is low-dose hydrochlorothiazide...

Is Low-Dose Hydrochlorothiazide Effective?P. F. ADRIAN MAGEE AND EDWARD D. FREIS

SUMMARY In a double-blind crossover study, 13 patients with pretreatment diastolic blood pres-sure between 95 and 109 mm Hg received nadolol, 80 mg/day, plus placebo of hydrochlorothiazide andnadolol, plus three different doses of active hydrochlorothiazide. Patients remained on each activeregimen for 3 weeks, with an intervening placebo period of 2 to 4 weeks. With 12.5 mg of hydrochloro-thiazide daily plus nadolol, there was no greater reduction of blood pressure than with nadolol alone.A dose of 25 mg of hydrochlorothiazide was associated with a significantly greater decrease in systolicbut not diastolic pressure, as compared with nadolol alone. A significantly greater reduction in bothsystolic and diastolic blood pressure was obtained only with the 50 mg/day dose of hydrochlorothia-zide. Extension to 6 weeks of treatment with 12.5 mg/day failed to lower the blood pressure more thanthe level seen at 3 weeks. These results suggest that in combination with nadolol, 12.5 mg of hydro-chlorothiazide per day has no significant antihypertensive effect. There was no evidence of a flat dose-response curve in the daily dose range of 12.5 to 50 mg. For most patients, a dose of 50 mg ofhydrochlorothiazide was required to lower both systolic and diastolic blood pressure significantlybelow the level obtained with nadolol alone. (Hypertension 8 [Suppl II]: II-135-II-139, 1986)

KEY WORDS • diuretics • hypertension • dosage

A LTHOUGH diuretics remain the cornerstone of/ \ antihypertensive therapy, they have come un-

A. \ - der increased attack in recent years, primarilybecause of hypokalemia and the fear that this sideeffect may contribute to myocardial infarction andsudden death. Two approaches have been utilized tominimize these assumed risks. One is to treat the hypo-kalemia with potassium supplements and potassium-sparing diuretics. The other is to reduce the dose of thediuretic to a level at which biochemical side effectswill be minimal.

Reduced doses of diuretics, particularly chlorthali-done,1-2 have been suggested by several authors whofavor the addition of a second drug instead of raisingthe dose of the diuretic. With respect to hydrochloro-thiazide, other investigators have claimed that doses assmall as 12.5 mg/day were as effective as 25 or 50mg/day.3 When used in combination with enalapril,doses of hydrochlorothiazide as small as 6.25 mg/daywere reported to be as effective as the more standard,larger doses.4 Other studies, however, indicate that thedose-response curve for hydrochlorothiazide is not flatuntil the dose reaches 50 mg/day or higher.5"7

From the Veterans Administration and Georgetown UniversityMedical Centers, Washington, D.C.

Address for reprints: Edward D. Freis, M.D., Senior MedicalInvestigator, Hypertension Research, VA Medical Center, 50 Ir-ving St. NW, Washington, DC 20422.

Because of the importance of diuretics in the treat-ment of hypertension and the conflicting informationregarding the optimal dosage, we conducted the pres-ent study to reinvestigate the relationship between dos-age and effectiveness in patients with hypertension.

MethodsThe study group consisted of 13 men with uncompli-

cated hypertension whose diastolic blood pressurewithout treatment was in the range of 95 to 109 mmHg. Their mean age ( ± S D ) was 56 ± 1 2 years.Twelve were black, and 1 was white. Fully informedconsent was obtained in every case. If a patient wasreceiving antihypertensive drugs at the time of initialscreening, these drugs were withdrawn. All patientswere given a placebo of nadolol and hydrochlorothia-zide during the first period of the study (pretreatmentplacebo period) and were seen again 2 and 4 weekslater (Figure 1). For a patient to be eligible to enter thenext phase of the trial, the average diastolic bloodpressure during the second and fourth weeks of thepretreatment placebo period had to be in the range of95 to 109 mm Hg. If the average diastolic pressure wasnot in this range, the pretreatment period was extendedfor an additional week, and the average pressure dur-ing the fourth and fifth weeks was used to determineeligibility. The average blood pressure for the second

11-135

by guest on July 14, 2018http://hyper.ahajournals.org/

Dow

nloaded from

11-136 1985 BLOOD PRESSURE COUNCIL SUPPL II HYPERTENSION, VOL 8, No 6, JUNE 1986

FIGURE 1. Flow chart showing the de-sign of the trial. The pretreatment placeboperiod was of 4 to 6 weeks' duration, eachactive drug regimen was followed for 3weeks, and the intervening placebo periodswere of 2 to 4 weeks' duration. The variousregimens were assigned in random order.HCTZ = hydrochlorothiazide.

Active First ' Second THrdPtaceo Run-in Regimen Placebo Crossover Placebo Crossover Placebo Crossover

Clinic Visits 0(Weeks)

6 7 11 12

PlaceboPeriod

Active Treatmentsin Random Order Double-Blind

(!) Nadolol 80mg/HCTZ Placebo(ii) Nadolol 80mg/HCTZ 12.5mg(Hi) Nadolol 80mg/HCTZ 25.0mg(iv) Nadolol 80mg/HCTZ 50.0mg

14 16 17 19 21 22

and fourth or fifth week of the pretreatment placeboperiod was 157/102 ± 19.5/4.4 mm Hg.

Subjects were assigned, in a double-blind fashion,to one of the four following regimens: 1) 50 mg ofhydrochlorothiazide plus 80 mg of nadolol once daily,2) 25 mg of hydrochlorothiazide plus 80 mg of nadololonce daily, 3) 12.5 mg of hydrochlorothiazide plus 80mg of nadolol once daily, or 4) placebo hydrochloro-thiazide plus 80 mg of nadolol once daily. The order ofadministration was determined in such a manner thatall possible sequences of the four regimens could beadministered in a double-blind manner.

Each treatment phase was of 3 weeks' duration (seeFigure 1). Clinic visits were scheduled for the secondand third weeks. The blood pressure measurements forWeeks 2 and 3 of each treatment were averaged sepa-rately. The placebo periods between the treatmentperiods were 2 weeks in duration. If the diastolic pres-sure at the end of the intervening 2-week placebo peri-od had risen to within 4 mm Hg or less of the baselinevalue (before random assignment), the patient enteredthe next treatment phase. If the blood pressure re-mained below this level after 2 weeks, the placeboperiod was extended for an additional 2 weeks, and if itwas still below the acceptable level, the patient waseliminated from the trial.

At each clinic visit the patient was asked whether hehad experienced any unusual symptoms or discomfortduring the interval since the last visit. He was alsoquestioned about specific side effects, including fa-tigue, faintness, weakness, and impotence. After thesubject had rested for at least 10 minutes in a quietroom, the blood pressure was recorded three timeseach in the lying, sitting, and standing positions by thenurse. The median of the three readings recorded ineach position was taken as representing the pressurefor a particular visit. Heart rate and body weight werealso recorded at each visit.

Laboratory studies included plasma potassium, uricacid, creatinine, cholesterol, and bicarbonate measure-ments and urinalysis. These values were determined atthe end of each treatment and each placebo phase.

Conventional statistical methods were used for thecalculations of mean values, standard errors, and stan-

dard deviations. The significance of differences be-tween paired observations was estimated with Stu-dent's r test. Values of p < 0 . 0 5 were consideredsignificant.

ResultsBlood pressure changes after 2 and 3 weeks of treat-

ment are presented in Figures 2 and 3. Systolic anddiastolic changes are shown separately for each treat-ment regimen. After 2 and 3 weeks of treatment with12.5 mg of hydrochlorothiazide plus 80 mg of nadolol,average reductions in blood pressure from the placebobaseline value were 11.3/5.0 and 11.7/8.9 mm Hg,respectively. These changes were not significantly dif-ferent from the reductions recorded after 80 mg ofnadolol plus hydrochlorothiazide placebo (11.1/7.7after 2 weeks and 11.7/10.3 after 3 weeks). The25-mgdose of hydrochlorothiazide also failed to lower dia-stolic blood pressure more than the nadolol control, theaverage reductions being 20.8/8.5 at 2 weeks and19.8/9.8 at 3 weeks. Unlike the 12.5-mg dose of hy-drochlorothiazide, the 25-mg dose did result in a sig-nificant fall in systolic pressure (p<0.05) . With the50-mg dose, the reduction in pressure at 2 and 3 weekswas 24.5/14.8 and 22.9/13.4 mm Hg, respectively.Systolic pressure was significantly reduced at 2 and 3weeks, as compared with the reductions obtained withnadolol alone. Diastolic pressure after the 50-mg doseof hydrochlorothiazide was significantly reduced at 2weeks, and the reduction approached significance at 3weeks.

The time required for diuretics to exert their fullantihypertensive effect is a controversial issue. In thepresent study blood pressure changes were not signifi-cantly different after 2 or 3 weeks of treatment. To ruleout the possibility of a longer term antihypertensiveresponse, a subgroup of seven patients remained on theregimen of 12.5 mg of hydrochlorothiazide plus 80 mgof nadolol for an additional 6 weeks. As shown inFigure 4, there was no further reduction of blood pres-sure from 3 to 6 weeks with this combination.

The pulse rate decreased with all regimens, rangingfrom — 13.4 beats/min with nadolol plus 12.5 mg of

by guest on July 14, 2018http://hyper.ahajournals.org/

Dow

nloaded from

LOW-DOSE HYDROCHLOROTHIAZIDE/Magee and Freis 11-137

Systolic Blood

Pressure Change

mm Hg

-100

-200

J IP<003

P«0JJ5

D

I

*rro>2 WEEKSTOATWNT

] AFTEH 3 WEEKSJTKEATKMT

P'OCOi

Nadolol 80mg Nadolol SOmg Nadolol SOmg Nadolol SOmgAlone + + +

HCTZ 123mg HCTZ 25.0mg H C T Z 5 0 0mg

FIGURE 2. Changes in systolic bloodpressure with nadolol alone and with nado-lol plus different doses of hydrochlorothia-zide (HCTZ), 2 and 3 weeks after treat-ment. The significance of the differencebetiwen nadolol plus placebo and nadololplus the various doses of HCTZ was deter-mined. The systolic change after nadololplus 12.5 mg of HCTZ was not significantlydifferent (NS)from the change after nadololalone. Reductions were significant with the25-mg and 50-mg doses of HCTZ.

Duitolic BloodPressure ChangemmHg

-5 0

-10 0

-15 0

4 IKS

• AFTER 2 WEEKSTREATMENT

• 0 0 2 5Nidolol SOmg Nadolol SOmg Nadolol 80mg Nadolol SOmg

Alone + + +HCTZ12 5mg HCTZ25 0mg HCTZSOOmg

] AFTER 3 WEEKS] TREATMENT

FIGURE 3. Changes in diastolic bloodpressure with nadolol alone and with nado-lol plus different doses of hydrochlorothia-zide (HCTZ). The only significant reduc-tion, as compared with nadolol alone,occurred with the 50-mg dose of HCTZ.

I Systole Chance

Dlastobc Change

Change in -5Blood PressuremmHg

-10

INadofcX BOmg Nadolol SOmg Nadolol SOmg

Alone HCTZ 123 mg HCTZ 12-5 mg3 Weeks 3 Weeks 6 Weeks

FIGURE 4. Changes in systolic and diastolic blood pressurewith nadolol plus placebo and nadolol plus 12.5 mg ofhydro-chlorothiaiide (HCTZ), after 3 weeks and 6 weeks of continuedtreatment. NS = not significant.

hydrochlorothiazide to —7.4 beats/min with nadololplus 50 mg of hydrochlorothiazide.

Small (not significant) decreases in serum potas-sium were seen with all doses of hydrochlorothiazideexcept the 50-mg dose, which was associated with asignificant decrease (0.23 mEq/L). In no case did theserum potassium fall below 3.5 mEq/L. Serum uricacid rose significantly with all diuretic regimens. Theelevations were 0.45, 1.60, and 1.48 mg/dl with 12.5,

25, and 50 mg of hydrochlorothiazide, respectively.The 25-mg and 50-mg regimens were associated withminor elevations in blood urea nitrogen and serumglucose, which were clinically unimportant but statis-tically significant for the 25-mg combination (p <0.01and 0.025, respectively). There were no significantchanges in serum calcium, creatinine, or cholesterol.

Six patients were dropped from the study. In three ofthe six the diastolic pressure failed to return to within 4mm Hg of the established baseline diastolic pressureduring the interim placebo period. Two of the six werenoncompliant, as judged by pill counts (see below),and one suffered a cerebrovascular accident during thestudy.

Side effects, which were infrequent, included impo-tence and nasal stuffiness. No patient was droppedfrom the study because of side effects.

Compliance, monitored by pill counts, was in gen-eral excellent. Patients whose pill counts fell below80% on two consecutive occasions were eliminatedfrom the study.

DiscussionThe small doses of hydrochlorothiazide were tested

in combination with a /3-adrenergic blocking drug,because advocates of small doses of diuretics generally

by guest on July 14, 2018http://hyper.ahajournals.org/

Dow

nloaded from

11-138 1985 BLOOD PRESSURE COUNCIL SUPPL II HYPERTENSION, VOL 8, No 6, JUNE 1986

recommend that they be used as an adjunct to enhancethe activity of another antihypertensive agent. Withboth drugs used together, the chances of controllinghypertension with small doses of the diuretic will bemuch greater than if the diuretic is given alone. Nado-lol was chosen because it has been shown to be highlyeffective, especially in combination with a diuretic.8

Nadolol also offers the convenience of a once-dailydose.

Used in combination with nadolol, the smallestdaily dose of hydrochlorothiazide that significantlylowered diastolic blood pressure, as compared withnadolol alone, was 50 mg. However, both the 25-mgand 50-mg doses of hydrochlorothiazide lowered sys-tolic pressure significantly. Other studies have shownthat the thiazide diuretics have a greater effect on sys-tolic than on diastolic pressure.5-8 This effect may besecondary to the reduction in extracellular volume andcardiac output that occurs with the administration ofdiuretics.9

Small doses of diuretics have been found to be effec-tive by other research groups. One of the first wasBengtsson et al., ' who found that a reduction in chlor-thalidone from 50 mg daily to three times per weekresulted in no change in diastolic blood pressure, al-though systolic pressure was moderately increased.Materson et al.10 found that 25, 50, and 75 mg of chlor-thalidone per day resulted in similar falls in bloodpressure. The reduction was smaller, however, with12.5 mg per day than with the larger doses. Similarly,Tweeddale et al. found no difference in the antihyper-tensive effectiveness of 25, 50, 100, and 200 mg ofchlorthalidone daily.'' Chlorthalidone has a longer du-ration of action than hydrochlorothiazide and is prob-ably more effective, milligram for milligram. Thus,the results obtained with small doses of hydrochloro-thiazide are not comparable to those obtained withsimilar doses of chlorthalidone.

The search for the smallest effective dose of diuret-ics has been motivated by the truism that the smallerthe dose, the fewer the side effects. This rationale isgenerally accepted. The controversy, however, con-cerns the minimal effective dose of hydrochlorothia-zide. Degnbol et al.6 determined the effects of dailydoses of 25, 50, 75, and 100 mg, with each dose givenfor a 6-week period in a crossover design without inter-vening placebo periods. They reported a fall of 11 mmHg in mean blood pressure after the 25-mg dose. Withthe higher doses, the fall in pressure was linearly relat-ed to the dose, but the greatest reduction occurred withthe 25-mg dose.

Three groups of investigators have reported on theeffects of very small doses of hydrochlorothiazide.Berglund and Anderson3 found average blood pressurereductions of 4/1 mm Hg with 12.5 mg daily, 4/2 mmHg with 25 mg, and 5/3 mm Hg with 50 mg. Theseresponses seem unusually small, especially for thehigher doses. For example, in the Veterans Adminis-tration study of ticrynafen versus hydrochlorothiazidea dose of 50 mg of hydrochlorothiazide reduced theaverage blood pressure by 16/10 mm Hg.l2 The differ-

ence was not due to racial effects, because 72% of thepatients were white.

MacGregor et al.'3 measured the effects of smalldoses of hydrochlorothiazide added to acebutolol. Themean fall in blood pressure was about 15% with eachdose of hydrochlorothiazide (12.5, 25, and 50 mgdaily). The reason for the discrepancy between theirresults and ours is not clear. The design of our studydiffered from theirs in that we incorporated an inter-vening placebo period between each active drug phase.This placebo period could be extended until the dia-stolic blood pressure returned to within 4 mm Hg of thebaseline control value; if this condition was not met,the patient was eliminated from the trial. A second andperhaps more important difference is that we incorpo-rated a control of /3-blocker with the placebo dose ofhydrochlorothiazide, whereas MacGregor et al.13 didnot assess the effects of the /3-blocker separately. Athird difference between the two studies is that weutilized the standard auscultatory method for measur-ing blood pressure, MacGregor et al. used a semiauto-matic ultrasonic method (Arteriosonde).

Andren et al.,4 the third group advocating very lowdoses of hydrochlorothiazide, added doses of 6.25,12.5, or 25 mg of the diuretic toenalapril, 10 mgor40mg. They found significant reductions in blood pres-sure in all groups, but there were no significant differ-ences among the groups. This double-blind random-ized study used parallel treatment groups instead of acrossover design. Again, our study differed in that weincorporated a control regimen containing the seconddrug plus placebo hydrochlorothiazide, which couldbe used for comparison with the addition of smalldoses of the diuretic.

Reports of the effectiveness of small doses of diuret-ics have led to the concept that the dose-response curvefor the antihypertensive effects of diuretics is essential-ly flat from very small doses to high doses. Other data,however, indicate a dose-related response even in thehigh-dose range of hydrochlorothiazide. For example,in the Veterans Administration study of propranololalone versus hydrochlorothiazide alone, 50% of thethiazide responders (diastolic pressure < 9 0 mm Hg)had hypertension controlled with the initial dose of 25mg twice daily. However, in 30% hypertension wasnot controlled until the dose was increased to 50 mgtwice daily, and 20% required 100 mg twice daily.Although such high doses are not recommended inroutine practice, the data nevertheless demonstrate thatincreased doses lead to greater antihypertensive effec-tiveness even up to high doses. Another investigatorreporting that the dose-response curve is not flat, evenat moderate doses, is Pederson,7 who found that 25 mgof hydrochlorothiazide twice daily failed to controlblood pressure in 15 of 20 patients with mild hyperten-sion. A dose of 50 mg twice daily lowered the pressurean additional 8/7 mm Hg. In patients with diastolicpressure between 110 and 115 mm Hg a dose of 25 mgof hydrochlorothiazide proved inadequate. A furtherreduction in pressure was obtained when the dose wasincreased to 50 mg daily.14 With doses of 50 mg per

by guest on July 14, 2018http://hyper.ahajournals.org/

Dow

nloaded from

LOW-DOSE HYDROCHLOROTHIAZIDE/Magee and Freis 11-139

day or less, biochemical side effects were minimal inthe above studies.

It would appear from the present study that very lowdoses of diuretics are ineffective in many patients andtherefore cannot be relied on in the effort to avoidbiochemical side effects. The problem is to strike anoptimal balance between antihypertensive effective-ness and side effects. The present results indicate that a12.5-mg dose of hydrochlorothiazide is significantlyless effective than higher doses and is too low to pro-vide adequate antihypertensive activity in most pa-tients, even when combined with a second drug. Nev-ertheless, since responsiveness to diuretics varies overa wide range, it is possible that individual patients willrespond to very small doses. Our data are inconsistentwith the concept that the dose-response curve for bloodpressure versus dose is flat at doses of 12.5 mg andhigher. The present results suggest that for most pa-tients at least 25 mg/day of hydrochlorothiazide is re-quired for the initial dose, which can be increased to 50mg/day if needed.

References1. Bengtsson C, Johnsson G, Sannerstedt R, Werko L. Effect of

different doses of chlorthalidone on blood pressure, serumpotassium and serum urate. Br Med J 1975;1:197—199

2. Thomson AE. Chlorthalidone in the long-term therapy of pa-tients with hypertension. Int Z Klin Pharmakol Ther Toxikol197O;3:21-25

3. Berglund G, Anderson O. Low doses of hydrochlorothiazide inhypertension: antihypertensive and metabolic effects. Eur JClin Pharmacol 1976;10:177-182

4. Andren L, Weiner L, Svensson A, Hannsson L. Enalapril witheither a "very low" or "low" dose of hydrochlorothiazide isequally effective in essential hypertension: a double-blind trialin 100 hypertensive patients. J Hypertension 1983;l(suppl2):384-386

5. Veterans Administration Cooperative Study on Antihyperten-sive Agents. Comparison of propranolol and hydrochlorothia-zide for the initial treatment of hypertension: I. Results ofshort-term titration with emphasis on racial differences in re-sponse. JAMA 1982;248:1996-2003

6. Degnbol B, Dorph S, Marner T. The effect of different diuret-ics on elevated blood pressure and serum potassium. Acta MedScand 1973;193:407-410

7. Pederson OL. Comparison of metoprolol and hydrochlorothia-zide as antihypertensive agents. Eur J Clin Pharmacol 1976;10:381-385

8. Veterans Administration Cooperative Study Group on Antihy-pertensive Agents. Efficacy of nadolol alone and combinedwith bendroflumethiazide and hydralazine for systemic hyper-tension. Am J Med 1983; 1230-1237

9. Wilson IM, Freis ED. Relationship between plasma and extra-cellular fluid volume depletion and the antihypertensive effectof chlorothiazide. Circulation 1959;20:1028-1036

10. Materson BJ, Oster JR, Michael VF, et al. Dose response tochlorthalidone in patients with mild hypertension. Clin Phar-macol Ther 1978;24:192-198

11. Tweeddale MG, Ogilive Rl, Ruedy J. Antihypertensive andbiochemical effects of chlorthalidone. Clin Pharmacol Ther1977;22:519-527

12. Veterans Administration Cooperative Study on Antihyperten-sive Agents. Comparative effects of ticrynafen and hydro-chlorothiazide in the treatment of hypertension. N Engl J Med1979;301:293-297

13. MacGregor GA, Banks RA, Markander ND, Baylis J, Roul-ston J. Lack of effect of beta-blocker on flat dose response tothiazide in hypertension: efficacy of low-dose thiazide com-bined with beta-blocker. Br Med J 1983;286:1535-1538

14. Anderson PD, Anderson HH, Hagman A, Henning R. Potas-sium sparing by amiloride during thiazide therapy in hyperten-sion. Clin Pharmacol Ther 1984;36:197-200 by guest on July 14, 2018

http://hyper.ahajournals.org/D

ownloaded from

P F Magee and E D FreisIs low-dose hydrochlorothiazide effective?

Print ISSN: 0194-911X. Online ISSN: 1524-4563 Copyright © 1986 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Hypertension doi: 10.1161/01.HYP.8.6_Pt_2.II135

1986;8:II135Hypertension. 

http://hyper.ahajournals.org/content/8/6_Pt_2/II135the World Wide Web at:

The online version of this article, along with updated information and services, is located on

  http://hyper.ahajournals.org//subscriptions/

is online at: Hypertension Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document.

Permissions and Rights Question and Answer Further information about this process is available in therequested is located, click Request Permissions in the middle column of the Web page under Services.the Editorial Office. Once the online version of the published article for which permission is being

can be obtained via RightsLink, a service of the Copyright Clearance Center, notHypertensionpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:

by guest on July 14, 2018http://hyper.ahajournals.org/

Dow

nloaded from