Correspondence
Angiotensin-Converting–Enzyme Inhibitors and Diuretics for Hypertension
N Engl J Med 2003; 349:90-93July 3, 2003
- Article
To the Editor:
The conflicting recommendations of the Second Australian National Blood Pressure Study (ANBP2) reported by Wing et al. (Feb. 13 issue)1 and the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)2 concerning the superiority of diuretics or angiotensin-converting–enzyme (ACE) inhibitors can be reconciled. In ALLHAT, the rates of the cardiovascular disease–related end points were higher in the ACE-inhibitor group than in the diuretic group because there were more hospitalizations for heart failure and more strokes. In ANBP2, there were no significant differences in either of these end points. The higher rate of stroke in the ACE-inhibitor group in ALLHAT may be explained by the higher mean blood pressure in that group than in the diuretic group. In ANBP2, there were no differences in blood pressure between the groups.
The excess cases of heart failure in the ACE-inhibitor group in ALLHAT occurred during the first year, which may have resulted from the discontinuation of diuretic therapy; 90 percent of the participants had been taking antihypertensive medications before enrollment. In ANBP2, only 62 percent of patients had been treated previously, and the rate of heart failure was marginally lower in the ACE-inhibitor group. Since ALLHAT provided no evidence of a reduction in mortality due to cardiovascular causes in the diuretic group, the conclusion that diuretics prevent heart failure may be unsound.
2 ReferencesThomas G. Pickering, M.D., D.Phil.
Mount Sinai Medical Center, New York, NY 10029-6574
thomas.pickering@msnyuhealth. org 1
Wing LMH ,Reid CM ,Ryan P , et al. A comparison of outcomes with angiotensin-converting-enzyme inhibitors and diuretics for hypertension in the elderly. N Engl J Med 2003;348:583-592
Full Text | Web of Science | Medline2
The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group
CrossRef | Web of Science
To the Editor:
We currently look to trials involving thousands of subjects, believing that larger numbers will reveal greater truths. But that is not guaranteed, as is clearly shown by the conflicting outcomes in ALLHAT1,2 and ANBP2,3 each of which compared an ACE inhibitor with a thiazide diuretic in patients with hypertension; this conflict was addressed in the editorial by Frohlich.4 Large numbers can almost guarantee statistical significance for negligible differences, and we “power” studies to account for this.4 Wing et al. report rates of a combined end point (itself a questionable maneuver) of 56.1 per 1000 patient-years in one group and 59.8 per 1000 patient-years in the other group. This difference of 3.7 per 1000 patient-years is all of 0.37 per 100 patients per year. The P value of 0.05 does not make this negligible difference more than it is. Similarly, differences of 2 or 3 percent in a biologic study such as ALLHAT have questionable meaning. So long as we allow P values to hypnotize us while we disregard the magnitude of differences,5 we will continue to overvalue the evidence on which we base our practice and will compound our confusion.
5 ReferencesLouis H. Krut, M.D.
St. Mary's Health Center, St. Louis, MO 63117
lkrut65258@aol.com 1
The ALLHAT Officers and Coordinators of the ALLHAT Collaborative Research Group
CrossRef | Web of Science2
Appel LJ . The verdict from ALLHAT -- thiazide diuretics are the preferred initial therapy for hypertension. JAMA 2002;288:3039-3042
CrossRef | Web of Science | Medline3
Wing LMH ,Reid CM ,Ryan P , et al. A comparison of outcomes with angiotensin-converting-enzyme inhibitors and diuretics for hypertension in the elderly. N Engl J Med 2003;348:583-592
Full Text | Web of Science | Medline4
Frohlich ED . Treating hypertension -- what are we to believe? N Engl J Med 2003;348:639-641
Full Text | Web of Science | Medline5
Krut LH . On the statins, correcting plasma lipid levels, and preventing the clinical sequelae of atherosclerotic coronary heart disease. Am J Cardiol 1998;81:1045-1046
CrossRef | Web of Science | Medline
To the Editor:
The choice of diuretic may explain the apparent superiority of ACE inhibitors in ANBP2. Available data suggest that hydrochlorothiazide — the diuretic used in ANBP2 — may be notably different from chlorthalidone. In the Multiple Risk Factor Intervention Trial, mortality due to coronary heart disease was lower with chlorthalidone than with hydrochlorothiazide.1 The mortality rate decreased when patients who had been receiving hydrochlorothiazide were switched to chlorthalidone. The majority of other studies with results supporting the use of thiazides have used chlorthalidone, including ALLHAT, the Hypertension Detection and Follow-up Program Study, the Systolic Hypertension in the Elderly Program trial, and the Treatment of Mild Hypertension Study.
Hydrochlorothiazide has an antihypertensive effect that dissipates after eight hours.2 A single 50-mg dose of chlorthalidone was found to control blood pressure for 24 hours at least as well as 50 mg of hydrochlorothiazide taken twice daily.3 Twenty-four–hour blood-pressure control and maintenance of a nocturnal “dip” may be necessary for optimal reduction of the risk of cardiovascular events, and such control may be attainable only with chlorthalidone.4 Although there has been no large randomized study directly comparing hydrochlorothiazide with chlorthalidone, these observations may lend insight into the ANBP2 trial, which did not show a benefit of hydrochlorothiazide-based diuretic therapy.
4 ReferencesAlicia M. Reese, Pharm.D.
Robert L. Talbert, Pharm.D.
Henry I. Bussey, Pharm.D.
University of Texas at Austin College of Pharmacy, San Antonio, TX 78229
reesea@uthscsa.edu 1
Mortality after 10 1/2 years for hypertensive participants in the Multiple Risk Factor Intervention Trial. Circulation 1990;82:1616-1628
CrossRef | Web of Science | Medline2
Pool PE ,Applegate WB ,Woehler T ,Sandall P ,Cady WJ . A randomized, controlled trial comparing diltiazem, hydrochlorothiazide, and their combination in the therapy of essential hypertension. Pharmacotherapy 1993;13:487-493
Web of Science | Medline3
Finnerty FA Jr . A double-blind study of chlorthalidone and hydrochlorothiazide in an outpatient population of moderate hypertensives. Angiology 1976;27:738-744
CrossRef | Web of Science | Medline4
Svensson P ,de Faire U ,Sleight P ,Yusuf S ,Ostergren J . Comparative effects of ramipril on ambulatory and office blood pressures: a HOPE substudy. Hypertension 2001;38:E28-E32
CrossRef | Web of Science | Medline
To the Editor:
ANBP2 and ALLHAT had conflicting results. However, three factors may have minimized the beneficial effect of ACE inhibitors in ALLHAT. First, the high or medium-level salt intake by patients in North America could have blunted most of the beneficial effects of ACE inhibitors in ALLHAT. Indeed, a high salt intake decreases the ability of ACE inhibitors to lower blood pressure and reduce proteinuria.1
Second, ALLHAT used a higher dose of diuretic than ANBP2, since chlorthalidone has a longer half-life than hydrochlorothiazide and a dose of up to 25 mg per day was given. This difference in dose may explain the better blood-pressure control with diuretics in ALLHAT, especially in the context of high salt intake.
Third, patients with more severe hypertension, who have a strong effect on the rates of end points, received more than one drug in both studies. The main second-line antihypertensive drugs in ALLHAT were beta-blockers, which are less synergistic with ACE inhibitors than with diuretics. Conversely, calcium-channel blockers were the main second-line agents in ANBP2, and they are very synergistic with ACE inhibitors. We therefore suggest that ACE inhibitors were used in a more favorable context in ANBP2 than in ALLHAT.
1 ReferencesVincent L.M. Esnault, M.D., Ph.D.
Nantes University Hospital, 44093 Nantes, France
vesnault@nantes.inserm. fr 1
Buter H ,Hemmelder MH ,Navis G ,de Jong PE ,de Zeeuw D . The blunting of the antiproteinuric efficacy of ACE inhibition by high sodium intake can be restored by hydrochlorothiazide. Nephrol Dial Transplant 1998;13:1682-1685
CrossRef | Web of Science | Medline
To the Editor:
Wing et al. report that the initiation of antihypertensive treatment with ACE inhibitors in older persons, particularly men, decreases the risk of cardiovascular events as compared with diuretic treatment. Frohlich weighs the results of ANBP2 and ALLHAT, discussing differences such as the populations of patients and the specific drugs used. We, who are among the authors of ALLHAT, find additional substantial differences.
There were more than 4 times as many cardiovascular events in ALLHAT as in ANBP2 (10 times as many cases of heart failure), and ALLHAT was a double-blind study. ANBP2 used a prospective, randomized, open-label, blinded–end-point design, raising the potential for bias in the reporting of events (the rates of some outcomes might have been “expected” to be lower with the ACE inhibitor).
Even given the possible biases, however, the results of ANBP2 are consistent with those of ALLHAT if one compares the upper confidence limit for the relative risks in ANBP2 with the estimates of relative risk in ALLHAT (Table 1Table 1
Blood Pressure and End Points in ANBP2 and ALLHAT.). In ALLHAT, the ACE inhibitor provided no advantage for any outcomes in white men or women, and the rates of heart failure in the ACE-inhibitor group were higher than those in the diuretic group. The Blood Pressure Lowering Treatment Trialists' Collaboration plans to perform aggregate analyses1; however, we believe that the totality of the evidence favors the diuretic.1 ReferencesBarry R. Davis, M.D., Ph.D.
University of Texas Health Science Center at Houston, Houston, TX 77030
davis@sph.uth. tmc. edu Jackson T. Wright, Jr., M.D., Ph.D.
University Hospitals of Cleveland, Cleveland, OH 44106Jeffrey A. Cutler, M.D., M.P.H.
National Heart, Lung, and Blood Institute, Bethesda, MD 208921
World Health Organization-International Society of Hypertension Blood Pressure Lowering Treatment Trialists' Collaboration
Web of Science | Medline
Author/Editor Response
The major concerns of the correspondents are the apparently conflicting results between our study (ANBP2) and ALLHAT. The studies differ considerably with regard to setting, population, selection of patients, entry criteria, study protocol, treatment regimens used, blood pressures achieved in each of the treatment groups, and primary outcome measures. Because of these substantial differences, we consider the trials essentially complementary rather than conflicting.
The differences between the study populations are particularly striking, reflecting the different ethnic mixes in the countries where the studies were performed. In ALLHAT, 32 percent of the patients were non-Hispanic blacks, 16 percent were Hispanics, and 47 percent were non-Hispanic whites, whereas in ANBP2, almost the entire study population was white (95 percent). As observed in ALLHAT, blacks have a less satisfactory blood-pressure response to ACE inhibitors than to other antihypertensive agents, particularly diuretics,1 which could have a significant effect on the outcome. ALLHAT also involved a higher-risk population of patients with hypertension than that in ANBP2; the patients in the latter study were essentially healthy older persons with hypertension.
The different entry protocols in the two studies may also have had an effect on the results. In ANBP2, treatment was stopped under medical supervision in all previously treated patients before screening for the study. Patients did not proceed into the study if any adverse response to the discontinuation of treatment was observed. In contrast, in ALLHAT, previously treated patients “continued any prior antihypertensive medications until they received randomized study drug.” As pointed out by Dr. Pickering, this practice may have led to the sudden discontinuation of diuretics in enough patients to contribute to the higher early prevalence of heart failure in the ACE-inhibitor group.
In contrast to ALLHAT, ANBP2 was a study of classes of drugs rather than of specific drugs in each class; nine different ACE inhibitors and six different thiazide diuretics were used. It is therefore not appropriate to suggest, as Reese et al. do, that “the choice of diuretics may explain the apparent superiority of ACE inhibitors” in ANBP2.
Both studies emphasize the importance of blood-pressure control in relation to outcome and the requirement in many patients for more than one agent in order to achieve this control. We agree with Dr. Frohlich that a combination of diuretics and ACE inhibitors might well be the most appropriate choice for the achievement of this goal in many patients, although such a combination was not specifically studied in either trial.
1 ReferencesLindon M.H. Wing, M.B., B.S.
Flinders University, Adelaide 5001, AustraliaChristopher M. Reid, Ph.D.
Garry L.R. Jennings, M.D.
Baker Heart Research Institute, Melbourne, VIC 8008, Australia
chris.reid@baker. edu. au for the ANBP2 Management Committee
1
Cushman WC ,Reda DJ ,Perry HM ,Williams D ,Abdellatif M ,Materson BJ . Regional and racial differences in response to antihypertensive medication use in a randomized controlled trial of men with hypertension in the United States. Arch Intern Med 2000;160:825-831
CrossRef | Web of Science | Medline
Author/Editor Response
In the space allotted for my editorial, my primary focus was ANBP2. Secondarily, I recognized the necessity for comparing ANBP2 with ALLHAT, since this comparison would be a major concern of readers. Deliberately excluded were comparisons of secondary outcomes between the subgroups in the two studies, since their complete descriptions and data were unavailable. Neither report detailed diagnostic criteria for cardiac failure or its outpatient or inpatient treatment. Table 5 in the ALLHAT report demonstrated no significant difference between the number of hospitalized patients treated with chlorthalidone and the number treated with lisinopril (P=0.11).
“Glaring differences,” in my analysis, were the racial makeup of the two study populations and the lack of equivalent pressure control between the diuretic group and the ACE-inhibitor group in ALLHAT but not in ANBP2. Indeed, the well-recognized responsiveness to diuretics in blacks, who accounted for 35 percent of patients in ALLHAT (whereas the population in ANBP2 was 95 percent white), was an important issue that could account for differences in blood-pressure reduction and the rate of stroke. Similar reasoning led me to exclude discussion of elevated glucose concentrations that developed in patients in ALLHAT who were treated with diuretics. I believe that the closing sentence of the letter from Davis et al. speaks for most readers in advocating the recommendation of initial diuretic therapy, and I look forward to follow-up publications from the ALLHAT investigators.
Edward D. Frohlich, M.D.
Ochsner Clinic Foundation, New Orleans, LA 70121
