Calcium Antagonists: Antagonists or Protagonists of Mortality in Elderly Hypertensives?

There are several fundamental issues that need to be considered in evaluating any antihypertensive agent or class of agents: The objective of treatment of hypertension is to prevent its cardiovascular complications—stroke, acute myocardial infarction, congestive heart failure—not just to lower an elevated blood pressure, All drugs have multiple actions, some beneficial and others perhaps harmful, and the sum of these mechanisms determines the overall health benefit to the patient, Reliance on a single so-called surrogate outcome, such as reduction in blood pressure, restricts attention to one action and ignores the others, some of which may be harmful, Therapies intended to be life-long require long-term safety data. Calcium antagonists (CAs), as a class, have to date not been properly evaluated in order to determine their long-term efficacy and safety in the treatment of hypertension. Small short-term randomized clinical trials documenting “surrogate efficacy” have secured regulatory approval. Evidence of important adverse secondary effects, particularly of short-acting dihydropyridine CAs in patients with coronary heart disease, 1 is emerging. The article by Pahor et al2 in this issue of the Journal adds important new information about the overall health effect of CAs in elderly hypertensives. In this cohort study, the investigators assessed drug use and other risk factors at baseline, followed subjects prospectively for events, and used a variety of techniques, including restrictions and multivariate adjustment, to minimize the effects of potential confounding factors. Although the methods of ascertaining events may have misclassified some of the non-fatal outcomes, the primary endpoint, total mortality, was assessed well and completely. The investigators have been appropriately cautious in their interpretation of the data. They had to control not only for traditional confounding factors such as age but also for the possibility of confounding by indication. If patients who receive CAs are more likely to die because of co-morbid conditions such as angina, the drug may simply be a marker of that prognostic factor rather than a true risk factor for mortality. One of the most powerful defenses against confounding by indication is the analytical approach called restriction. When the analyses were restricted to hypertensives who had no CHD and no potential contraindications to beta-blockers, all three CAs, including verapamil, appear to increase the risk of mortality. In the end, the ability to detect differences among CAswill require clinical trials. The randomized clinical trial (RCT) is the gold standard for assessing the effects of an intervention. The available data from RCTsevaluating the overall health benefit of CAsin hypertension is woefully inadequate. The two largest trials available to date—the Multicenter Isradipine Diuretic Atherosclerosis Study (MIDAS)3 from the US and the Evaluation Group of Long-term Antihypertensive Treatment (GLANT)4 from Japan—included only 1326 and 1017 person-years, respectively. Although this is only 5 to 10% of the number of person-years typically required, preliminary findings of these two trials have raised concerns about the safety of these drugs. In MIDAS, the risk of major cardiovascular complications was approximately 80% higher in patients treated with isradipine than in those treated with hydrochlorothiazide, a difference that was, however, of only borderline significance.3 See also p 1191 In GLANT, the proportion of cerebrovascular accidents (CVA) was significantly higher (P < .02) in the group receiving either of two dihydropyridine CAs compared with the group receiving an angiotensin-converting enzyme (ACE) inhibitor even though the blood pressure reduction was greater in the CA group.4 The disparity between the blood pressure findings and the CVA findings illustrates the potential fallacy of relying on surrogate outcomes. Caution is advised in interpreting the GLANT data until the full trial report is published. Because of the lack of adequate, well controlled RCTs, and mindful of the potential limitations of these sources, we have to consider other sources of information regarding the possible health benefits of CAs. Replications are critical to the advancement of knowledge. The history of medicine is full of unconfirmed “breakthroughs.” The report by Pahor et al. in this issue, stimulated by a report from a case-control study5 arrives essentially at the same overall conclusion. Short-term CAsin high doses appear to have adverse health effects in older hypertensives. Psaty et al5 reported on the association between risk of acute myocardial infarction and use of the four major classes of antihypertensive agents. Pahor et al2 extended the findings to all-cause mortality, incident coronary heart disease, and hospitalized congestive heart failure (CHF). After adjustments for potential confounding factors, short-acting nifedipine emerged with the worst record-a 70% excess risk of mortality (RR = 1.7, 95% CI 1.1-2.7) compared with beta-blockers. Diltiazem showed a trend toward a nonsignificant 30% excess (RR = 1.3,95% CI 0.8-2.1), whereas verapamil and ACE inhibitors were similar to beta-blockers (RR = 0.8, 95% CI 0.4-1.4 and RR = 0.9, 95% CI 0.6-1.4, respectively). This differentiation among the CAs is consistent with that suggested by a meta-analysis of trials in patients with coronary disease.6 The dihydropyridines appeared to increase all-cause mortality (RR = 1.16, 95% CI 0.99-1.35), but the effect of the nondihydropyridine CAs was neutral (RR = 0.95,95% CI 0.82-1.09). The 2- to 3-fold excess risk of CHF applicable to all three CAs reported by Pahor et al. is alarming, but perhaps not surprising, since many agents of this class have negative inotropic effects. Inasmuch as CHF is a common complication of hypertension particularly in older patients, the wisdom of placing older patients with uncomplicated hypertension on a CA can be questioned if their risk of developing CHF is high. The article by Pahor et al. and the other reports either published or just reported have raised several questions: Conclusive data from RCTs are not available so the jury is still out. However, the evidence is mounting. For the dihydropyridine CAs, the consistency of the findings from the two largest trials, MIDAS3 and GLANT4 the case-control study5 and the cohort study2 is very suggestive. The similarity of these observations in patients with hypertension and in patients with coronary heart disease1,6 is also noteworthy. Scientifically, what strength of evidence is required to conclude that an agent is definitely harmful? For ethical reasons, the common view is that evidence of treatment efficacy should be clear and compelling although the evidence required to establish harm may be less stringent. Today we cannot conclude that CAs have an overall effect that is harmful in hypertensives, but the limited trial and epidemiologic data suggest that the CAs evaluated in these studies, largely the immediate-release formulations, may increase the risk of major adverse clinical events and mortality. Because the CAs have not been proven to reduce the risk of the cardiovascular complications of hypertension, and they may or may not have significant adverse effects, and because proven treatment alternatives are available, clinicians ought to use CAs cautiously. While awaiting the results of ongoing long-term trials it appears prudent to adhere to the JNC V guidelines7 and not only use diuretics and beta-blockers as first-line drugs but also reserve CAs for third-line agents or when diuretics and beta-blockers have failed or cannot be tolerated. Although CAs are normally well tolerated in older hypertensives, this patient group, which has a high prevalence of clinical and subclinical coronary disease, may be at a higher risk of adverse reactions to CAs than are middle-aged hypertensives. The recommendations should be re-evaluated as new evidence from RCTs emerges. Most of the alarming data on CAs are from studies of short-acting formulations, in particular the dihydropyridine, nifedipine. It appears that the adverse findings may apply to other short-acting dihydropyridines, including felodipine and isradipine. Data on slow-release formulations and other compounds are lacking. It is clear that there are major pharmacologic differences among the three subclasses of CAs and these differences may affect their safety and efficacy. The potentially adverse proischemic effect demonstrated for short-acting dihydropyridines, for example, has not been described for the nondihydropyridines. Among the CAs, verapamil appears to have the best record for efficacy and safety. Several parties have contributed to the current situation. Regulatory agencies approved CAs based on surrogate efficacy without adequate evidence of long-term safety. Pharmaceutical companies used skillful marketing, based on concepts such as individualizing treatment and mechanisms of action, to promote these agents and ignored for more than a decade calls for large-scale RCTs to evaluate the effect of these drugs on cardiovascular endpoints. Clinicians who have chosen to prescribe these drugs without proper evidence of health efficacy and long-term safety are also responsible. Several large long-term RCTs of various CAs in hypertension are now underway. Regrettably, it will be many years until we know with certainty the proper role of these agents in the treatment of hypertension.