Return of Elevated Blood Pressure after
Withdrawal of Antihypertensive Drugs

By VETERANS ADMINISTRATION CooPERATIVE Stupy Group

ON ANTIHYPERTENSIVE AGENTS"

SUMMARY

The rate at which arterial pressure rises after discontinuing active treatment was investigated in a group
of 86 hypertensive patients who received treatment with hydrochlorothiazide, reserpine and hydralazine for
two years or longer and whose diastolic pressures averaged below 96 mm Hg for the last year of treatment.
Sixty patients were assigned double-blind to placebos and 26 were continued on active drugs.

Forty-two of the placebo group of patients were removed over an 18 month follow-up because of return of
elevated blood pressures, 39 being removed in the first six months. Six patients in the placebo group and
none in the treated group were removed because of morbid events. Nine or 15% of the placebo patients
remained normotensive. The rate of rise in arterial pressure in the placebo group appeared to be related
directly to the height of the pressure prior to initiation of active treatment and inversely to the age of the

patients.

Serum uric acid fell significantly while serum potassium rose significantly after active treatment was dis-
continued. The glucose tolerance test changed slightly in a direction toward normal while serum creatinine

showed no significant change.

Additional Indexing Words:

Hypertension
Antihypertensive drug withdrawal

paar RESETTING of blood pressure control
mechanisms can occur is suggested by observa-
tions of McCubbin and his associates.' They
demonstrated upward resetting of the baroreceptors
in dogs subjected to experimental renovascular
hypertension.

The available data in man on modification of
hypertension following long term treatment with an-
tihypertensive drugs are conflicting. Page and Dustan’
initially reported that 9 of 27 patients in whom treat-
ment had been discontinued remained essentially nor-
motensive without drug therapy for periods varying
from six months to five years of observation. In a later
review of 65 patients, however, the same authors
found that diastolic hypertension reappeared in all
but two patients.* In general, the earliest to manifest
return of the hypertension were the patients with
malignant or with renal hypertension while the

 

From participating Veterans Administration hospitals in
Washington, D.C.; West Haven, Conn.; Salt Lake City; Bir-
mingham; Nashville; Jackson, Miss.; Iowa City; Brooklyn;
Oklahoma City; Richmond; West Roxbury; Pittsburgh; San Juan,
PR; St. Louis; Dayton; Memphis; Allen Park.

*For a complete list of the participants in this study, see the
addendum at the end of the article.

Address for reprints: Edward D. Freis, M.D., Senior Medical
Investigator, Veterans Administration Hospital, 50 Irving Street
NW, Washington, D.C. 20422.

Received January 16, 1975; accepted for publication February 3,
1975.

Circulation, Volume 51, June 1975

Placebo effects in hypertension

1107

Return of hypertension

longest remissions were seen in the patients with es-
sential hypertension.

Perry and his associates were able to withdraw
drug treatment in 16 or 5% of 316 severely hyperten-
sive patients without inducing a rise in blood pressure.
Eleven of these patients had benign essential
hypertension and five originally had malignant
hypertension. Remissions were more frequent among
patients who had hydralazine toxicity during treat-
ment.

Thurm and Smith’ discontinued antihypertensive
drugs in 69 patients. They found that 16 or 23%
remained normotensive for periods of 10 to 42 months
after discontinuing treatment. The patients with the
mildest elevations of blood pressure and the least ap-
parent vascular disease were the most apt to remain
normotensive, but there was no correlation with the
known duration of the hypertension or with the length
of treatment. Unlike the previous investigators,
Thurm and Smith concluded that long term remission
of hypertension is not an infrequent occurrence in es-
sential hypertensives treated with antihypertensive
drugs.

The reason for the discrepant results reported by
these various authors may reside in the type of
hypertensive patients under study. Patients with
severe hypertension were preponderant in the first
two studies whereas Thurm and Smith investigated
patients with mild and moderate hypertension.
1108

Further studies seemed indicated in an attempt to
resolve this question.

Materials and Methods

When the Veterans Administration Cooperative Study on
morbidity® was terminated there were a number of treated
patients whose blood pressures had been controlled at nor-
motensive levels for a period of two years or longer. This
group of patients was entered into the present study which
was designed to re-examine the question of re-establishment
of normal pressure in previously treated hypertensive
patients.

The design of the original trial from which the 86 patients
were drawn has been described in previous com-
munications.® 7 The original trial, in which all of these 86
patients were assigned to the active treatment group, will be
designated as trial A and the present trial, in which the
effect of withdrawal of treatment was assessed, will be called
trial B. In trial B the 86 patients were subdivided by
stratified randomization so that 60 patients or 70% were
assigned double-blind to placebos and 26 patients or 30%
were continued on active medications. The latter treatment
as in trial A consisted of hydrochlorothiazide, reserpine and
hydralazine.®

All of the patients who entered trial A were male veterans
with established hypertension. All were hospitalized prior to
treatment. Their blood pressure was measured in the
hospital at least three times daily for six days. The average of
their readings of diastolic blood pressure from the fourth
through the sixth day of hospitalization was 90 mm Hg or
higher and the average of the diastolic blood pressures
recorded in the outpatient clinic while receiving placebos
before randomization into trial A ranged between 90 and
129 mm Hg. Twenty-seven patients exhibited initial
diastolic blood pressures in the range of 115-129 mm Hg
and 59 were in the range of 90-114 mm Hg prior to entering
trial A. Forty-four patients entering trial B were black and
42 were white. Their average age was 52.2 years for the
placebo group and 52.8 years for the treated patients at the
time of randomization into trial B.

The characteristics of the patients prior to receiving any
treatment, that is prior to entering trial A, are indicated in
table 1. The average pressures taken during the last two out-
patient visits preceding their randomization into trial A were
171/109 mm Hg for the 60 patients randomized onto
placebos in trial B and 171/112 mm Hg for the 26 patients
continued on active drugs. The average severity scores were
not significantly different in the 60 placebo treated and 26
actively treated patients (table 1). The duration of known
hypertension prior to any treatment was 6.3 years in the
placebo group and 7.4 years in the actively treated patients.
The average duration of active treatment prior to entering
the present trial was 4.6 years for both groups. The blood
pressure during the last year of treatment averaged 129/82
in the patients randomized to placebos and 124/80 mm Hg
in the patients who continued on active treatment. All 86
patients had been on active treatment with good control for
a minimum of two years prior to randomization.

Screening of patients for randomization into the present
trial began on October 1, 1970. The first visit on or subse-
quent to this date was called the eligibility visit. To be eligi-
ble the patient had to meet the following criteria: 1) The
diastolic blood pressures for the two preceding visits plus the
eligibility visit averaged 90 mm Hg or less. 2) No diastolic
blood pressures above 95 mm Hg were recorded during any
of these three visits. 3) The average of all diastolic blood

VA COOPERATIVE STUDY GROUP

pressures during the preceding 12 months was 95 mm Hg or
less.

Patients who had major cardiovascular complications in
the past were excluded. These complications are defined in a
prior report’ and include stroke, myocardial infarction, con-
gestive heart failure, renal failure and others. Also excluded
were patients who exhibited violations of pill counts on
more than two clinic visits during the preceding year.
Patients who had been transferred to drugs other than
the hydrochlorothiazide-reserpine-hydralazine combination
used in the original morbidity trial also were excluded. All
patients were informed that they may be transferred to inert
tablets but they would be replaced on active treatment if the
hypertension became re-established. Written informed con-
sent was obtained in all cases.

Blood pressures were measured by the nurse or clinic
assistant in the right arm using a mercury manometer after
the patient had been resting supine in a separate examining
room for 15 minutes. Both the point of muffling and of dis-
appearance were recorded but only the values at dis-
appearance are reported in this paper. Blood pressures then
were recorded with the patient sitting and standing. Subse-
quently, blood pressure was measured in the right arm by
the clinic physician in the sitting position only. The latter
values form the basis for the present report.

The end point criteria for removing a patient from the
trial following randomization were as follows: 1) One visit
with diastolic levels above 129 mm Hg. 2) Any two visits
with diastolic levels above 114 mm Hg. 3) Any three visits
with diastolic levels above 104 mm Hg. 4) Any five visits
with diastolic levels above 94 mm Hg. 5) Any major car-
diovascular complication (class A or B morbid event or treat-
ment failure).”

Patients were seen at 1, 3, 7, 11, 15, 19 and 23 weeks
following randomization. At the end of the 23rd week the
diastolic blood pressures recorded at all of these visits were
averaged. If the average exceeded 89 mm Hg the patient
was removed from the trial. Otherwise he continued in the
study for an additional 12 month period unless his blood
pressure rose to levels which met the criteria for termination
as indicated above. The blood pressures taken during the
first six month period following randomization were dis-
regarded in making the decision to terminate. Pill counts
were carried out throughout the trial.

Table I
Characteristics of Patients Prior to Entering Trial A*

 

 

Placebo groupt Active groupt

 

Characteristic (mean) (mean)
Height (in.) 68.3 68.9
Weight (Ib) 174.8 186.8
Duration known hypertension (yrs) 6.3 7.4
Average clinic systolic pressure 171.0 171.0
(mm Hg)
Average clinic diastolic pressure 108.8 111.6
(mm Hg)
Severity scores:
Optic fundi (0-4) 1.1 1.1
Cardiac (0-4) 0.7 0.8
Central nervous system (0-4) 0.4 0.2
Renal (0-4) 0.3 0.6

 

*Trial A indicates original morbidity study.
tindicates assignment in trial B (present study).

Circulation, Volume S!, June 1975
BP RISE AFTER DRUG WITHDRAWAL

The effect of withdrawal of long term thiazide treatment
on certain blood constituents also was determined. At the
third week after randomization blood was drawn for fasting
blood sugar, uric acid, serum potassium and creatinine. The
patients were also given 100 grams of glucose orally and
samples for blood sugar were drawn at 1, 2 and 3 hours
thereafter to compare with the glucose tolerance tests
carried out during active treatment in trial A.

Results
Changes in Blood Pressure

During the entire postrandomization follow-up
period of 72 weeks, 51 of the 60 placebo treated
patients were removed from the trial. Of this number
42 were terminated because of return of increased
arterial pressures, six because of major cardiovascular
complications and three for reasons unrelated to their
blood pressure. Among the 26 actively treated
patients there were no morbid events and only one
patient was removed from the trial because of a
gradually rising blood pressure.

Thirty-nine of the 42 placebo treated patients who
were removed because of significant increase in blood
pressure attained the criteria for termination from
trial B within the first six months following ran-
domization. During this period three patients met the
terminating criterion of two clinic visits with diastolic
levels above 114 mm Hg, 12 had three visits with
diastolic pressures above 104 mm Hg and 16 patients
had five visits with diastolic reading above 94 mm Hg.
Six patients were removed because their diastolic
blood pressure averaged 90 mm Hg or above during
the initial six months following randomization. Two
additional patients were removed after four visits
because all of their diastolic readings were 98 mm Hg
or higher and each had two diastolic readings of 109 or
higher preceding termination. Because of this clear
evidence of return of elevated pressures they were
removed from the trial even though their readings did
not fit any of the listed criteria required for termina-
tion.

The majority of the patients were removed during
the second to sixth postrandomization month rather
than earlier because the levels of diastolic elevation at-
tained were only moderate; therefore, three to five
clinic visits were needed to meet the criteria required
for termination. At 23 weeks after randomization
78.9% of the placebo patients had exhibited at least
two consecutive readings of diastolic blood pressure of
95 mm Hg or above (fig. 1). The 95% confidence
limits at 28 weeks were 67.4% and 90.4% of the
placebo group developing terminating criteria. Only
three patients developed terminating blood pressure
criteria after the first six months of follow-up (table 2).
The terminating criterion in these latter patients was a
diastolic blood pressure above 94 mm Hg on five of
the clinic visits. They were removed after 11, 12 and

Circulation, Volume 51, June 1975

1109

14 months respectively following randomization.
Nine or 15% of the patients in the placebo group as
well as 21 of the 26 patients in the treated group
remained in the trial until its termination at 72 weeks
of follow-up. The nine remaining placebo treated
patients had an average blood pressure of 153/102
mm Hg prior to their original treatment in trial A. The
48 patients who were removed from the study because
of elevated blood pressure or cardiovascular complica-
tion had a higher average pretreatment blood pressure
of 174/110 mm Hg. After 12 months of treatment the
average blood pressure was 131/83 mm Hg in the
nine patients who remained normotensive on placebos
and 133/87 mm Hg in the 48 who returned to
hypertensive levels or had a cardiovascular event.

Morbidity

As indicated above six patients in the placebo group
developed a cardiovascular complication during the
postrandomization period. No morbid event occurred
in the smaller group of patients who remained on ac-
tive drugs. However, the difference was not significant
using the Chi square test. The six cardiovascular com-
plications were fatal myocardial infarction in one,
nonfatal congestive heart failure in three, atrial
fibrillation in one and right bundle branch block in
one patient.

Characteristics Associated with Rate of Return of Hypertension

The rate of increase in pressure also was analyzed
using a more sensitive index then the termination
criteria for estimating the re-establishment of elevated
pressures. This index, which was called the

100 -
90 +
80 +
70+ a

60+ /
sor /
40+

30/

20
10

oo

PLACEBO _.
a

a!

ot

 

Percent with Return of Hypertension

ACTIVE , .
—

ee

 

o—e— eo °

—T ! 1 ! 1 I L 1 t !

6 12 18 24 30 36 42 48 54 60
Post-Randomization (weeks)

Figure 1

 

Cumulative percent of patients attaining diastolic blood pressure of
95 mm Hg or higher on two successive clinic visits ("2 X 95+’) is
shown on the ordinate. Time after randomization is shown on the
abscissa, At six weeks after randomization 51% of the placebo group
demonstrated ‘2 X 95+” and at six months 82% had reached this
level.
1110

VA COOPERATIVE STUDY GROUP

Table 2

Terminating Events in Placebo Treated Patients

 

 

Number of patients terminated

 

 

Time from Cumulative percent
randomization Elevated blood Cardiovascular Noncardiovascular terminated for cardiovascular
in weeks pressure complications reasons Total causes
1 0 0 0 0 0
3 2 0 1 3 3
7 9 1 0 10 20
11 9 0 0 9 36
15 2 3 0 5 44
19 6 1 1 8 57
23 11 0 0 11 76
27 0 0 0 0 76
36 0 0 0 0 76
45 1 0 0 1 78
54 I 0 0 1 79
63 1 0 1 2 82
72 0 1 0 1 84
Totals 42 6 3 51 84

 

“9% 95+” criterion, indicated the time from ran-
domization at which the diastolic blood pressure was
95 mm Hg or higher on two consecutive clinic visits.
The percentage of patients reaching “2 X 95+’ over
time is shown in figure 1. The curve for the placebo
group is seen to rise steeply initially and then more
gradually in a roughly parabolic shape. At six weeks
51% of the patients had attained “2 X 95+” and at six
months 82% had reached this level of blood pressure.

Various clinical features of the placebo group were
examined in order to determine the characteristics
which correlated with an early return of elevated
pressure. Characteristics such as blood pressure and
target organ damage (severity scores) prior to entering
trial A were assessed with respect to the percentage of
patients who attained “2 X 95+” return of hyperten-
sion three, 11 and 23 weeks after randomization in the
present trial.

The most significant characteristic was the average
clinic diastolic blood pressure prior to entrance into
trial A, the median value of which was 107.5 mm Hg.
Using a pretreatment average diastolic blood pressure
of 108 mm Hg, 20% of patients below this level at-
tained “2 X 95+’”’ at three weeks as opposed to 55%
of the group with initial diastolic pressure of 108 or
higher (table 3). Using a one-tailed Chi-square test
the difference was significant at all of the postran-
domization periods examined with P < 0.01 at three
weeks, P < 0.005 at 11 weeks and P < 0.05 at 23
weeks. The correlation between pretreatment
diastolic blood pressure and the average of the
readings at 19 and 23 weeks after withdrawing treat-
ment is shown in figure 2. If the patient was
withdrawn from the study prior to that time the
diastolic blood pressures taken on the last two visits
prior to withdrawal were used to determine the

average reading. The patients with higher pretreat-
ment diastolic blood pressures also had greater blood
pressure increases after withdrawing treatment
(r = 42).

With respect to pretreatment average clinic systolic
blood pressure, the patients with systolic levels of 167
mm Hg or higher showed a greater return to
“2X 95+” than those with lower systolic readings.
However, the difference was only significant
(P < 0.05) at the 11th week postrandomization and
not at three or 23 weeks.

Significantly elevated pressures returned more
quickly in black patients than in white and in patients
under age 50 as opposed to those age 50 and above

 

 

S45b

s . e
5 _

$ 35+ Lot
= a *.* .

§ 25+

E _

@ 15+

- _

& 57

o _

2 5) *

ce bop tf pb tt

 

 

 

90 100 110 120 130 140
Pre-Treatment
Figure 2
Change in diastolic blood pressure after discontinuation of
treatment is shown on the ordinate. Level of diastolic blood
pressure prior to beginning treatment in trial A is indicated on the

abscissa. The degree of rise after discontinuing active drugs is cor-
related with the level of pretreatment diastolic blood pressure.

Circulation, Volume 51, June 1975
BP RISE AFTER DRUG WITHDRAWAL

lll

Table 3
Relationship of Various Characteristics to Rate of Return of Elevated Blood Pressure

 

 

Percent of placebo patients reaching
“2X 95+" by following week

 

 

No. of after randomization
Characteristic patients 3 11 23
Age (yrs)
< 49 25 54 71 82
50+ 35 26 60 77
Race
White 30 24 59 70
Black 30 50 70 87
Duration known hypertension
< 5 yrs 34 32 65 79
5+ yrs 26 44 64 79
Average clinic systolic blood pressure (mm Hg)
(pretreatment )
< 166 31 29 52 72
167+ 29 46 79 86
Average clinic diastolic blood pressure (mm Hg)
(pretreatment)
< 107 30 20 47 68
108+ 30 55 83 90
Total severity score
(pretreatment)
2-7 39 42 68 83
8-15 21 29 57 71
Duration of treatment (yr)
2-3.9 23 48 78 83
4-7 37 31 56 77
Average systolic level for last year of
treatment (mm Hg)
< 126 31 39 65 86
127+ 29 36 64 71
Average diastolic level for last year of
treatment (mm Hg)
< 82 34 35 62 81
838+ 26 40 68 76

 

(table 3). The differences were significant (P < 0.05)
at the third week following withdrawal of active treat-
ment but not at the 11th or 23rd weeks. The racial
difference probably was not associated with race, per
se, but rather with the level of the pretreatment
diastolic blood pressure which was higher in the black
patients. The diastolic blood pressures of the black
patients prior to entering into trial A averaged 112
mm Hg, whereas those of the white patients averaged
only 106 mm Hg. The difference was significant
(P < 0.025). No correlation was found, however,
between age and pretreatment blood pressure. The
correlation coefficient was r = —0.12. Therefore, age,
per se, appeared to have some influence on the rate of
return of elevated pressures, in that the younger the
age the greater the tendency for the high pressures to
become re-established after withdrawal of treatment
regardless of pretreatment level of blood pressure.

Circulation, Volume 51, June 1975

Those who exhibited a fall of 20 mm Hg or more
after treatment in trial A exhibited a significantly
(P < 0.05) greater return to “2X 95+ ”at three
weeks following withdrawal of treatment than those
who exhibited lesser falls of diastolic pressure during
trial A. This result, however, probably is associated
with the fact that those with the highest blood
pressures prior to randomization showed the greatest
falls after treatment in trial A. This result is shown in
figure 3 which displays the relationship between
pretreatment diastolic blood pressure and the degree
of fall at the eighth month following the beginning
of treatment. The two were strongly correlated
(r = 0.72).

Changes in Blood Chemistries

The changes in serum uric acid, potassium and
creatinine three weeks after randomization into trial B
1112

 

oi
oi
+

Fall with Treatment
T

 

 

 

.
| 4 Jt | | i | I I L

90 100 110 120 130 140
Pre-Treatment
Figure 3

 

Level of diastolic blood pressure prior to treatment in trial A is
shown on the abscissa. The subsequent fall with treatment is in-
dicated on the ordinate. The two are strongly correlated (r = 0.72).

are shown in table 4. Comparison was made with the
values obtained at the last annual examination in trial
A, while all patients were undergoing active treatment
prior to randomization into trial B. Uric acid values
were obtained on 51 of the 60 patients who were
assigned to placebos and on 24 of the 26 patients who
were continued on active treatment. Serum uric acid
fell significantly (P < .0005) in the placebo group
from a mean of 7.5 to 6.5 mg/100 ml. Serum
potassium rose significantly (P< .0005) after
withdrawal of active treatment from a mean value of
3.7 mEq/L to 4.2 mEq/L. Neither serum uric acid nor
serum potassium changed significantly in the patients
who continued on active treatment. Serum creatinine
did not change significantly in either the placebo or
actively treated groups of patients.

Glucose tolerance tests were available in 51 of the
60 patients transferred to placebos and in 22 of the 26
patients continued on active treatment. These
patients were not prepared with special diets prior to
the glucose tolerance tests. The results obtained three
weeks after randomization were somewhat in-
conclusive (table 5). There were no significant
changes in the actively treated patients. In the

VA COOPERATIVE STUDY GROUP

placebo group there were significant falls (P < 0.05)
in the fasting and three hour values but not in the one
and two hour values or in the sum of the fasting
through three hour values.

Discussion

Following substitution of placebos, significantly
elevated blood pressure reappeared only gradually in
the majority of patients. Using two successive clinic
visit readings of 95 mm Hg or higher as the criterion
for re-establishment of hypertension, 37% of patients
receiving placebos met this criterion three weeks after
treatment was stopped, 66% at 12 weeks, and 86% at
one year. However, nine percent of the patients
remained normotensive for the entire one and one-
half years of observation.

These results are intermediate between those
reported by Thurm and Smith® and the findings
observed by others.* * Thurm and Smith observed a
gradual return of hypertension in the majority of their
patients while 23% maintained normal blood
pressures for 10 to 42 months after discontinuing ac-
tive treatment. On the other hand, Dustan and
associates? and Perry and his coworkers,* who
observed primarily patients with severe hypertension,
usually found a rapid return of the elevated blood
pressure and only a small percentage of patients
remaining normotensive.

The present observations indicate that arterial
pressure is capable of modification following long
term treatment. Were this not the case the elevated
pressures would reappear promptly, whereas in most
patients its re-establishment was delayed over periods
of weeks and months. Nevertheless, in the majority of
cases the elevated pressures eventually returned in-
dicating that the modification is not permanent except
possibly in a few. What the factors may be that even-
tually cause the arterial pressure to increase are not
clear from this study. A similar fall and gradual return
of hypertension also may be seen after a major com-
plication such as myocardial infarction or cerebral
vascular accident or after major surgery.

A search was made for parameters that might be in-
fluential in determining the rate at which the arterial

Table 4

Changes in Serum Potassium, Uric Acid and Creatinine Three Weeks after Entering Trial B

 

 

Placebo group Active drug group

 

 

Mean in Mean in Mean in Mean in
Measurement No. trial A trial B pe No. trial A trial B pr
Potassium (mliq/L) 45 3.7 4.2 <.0005 22 3.8 3.8 NS
Uric acid (mg/100 ml) 51 7.0 6.5 <.0005 24 7.2 7.2 NS
Creatinine (mg/100 ml) 50 1.30 1.27 NS 23 1.39 1.33 NS

 

*Student’s ¢-test was calculated to test the significance of the difference between means for paired observa-~
tions. The P value was determined using a one-tailed test. NS indicates not significant.

Circulation, Volume 51, June 1975
BP RISE AFTER DRUG WITHDRAWAL

1113

Table 5
Changes in Glucose Tolerance Test Three Weeks after Entering Trial B

 

 

Placebo group

Active drug group

 

 

Blood glucose Mean in Mean in Mean in Mean in

(mg/100 ml) No. trial A trial B P No. trial A trial B P
Fasting 51 112 105 <.05 24 112 113 NS
1 hr after glucose 51 184 187 NS 23 168 177 NS
2 hr 51 171 159 NS 23 148 148 NS
3 hr dl 141 122 <.05 22 119 114 NS
Sum of values 51 608 573 NS 22 534 546 NS

 

pressure increases and for the characteristics that lead
to a sustained normalization of the blood pressure
despite withdrawal of treatment. The most impressive
correlation with the rate of increase in arterial
pressure was the level of pretreatment blood pressure,
particularly the diastolic pressure. The correlation was
direct; the higher the level of pressure before in-
stituting treatment, the more quickly in general was
the pretreatment pressure re-established after discon-
tinuing treatment. The patients who remained nor-
motensive throughout the one and one-half years of
follow-up came from the group who had the mildest
severity of hypertension prior to treatment.

The other parameter which correlated with the rate
of increase in pressure was age. The blood pressure of
younger patients tended to rise more quickly to
hypertensive levels than that of older patients
irrespective of the level of blood pressure. Other
parameters which appeared to correlate with the rate
of increase in pressure could be explained by the fact
that pretreatment levels of blood pressure tended to
average higher in certain groups of patients. For ex-
ample, black patients exhibited an earlier increase in
pressure than whites but their pretreatment blood
pressures also averaged significantly higher. The same
explanation pertains to the observed direct correlation
between degree of blood pressure reduction with
treatment and rate of increase in pressure following
discontinuation of treatment. The patients showing
the greatest fall in blood pressure with treatment were
those with the highest pretreatment levels of blood
pressure. Level of pretreatment blood pressure and
age, therefore, remain as the only primary correlates
with the rate of return of increased pressure.

It should be emphasized that all of the patients had
established hypertension since patients with
borderline hypertension whose diastolic blood
pressures fell to normal during hospitalization were
excluded. If such borderline hypertensive patients had
been included in the trial it is possible that a higher
percentage of patients would have remained nor-
motensive during the follow-up period.

The changes in blood chemistry values following
withdrawal of treatment can be accounted for by the

Circulation, Volume 51, June 1975

effect of withdrawing treatment with hydro-
chlorothiazide. Serum uric acid fell significantly,
serum potassium rose significantly and glucose
tolerance tended to return toward normal. There was
no significant change in serum creatinine indicating
that in this group of patients with normal or nearly
normal renal function there was no evidence of
depression of renal function as judged by serum
creatinine values during the period of reduced blood
pressure.

Addendum

Members of Study Group: Edward D. Freis, MD, Chairman:
Massimo Calabresi, MD; C. Hilmon Castle, MD; Eugene C. Clark,
MD, William S. Coppage, Jr., MD; Leo Elson, MD; Annette Fitz,
MD; Rudph E. Fremont, MD; Edward D. Frohlich, MD; Arthur S.
Gear, MD; David Littman, MD; Donald McCaughan, MD; Milos
Ulrych, MD; Eli A. Ramirez, MD; H. Mitchell Perry, Jr. MD;
James T. Taguchi, MD; J. R. Thomas, MD; Frederick N. Talmers,
MD.

Consultants and Associates; Walter M. Kirkendall, MD; David
W. Richardson, MD; Harold W. Schnaper, MD; Paul D. Williams,
M.S.; Lawrence W. Shaw, A.M.; Mary C. Kyle, M.S.; Bert Silver-
man.

References

I. McCussin JW, Green JH, Pace TH: Baroreceptor function in
chronic renal hypertension. Circ Res 4: 205, 1956
2. Pace IH, Dustan HP: Persistence of normal blood pressure
after discontinuing treatment in hypertensive patients. Cir-
culation 25: 433, 1962
3. Dustan HP, Pace IH, Tarazi RC, Frouuicn ED: Arterial
pressure responses to discontinuing antihypertensive drugs.
Circulation 37: 370, 1968
4. Perry HM Jr, ScuroeperR HA, Catanzaro FJ: Studies on the
control of hypertension. VIII. Mortality, morbidity and
remissions during twelve years of intensive therapy. Circula-
tion 33: 958, 1966
5. Tuurm RH, Smrra WM: On resetting of “‘barostats’ in
hypertensive patients. JAMA 201: 301, 1967
6. Veterans Administration Cooperative Study Group on
Antihypertensive Agents: Effects of treatment on morbidity
in hypertension. I. Results in patients with diastolic blood
pressures averaging 115 through 129 mm Hg. JAMA 202:
1028, 1967
. Veterans Administration Cooperative Study Group on
Antihypertensive Agents: Effects of treatment on morbidity
in hypertension. II. Results in patients with diastolic blood
pressure averaging 90 through 114 mm Hg. JAMA 213:
1148, 1970

~l