Journal of Cardiovascular Pharmacology
7:S112-S116 0 1985 Raven Press, New York

Choice of Initial Treatment

Edward D. Freis

Veterans Administration Medical Center and Georgetown University Medical Center, Washington, DC

Summary: An important consideration in the choice of
initial treatment is race. In a Veterans Administration
study nadolol reduced blood pressure more in whites than
in blacks, while the reverse was true with hydrochloro-
thiazide. Combining both drugs enhanced antihyperten-
sive effectiveness and abolished the racial difference. The
results of this and other studies suggest that for first drug
selection, beta-blockers are indicated in whites and di-
uretics in blacks. Beta-blockers are also indicated in all
patients with prior myocardial infarction or with tachy-
cardia. Thiazides are also used in combination with other
antihypertensive drugs and in patients with heart failure.
Reluctance to use thiazide diuretics stems from the pos-
sibility of hypokalemia-induced arrhythmias and long-
term elevations of serum cholesterol. However, a causal

The choice of primary treatment has become a
matter of controversy. The long established pri-
macy of the thiazide diuretics is being challenged
principally by the growing popularity of the beta-
adrenergic blocking drugs. This trend has been in-
fluenced, on the one hand, by the opinion that
beta-blockers may protect the heart against fatal ar-
rhythmias and on the other, by recent criticisms
concerning the thiazide diuretics. The principal crit-
icisms are the possibility of fatal cardiac arrhyth-
mias secondary to diuretic-induced hypokalemia
and the long-term increased risk of coronary artery
atherosclerosis resulting from raised serum choles-
terol concentrations. This review evaluates the ev-
idence for the relative effectiveness and long-term
risks of treatment with beta-blockers as compared
with diuretics in primary treatment and indicates
briefly some important step 2 approaches.

BETA-BLOCKER VERSUS THIAZIDE
   AS INITIAL TREATMENT

Is the most effective drug for the initial or primary
treatment of hypertension a beta-blocker or a di-
uretic? Several of the recent Veterans Administra-
tion trials have addressed this question (l-3). In

relationship between hypokalemia and the incidence of
arrhythmias is not well supported by physiologic or clin-
ical evidence. Elevation of cholesterol appears to be tran-
sient, reverting back to pretreatment levels after 6-12
months of treatment. An alternative regimen which is
both highly effective and well tolerated is the combination
of small doses of both a thiazide diuretic and captopril.
Perhaps less well tolerated, but useful where cost is the
major consideration, is a thiazide followed by small doses
of reserpine, if needed; this is an effective, low-cost treat-
ment. Calcium channel blockers appear promising but re-
quire further evaluation. Key Words: Initial treatment-
Antihypertensive drugs-Thiazide toxicity-Beta-
blockers.

one such trial the beta-blocker nadolol was com-
pared with the diuretic bendroflumethiazide and
with the combination of the two (1). Both drugs are
very long acting, permitting once daily dosage. Na-
dolol was of interest not only because of its long
action but also because it exhibits no first-pass phe-
nomenon and it induces renal vasodilatation. There
were 365 patients who were randomized into the
trials with pretreatment diastolic blood pressure in
the range of 95- 114 mm Hg. Nadolol was titrated
from 80 to 240 mg once daily and bendroflumethia-
zide was increased from 5 to 10 mg. The reduction
in diastolic blood pressure was essentially the same
with the two drugs, averaging 12.4 mm Hg with
nadolol and 12.9 mm Hg with bendroflumethiazide.
Systolic blood pressure was reduced somewhat
more with the diuretic than with the beta-blocker.
The reduction with the combination was consider-
ably greater than with either drug alone, the dia-
stolic reduction averaging 17.9 mm Hg.

There was a significant racial difference in the
response to nadolol. Diastolic blood pressure in
whites was reduced by 15.6 mm Hg compared with
only 9.6 mm Hg in blacks. The percentage of pa-
tients whose diastolic blood pressure was controlled
below 90 mm Hg exhibited an even more striking

Address correspondence and reprint requests to Dr. Freis at Veterans Administration Medical Center, 50 Irving St. N.W., Washington,
DC 20422.

s112


HYPERTENSION: INITIAL TREATMENT

s113

difference, with 77% of whites controlled compared
with 31% of blacks. The racial difference was abol-
ished and effectiveness was enhanced in both
groups when the drugs were combined, with 85%
of both blacks and whites achieving control of di-
astolic blood pressure.

Subjective side effects were uncommon with ei-
ther drug but were less with nadolol than with ben-
droflumethiazide. The usual biochemical side ef-
fects of the thiazide diuretics were present with
bendroflumethiazide,

A related trial by the Veterans Administration
group compared hydrochlorothiazide alone with
propranolol alone in 683 patients with mild and
moderate hypertension (2,3). With hydrochlorothi-
azide there were fewer patients withdrawn because
of raised blood pressure, fewer required an increase
in dose and, after discontinuation of treatment,
fewer raised their blood pressure to above pretreat-
ment levels. In this study, also, there were racial
differences in response similar to the nadolol trial.
Hydrochlorothiazide lowered blood pressure in
both groups, but more in blacks than in whites,
whereas the reduction in blood pressure after pro-
pranolol was significantly greater in whites than in
blacks. Similar results have been reported by other
investigators (4,.5).

SIGNIFICANCE OF THIAZIDE-INDUCED
      HYPOKALEMIA AND
    CHOLESTEROL ELEVATION

The recent decrease in the popularity of thiazide
and related diuretics as primary treatment has re-
sulted partly from the suspicion that they exert two
insidious and potentially lethal side effects. One is
that hypokalemia predisposes to cardiac arrhyth-
mias, including ventricular fibrillation and sudden
death. The other is that the moderate rise in cho-
lesterol concentration produced by the diuretics ag-
gravates and accelerates coronary artery athero-
sclerosis. These beliefs are attractive because they
might explain the failure of most of the clinical
trials, all of which incorporated diuretics, to show
a significant reduction in the incidence of coronary
heart disease. Nevertheless, neither accusation
seems justified, or at least each remains unproved.

The reasons why the hypokalemia hypothesis is
poorly supported are as follows: Firstly, hypoka-
lemia after diuretics does not imply potassium de-
pletion, as is commonly stated (6,7). A review of
the many studies of changes in total body potassium
after thiazides indicates that, with the exception of
only a few contrary reports, the total body losses
of potassium are in the range of 5 to 7%, far below
the level that could be considered biologically im-
portant. The hypokalemia represents in part a
movement of potassium from the extracellular fluid

into cells secondary to the volume depletion. There-
fore, there is little change in intracellular concen-
tration despite the extracellular hypokalemia.

Secondly, the sensitivity of the heart to depolar-
ization depends in part on the ratio of the concen-
tration of potassium inside the cells to the concen-
tration outside (6,7). An increase in this ratio raises
the threshold to depolarization, that is, it reduces
irritability. Therefore, thiazide-induced hypoka-
lemia should make the heart more resistant to ar-
rhythmias, since the ratio will be increased by the
decrease in extracellular potassium concentration
with little change in the intracellular concentration.

Thirdly, recent studies with 24- or 48-h moni-
toring of the electrocardiogram refute previous
work claiming a relation between hypokalemia
caused by thiazides and the incidence of ar-
rhythmia. Papademetriou, in one laboratory, has re-
cently found that induction of hypokalemia failed to
increase the frequency of arrhythmic activity as
compared with the pretreatment state (8). Also, cor-
rection of the hypokalemia with potassium supple-
ments and/or triamterene did not reduce the fre-
quency of arrhythmias (9). Lief et al. also investi-
gated this question using 48-h monitoring before
and after inducing hypokalemia (10). They also
found no increase in at-rhythmic activity during the
period of hypokalemia.

The principal support for the concept that hypo-
kalemia secondary to thiazides increases ar-
rhythmic activity comes from Holland et al. (1 I),
who found an increase in arrhythmias in 7 of 21
patients subjected to 24-h monitoring. The validity
of this conclusion is questionable, however, be-
cause of the experimental design. Holland accepted
only patients who showed fewer than six ventricular
premature beats per hour. However, there is con-
siderable spontaneous variability in the frequency
of ectopic activity from day to day. When Holland
selected only those with minimal arrhythmic ac-
tivity in a single 24-h monitoring he increased the
chances of observing greater activity on the second
or post-treatment monitoring simply on the basis of
spontaneous fluctuation or deviation toward the
mean.

Fourthly, the Multiple Risk Factor Trial (MRFT)
is often referred to as providing evidence that thia-
zide-induced hypokalemia may be responsible for
the increased incidence of sudden death found in a
subgroup of the trial (12). This evidence also is not
very convincing. First, the correlation was found
on hindsight by examining multiple subsets after the
study was completed. This approach carries a high
risk of finding one or two positive correlations
simply on the basis of chance. Second, and more
importantly, they found no correlation between the
presence of hypokalemia and sudden death. Data
from the Medical Research Council trial of Great
Britain also failed to show any correlation between

J Cardiovasc Pharmacol, Vol. 7 (Suppl. I), 1985


s114                           E. D. FREIS

hypokalemia induced by thiazides and the incidence
of arrhythmias (13).

It would appear from the above brief review that
the concept of thiazide-induced hypokalemia as a
cause of ventricular fibrillation and sudden death is
poorly supported. This conclusion, however, ap-
plies only to hypertensive patients without overt
heart disease. Thiazides do increase digitalis-in-
duced arrhythmias, and their safety in the treatment
of patients with overt heart disease, where there
may be losses of intracellular potassium, has not
been either proved or disproved by well controlled
trials (67).

The second objection to the use of thiazide di-
uretics has concerned their effects on serum cho-
lesterol. During the early months of treatment with
thiazides serum cholesterol rises to a modest degree
(14,15). Even though the rise is relatively small, it
might increase the risk of coronary artery athero-
sclerosis if the increase persisted over many years.
Nevertheless, other evidence indicates that the rise
reverts back to pretreatment levels over the long
term. When serum cholesterol concentrations are
measured after one or two years of thiazide treat-
ment in large drug-intervention trials the cholesterol
level is unchanged from the pretreatment control
(16,17).

Since the large-scale preventive trials often added
other antihypertensive drugs, possibly the addition
of these agents might have moderated the choles-
terol-raising effects of the thiazides. For example,
prazosin (18), reserpine (19), and hydralazine (20)
have all been reported to prevent the rise in serum
cholesterol concentrations induced by thiazides.
However, Alcazar et al. (21), who employed hydro-
chlorothiazide and amiloride without other drugs,
found a rise of serum cholesterol at 1-3 months of
treatment, which returned to baseline levels at 6
months to 2 years of follow-up. The Veterans Ad-
ministration study of hydrochlorothiazide versus
propranolol (8) used no other antihypertensive
agents and no dietary interventions in 343 patients
treated with hydrochlorothiazide alone. During the
first 3 months of treatment serum cholesterol was
raised but at 12 months it had fallen to slightly
below the baseline value. Further evidence, there-
fore, indicates that the rise in serum cholesterol
concentration is short lived and, therefore, could not
be an important factor influencing mortality from
coronary heart disease in hypertensive patients.

There is a trend at present to prescribe quite
small doses of diuretics such as 25 mg or even 12.5
mg hydrochlorothiazide once daily. The principal
reasons- to avoid hypokalemia and a rise in
plasma cholesterol concentration-are not well jus-
tified, as indicated above. It is also doubtful
whether such doses are effective in most patients.
This question was examined in the Veterans Ad-
ministration Study of propranolol versus hydro-

J Cardiovasc Pharmacol, Vol. 7 (SuppI. I), 1985

chlorothiazide (2). Doses of hydrochlorothiazide
alone were titrated from 25 mg twice daily, the time
interval between dosage increments being about 2
months. Of the patients whose diastolic blood pres-
sures were reduced below 90 mm Hg, about half
responded to the dose of 25 mg twice daily, 30%
required 50 mg twice daily, and the remaining 20%
needed 100 mg twice daily. If, as the evidence sug-
gests, diuretics lower blood pressure by volume re-
duction, it is difficult to see how the very small doses
employed by some physicians can be effective. As
with other antihypertensive drugs, titration of doses
with thiazides should be based on blood-pressure
response versus valid side effects rather than on
prevention of hypokalemia, rise of the plasma cho-
lesterol concentration, or other unsubstantiated
claims of toxicity

OTHER DRUG TREATMENTS

Thiazides and beta-blockers are not the only
drugs which should be considered for primary care.
Several newer agents deserve consideration, The
calcium channel blockers may become candidates
but there is still insufficient experience with them
to make such a judgment at present. Minoxidil
given with a diuretic is probably the most effective
treatment for severe hypertension with renal
failure.

Captopril

The angiotensin-converting enzyme inhibitors
probably represent the leading new candidates for
primary or step 2 treatment. It is appropriate, there-
fore, to review the results of the recent Veterans
Administration trial of captopril with and without a
diuretic (22,23). Doses of captopril alone of 12.5,
25, and 50 mg three times daily or 37.5 mg twice
daily or placebo were randomly assigned, double
blind to 475 patients with pretreatment diastolic   '

blood pressures of 92-109 mm Hg. After 7 weeks
hydrochlorothiazide, 25 mg twice daily, was added
in two-thirds of the patients and in all of the placebo
group. Treatment was then continued for 8 to 10
months.

Small doses appeared to be as effective as the
large doses after 7 weeks of treatment, although
there was some waning of the response to the
lowest (12.5 mg) dose after 8 months of treatment.
Reductions in diastolic blood pressure averaged 8
to 17 mm Hg with the various doses. Addition of
hydrochlorothiazide considerably enhanced the an-
tihypertensive effect, with diastolic reductions av-
eraging 16 to 19 mm Hg below pretreatment levels
with the various doses of captopril. These reduc-
tions were significantly greater than with hydro-
chlorothiazide alone. With this combined therapy
there were no significant differences between any
doses of captopril, including the 12.5 mg dose.

There were no life-threatening toxic effects. Re-


HYPERTENSION: INITIAL TREATMENT

s115

versible proteinuria appeared in two patients taking
the active drug and in one taking placebo. Revers-
ible loss of taste occurred in two patients. A reduc-
tion of white blood count from a pretreatment value
of 3.6-2.6 x 1,000/mm3 after captopril was found
in one alcoholic patient and was probably not drug
related. A few patients developed rashes or urti-
caria. Overall, however, the drug was extremely
well tolerated. The usual complaints that may occur
with other antihypertensive drugs, such as loss of
energy, weakness, or impotence, were rarely noted.
In fact, many patients experienced a sense of well
being.

The favorable subjective profile makes captopril
a promising candidate for primary treatment. In ad-
dition, the drug does not appear to be toxic when
given in small doses to patients with mild to mod-
erate hypertension and it is effective in lowering
blood pressure in these doses, particularly when
combined with a thiazide diuretic.

Thiazide plus reserpine

Another effective treatment for hypertension is
the combination of a small dose of the long-acting
drug reserpine with a long-acting diuretic. This
choice can be considered when cost is a major ob-
stacle to effective treatment, such as in countries
where funds for medical care are severely re-
stricted. Moreover, because the patients in such
countries are often poorly educated, compliance
may be a major problem and once-daily dosage be-
comes particularly important. A long-acting diuretic
should be tried first such as bendroflumethiazide or
chlorthalidone which can be given once daily. If this
is ineffective a combination tablet of a long-acting
diuretic with reserpine can then be substituted.

Reserpine has considerable antihypertensive ef-
fect when given with a diuretic. The principal ob-
jection to its use is the possibility of subjective com-
plaints. Nevertheless, the incidence of side effects
is less with lower doses. In the recent Veterans Ad-
ministration study of reserpine (24) it was shown
that doses as low as 0.125 mg per day were as ef-
fective as the standard dose of 0.25 mg per day.
Thus, a combination tablet of 0.1 or 0.125 mg re-
serpine plus 25-50 mg chlorthalidone or 5-10 mg
bendroflumethiazide should provide effective treat-
ment when cost is a major consideration.

CONCLUSION

Diuretics have not been replaced in the step 1
position but their range of application has been
somewhat narrowed. This is not because of hypo-
kalemia or hypercholesterolemia but because of the
possibly greater effectiveness of some beta-
blockers in white patients. Beta-blockers also are
preferred as primary treatment in patients with
prior myocardial infarcts and in patients with rapid
heart rates. Diuretics are the drug of first choice in

black patients, in all hypertensive patients with a
history of congestive heart failure, and in patients
with renal failure where loop diuretics may be in-
dicated. They are also preferred over other drugs
for use in combination with captopril, reserpine, or
any step 2 drug. Captopril provides an effective and
well tolerated alternative primary treatment of hy-
pertension.

1.

2.

3.

4.


5.

6.

7.


8.

9.

IO.

Il.


12.

13.

14.


15.

16.

17.


18.

19.

REFERENCES

Veterans Administration Cooperative Study Group on An-
tihypertensive Agents. Efficacy of nadolol alone and com-
bined with bendroflumethiazide and hydralazine for sys-
temic hvoertension. Am J Cardiol 1983;52:1230-7.
Veterans'Administration Cooperative Study Group on An-
tihypertensive Agents. Comparison of propranolol and hy-
drochlorothiazide for the initial treatment of hypertension.
I. Results of short-term titration with emphasis on racial
differences in response. JAMA 1982;248:1996-2003.
Veterans Administration Cooperative Study Group on An-
tihypertensive Agents. Comparison of propranolol and hy-
drochlorothiazide for the initial treatment of hypertension.
II. Results of long-term treatment. JAMA 1982;248:
2004-11.

Seedat YK. Trial of atenolol and chlorthalidone for hyper-
tension in black South Africans. Br Med J 1980;281: 1241-3.
Moser M, Lumm J. Comparative effects of pindolol and hy-
drochlorothiazide in black hypertensive patients. Angiology
1981;32:561-6.
Kassirer JP, Harrington JT. Diuretics and potassium metab-
olism: a reassessment of the need, effectiveness and safety
of potassium therapy. Kidney Int 1977; 11:505- 15.
Harrington JT, Isner JM, Kassirer JP. Our national obsession
with potassium. Am J Med 1982;73:155-9.
Panademetriou V. Price MB, Notargiacomo A, Fletcher R,
F&is ED. Effect of thiazide therapy on ventricular ar-
rhythmia in patients with uncomplicated hypertension. C/in
Res 1984,32:337A.
Papademetriou V, Fletcher R, Khatri IM, Freis ED. Di-
uretic-induced hypokalemia in uncomplicated systemic hy-
pertension: effect of plasma potassium correction on cardiac
arrhythmias. Am J Cardiol 1983;52:1017-22.
Lief PD, Belizon I, Mates J, Bank N. Diuretic-induced hy-
pokalemia does not cause ventricular ectopy in uncompli-
cated essential hypertension. Kidney Int 1984;25:203.
Holland OB, Nixon JV, Kuhnert I. Diuretic-induced ven-
tricular ectopic activity. Am J Med 1981;70:762-8.
Multiple Risk Factor Intervention Trial Research Group.
Multiple Risk Factor Intervention Trial. Risk factor changes
and mortality results. JAMA 1982;248: 1465-77.
Medical Research Council Working Party on Mild to Mod-
erate Hypertension. Ventricular extrasystoles during thia-
zide treatment: substudy mild hypertension trial. Br Med J
1983;287:1249-53.
Ames RP, Hill P. Elevation of serum lipids during diuretic
therapy of hypertension. Am J Med 1976;61:748-57.
Grimm RH Jr, Leon AS, Hunninghake DB, Leng K, Hannon
P, Blackburn H. Effects of thiazide diuretics on plasma lipids
and lipoproteins in mildly hypertensive patients. A double
blind controlled trial. Ann Int Med 1981;94:7- 11.
Amery A, Birkenhager C, Bulpitt D, et al. Influence of an-
tihypertensive therapy on serum cholesterol in elderly hy-
pertensive patients. Acta Cardiol 1982;27:235-44.
Kannel WH, Gordon T, McGee D. Diuretics and serum cho-
lesterol. Lancer 1977:i:l362-3.

Kirkendall WM, Hammond JJ, Thomas JC, Overturf ML,
Zama A. Prazosin and clonidine for moderately severe hy-
pertension. JAMA 1978;240:2553-6.
Burris JF, Freis ED, Rose W, Thomas JR. Reserpine coun-

J Cardiovasc Pharmacol, Vol. 7 (Suppl. 1), 1985


S116                          E. D. FREIS

teracts thiazide effects on serum cholesterol. Clin Res
1983;31:327A.

20. Burris JF, Kukich S, Yao LLY, Freis ED. Favorable effects
of reserpine and hydralazine on serum lipids. C/in Res
  1983;31:330A.

21. Alcazar J, Ruilope L, Ladron de Guevara P, Navarro A.
Sanz J, Borrientos A, Rodicio J. Interrelationship between
uric acid, cholesterol and triglycerides in essential hyperten-
sion. (Meeting Abstract No. 8) Proceedings of the Ninth
Meeting of the International Society for Hypertension,
Mexico City, 1982.

22. Veterans Administration Cooperative Study Group on An-

tihypertensive Agents. Captopril: evaluation of low doses,
twice daily doses and the addition of diuretic for the treat-
ment of mild to moderate hypertension. C/in Sci
1982;63:443-5s.

23. Veterans Administration Cooperative Study Group on An-
tihypertensive Agents. Low-dose captopril for the treatment
of mild to moderate hypertension. Hypertension
  1983;5(suppl 3): 139-44.

24. Veterans Administration Cooperative Study Group on An-
tihypertensive Agents. Low doses v standard dose of reser-
pine: a randomized, double-blind, multiclinic trial in patients
taking chlorthalidone. JAMA 1982;248:2471-7.

J Cardiovasc Pharmacol, Vol. 7 (Suppl. I), 1985