All hats off to ALLHAT

The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) constitutes a triumph for diuretics, hard end-points and the value of blood pressure lowering per se. It also constitutes a triumph for the National Heart, Lung and Blood Institute of the National Institutes of Health. Conceived in the early 1990s, when passions ran hot on the relative value of newer blood pressure lowering drugs versus diuretics and beta-blockers, ALLHAT delivers the goods, and provides a clear result with some very clear answers. Similar to all good studies, it also leaves some questions unanswered and poses some new questions. ALLHAT was a randomized, double-blind, prospective comparative study, designed to determine whether the incidence of the primary outcome, fatal coronary heart disease or non-fatal myocardial infarction differed between treatment initiated with a diuretic (chlorthalidone) versus treatment initiated with a calcium antagonist (amlodipine), an angiotensin-converting enzyme (ACE) inhibitor (lisinopril) or an alpha-adrenergic blocker (doxazosin) 1. Over 42 000 hypertensive subjects aged 55 years or older were randomized into the trial between 1994 and 1998 2, and following the discontinuation of the doxazosin treatment arm in some 9000 subjects in January 2000 3, the results in the remaining three arms of the study have now been published 4. The results of a comparison of the effects of pravastatin versus usual care, in a subset of approximately 10 000 subjects with mild-to-moderate elevations of serum cholesterol, have also been published 5. The value of diuretic-based therapy in hypertensive subjects was established well before ALLHAT, in many studies confirming their benefit in reducing stroke, coronary heart disease and heart failure 6–15. Main messages Success of diuretics The outstanding feature of the results of ALLHAT was that across a large number of comparisons for both the primary and secondary outcomes, versus a calcium antagonist, an ACE inhibitor or an alpha-blocker, the diuretic was either equally effective, as was the case for each comparison for the primary outcome, or in some instances superior, as seen with some comparisons for secondary outcomes 3,4. In no comparison did the diuretics come off second best. In large measure, the success of the diuretic appears to reflect its success in lowering blood pressure, and particularly its success in lowering systolic blood pressure, which was superior to that of the ACE inhibitor or the alpha-blocker by 2–3 mmHg and to that of the calcium antagonist by 1 mmHg. Looking back to the early days of diuretic therapy, these drugs were largely seen as good adjunctive therapy in the 1960s, and were used either to potentiate more potent adrenergic neurone blockers or as the first step and foundation for antihypertensive therapy in the 1970s, when ‘step therapy’ was in vogue. Many recent studies, now topped off by ALLHAT, make it clear that one of the major reasons for the benefits seen with diuretics is their potency in blood pressure lowering, and particularly their efficacy in lowering the systolic blood pressure, which has increasingly been seen as the major determinant of cardiovascular risk. This potency helps explain the success of diuretics across age groups, stretching to elderly subjects with essential hypertension or with systolic hypertension 6–8 and now, as seen with the ALLHAT results, to Black and non-Black subjects, to diabetic and non-diabetic subjects, as well as to those above and below 65 years of age 4. Importance of lowering blood pressure per se, especially systolic blood pressure One of the most striking features of ALLHAT is that there was absolutely no difference in the frequency of the primary outcome (fatal coronary heart disease or non-fatal myocardial infarction) nor in all-cause mortality between any of the treatment groups, including the doxazosin group 3,4. The cumulative event curves for the primary outcome, comparing chlorthalidone and each of the other three drugs, could not be closer 3,4. The equivalence of the efficacy of newer drugs with that of the diuretic in preventing the primary outcome, is further confirmation of the value of these newer agents in hypertensive patients. Thus, one clear message is that effective blood pressure lowering is the key and that this transcends the choice of drugs. Another striking feature is that the superiority of the diuretic, chlorthalidone, in relation to some of the secondary outcomes, was most striking for comparisons where it had a clear advantage in lowering the systolic blood pressure (i.e. in comparisons with the ACE inhibitor and the alpha-blocker) 3,4. In each of these comparisons, the superiority of the diuretics was seen with heart failure, with stroke and with combined cardiovascular disease 3,4. The steeper slope for the relationship between blood pressure and the risk of stroke than for that with the risk of coronary heart disease could well have contributed to these differences 10. It seems likely that the difference in systolic blood pressure contributed to the difference observed in the incidence of these secondary outcomes between the diuretic and the ACE inhibitor or the alpha-blocker. It is interesting that in the comparison of chlorthalidone versus the calcium antagonist, amlodipine, the outcomes were almost uniformly equivalent, except for heart failure where chlorthalidone was superior 4. In this instance, the reductions in blood pressure were very much closer, with a 1 mmHg advantage to the diuretic in lowering the systolic pressure offset by a 1 mmHg advantage to the calcium antagonist in lowering the diastolic pressure 4. Importance of hard end-points It is instructive to reflect on the trends in the recommendations of national and international guidelines and of authoritative commentators on the treatment of hypertension over the past 15–20 years. In the late 1980s, there was great concern over the metabolic side-effects of diuretics drugs, including their tendency to cause potassium loss, exacerbate glucose intolerance and cause dislipidaemia. Many were anxious that these side-effects might have contributed to the smaller than expected reductions in coronary heart disease reported in the major trials of blood pressure lowering therapy 9,10. At the same time, favourable effects on lipid profiles, on insulin resistance or on the atherosclerotic process in animal studies, observed with newer agents such as alpha-blockers, ACE inhibitors and calcium antagonists, heightened a tendency to favour the newer drugs over diuretics and/or beta-blockers. This trend was reflected in clinical practice worldwide, even though there were no data available at that time to demonstrate reductions in stroke or heart attack with the newer agents. This was evident with the 1989 and 1993 guidelines published by the World Health Organization – International Society of Hypertension Guidelines Committee 11,12, which recommended that practising clinicians could choose to treat hypertensive patients with any of the main classes of blood pressure lowering drugs because the evidence suggested that the benefit was largely attributable to the blood pressure lowering per se 12. The recommendations of the Joint National Committee (JNC) in the USA were similar until the publication of JNC V in 1993 13 and JNC VI in 1997 14, which came down clearly in favour of diuretics and beta-blockers both because of the lack of hard end-point evidence for the newer drugs, and because of the lower costs of the older agents 13,14. The scepticism regarding ‘softer’ surrogate end-points, and the preference for ‘hard’ end-points in making comparisons between classes of anti-hypertensive drugs, is clearly vindicated by many studies published in the past 6–7 years 15 and now by the results of ALLHAT, which demonstrate that diuretics are at least as effective as newer agents and restore them to a pre-eminent place in the armamentarium for lowering blood pressure 3–4. Generalizability of ALLHAT results within the USA Black Americans made up approximately 35% of the ALLHAT population, and most of the messages hold true for both the Black and the non-Black subjects in the study. The main conclusion of equivalence in the primary end-points holds true for all subgroups reported: those above and below 65 years of age, men and women, Black and non-Black subjects and diabetic and non-diabetic subjects. The main differences, such as the greater benefit in the prevention of heart failure with chlorthalidone compared to amlodipine or to lisinopril, also hold true across all subgroups 4. One exception is the result for stroke obtained in the comparison between chlorthalidone and lisinopril, where the 15% overall marginal advantage to chlorthalidone appears to be entirely attributable to a 40% benefit in the Black cohort, with no detectable difference at all in the non-Black subjects. This tendency for the ACE inhibitor to be less effective than the diuretic in Black subjects is seen with all the major outcomes, primary and secondary 4, and is consistent with many suggestions over the years that ACE inhibitors and beta-blockers might be less effective than diuretics and calcium antagonists in Black subjects 14. Some comments and caveats It was disappointing that the results of the pravastatin versus ‘usual care’ comparison in hypercholesterolaemic, hypertensive patients were inconclusive 5, possibly reflecting the small difference between the groups in the reduction of serum cholesterol. While the results of ALLHAT correct a lack of evidence in Black Americans, they do not provide any evidence regarding the relative value of the drugs tested in Asian subjects, who contribute a very large fraction to the global burden of cardiovascular disease. Over 5000 ‘hispanic’ subjects were included in the study, but this was insufficient to separately assess the difference in outcome in hispanic and non-hispanic subsets. The direct comparison of the results obtained with the ACE inhibitor and the calcium antagonist are also of great interest and it is hoped these will soon be published in a separate paper. ALLHAT does not provide any evidence on the relative value of beta-blockers, or of angiotensin receptor blockers (ARBs). While early comparisons of beta-blockers and diuretics did not suggest there was any significant advantage either way, it may be that these studies were not sufficiently powered to provide a result 16–19. On the other hand, the LIFE study reports a 25% advantage for losartan, an ARB, compared to atenolol, a beta-blocker, in preventing stroke in hypertensive subjects, even though the degree of blood pressure lowering was similar between the two drugs 20. This result requires confirmation. The premature discontinuation of the doxazosin arm 3 appears even more regrettable because it is unlikely that there will be further evaluation of alpha-blocking drugs in major prospective studies in the foreseeable future. It is interesting that the initial wide separation between the cumulative event curves for heart failure in the doxazosin and chlorthalidone arms, which was subjected to extensive comment, is repeated to a lesser extent in the comparison between chlorthalidone on the one hand and amlodipine and lisinopril on the other. It might also be that the clear superiority of chlorthalidone over lisinopril in preventing heart failure would not have been predicted by many observers. It may be that the marked success of the ACE inhibitors in the treatment of heart failure has often been evaluated on a background of free use of diuretics, and that the two groups of drugs have not been subjected to extensive head-to-head comparison. The rate of decline of glomerular filtration rate was also no better with the ACE inhibitor than with the other two drugs. Indeed, despite much weight of opinion suggesting that antagonism of the renin–angiotensin system may be specifically advantageous for the preservation of renal function, the performance of lisinopril in this respect appeared to be equivalent to that of chlorthalidone and not quite as good as that of amlodipine. Finally, it must be stressed that while ALLHAT is certainly the largest and most impressive comparison available of the effects of different blood pressure lowering drugs, there is a wealth of other evidence in this area. Indeed, the combined evidence presented by the Blood Pressure Treatment Trialists collaboration in 2000 15, taken with the results of subsequent relevant trials, will provide comparison of ACE inhibitors versus ‘conventional therapy’ with diuretics and/or beta-blockers in a further cohort in excess of 30 000 patients, quite apart from ALLHAT. In the same way, there is an additional cohort of over 20 000 patients available for comparison of calcium antagonists with conventional therapy. Given the trends observed in the report of the Trialists Collaboration 15 and those observed in ALLHAT 4, which often had minor trends in different directions, the combined overview of all the available evidence, including ALLHAT, will be remarkably informative. It may well reinforce the messages that there is little to choose between the major groups of drugs in reducing the burden of cardiovascular disease and that diuretics are at least as effective as the other available drug groups. Take home message for the practising clinician – a return to ‘step therapy' The results of ALLHAT demonstrate that diuretics can withstand comparison with ACE inhibitors, calcium antagonists and alpha-blockers. They throw no light on comparison with ARBs. However, the sum of the evidence now available suggests that a return to a simpler routine for the initiation and maintenance of therapy with blood pressure lowering drugs is needed. Indeed, ALLHAT could herald a return to STEP THERAPY along the following lines. Step 1 Unless there are particular reasons to the contrary, start with a low-dose diuretic, on a background of non-pharmacological measures. Step 2 In many patients, achievement of goal blood pressure will require a second drug and the practitioner should choose among ACE inhibitors, ARBs, beta-blockers and calcium antagonists (listed in alphabetical order), according to the clinical situation. It must be stressed that effective combination therapy will usually be needed to achieve good control of blood pressure, both in individuals and in populations. In this context, one of the key messages of ALLHAT is that diuretics should be one component of combination therapy in the vast majority of hypertensive patients, or in other patients with high cardiovascular risk requiring a blood pressure lowering regimen. Finally, the clear evidence from ALLHAT favouring low-dose diuretics, coupled with the lower cost of these drugs compared to newer agents, provides a welcome message for hypertension control worldwide and particularly in low and middle income countries.