Reprinted from the Milbank Memorial Fund Quarterly, July, 1969, Vol. XLVIUI, No. 3, Part 2.
Published by the Milbank Memorial Fund, 40 Wall Street, New York 10005.

EFFECT OF ANTIHYPERTENSIVE TREATMENT
ON MORBIDITY

EDWARD D. FREIS

Clinical trials evaluating the effects of treatment with antihyperten-
sive drugs on the prevention of the complications of hypertension have
been limited almost exclusively to patients with severe hypertension.
In malignant hypertension the life expectancy of the untreated patient
seldom exceeds one or two years. The prolongation of life of many
patients to five years or longer with antihypertensive drug treatment has
provided convincing evidence of the benefits of drugs in this condition.1-*

The course of the disease, however, is much more variable in the less-
severe forms of hypertension. Therefore, the effectiveness of treatment
with respect to morbidity and mortality is more difficult to evaluate.
In the absence of adequately controlled, prospective trials some authori-
ties have questioned the benefit from antihypertensive drugs on mor-
tality.*-*

Recently, this need for definitive data on the efficacy of treatment in
essential hypertension has been partially fulfilled by several prospective
controlled clinical trials. The first is a study by Hamilton in England’
and the second is the Cooperative Study on Antihypertensive Agents
carried out by the Veterans Administration.

In Hamilton’s study 61 patients without evidence of prior cardio-
vascular complications, but with diastolic blood pressures of 110 mm
Hg or higher as measured in the clinic, were alternately assigned to
antihypertensive drug treatment or placebos, Over an eight-year period
of follow-up, 16 of the untreated patients had complications as com-
pared to five of the treated group. Of the latter, the blood pressure was
poorly controlled in four patients. Thus, of the treated patients whose

153
diastolic pressures were reduced to less than 110 by active treatment
only one had a severe complication. The most common causes of
morbidity in Hamilton’s study were strokes and enlargement of the
heart usually associated with congestive heart failure.

The Veterans Administration Cooperative Study was unique in sev-
eral aspects of experimental design. It presents the most convincing
evidence for the effectiveness of treatment. The patients were all males
with diastolic blood pressures averaging between 90 and 129 mm Hg
both in the hospital and in subsequent outpatient follow-up of two to
four months when all received only placebos. Thus far, results have
been published only on the subgroup whose diastolic pressures were
115 mm Hg or above prior to randomization.

No attempt was made in this study to eliminate patients who had any
evidence of prior cardiovascular complications. The exclusions were
patients with a prior history of Grade III or IV changes in the optic
fundi, cerebral or subarachnoid hemorrhage, acute hypertensive en-
cephalopathy, congestive heart failure uncontrolled by digitalis and
‘ mercurials, dissecting aneurysm and uremia. Patients were included in
the study who had prior myocardial infarctions or cerebrovascular
thrombosis (as opposed to hemorrhage), cardiomegaly as determined
by roentgenography or left ventricular hypertrophy as defined by elec-
trocardiography.

A number of problems that are unique to long-term therapeutic trials
in outpatients were considered in the design of the study. One of the
most important is the problem of dropouts. The larger the number of
dropouts the less valid the results. Because it is impossible to determine
the reason for defaulting no assumptions can be made as to their out-
come. Prior experience of the cooperative study group had indicated
the greatest percentage of dropouts occurred in the first several months
after the patients entered the trial. To avoid this a prerandomization
trial period was instituted in which the patients were required to adhere
to regular clinic attendance for a period of two to four months.

A second equally obvious but often overlooked problem in outpatient
trials is the failure of the patients to take their medications with suffi-
cient regularity to provide a consistent therapeutic effect. Identification
of such patients is often difficult. In the Veterans Administration’s trial
all patients received known placebos during the prerandomization trial
period. Pill counts were carried out at each clinic visit. In addition,
riboflavin, which produces fluorescence of the urine in ultraviolet light,
was incorporated as a marker in the tablets. A urine sample was exam-

154
ined under an ultraviolet light at each visit. Only those patients who
exhibited both urinary fluorescence and acceptable pill counts during
two successive visits of the trial period were accepted into the study.

Another problem was to assure that the treatment and placebo
groups of patients were comparable not only with regard to known
prognostic criteria but also with respect to presently unknown factors
that may influence the risk of developimg complications. Standard
randomization procedures were used to assure comparability, relying
on the laws of chance to provide two groups at equal risk at the time
of randomization. The success of this procedure is attested to by the
fact that the two groups were practically identical by all known prog-
nostic indices such as age, race, body weight, average blood pressure,
fundi grades, electrocardiogram changes, heart size and renal changes.

Another crucial part of the trial was to assure that the blood pressure
was reduced to normotensive or nearly normotensive levels in the great
majority of the treated patients. Therapy also should be reasonably
standardized in the various participating clinics both with regard to
drugs and doses. The active drug regimen decided upon was based on
experience gained in previous comparative effectiveness drug trials of
the Cooperative Study.’? Results of these trials indicated that additive
hypotensive effects were obtained by combining antihypertensive drugs.
Diuretic doses of a thiazide plus reserpine plus hydralazine had, on the
average, three times the antihypertensive effectiveness of similar doses
of each drug used separately.12 The doses of each drug decided upon
were 50 mg hydrochlorothiazide plus 0.25 mg reserpine combined in a
single tablet given twice daily. Hydralazine was made up separately in
25 and 50 mg tablets, the smaller dose being given three times daily
initially, followed by the 50 mg dose if the diastolic blood pressure did
not fall below 90. Further increases in either the hydrochlorothiazide-
reserpine tablet or the hydralazine tablets were not permitted. How-
ever, the doses could be reduced in the event of an excessive antihyper-
tensive effect; hydralazine could be discontinued and if hypotension
persisted the thiazide-reserpine tablet could be reduced.

The placebo and active drugs were made up in identically appearing
tablets. Three different code numbers were assigned to each regimen.
The code numbers contained four digits to make them more difficult
for the participants to remember. Another group of “special’’ tablets
were made up. Some of these contained hydrochlorothiazide without
reserpine and the others contained reserpine without the thiazide.
When a severe side effect occurred, presumably caused by either drug,

155
the participant requested the substitute “special” tablets that did not
contain the offending agent. Of course, in some cases one placebo was
substituted for another, but in other instances the substitution was useful
in maintaining the patient on at least two of the three active agents. As
in the trial period, riboflavin was added to both the active drugs and
placebos used in the treatment period so that regular checks could be
made on the patients’ adherence to the program. In addition to the
fluorescence tests pill counts also were carried out at each clinic visit.

The three-drug regimen proved to be quite successful in controlling
the blood pressure in the actively treated group. After four months of
active treatment the reduction of blood pressure from prerandomization
levels in the 73 patients receiving active drugs averaged 43 mm Hg
systolic and 30 mm diastolic. The mean blood pressure for the group
was 186/121 mm Hg in the clinic during the prerandomization trial
period; it fell to 143/91 mm following the institution of active treat-
ment. In the 70 placebo-treated patients the average prerandomization
blood pressure was 187/121 mm Hg and this remained essentially
unchanged following randomization.

A significant difference in the rate of organic complications developed
within a relatively short period in these “high-risk” patients whose pre-
randomization clinic blood pressures averaged between 115 and 129
mm Hg. The average period of follow-up of the placebo-treated pa-
tients prior to termination of the study was only 15.7 months. During
this period morbid events occurred in 27 of the 70 patients. The follow-
up period averaged somewhat longer (20.7 months) in the actively
treated group. The difference in duration of follow-up in the two
groups was because of the termination of many placebo-treated patients
at an early period because of serious complications,

Only two morbid events occurred in the actively treated group dur-
ing this period.

Twenty-one terminating events occurred in the placebo-treated pa-
tients. Four of these were deaths caused by dissecting aortic aneurysm
in two, ruptured abdominal aortic aneurysm in one and “sudden
death” in the fourth. The distribution of deaths is unusual because of
the high percentage of aortic catastrophes. In the 17 nonfatal but
terminating events the most frequent causes were the development of
accelerated hypertension associated with hypertensive neuroretinitis,
congestive heart failure that could not be controlled with digitalis, salt
restriction and mercurial diuretics, and persistent excessive hyperten-
sion. Also, two instances of severe cerebrovascular accidents occurred

156
and two patients exhibited signs of renal failure. Only one terminating
event occurred in the actively treated group. This was a patient who
developed hyperglycemia on the thiazide-reserpine combination tablet.
The “special” tablet containing only reserpine was substituted, but the
patient then developed a depression.

Six events in the placebo-treated patients were nonterminating.
These consisted of two patients who developed myocardial infarction,
two who exhibited thrombotic cerebrovascular events and one who
developed congestive heart failure responding to digitalis. The single
nonterminating event in the active treatment group occurred in a 68-
year-old man who had hypotensive levels of blood pressure accom-
panied by a left-sided hemiparesis.

Fortunately, this clearcut result favoring treatment was not obscured
by an excessive dropout rate. Of the 143 patients, 12 dropped out, an
incidence of less than ten per cent. Further, of this number the dis-
tribution was almost equally divided between the two regimens, seven
having been assigned to placebos and five to active drugs. Hf the worst
possible assumption is made regarding the dropouts, namely that all of
the actively treated patients had developed morbid events as opposed
to none in the placebo group—a most unlikely assumption—the result
would still be significant at a probability of less than one in 1,000 that
the result favoring treatment could have occurred by chance.

Although essential hypertension is generally regarded as a slowly
progressive disorder, such was not the case in these male patients with
clinic diastolic blood pressures of 115 mm Hg or higher. A surprising
number of placebo-treated patients showed evidence of progression to
the accelerated phase of hypertension and probably would have gone
on to develop fatal malignant hypertension if known active treatment
had not been started. The combined results of the VA and Hamilton’s
study also indicate significant protection against vascular catastrophes
such as dissecting aortic aneurysm and cerebrovascular hemorrhage.
It also was apparent that congestive heart failure was prevented by the
antihypertensive regimen.

On the other hand the incidence of atherosclerotic events such as
myocardial infarction and cerebrovascular thrombosis did not occur in
sufficiently high incidence in either group to make an informed judg-
ment regarding the possible protective effect of antihypertensive drug
treatment. In treated patients, myocardial infarction including “sud-
den death” has become the leading cause of mortality according to
Hodge and Smirk,* far exceeding congestive heart failure, cerebro-

157
vascular hemorrhage and uremia, which formerly were the leading
causes of death. Whether the fotal incidence of atherosclerotic events
has been reduced by antihypertensive treatment cannot be deduced
from such uncontrolled observations. It seems likely, however, that the
effect of antihypertensive treatment, if any, on the prevention of accel-
erated atherosclerosis is not as great as the prevention of purely hyper-
tensive complications.

In view of the high risk associated with diastolic elevations in the
110-120 mm Hg range and the conclusive evidence favoring treatment
in this group it would appear that the early identification and treatment
of such patients is an important and immediate public health problem.

The evidence demonstrating the benefits of treatment is more soundly
based than it is in most other chronic diseases. In addition, hyperten-
sion of this degree of severity is not uncommon. The justification and
need for a more positive approach to this problem should be apparent
to all physicians and agencies concerned with national health.

REFERENCES

1 Perry, H. M., Jr. and Schroeder, H. A., The Effect of Treatment on Mor-
tality Rates in Severe Hypertension: A Comparison of Medical and Surgical
Regimens, Archives of Internal Medicine, 102, 418-425, 1958.

2 Dustan, H. P., et al., The Effectiveness of Long-Term Treatment of Malig-
nant Hypertension, Circulation, 18, 644-651, 1958.

3 Harrington, M., Kincaid-Smith, P. and McMichael, J., Results of Treat-
ment in Malignant Hypertension, British Medical Journal, 2, 969-980, 1959.

4 Sokolow, M. and Perloff, D., Five-Year Survival of Consecutive Patients
with Malignant Hypertension Treated with Antihypertensive Agents, American
Journal of Cardiology, 6, 858-863, 1960.

5 Bjork, S., et al., The Effect of Active Drug Treatment in Severe Hyperten-
sive Disease, Acta Medical Scandinavica, 169, 673-689, 1961.

® Mohler, E, R. and Freis, E, D., Five-Year Survival of Patients with Malig-
nant Hypertension Treated with Antihypertensive Agents, American Heart Jour-
nal, 60, 329-335, 1960.

7 Perera, G. A., Antihypertensive Drug Versus Symptomatic Treatment in
Primary Hypertension: Effect on Survival, Journal of the American Medical
Association, 173, 11-13, 1960.

® Goldring, W. and Chasis, H., Antihypertensive Drug Therapy: An Ap-
praisal, Archives of Internal Medicine, 115, 523-525, 1965.

* Relman, A. S., in Ingelfinger, F. J. and Relman, A. S., CONTROVERSY IN

INTERNAL MepicrneE, Philadelphia, W. B. Saunders Company, 1966, pp. 101-
102,

158
10 Hamilton, M. in Gross, F., Antihypertensive Therapy: Principles and
Practice, an International Syposium, New York, Springer-Verlag New York,
Inc., 1966, pp. 196-211.

11 Veterans Administration Cooperative Study Group on Antihypertensive
Agents, Effects of Treatment on Morbidity in Hypertension: Results in Patients
with Diastolic Blood Pressures Averaging 115 Through 129 mm Hg, Journal of
the American Medical Association, 202, 1028-1034, 1967.

12 » Double Blind Control Study of Antihypertensive Agents: III.
Chlorothiazide Alone and in Combination with Other Agents, Archives of In-
ternal Medicine, 110, 230-236, 1962.

13 Hodge, J. V. and Smirk, F. H., The Effect of Drug Treatment of Hyper-
tension on the Distribution of Various Causes: A Study of 173 Deaths among
Hypertensive Patients in the Years 1959 to 1964 Inclusive, American Heart
Journal, 73, 441, 1967.

 

Edward D. Freis, M.D., is on the staff of the Veterans Administration Hospital and the
Department of Medicine, Georgetown University, Washington, D.C.

159