Treatment of Hypertension: Does Choice of Drug Therapy Make a Difference?

Abstract & Commentary

Synopsis: Doxazosin was less effective in preventing major CV events, particularly CHF, than the individuals who received the diuretic chlorthalidone.

Source: Davis BR, et al. Ann Int Med. 2002;137: 313-320.

Randomized clinical trials in hypertension over the past decade have suggested that specific drug classes used for the treatment of hypertension may make a difference in clinical outcomes, irrespective of the efficacy of blood pressure lowering. All available active hypertension agents will lower blood pressure significantly in two thirds of individuals who receive monotherapy. However, the majority of hypertensive patients require 2 or more drugs, particularly diabetics. Thus, it has been difficult to know if choice of a specific agent makes a difference. Meta-analyses have suggested that differences in outcomes are related to drug class, with ACE inhibitors and beta-blockers associated with decreases in coronary events, while calcium channel blockers appear to lower stroke risk. Recently, the NHLBI ALLHAT study (Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial) issued an alert, discontinuing the alpha blocker doxazosin (dox) from the active drug regimens because of an increase in "major cardiovascular events." The present report is an analysis of the data relating to that decision, and confirms the original policy of the steering committee to halt the dox arm of ALLHAT (January 2000). The initial data demonstrated an increase in major cardiovascular disease, particularly congestive heart failure (CHF), with a 25-100% increased risk compared to the diuretic chlorthalidone (chlor) alone. Furthermore, it was concluded that it was unlikely that dox could ever show a benefit over chlor for the primary study end point of nonfatal CAD or nonfatal MI. The ALLHAT Trial enrolled 42,418 patients who were randomized to chlor vs. amlodipine, lisinopril, or doxazosin. Secondary end points included other major CAD events and stroke. (Another component of ALLHAT as yet to be reported, is assessing the use of a pravastatin vs usual care). Eligible individuals were older than 55 years of age, with a blood pressure of ³ 140/90 mm Hg, and one additional major CAD risk factor (prior MI, ECG LVH, diabetes, smoking, or low HDL cholesterol). Enrollment took place between 1994-1998. Hypertensive patients were added in a step approach with a target of blood pressure (BP) end point of < 140/90 mm Hg. Chlor was given at 12.5 and 25 mg; dox was dosed at 2, 4, and 8 mg depending on the BP response. If optimal BP control was not achievable with one agent, other drugs could be added, including reserpine, clonidine, atenolol, or hydralazine, at the discretion of Davis and colleagues.


After 3.3 years of follow-up, the dox arm was discontinued because of a significant increase of CHF events vs. chlor. Approximately two thirds of subjects in both cohorts required additional antihypertensive therapy over the study drug. Individuals in the dox arm did worse whether they received a second or third hypertensive agent. There was a 2.5 to 3-fold increased risk with dox monotherapy, somewhat less in those who took an open-label additional drug. The increase in CHF with dox did not appear to be related to BP or blood pressure control. Patients who received chlor at either dose had less CHF. Dox patients had a slightly greater systolic BP on treatment than chlor by 2-3 mm Hg, with no differences in diastolic BP. Subjects with a lower BP on dox actually had a higher risk of CHF. The study concluded that dox was less effective in preventing major CV events, particularly CHF, than the individuals who received the diuretic chlorthalidone. No conclusion could be made as to whether the diuretic was a more potent active control than dox, or there was harm induced by dox. Dox patients were one third more likely than chlor to receive additional antihypertensive therapy. Time to development of heart failure was shorter with dox than chlor. Davis et al conclude that dox causes excessive risk for heart failure in high-risk hypertensive individuals. They state that "the degree of difference in blood pressure does not explain the increased risk for heart failure in the dox group compared to chlor." They emphasize that chlor resulted in a 50% decrease of CHF in patients with isolated diastolic hypertension in the SHEP trial (Kostis JB, et al. JAMA. 1997;278:212-216).

Comment by Jonathan Abrams, MD

This study, along with the HOPE, LIFE, and PROGRESS data, suggests that the type or class of antihypertensive agent is important, and that small differences in blood pressure achieved in a trial do not seem to account for differences in clinical outcomes. For instance, in the HOPE trial, where many of the patients did not have hypertension, it was concluded that the stroke reduction achieved with the ACE inhibitor ramipril vs. placebo were unrelated to the small blood pressure decrement in the active treatment arm. In the recently published PROGRESS trial, a post-CVA-TIA trial using an ACE inhibitor and a diuretic, combination therapy resulted in a greater decrease in subsequent brain events, and it appeared that the addition of the diuretic to the ACE inhibitor was critical for a beneficial clinical outcome. Greater BP reduction was associated with a greater reduction in stroke; however, stroke risk was decreased even in nonhypertensive patients, particularly with the combination, whereas single drug with perindopril therapy was not associated with a decrease in stroke. Thus, this study suggests that even in the absence of hypertension, the combination of an ACE inhibitor and a diuretic, which was associated with BP lowering in normotensives, results in maximal benefit for recurrent stroke. In the recently published LIFE trial (Abrams J. Clinical Cardiology Alert. 2002;21[6]:46-47), losartan was found to decrease stroke rates more than atenolol, in spite of identical blood pressure lowering with both agents. These individuals were high risk, as they had ECG and/or echo LVH. Furthermore, in the diabetic substudy of LIFE, there was broader risk reduction, including coronary artery events, which were not reduced in the overall LIFE trial.

What can be made from these recent trials, as well as other data in the literature? It would appear that in high-risk hypertensive individuals, particularly diabetics, an ACE inhibitor should be part of the drug regimen. In addition, the use of diuretics, long confirmed to be associated with decreased event reduction in hypertension alone or in combination with a beta-blocker, should be standard antihypertensive therapy. Alpha blockers have no role in the sole treatment of hypertension, but can be used in individuals with prostatic hypertrophy in combination with other antihypertensives. Calcium channel blockers have been shown to be effective in systolic hypertension of the elderly, and in hypertensive therapy meta-analyses decreased stroke rates, but showed some increase in myocardial infarction. Thus, these agents should not be used as first- line therapy in individuals without known CAD and isolated systolic hypertension. Beta-blockers are effective in all individuals, but not necessarily as first-line therapy. There remains some controversy about the use and efficacy of beta-blockers in the elderly, but recent commentaries support such a policy. Nonetheless, the LIFE trial results suggest the use of an ARB over a beta-blocker. Other trials are needed to confirm this since beta-blockers do prevent coronary artery events. In individuals with any degree of renal insufficiency, an ACE or an ARB should be used as initial therapy.

Finally, most of the data that have accumulated over the past decade indicates that in treating hypertension, lower is better. Target BP should be no higher than 145/85 mm Hg, and in individuals with multiple risk factors or diabetes, optimal blood pressure control of < 135/80 mm Hg is the name of the game.

Dr. Abrams is Professor of Medicine Division of Cardiology University of New Mexico, Albuquerque.