Editorials

Calcium channel blockers

The jury is still out on whether they cause heart attacks and suicide 

Calcium antagonists are used extensively for treating high blood pressure and angina. Since 1995 they have been accused of causing myocardial infarcts, cerebrovascular events, cancer, bleeding, depression, and suicide by mechanisms that include pro-ischaemic, pro-arrhythmic, negative inotropic, hypotensive, and reflex sympathetic effects (cardiovascular events); inhibition of apoptosis (cancer); inhibition of platelet aggregation and the normal vasocontrictive response to bleeding (bleeding); excess hypotension or interference with neurones and receptors involved in mood regulation (depression). A recent review of the evidence recommended no change to current guidelines and clinical practice,¹ but since then a report of raised suicide rates among patients taking calcium antagonists has been published,² together with three randomised controlled trials suggesting that myocardial infarcts might be increased in diabetics on calcium antagonists.^3-5

Much of the evidence concerning risk of cardiovascular events, cancer, and bleeding, both published and unpublished, was reviewed in 1996-7 by an ad hoc subcommittee of the Liaison Committee of the World Health Organisation and the International Society of Hypertension.¹ The principal conclusions were that the available evidence did not prove the existence of either beneficial or harmful effects of calcium antagonists on the risks of major coronary heart disease events and that there was no good evidence for adverse effects of calcium antagonists on either cancer or bleeding risks. The committee commented that the bulk of the evidence for adverse effects was derived from observational studies or small randomised clinical trials and pointed to a "clear failure of pharmaceutical companies, regulatory authorities, and clinical researchers to ensure the timely conduct of studies, involving both large numbers of cases and random assignment of treatments." Many such large randomised clinical trials are now in progress, but reliable detection of any modest adverse or beneficial effect of calcium antagonists is not expected until early in the next century. Do these recent published studies alter the picture enough to suggest clinicians should change their practice?

Prompted by several studies suggesting a link between calcium channel blockers and depression, Lindberg et al investigated the associations between use of cardiovascular drugs and suicide in Sweden.² In a cross sectional ecological study they found a significant correlation between the rates of use of calcium channel blockers and age adjusted suicide rates in 152 of Sweden's 284 municipalities (r=0.29; P<0.001). Furthermore, in a population based cohort study in one municipality they reported that the relative risk of suicide, adjusted for differences in age and sex, among users of calcium channel blockers was 5.4 (95% confidence interval 1.4 to 20.5) compared with users of other antihypertensive agents. It is noteworthy that there were only 9 suicides in the cohort study (5 users of calcium antagonists and 4 non-users). The authors concluded that use of calcium channel blockers may increase the risk of suicide.

These two studies were observational: treatment was provided to individual municipalities (ecological study) and individual patients (cohort study) on the basis of clinical indication, rather than random assignment. In these circumstances the potential for systematic errors, principally due to confounding by indication, is great. Calcium channel blockers have been reported to be used more commonly in sicker patients,⁶ so the differences in suicide rate could have been due simply to differences between the types of patients given each type of drug.

The Multicenter Isradipine Diuretic Atherosclerosis Study (MIDAS) was a multicentre, randomised, double blind, controlled trial in 883 hypertensive patients, comparing the effect of isradipine and hydrochlorothiazide on the progression of early atherosclerosis in carotid arteries. ^{3 7} After three years' follow up there was no difference in the primary endpoint between the two treatments. There was, however, a trend for an increased incidence of major vascular events (myocardial infarction, stroke, congestive heart failure, angina, and sudden death) in patients taking isradipine compared with with those taking hydrochlorothiazide (25/ 442 v 14/441; P=0.07). In a recent reanalysis of the trial the increase in major vascular events associated with the use of isradipine appeared to be largely confined to patients with impaired glucose metabolism.³ Patients with a glycosylated haemoglobin greater than 6.6% and randomised to isradipine had more than double the risk of an event than those randomised to diuretic (15/199 v 6/216; P=0.04).

In the Fosinopril Amlodipine Cardiovascular Events Trial (FACET) the relative benefits of fosinopril and amlodipine were compared in 380 hypertensives with non-insulin dependent diabetes.⁴ The patients receiving fosinopril had a significantly lower risk of major cardiovascular events (fatal or non-fatal acute myocardial infarction, fatal or non-fatal stroke, hospitalised angina) than those receiving amlodipine (14/189 v 27/191; P=0.03).

The Appropriate Blood Pressure Control in Diabetes (ABCD) trial is a prospective, randomised, blinded trial comparing the effects of moderate control of blood pressure with those of intensive control on the incidence and progression of diabetic nephropathy, neuropathy, retinopathy, and cardiovascular events. The study also compared nisoldipine with enalapril as first line antihypertensive agents in terms of the prevention and progression of complications of diabetes. The primary end point was the change in 24 hour creatinine clearance. Secondary end points included cardiovascular events, retinopathy, clinical neuropathy, urinary albumin excretion, and left ventricular hypertrophy. The recent report in the New England Journal of Medicine concerns only data on a secondary end point (myocardial infarction) in the subgroup of patients who had hypertension (n = 470).⁵ After five years' follow up the data safety and monitoring board recommended that nisoldipine treatment should be terminated in the hypertensive patients, as in this subgroup it was associated with a higher incidence of fatal and non-fatal myocardial infarction than enalapril (25/235 v 5/235: P<0.001).

Commenting in the Lancet on the results of the above three trials, Pahor et al recommended that angiotensin converting enzyme inhibitors and low dose diuretics, rather than calcium antagonists, should be the preferred first line agents for hypertensive patients with impaired glucose metabolism or diabetes.⁸ However, these trials are some distance from definitively proving deleterious effects of calcium antagonists in diabetes. The trials were relatively smallin aggregate a total of 92 cardiovascular events occurred in 1265 patientsand hence prone to random errors. More than half of the original cohort of the ABCD trial discontinued their assigned study medication before completion of the study, raising the possibility of systematic bias. In both the MIDAS and ABCD studies cardiovascular events were secondary end points, and the apparent adverse effects were identified only by subgroup analyses. The authors of the ABCD trial themselves commented that their results should be interpreted cautiously and would require confirmation. In both the ABCD and FACET studies long acting calcium channel blockers were compared with angiotensin converting enzyme inhibitors; without a placebo group it is impossible to say whether these studies show harmful effects of calcium antagonism or beneficial effects of angiotensin converting enzyme inhibition. Interestingly, in the ABCD trial the rate of myocardial infarction among patients with non-insulin dependent diabetes who were randomly assigned to treatment with nisoldipine was similar to that seen in historical controls.

Some researchers and clinicians clearly consider that the evidence against calcium antagonists is sufficient to advise the use of alternative types of drug where possible. In considering this evidence, however, we should remember the "cholesterol controversy" of the early 1990s. At that stage evidence from observational studies suggested an association between low cholesterol concentration and increased non-cardiovascular morbidity and mortality.⁹ Several meta-analyses of randomised clinical trials reported negative effects of lipid lowering interventions on non-coronary heart disease mortality.¹⁰ Lipid lowering strategies could have been abandoned on the basis of these studies. The benefits associated with effective lipid lowering by statins, in terms of reductions of both coronary events and all cause mortality, seen in the large, adequately powered, randomised lipid lowering trials4S, WOSCOPS, and CARE^11-13illustrate clearly that this would have been a mistake.

When used in hypertension or angina, calcium channel blockers probably either have no effect on risk or cause modest harm or modest benefit. Reliable detection of a 20% increase or decrease in risk requires studies in which at least 1000 patients develop the relevant event during follow up. Two trials are currently under way of sufficient size and duration to definitively confirm or rule out modest effects on cardiovascular risk and all cause mortality. The Antihypertensive and Lipid Lowering treatment to prevent Heart Attack Trial (ALLHAT) is a comparison of first line therapy with amlodipine, lisinopril, doxazosin, or chlorthalidone; the Anglo-Scandinavian Coronary Outcomes Trial (ASCOT) is a comparison of the effect of two therapeutic regimens,  blockers with or without diuretics versus calcium antagonists with or without angiotensin converting enzyme inhibitors, on non-fatal myocardial infarction and fatal coronary heart disease in hypertensive patients at high risk of cardiovascular events. Hence, I would advocate no change in current clinical practice on the basis of non-randomised observational studies, or subgroup analyses of small clinical trials not specifically designed to assess morbidity and mortality. The conclusions of the 1997 WHO and International Society of Hypertension committee remain valid¹ until we have the results of large randomised trials.

I have been and will be involved in trials of all classes of antihypertensive agenets, including the ASCOT trial.