136 146 135 (HO) (140) (138) (146) (142) • Median = mean of medians fiom the rclatixe weight groups: < 90,90-100.100-110. 110-120. 120-130, > 130. •* Values based on n < 30. Winkelstein W, Kagan A, Kato H, Sacks S: Epidemiological Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Blood Pressure Distributions. Presented at 101st Annual Meeting, American Public Health Association, San Francisco, California, November, 1973. 539 TABLE 50 180 MEN -----US. 1960-62 —— Framingham. Mass —— Taipei. Taiwan Japanese Nutrition Taiwan, 1976 en I 140 K 3 w. __ a. -a o o m SYSTOLIC 100 601- ........ _—^ ^ DIASTOLIC L 20 30 40 50 60 Age (in years) 7d 60 Mean blood pressure, by.age for men: five surveys. Tseng WP. Epidemiological study of hypertension and stroke in Taiwan. In: Prophylactic Approach to Hypertensive Diseases, Y. Yamori et al. (eds.) , Raven Press, New York, 1979 540 TABLE 51 Prevalence of Hypertension for Japanese Males by Age and Geographical Location Age Japan Hawaii California Definite Hypertension (Prevalence/1000) Japan Hawaii California Borderline Hypertension (Prevalence/1000) 45-49 139 142 234 163 222 306 50-54 194 183 286 203 221 280 55-59 255 199 263 199 252 253 60-64 280 247 384 239 236 292 65-69 318 352 423 278 288 245 Age-Adjusted Rate 223 194 315 215 235 285 No. of men 2127 7998 1795 2127 7998 1795 Winkelstein W: Cooperative studies of blood pressure in Japanese in Japan, Hawaii and the United States. In: Epidemiology and Control of Hypertension, edited by Oglesby Paul, Stratton Intercontinental Medical Book Corporation, 1975. 541 TABLE 52 Prevalence of left-ventricutar toyfwrwaffcy tad hypawwivt n«__n liiwi— anwug Jap__» mm rmi. br geographical loom*. Af*a__juswd provaiancsMOOO a*-* lasaA to-2127) Hawaii to - 7991) California to-1795) LaA-venincuUr hypertrophy • Hypertensive heart tiitmn * 144 5.1 4.1 44 ■ Uft-vcmricular hypertrophy - Mmimmu Cartes 3-4 »*_» 4-1, 4-2 or 4-3 piuc 5-1. 5-2 or • Definite hypertension (sy«ohc > M» wwwi95 mm Hg and Screen 2 DBP >90 rain Hg by Color Abnormality Q-QS ST Depression (A.1-4.3) Negative T Waves (5.1-5.3) A-V Conduction Defects (6.1-6.3) Ventricular Conduction Defects (7.1-7.8) Rhythm (8.1-8.6) Low QRS (9.1) ST Elevation (9.2) Lai'c Axis deviation (£30°) Rifcht Axis deviation (>+120°) liigh R - left (3.1) High R - right (3.2) Hi£h R (3.3) 5.1-5.3^ or left axis deviation ^-20°, or 3.1 or 3.3 .Any Abnormality Non- black Black (..=623) (N -6367) Total No. (N-6995) No. Percent No. Percent Percent 18 2.9 95 1.5 113 1.6 20 3.2 168 2.6 188 2.7 60 9.6 304 4.8 364 5.2 11 1.8 74 1.2 85 1.2 24 3.8 423 6.6 447 6.4 27 4.3 242 3.8 269 3.9 0 0.0 2 0.0 2 0.0 56 8.9 70 1.1 126 1.8 17 2.7 208 3.3 225 3.2 1 0.2 7 0.1 8 0.1 81 12.9 248 3.9 329 4.7 0 0.0 0 0.0 0 0.0 150 23.0 430 6.C 580 8.3 275 43.8 1096 17.2 1371 19.6 328 52.3 1769 27.8 2097 30.0 MRFIT Study, Unpublished Data. 544 150 140 H ~ 130-1 X | 120 T 110 in 8 100 H S 90 o CD 80 70 -I 60 TABLE 55 Males ..o Urban •- Rural o i i i " t i » i i i 15- 20- 25- 30- 35- 40- 45- 50- 55- Age in years Mean systolic and diastolic l)lood pressure of urban and rural males. 160 150 140 a» x 130 -I E 1 120 ^ S 110 * 100 i CO 80 70 60 Females Diastolic ^-° Urban Rural o< 15- 20- 25- 30- 35- 40- 45- 50- 55- Age in years Mean systolic and diastolic blood pressure of urban and rural females. Oviasu VO, Okupa FE. Occupational factors in hypertension in the Nigerian African. J Epidemiol Community Health 19 79; JJ.2/4-2A8, 545 TABLE 56 <, .59/ <, Mmm. Hg 16O-179/95-109 >ieo/ >no BM BW Risk of stroke or ischemic heart dis- ease according to blood pressure on entry—age- adjusted. In this figure and in the subsequent ones, the numbers within each bar indicate the population at risk in the denominator and the subjects developing either stroke or ischemic **art disease in the numerator. Heyman A, Karp HR, Heyden S, et al: Cerebrovascular Disease in the 128?949-955PUlati°n °f ***** C°Unty' GeorSia« ^ Intern Med 1971; 546 TABLE 57 Total Fetal and Nonfatal Stroke** (fatal Hortmtal Total Group • 8_Mt 6Uo* aa Mwt aa Strot •a SC RC 6C RC 8C RC «C RC Blaofc man 1.064(42) 1.084(70) 7 14 26 28 32 40 Biaa* mommn 1.344 (CM) 1.364(91) 8 18 18 27 24 46 -Vtirta man 1.692(60) 1.881(130) 9 12 18 36 27 47 WrtMt# woman (.186(80) 1.168(124) 6 a 14 18 IS IS Total biacka 2.406(00) 2.438(161) 16 32 41 63 68 88 Total «*ft*«a 3.077(149) 3.017(264) 14 20 32 63 48 73 Total 6.486(246) 6,466(416) 29 62 73 toe 102 166 'Data lor Slapped Cera (SC) an* RetarraO Cera (RC) participant* during *»■ yaar foHowup. by raca-aam groupe fMumbara to parentheeea tndtoata untuwna In whom nonfatal atrofca lno.8__.tn mi not setamwneo. Five-Year Findings of the Hypertension Detection and Follow-up Program. III. Reduction in Stroke Incidence Among Persons with High Blood Pressure. JAMA 1982;247:633-638. 547 TABLE 58 Number of Deaths by Cause for MRFIT Black and White Men Black Men White Men Cause of Death (ICD 9th Revision) Total With Death Certificates All Cardiovascular Diseases Cerebrovascular Diseases (430-438) Myocardial Infarction (410) Other Ischemic Heart Disease (411-414) Hypertensive Heart Disease (402) Other Hypertensive Disease (401,403-405) Other Cardiovascular Disease (390-459 Exclusive of above) All Non-cardiovascular Diseases Genitourinary Diseases (580-629) ' Diabetes Mellitus (250) Neoplastic Diseases (140-239) Gastrointestinal Diseases (520-579) Respiratory Diseases (460-519) Infectious Diseases (001-139) Accidents, Suicides and Homicides (800-999) Other Disease Rate/ + No, 1000 Percent 450 19.2 100.0 203 8.6 45.1 30 1.3 6.7 17 0.7 0.0 3.8 0.2 48 2.0 10.7 Rate/ + No. 1000 Percent 4602 14.1 100.0 2226 6.8 48.4 152 0.5 3.3 78 3.3 17.3 1225 3.8 26.6 29 1.2 6.4 483 1.5 10.5 26 0.1 0.0 333 1.0 0.6 0.1 7.2 247 10.5 54.9 2376 7.3 51.6 0 0.0 0.0 14 0.0 0.3 5 0.2 1.1 37 0.1 0.8 128 5.4 28.4 1440 4.4 31.3 15 0.6 3.3 155 0.5 3.4 12 0.5 2.7 124 0.4 2.7 4 0.2 0.9 16 0.0 0.3 63 2.7 14.0 455 1.4 9.9 20 0.9 4.4 135 0.4 2.9 ♦ Percent of total number of deaths in race group. Neaton JD, Kuller LH, Wentworth D: Total and Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol M^X?,™^™?}^ B1°°d ?ressure Among Black and White Men Followed 5 Years. 548 TABLE 59 "O cd *_» (/) 73 5< I? O) 3 < O W o o o> a O q) cd > in cd cd 91) 7,491 66.322 Neaton JD, Kuller LH, Wentworth D: Total arid Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol Concentration and Diastolic Blood Pressure Among Black and White Men Followed Five Years. Table 60 Comparison of Logistic Regression Coefficients'1' For Diastolic Blood Pressure For All Cause and Cause Specific Mortality for Each Race Group Bl_ tck Whit :e Difference In Coeff Coeff SE Coeff SE SE All Cause (1) .0195 .0033 .0170 .0013 .0025 .0035 Mortality (2) .0202 .0035 .0171 .0013 .0031 .0037 CVD Death (1) .0261 .0051 ,0301 .0019 -.0040 .0054 (2) .0299 .0055 .0322 .0020 -.0023 .0058 CHD Death (1) .0188 .0072 .0263 .0021 -.0075 .0075 (2) .0244 .0078 .0289 .0024 -.0045 .0082 Death From (1) .0623 .0105 .0372 .0068 .0251* .0125 Cerebrovascular Disease (2) .0624 .0109 .0324 .0072 .0300* .0131 (1) All Participants in Race Group. (2) Excludes Those Participants Who Reported Previous Hospitalization for a Heart Attack or Taking Medication for Diabetes. + Estimated for Fixed Age, Serum Cholesterol and Cigarettes per Day. * P < .05 Neaton JD, Kuller LH, Wentworth D: To'al and Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol Concentration, and Diastolic Blood Pressure Among Black and White Men Followed 5 Years. 550 TABLE 61 no BO 70 50 30 I 10 a a _. S 110 1 * 70 30 18 5T6 ____. -I II -1 | No ECG Abnormality I ECG Abnormality Stroke T?t?B II io _Tf 11 t i- m He«n Oismm 11 ]__.___. ii WM WW BM BW Risk of stroke or ischemic heart dis- ease according to ECG abnormalities on en- try—age-adjusted. Heyman A, Karp HR, Heyden S, et al: Cerebrovascular Disease in the Biracial Population of Evans County, Georgia. Arch Intern Med 1971; 128:949-955. 551 TAULE bl Five year mortality per 1000 in the absence of ECG abnormalities: HDFP participants with entry DBPof 90 io 104 mm Hg who were not on antihypertensive medication and were free of specific end organ damage at baseline* White men White women Black men Black women Total SC RC SC RC SC RC SC RC SC RC Sample size (n) 777 745 401 408 189 169 237 284 1604 1606 No. of deaths All causes 22 29 12 14 14 17 6 16 54 76 CVD 11 12 5 4 4 6 2 11 22 33 CHD 8 11 3 4 3 5 2 5 16 25 Mortality rate All causes 28.3 38.9 29.9 34.3 74.1 100.6 25.3 56.3 33.7 47.3 CVD 14.2 16.1 12.5 9.8 21.2 35.5 8.4 38.7 13.7 20.5 CHD 10.3 14.8 7.5 9.8 15.9 29.6 8.4 17.6 10.0 15.6 95% Confidence intervals for (SC-RC) rate All causes (-28.7, 7.5) (-28.7, 19.9) (-85.3, 32.2) (-64.5, 2.4) (-27.3, 0.0) CVD (-14.2. 10.3) (-11.8, 17.1) (-49.0, 20.3) (-55.6, -5.0) (-15.8, 2.1) CHD (-15.7, 6.7) (-15.1, 10.4) (-44.9. 17.4) (-28.4, 10.1) (-13.4, 2.2) CHD = coronary heart disease; CVD = cardiovascular disease; SC = stepped care; RC = referred care. ''End organ damage includes history of definite myocardial infarction, stroke, intermittent claudication, angina pectoris, or elevated serum creatinine (^ 1.70 mg/dl). Five year mortality per 1000 io the presence of ECG abnormalities: HDFP participants with entry DBP of 90 to 104 mm Hg who were not on antihypertensive medication and were free of specific end organ damage at baseline* White men White women Black men Black women Total SC RC SC RC SC RC SC RC SC RC Sample size (n) 313 318 168 198 232 298 220 216 933 1030 No. of deaths All causes CVD CHD 21 14 11 23 12 7 7 3 2 11 8 2 18 4 4 28 15 8 8 5 3 16 5 1 54 26 20 78 40 18 Mortality rate All causes CVD CHD 67.1 44.7 35.1 72.3 37.7 22.0 41.7 17.9 11.9 55.6 40.4 10.1 77.6 17.2 17.2 94.0 50.3 26.8 36.4 22.7 13.6 74.1 23.1 4.6 57.9 27.9 21.4 75.7 38.8 17.5 95% Confidence intervals for (SC-RC) rates All causes CVD CHD (-45.0, (-24.0, (-12.9, 34.5) 38.0) 39.1) (-57.8, (-56.5, (-19.7, 30.1) 11.4) 23.3) (-64.1, (-63.0, (-34.5, 31.4) -3.1) 15.2) (-80.5, (-28.5, (-8.8, 5.1) 27.7) 26.8) (-39.9, (-26.8, (-8.3, 4.2) 4.9) 16.2) CHD = coronary heart disease; CVD = cardiovascular disease; SC = stepped care; RC = referred care. AEnd organ damage includes history of definite myocardial infarction, stroke, intermittent claudication, angina pectoris, or elevated serum creatinine (2:1.70 mg/dl). The Hypertension Detection and Follow-up Program Cooperative Research Group. The effect of antihypertensive drug treatment on mortality in the presence of resting electrocardiographic abnormalities at baseline: The HDFP experience. Circulation 1984:70:996-1003. 552 TABLE 63 Five-year mortality rates in black and white males by education and presence of LVH at baseline for all HDFP stratum I participants and those not receiving medication at baseline: Referred-care males ages 40 to 69 years, with entry diastolic blood pressure of 90 to 104 mm Hg* % Mortality (Deaths) N Crude Age-adjusted^ LVH-t LVH+ LVH- LVH+ LVH- LVH+ All stratum I participants WM > HS (18) 446 WM - HS (26) 402 WM < HS (42) 335 BM < HS (56) 344 Participants not receiving medication at baseline WM > HS (13) 347 WM - HS (21) 315 WM < HS (22) 260 BM < HS (41) 268 ( 1) 8 4.0 12.5 4.5 10.2 ( 2) 10 6.5 20.0 7.0 21.9 ( 7) 16 12.5 43.8 10.6 28.6 (11) 32 16.3 34.4 15.4 32.3 me ( 1) 8 3.8 12.5 4.4 10.2 ( 0) 7 6.7 0.0 7.4 0.0 ( 5) 12 8.5 41.7 7.2 12.4 ( 6) 18 15.3 33.3 14.9 33.2 WM - White males; BM - black male*; HS - high school education completed. •From Tyroler HA: Race, education, and 5-year morulity in HDFP stratum I referred-care males. In Gross F, Strasser T, editors: Mild hypertension: Recent advances. New York, 1983, Raven Press. tAge-adjusted rates by direct method. Standard population is age decade distribution of all white males aged 40 to 69 years with entry diastolic blood pressure of 90 to 104 mm Hg. JLVH determined by ECG. LVH+ defined as major LVH by. Minnesou Code; LVH- defined as all others. Tyroler HA: Overview of risk factorsAfo^™ary heart disease in black populations. Am Heart J 1984;108:658-660. 553 TABLE 64 STROKE INCIDENCE 6r < UJ o o o UJ o z UJ Q O s ~^\ THROMBO-EMBOUC HEMORRHAGIC ill <\2\ 122-134 135-149 150+ SYSTOLIC BLOOD PRESSURE (mmHg) Incidence of stroke by quartile of systolic blood pressure. Kagan A, Popper JS, Rhoads GG: • Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 554 TABLE 65 TEST <»30 1-0- 140- 160 180' 139 159 W9 SYSTOLIC 8LO0 Stmpto 0*fr«wl>on l%l JA'AN n3 13.0 339 13.1 • < HA»»A • 2(3 36 7 23 1 ■ ft 35 T-« Relative risk of i/ilnu nmial firm- orrhage ami thromboembolic stroke by systolic blood pressure ami country. 130- uo- 139 IM ISO- 179 •» <0O5 St* 0 06 < * < 0 10 NS » 0 10 Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 555 TABLE 66 CTs Table 3. Mortality from cerebrovascular disease (CVD) in persons aged 15 and over during 19-year period according to the level of blood pressure at entry Normotensive Borderline Hypertensive Total Male Female Total Total CVD Total CVD Total No. % No. % 5923 1179 738 7840 71 42 126 239 1.2 3.6 17.1 3.0 CVD No. % 7 282 41 0.6 13205 112 0.8 1 603 35 2.2 2 782 77 2.8 1228 108 8.8 1966 234 11.9 10 113 184 1.8 17 953 423 2.4 Tseng WP. Epidemiological study of hypertension and stroke in Taiwan. In: Primary Health Care in the Making, U. Laaser, R. Senault, H. Viefhues (eds.), Springer-Verlag, Berlin, 1985. TABLE 67 Stroke Risk of stroke or ischemic heart disease according to level of serum cho- lesterol on entry—age-adjusted. Heyman A, Karp HR, Heyden S, et al: Cerebrovascular Disease in the Biracial Population of Evans County, Georgia. Arch Intern Med 1971; 128:949-955. 557 TABLE 68 00 •a (1) 0) CO (0 a) 1 0) en u Q < 1 k_ 01 cd O) ~i < o v—.* Ui o as o > o 1— o -Q _-. a> 0) __ oj a O a> _- CO ca 0) GC >- _c •«-• m c0 a> Q • Black Men o White Men (<182) No. Black Men: 5,947 No. White Men: 62,614 III IV (182-202) (203-220) (221-244) Serum Cholesterol Quintile (mg/dl) 4,835 65,222 4.214 64,277 4,117 66,011 V (>244) 4,377 67,260 Neaton JD, Kuller LH, Wentworth D: Total and Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol Concentration and Diastolic Blood Pressure Among Black and White Men Followed 5 Years. TABLE 69 Comparison of Logistic Regression Coefficients'*' For Serum Cholesterol Concentration For all Cause and Cause Specific Mortalilty For Each Race Group Bl_ tck Whit :e Difference In Coeff Coeff SE Coeff SE SE All Cause (1) .0016 .0010 .0028 .0003 -.0012 .0010 Mortality (2) .0013 .0011 .0024 .0004 -.0011 .0011 CVD Death (1) .0054 .0014 .0072 .0004 -.0018 .0014 (2) .0059 .0015 .0073 .0005 -.0014 .0015 CHD Death (1) .0071 .0017 .0079 .0005 -.0008 .0018 (2) .0079 .0017 .0082 .0005 -.0003 .0018 Death From (1) -.0034 .0046 .0023 .0020 -.0057 .0050 Cerebrovascular Disease (2) -.0032 .0047 .0021 .0021 -.0053 .0051 (1) All Participants in Race Group (2) Excludes Those Participants vho Reported Previous Hospitalization for a Heart Attack or Taking Medication for Diabetes. + Estimated for Fixed Age, Diastolic Blood Pressure and Cigarettes per Day Neaton JD, Kuller LH, Wentworth D: Total and Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol Concentration, and Diastolic Blood Pressure Among Black and White Men Followed 5 Years. 559 TABLE 70 STROKE INCIDENCE 6r a: < UJ >- o o o UJ o z UJ Q O ^] THROMBO-EMBOUC ■H HEMORRHAGIC < 193 194-216 217-241 242+ CHOLESTEROL (mg/dl) Figure 7. Incidence of stroke by quartile of serum iholcslcrot. Kagan A, Popper JS, Rhoads GG: Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 560 TABLE 71 Cerebral Hemorrhage* 3- 2.53 0.35 2 1.98 1 - 0.89 0.63 0 - I 129 130 159 160 199 200 / 229 230 Serum Cholesterol (mg/100ml) 3- 2- 1 - 0-1 Cerebral Infarction 1.43 1.04 1.00 1.12 0.59 Standardized morbidity ratio for stroke by the level of serum cholesterol at entry (both sexes, _> 40 years of age at entry, 10-year follow-up, Taisho. Japan). 130 160 200 230 . 1 . . . 129 159 199 229 Serum Cholesterol (mg/100ml) • The decreasing trend (slope) is significantly different from zero at the p < 0.05 level. Tanaka H, Ueda Y, Hayashi M, et al: Risk factors for cerebral hemorrhage and cerebral infarction in a Japanese rural community. Stroke 1983;13(1):62-73. 561 TABLE 72 Axe-adjusted* Mean* of Selected Variables \iilh ami \\llhontStroke — Definite and Passible Cases Japan Hawaii Non- Non- Variable strokc ICH T-E stroke ICH T-_: Systolic MikkJ pressure (mm Hg) 135.2 163.4**" IM 4— I3S.X 148.9" 154.1*" Diastolic blood pressure (mm Hg) 84.4 Y7.2** ^C.}— 81.9 92.5*** 88.8*** Serum cholesterol dug.'100 ml I 190.8 1X3.7 isio 217.1 207.4 Sug. 218.5 Hematocrit (CJ) 43.3 44.2 44Z 44.5 44.2 44.6 Relative body weight (CJ) 101.7 108.7 107 7 III.8 115.8 111.7 Cigarette/day 13.7 10.3 18.5*** 9.6 12.1 11.7 Calorie 2251 2144 20.^ Sue 2177 2020 2063 Animal protein (g) 42.2 28.1 3?.ir 66.9 54.3** 57.4** Vegetable protein (g) 38.0 44.7 36 S 23.3 24.9 21.9 Saturated fat (g) 17.7 10.9 9.0* 55.1 41.1*** 53.3 Unsaturated fat (g) 22.1 20.7 19.4 24.5 29.6 20.7* Simple carbohydrate (g) 63.0 67.0 55* 86.9 74.1 86.4 Complex carbohydrate (g) 283.9 270.0 2si: 166.7 170.9 159.0 Alcohol (g) 31.1 40.6 24 7 12.2 13.0 14.2 Calorics'body weight (kg) 41.0 38.0 »$•« 35.6 31.8 33.7 *Agc adjustment constants Japan Hawaii 45-54 0.216 45-49 0.120 55-59 0.267 50-54 0.233 60-64 0.262 55-59 0.211 65-69 0.255 60-64 0.256 Sug.:/. < 0.10 65-69 0.180 *: p < 0.05 **: p < 0.01 •••:/» < 0.001 Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii Stroke 1984;15:15-23. u" 562 TABLE 73 Standanlized Multiple b»xi\iic Function Coefficients — Definite and Possible Stroke Cases Standardized coefficient Absolute coeflicienl Variable Japan Hawaii Japan Hawaii Tcst+ Systolic MikhJ pressure (mm Hg» 0.73* 0.45* 0.030 0.021 NS Serum cholesterol (mg/l(X) ml) -0.12 -0.06 - 0.003 -0.1X12 NS Relative body weight (r/f) -0.01 -0.12 -0.001 -0.001 NS Hematocrit (CJ-) 0.24 Sug. -0.06 0.066 -0.019 NS Proteinuria 0.19 Sug. 0.14* 0.384 1.046 NS LVH in l-XG -0.08 0.26* -0.401 0.880 * Alcohol (g) -0.11 0.12 -0.002 0.005 NS Saturated fat (g) -0.22 0.02 -0.041 0.002 NS Animal protein (g) -0.27 Sug. -0.27* -0.077 -0.062 NS Cigarettes per day 0.14 -0.12 0.012 0.009 NS Calorics/body weight (kg) -0.22 -0.17* -0.019 -0.013 NS Age 0.39* 0.30* 0.069 0.053 NS tTcsts for difference of absolute coefficient between two countries NS:/» > 0.10 Sug.: p < 0.10 *: p < 0.05 Takeya Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan,, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 563 TABLE 74A C/V.V Disease i \. \o CXS Disease at Autopsy Age adjusted me ;ths No disease (191) Brain infarction (57) Hemonhage (Examination variables BMI (wt/hn 23.5 24.5 23.7 Cigarcttcs/day 12.9 11.7 19.8* Scrum cholesterol 216 235+ 202 Scrum triglyceride 246 220 214 Scrum uric acid 6.2 6.1 6.7 1-hr. pp. glucose 178 199 156 Hematocrit 44.7 45.8* 44.1 FVC 3.1 2.93 3.29 Systolic BP 139 146 153+ Diastolic BP 83 87* 95* Diet variables Calories 2311 2032* 2322 Alcohol (g/d) 20.3 13.5 29.1 Animal protein (g) 69.4 67.7 59.6 Vegetable protein (g) 23.0 20.4 22.7 Saturated fat (g) 59.6 53.3 51.2 Unsaturated fat (g) 23.9 23.5 27.9 Starch (g) 167 136* 148 Cholesterol (mg) 565 512 463 Cf Calorics, protein 16.2 17.7* 14.9 Ct Calorics, fat 32.6 33.3 30.6 <57> (22) Examination variables Body mass index 23.9 24.5 23.8 Cigarcttcs/day 9.7 12.4 20.5* Serum cholesterol 218 233+ 202* Scrum triglyceride 234 237 221 Uric acid 596 6.15 6.78+ Glucose 158 196$ 153 Hematocrit (Ct) 44.7 46.0* 44.3 FVC 3.25 2.99? 3.36 Systolic BP 133 144* 1528 Diastolic BP 82 87+ 958 Diet variables Calories 2301 2091* 2394 Diet alcohol (g/d) 12.9 13.8 29.9+ Animal protein 71.6 70.1 62.4 Vegetable protein 23.7 20.4* 22.9 Saturated fat 60.1 56.3 54.2 Unsaturated fat 26.5 23.7 28.7 Starch 164 139+ 150 Sugar 46.1 38.4 52.2 Cholesterol 549 537 478 Cfc Calorics, protein 16.7 17.7 14.9* He Calorics, fat 33.5 34 31.2 tt Calories. CHO 46.3 44.3 45.5 Ic Calorics, alcohol 3.57 3.94 7.98+ Sodium 2914 2586 2646 *p sJ 0.05. +/> *J 0.01. %p *J 0.001. Ip *J 0.0001. Signifi- cance tests are based on the contrast to the living group. Stemmermann GN, Hayashi T, Resch JA, et al: Risk Factors Related to Ischemic and Hemorrhagic Cerebrovascular Disease at Autopsy: The Honolulu Heart Study. Stroke 1984;15:23-28 565 TABLE 75 Percent of the United States population with diabetes, 1976-80* Age (Years) 20-74 20-44 45-54 55-64 65-74 White 6.2 1.7 8.2 11.9 16.9 Male 5.3 1.0 7.7 9.1 18.1 Female 7.0 2.2 8.5 14.5 16.1 Black 9.6 3.1 12.9 20.8 25.9 Male 8.5 2.8 11.1 14.4 29.4 Female 10.5 3.5 14.5 25.4 23.1 *Sum of percent of persons with a physician-diagnosed medical history of diabetes and of undiagnosed diabetes using NDD6 criteria. SOURCE: Hadden, WC, and Ml Harris. Prevalence of diabetes and Impaired glucose tolerance and plasma glucose In the United States population, NCHS Vital and Health Statistics, Series 11, In preparation. Diabetes in America, Diabetes Data Compiled 1984. U.S. Department of Health and Human Services, NIH Publication No. 85-1468, August, 1985. 566 TABLE 76 25Cr OHAGcflSTOWN EZ_ SAVANNAH WHITE WHITE BLACK MALES FEMALES MALES BLACK FEMALES ALL GROUPS Men* cholesterol and blood giucxxs* determinations and 95% confidence limits (±1.96 standard error) by color and sex for each city. 1972-1974 Stolley PD, Kuller LH, Nefzger MD, Tonascia S, et al: Three-Area Epidemiological Study of Geographic Differences in Stroke Mortality II. Results. Stroke 1977;8:551-557. 567 TABLE 77 Diob«l«s Median Vessel Scores al Given Age 40 ■ 5 30 20 ■ 10 « • 0-_UI«_ \ _. » • NS OiObfllflt )< NS OiOMIflt ••* DioMIM ■-• NO OioMMl 0 ^v- 30 40 SO 60 70 80 90 Median vessel scores for diabetic and nondiabetic males and females. In the fifth and sixth decade* the diabetic female has higher median vessel scores than even the diabetic male. Source: Ila, G.C. Flora, A.B. Baker, R.B. Loewenson and A.C. Klassen: A Comparative Study of Cerebral Atherosclerosis in Males and Females Circulation, Volume XXXVIIT, November, 1968. 568 TABLE 78 Death Reft end Mortality Ratio* tor CHD and Stroke by History oi Diabetes and High Blood Pressure by Sex and Age CHD Stroke Neither High Diabetes Both Neither High Diabetes Both Diabetes Blood but not Diabetes Diabetes Blood but not Diabetes nor High Pressure High and High nor High Pressure High and High Age Blood but not Blood Blood Blood but not Blood Blood Sex Group Pressure Diabetes Pressure Pressure Pressure Diabetes Pressure Pressure Death Ra tes 40-49 152 423 400 1.078' 17 68 ... ... M«?n 50-59 390 910 993 1.995 44 227 126 ... 60-69 803 1.735 1,950 1.891 163 507 258 729» 70-79 1.788 2.904 3.090 5.499 604 1.311 839 2.011* 40-49 15 81 182» 12 72 ... women 50-59 57 196 411 550 28 91 83* ... 60-69 228 504 1,095 1.378 86 244 320 440 70-79 798 1.397 2.412 2,650 393 737 837 779 Mortality R ados 40-49 1.00 2.78 2.63 7.09* 1.00 4.00 ... ... Men 50-59 1.00 2.33 2.55 5.12 1.00 5.16 2.86 ... 60-69 1.00 2.16 2.43 2.35 1.00 3.11 1.58 4.47* 70-79 1.00 1.62 1.73 3.08 1.00 2.17 1.39 3.33» 40-49 1.00 5.40 12.13* 1.00 6.00 ... ... W-men 50-59 1.00 3.44 7.21 9.65 1.00 3.25 2.96* ... 60-69 1.00 2.21 4.80 6.04 1.00 2.84 3.72 5.12 70-79 1.00 1.75 3.02 3.32 1.00 1.88 2.13 1.98 * R_ites based upon only five to nine deaths. Source: Illb, E.C. Hammond and L. Garfinkel: Coronary Heart Disease, Stroke, and Aortic Aneurysm. Arch. Environ. Health, Vol. 19, August, 1969. 569 TABLE 79 Uj 130- > \ 120- O OQ 90- < o UJ u a u 80- 70- 60- 50- 40- 30- 20- 10- 0 210- 150 uoH I I ASSENT .RESENT Z Z % WOMEN 50-59 60-69 70-79 50-59 60-69 70-79 AGE AT EXAM Average annual incidence of atherothrombotic brain infarction according to diabetic status, age at biennial examination, and sex; 22-year follow-up. Diabetes is significantly associated with ABI Incidence, p < 0.05 in men and p < 0.001 in women. Source: Illb, P.A. Wolf, T.R. Dawber, H.E. Thomas, T. Colton and W.B. Kannel: Epidemiology of Stroke. Advances in Neurology, Vol. 16. 570 TABLE 80 Number of deaths in which cerebrovascular disease was listed on the death certificate and percent that also listed diabetes, arteriosclerotic heart disease or hypertensive disease in Baltimore City and 8 counties in Maryland based on multiple-cause tabulation of death certificates for I960* Af e-edJiuUd to total cerebrovascular dlisaie <_4__U_« m« 50-7Bf Race* M* No. CVD DiajnosU Perceataf a with diabetes Percentage with hyper-tensive disease j Percentage with ASHD Balti-more Em Urn WaUrn Bald-store Eastern Western Balti-more Eaetem Western Balti-more Easier* Wssisra WM WF NM NF 321 300 164 183 166 164 92 68 197 209 6.6 9.8 4.9 10.9 8.1 11.0 2.2 14.1 6.0 13.0 27.9 36.2 38.7 46.1 21.0 32.9 26.6 36.4 26.6 34.3 33.6 31.9 23.7 30.7 21.2 19.6 18.1 16.2 18.6 24.1 Toul 967 460 406 * 1960, 1961, 1962 death* in oouaties. t Excludes deaths due to cerebral aneurysm. Source: tta t Kuller and R. Seltser: Cer-b^vascular Disease Mortality In ^f^ ^"?, «, Tournal »* F.nidemiology, Vol. 86, No.2. 571 TABLE 81 Distribution of Other Diseases Listed in Hospital Chart—Stroke Cases by Area (Ages 45-69) Diseases At Hypertei nsion ASHD Diabetes least Area Total stroke cases No. % No. % No. % No. % Miami 531 159 29.9 206 38.8 91 17.1 .289 54.4 Denver 293 81 27.6 91 31.1 36 12.3 135 46.1 Kansas 130 42 32.3 32 24.6 15 11.5 62 47.7 Ssattle 365 90 24.7 82 22.5 44 12.1 142 38.9 Buffalo 532 224 42.1 237 44.5 143 26.9 348 65.4 North Carolina 214 88 41.1 59 27.6 36 16.8 123 57.5 Georgia 93 30 32.3 31 33.3 15 16.1 54 58.1 South Carolina 162 37 22.8 13 8.0 8 4.9 37 22.8 Total 2,320 751 32.4 751 32.4 388 16.7 1,190 51.3 Excludes deaths certified by medical examiner (five), cases discharged alive from hospital but subsequently found dead in mortality study (40), and stroke on past history only. Ib. L. Kuller « person .Peterson 1 -ssel.^.^lets^.^^. . Paegel. M. Saslaw C. Sisk J. Wilbe , dlsease Morbidity s Source: lib A*. LUienfeld,R. Seltser, Nationwide c Stroke 1:86-99, 1970. Study. 572 TABLE 82 hrequency oi First Stroke in the Di.tlu-lic Cohort, Rochester, Minnesota, 1945-1969* Patients Person-years nt risk Relative risk Observed [vp.'Cted. (obs/exp) Hypri tensive 22 10.7 2.1 36 2J.5 1.5 5ti 14 J Nonhvpi-nensive 1.7 3-4 M.J 1.1 21 17.7 1.2 5. III 0 11 P value -IJ "u conndence intervals oi obs^exp Males Females Total Males Females T.n.il 1,136 2,571) J,712 4,29-1 3.172 7.-4I.D <0.001 <0.05 <0.00l NS NS NS 1.3-J.1 1.1-2.1 1.3-2.2 O.B-1.6 0.7-1 H 0.1M 5 *Ol)s = observed; exp = ex|>ectetl; NS = not sigmlicanl. Source: Ilia, M.E. Roehmholdt, P.J. Palumbo, J.P. Whisnant and L.R. E^eback: Transient Ischemic Attack and Stroke. Mayo Clin. Proc, Jan 1983, Vol. 58 573 TABLE 83 STROKE INCIDENCE 6r CL < UJ >- o o o UJ o UJ ^ Q ~^\ THROMBO-EMBOUC ■H HEMORRHAGIC ■ I <£I22 122-149 150-188 189+ GLUCOSE (mg/dl) Incidence of stroke by quartile of serum glucose I hour after a SO gin oral had. Kagan A, Popper JS, Rhoads GG: Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 574 TABLE 84 Age-adjusted Death Rates per 1000 (in 9 years) by Level of Alcohol Consumption Alcohol (oz/mo) CHD Thrombo-embolic stroke Hemorrhagic stroke 0 24.6 3.7 3.0 1-6 16.6 5.2 3.2 7-15 20.9 3.0 7.3 16-39 7.8 1.8 6.3 40-59 5.9 3.7 5.5 60+ 12.8 5.1 12.7 Abbreviation: CHD = coronary heart disease. Kagan A, Yano K, Rhoads GG, et al. Alcohol and Cardiovascular Disease: The Hawaiian Experience. Circulation 1981;64(Suppl III):III-27-III-31. 575 TABLE 85 STROKE INCIDENCE UJ o 6r cr < in >- 4h •*» o o o UJ Q O 2 | 1 THROMBOEMBOLIC ■ I HEMORRHAGIC JJ ill 1-16 17 + ALCOHOL CONSUMPTION (ml/day) Incidence of stroke by tertilc of alcohol con- sumption. Kagan A, Popper JS, Rhoads GG: Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 576 TABLE 86 -J -J 0) UJ Ui en CO 3 a> en *a a < i -_. 0) a en D < O s__/ CO o ivl o > o o _- »— _D 0) LI O 0) *-• (0 a) CO >- _c #^« IO ca a> Q • Black Men o White Men 1-15 16-25 26-35 No. of Cigarettes Reported (per day) No. Black Men: 11,748 No. White Men: 208.481 4.730 21.184 4.905 38.399 1.168 27.044 > 35 939 30.276 Neaton JD, Kuller LH, Wentworth D: Total and Cardiovascular Mortality in Relation to Cigarette Smoking, Serum Cholesterol Concentration and Diastolic Blood Pressure Among Black and White Men Followed Five Years. TABLE 87 180 160 140 120 100 BO I M £ 20 C 180 s E 160 140 120 Men <. 40 _i35 41-49 ___j 36-47 2. SO B8 > 48 100- 8o|— SO 40 20 Women Stroke ____£ Risk of stroke or ischemic disease according to level of hematocrit—age-ad- justed. Heyman A, Karp HR, Heyden S, et al: Cerebrovascular Disease in the Biracial Population of Evans County, Georgia. Arch Intern Med 1971;128:949-955 578 TABLE 88 DEFINITE THROMBO-EMBOLIC STROKE* ■■ HYPERTENSIVE | | NOT HYPERTENSIVE 0 12 3+ 0 12 + NUMBER OF Rl$K FACTORS NUMBER OF OTHER RISK FACTORS #• / CASE OMITTED BECAUSE OF MISSING DATA Incidence of thromboembolic stroke according to number of risk factors: hypertension (BP ^160/95); cigarette smoking: hyperglycemia (>170 mg/dl. I h after 50 gm glucose, p.o.). increased hematocrit Ck48%); and/or left ventricular hypertrophy or strain on the ECG. lumbers within bars denote the number of cases. Kagan A, Popper JS, Rhoads GG: Factors Related to Stroke Incidence in Hawaii Japanese Men. The Honolulu Heart Study. Stroke 1980;11:14-21. 579 TABLE 89 % 30 r 25 20 10 Z2HDFP ■ IHI 255 WHITE BLACK BLACK FEMALE MALE FEMALE Change in prevalence of elevated diastolic blood pressure: HDFP (1973-74) and IHI (1977-78) (DBP _> 95 mm Hg). Apostolides A, Cutter G Krauc ttt __«- „-v T In_or._a._on o High Blo^L^ ££_ B^.Inl^^T Hypertension 1980;2:708-713. en iy/J and 1978. 580 TABLE 90 Secular trends in hypertension control Study Minnesota Heart Survey (10) Date Anglos 1973-1974 Per cent under control* Men 29% Women 53% Hypertension Detection and Follow-up Program (23) 1973-1974 28% 52% Impact of Hypertension Information Study (8) 1977-1978 44% 69% Minnesota Heart Survey (10) 1980-1981 72% 81% San Antonio Heart Study 1979-1982 87% 97% Laredo Study (12) Mexican Americans 1979 37% 77% San Antonio Heart Study 1979-1982 64% 87% * Expressed as per cent of total hypertensives in the community, including both treated and untreated. + These studies also present data on hypertension control in blacks which indicate that black women are better controlled than black men and that the level of control in the black community also improved from 1973-1974 to 1977-1978. Franco LJ, Stern MP, et al: Prevalence, Detection and Control of Hypertension in a Bi-Ethnic Community: The San Antonio Heart Study. Am J Epidemiol (In Press). 581 TABLE 91 Baseline Body Mass Index (BMI) and 5-Year Mortality Ln 00 to Quintile of Stepped- care participants Refenred- care participants Age-sex-race- Age -sex-race- [(SC-RO/RC) body mass Sample Number Rate adj. rate Sample Number Rate adj. rate SC - x 100" index size of deaths per 1,000 per 1,000 size of deaths per 1,000 per 1,000 RCfl (%) All participants* <23.89 1,139 105 92.2 90.2 1,040 141 135.6 127.5 -37.3 -29.3 23.89-26.42 1,112 71 63.8 65.5 1,066 76 71.3 72.0 -6.5 -9.0 26.43-28.78 1,043 50 47.9 44.6 1,135 81 71.4 70.6 -26.0 -36.8 28.79-32.25 1,090 59 54.1 55.4 1,099 59 53.7 54.5 0.9 1.7 2=32.26 1,090 62 56.9 59.9 1,094 59 53.9 62.5 -2.6 -4.2 DBPf 90- 104 mm Hg and not on antihypertensive drugs at basel ine^ <23.89 606 44 72.6 72.8 593 66 111.3 113.3 -40.5 -35.7 23.89-26.42 638 33 51.7 60.6 615 42 68.3 76.8 -16.2 -21.1 26.43-28.78 588 26 44.2 38.3 639 34 53.2 62.1 -23.8 -38.3 28.79-32.25 574 27 47.0 52.5 577 29 50.3 53.0 -0.5 -0.9 2*32.26 494 23 46.6 47.2 519 24 46.2 56.6 -9.4 -16.6 0 Based on age-sex-race-adjusted rates. b \2 for homogeneity = 7.63 (P = 0.12); x2 for SC-RC difference in mortality adjusted for baseline BMI = 8.43 LP < 0.005). r DBP = Diastolic blood pressure. d x for homogeneity = 2.40 (P = 0.66); x2 for SC-RC difference in mortality = 5.10 LP = 0.025). The Hypertension Detection and Follow-up Program Cooperative Group. Mortality findings for stepped-care and referred-care participants in the Hypertension Detection and Follow-up Program, stratified by other risk factors. Prev Med 1985;14:312-335. TABLE 92 MtrfNIkWin. |cMTnf .imoi 3 £ M»*H«lfN«'ON (.Hill | SI I KOI © e © .mail vfssr_ scimosis *T»€«OSCtt«'jSiS OF CiMCl f *.(.n_ ** irUt'ltn. moOr. ml- h~tst~r •** ^ fl' Risk (blood pressure. tUolesteroh of cerebral in- farction. Takeva Y, Popper JS, et al: Epidemiologic Studies of Coronary Heart Disease and Stroke in Japanese Men Living in Japan, Hawaii and California: Incidence of Stroke in Japan and Hawaii. Stroke 1984;15:15-23. 583 TABLE 93 Risk factor influence on distribution of stroke and coronary heart disease between geographic areas and ethnic groups with associated topography of vascular involvement and proposed model populations Risk factor Disease Class Blood lipid levels Blood pressure Stroke Coronary heart disease Model population Incidence Site of vascular lesion Incidence 1 2 3 4 High High Low Low High Low High Low High Intermediate High Low Both intracranial and extracranial Mainly extracranial Mainly intracranial and intracerebral Few or none High High Low Low USA black USA white Japan Guatemala Kuller L, Reisler DM: An explanation for variations in distribution of stroke and arteriosclerotic heart disease among populations and racial groups.' Am J Epidemiol 1971;93:1-9. 584 *U.S. GOVERNMENT PRINTING OFFICE:1986-620-638:^0716