Journal of Human Hypertension (2001) 15, 85–87 2001 Nature Publishing Group All rights reserved 0950-9240/01 $15.00 www.nature.com/jhh COMMENTARY Calcium channel blockers in hypertension: the debate reawakens GYH Lip and DG Beevers University Department of Medicine, City Hospital, Birmingham B18 7QH, UK Keywords: calcium channel blockers; meta-analysis; mortality; morbidity Ever since the publication of the Puget Sound phar- the confidence of any one study to be sure of these macosurveillance study in 1995,1 there has been a such differences.7,8 continuing debate on the effectiveness or safety of Then as the millennium year ground to its sodden the calcium channel blockers (CCBs) in the treat- halt (in England at least) came a ‘rush’ of meta- ment of hypertension and heart disease. Various analyses. To highlight a few, The Lancet published retrospective case-control studies had raised the two papers on 9 December9,10 and a ‘mini’ meta- possibility that these drugs may cause haemorrhage, analysis in their correspondence columns on 2nd suicide, cancer and excess cardiovascular morbidity December.11 Another appeared in the British Journal and mortality.2 Several observational studies and of Cardiology in November.12 There probably have some individual randomised trials in hypertension been many others. Perhaps meta-analyses begat have even suggested that, compared with other further meta-analyses. drugs, CCBs may be associated with a higher risk of The Blood Pressure Lowering Treatment Trialists’ coronary events, despite similar blood pressure con- Collaboration is a programme of overviews of trol. randomised trials which was established to investi- Many of these anxieties seemed to have been gate the effects of angiotensin-converting enzyme allayed by the placebo-controlled SYST-EUR Trial3 (ACE) inhibitors, CCBs, and other blood pressure- where the CCB, nitrendipine, bought about a major lowering drugs on mortality and major cardiovascu- reduction in stroke and a large but nonsignificant lar morbidity.9 Their overview of placebo-controlled fall in heart attacks, with no excess of cancer of trials of ACE inhibitors (four trials, 12124 patients haemorrhage. Similarly, the use of the CCB, felodip- mostly with coronary heart disease) revealed ine, was beneficial in the HOT study.4 Was the reductions in stroke (30% [95% CI 15–43]), coron- debate over? So it seemed, but the focus of the argu- ary heart disease (20% [11–28]), and major cardio- ment changed slightly. No one doubted that CCBs vascular events (21% [14–27]), which the meta- were better than placebo, but how do they compare analysis of placebo-controlled trials of CCBs (two with other agents? The STOP-H2 trial perported to trials, 5520 patients mostly with hypertension) show evidence of equivalence of calcium channel showed reductions in stroke (39% [15–56]) and blockers (felodipine or isradipine) with the older major cardiovascular events (28% [13–41]). In the agents (diuretics and beta-blockers) and was deemed overview of trials comparing blood pressure-lower- by a Lancet commentary to be ‘a draw’.5,6 However, ing strategies of different intensity (three trials, there was still the tendency for the CCBs to be 20408 patients with hypertension), there were ‘slightly better’ than other agents at preventing reduced risks of stroke (20% [2–35]), coronary heart stroke but to be ‘slightly less effective’ at preventing disease (19% [2–33]), and major cardiovascular myocardial infarction and heart failure, with no dif- events (15% [4–24]) with more intensive therapy. In ference in all-cause mortality. Two more trials with the trials that compared CCB-based regimens with CCBs which were published in 2000 (INSIGHT and diuretic-based or beta-blocker-based regimens, NORDIL) broadly produced similar findings but stat- patients assigned CCB-based therapy had a signifi- istical and methodological considerations limited cant 13% lower risk of stroke (95% CI 2–23) than among those assigned diuretic-based or beta- blocker-based therapy. Additionally, there was a Correspondence: Dr GYH Lip, University Department of Medi- 12% greater risk of coronary heart disease events of cine, City Hospital, Birmingham B18 7QH, UK. E-mail: g.y.h.lipȰbham.ac.uk borderline significance (0–26) amongst those assigned CCB-based therapy. Calcium channel blockers in hypertension GYH Lip and DG Beevers 86 The meta-analysis by Psaty et al10 of nine eligible as second-line agents. In reality, more than half of trials (27743 participants), found that CCBs and all hypertensive patients need an additional drug, other drugs achieved similar control of both systolic and a third require three drugs or more, to bring and diastolic blood pressure. Compared with about adequate control of blood pressure. patients assigned diuretics, beta-blockers, ACE Perhaps one may extend the debate by speculating inhibitors, or clonidine (n = 15 044), those assigned on the mechanisms whereby the CCB’s may differ CCBs (n = 12699) had a significantly higher risk of from other antihypertensive agents, either ben- acute myocardial infarction (odds ratio 1.26 [95% CI efically or adversely.2,13 The CCB’s might be pro- 1.11–1.43], P = 0.0003), congestive heart failure arrhythmic, pro-haemorrhagic and may also cause (1.25 [1.07–1.46], P = 0.005), and major cardiovascu- coronary steal; they may also encourage apoptosis lar events (1.10 [1.02–1.18], P = 0.018). There was and activate the renin system. However, there are no significant difference for the outcomes of stroke equal arguments in their favour, particularly as they and all-cause mortality. may be antithrombotic. Sadly, laboratory-based The ‘mini’ meta-analysis of the INSIGHT, NORDIL scientists can provide theoretical mechanisms and STOP-H2 trials11 also showed a trend in favour hereby anything might happen and other scientists of CCBs for cerebrovascular events (with a 14.7% can equally speculate why the same thing might reduction), and in favour of diuretics and beta- not happen. blockers for coronary events, with CCBs use As with the last calcium channel blocker crises, resulting in a 20% increase in myocardial infarc- there will be some anxiety amongst hypertensive tions. In the ‘solid’ meta-analysis of eight trials by patients when the ‘anti-CCB’ papers are reported in Rafflenbeul,12 for the end point of acute myocardial the newspapers or become available on the INTER- infarction, there was a relative difference of 19% in NET. This also follows on from the recent concerns favour of the diuretics, a 14% reduction in stroke over the use of doxazosin in the Antihypertensive favouring CCBs (vs diuretics/beta-blockers), whilst and Lipid Lowering treatment to prevent Heart heart failure end points were no different. Attack Trial (ALLHAT) study, with an increased risk These overviews, armed with large numbers, seem of heart failure and stroke.14 Clinicians have to have to confirm the same trends of the three major trials, a sensible answer when questioned by their patients that there was no difference between CCBs and other about the pros and cons of their antihypertensive therapies on all-cause mortality, but a trend, which therapy. Perhaps the best response is to emphasise reaches statistical significance, for an inferior effect that the highest priority is to control the blood press- on heart disease and a superior effect on stroke. It is ure, and that there are subtle differences in outcome clear that treating blood pressure is beneficial, with comparing the CCB’s with other agents, but these are strong evidence of benefits of ACE inhibitors and very controversial and may be unproven. We await CCBs by the overviews of placebo-controlled trials. the results of proper prospective randomised trials, If the question is no longer ‘do we treat hyperten- such as ALLHAT and ASCOT (Anglo-Scandinavian sion?’, but ‘how to treat?’ and ‘who to treat?’, some Cardiac Outcomes Trial) with increased interest. uncertainties arise. The meta-analyses suggest that there is weaker evidence of differences between treatment regimens of differing intensities and of Conflict of interest statement differences between treatment regimens based on DGB and GYHL are local co-principal investigators different drug classes. The meta-analysis by Psaty et for ASCOT in Birmingham, UK. Both have received al10 even goes as far as suggesting that CCBs are research funding and honoraria for educational sym- inferior to other types of antihypertensive drugs as posia, meetings, etc from various antihypertensive first-line agents in reducing the risks of several drug manufacturers. major complications of hypertension, and conclude that ‘the longer-acting calcium antagonists cannot be recommended as first-line therapy for hypertension’. References Because meta-analyses look impressive, clinicians 1 Psaty BM et al. The risk of myocardial infarction asso- tend to believe them, but in fact they have many ciated with antihypertensive drug therapies. JAMA problems. Some of the trials included were designed 1995; 274: 620–625. to demonstrate drug equivalence, whilst others 2 Lip GYH, Beevers DG. Are calcium antagonists safe in hoped to detect drug superiority. Some were open hypertension? Postgrad Med J 1996; 72: 193–194. label studies with the risk of bias in randomisation. 3 Staessen JA et al. Randomised double-blind compari- There were also major differences in the methods of son of placebo and active treatment for older patients all trials included. Furthermore, the meta-analyses with isolated systolic hypertension (SYST-EUR Trial). themselves may be prone to bias as to which end Lancet 1997; 350: 754–764. 4 Hansson L et al. Effects of intensive blood-pressure points are examined. There is also the problem that lowering and low-dose aspirin in patients with hyper- they rely heavily on the larger trials, some of which tension: principal results of the Hypertension Optimal can be criticised for their design. Finally, the biggest Treatment (HOT) randomised trial.