Angiotensin Receptor Blockers, Cancer, and Smoking

Concern that antihypertensive drugs might be associated with increased risk of cancer has a long history, beginning with a report that linked use of reserpine with breast cancer.1 That report was not supported by subsequent surveys.2, 3 Case control associations have also suggested that diuretics were related to renal cancer.4 Calcium channel blockers had been associated with cancer,5, 6 again with little support after other studies were analyzed.7-9 In the very large Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT), no association was found between cancer and the three drug limbs of the trial: a thiazide-type diuretic, an angiotensin-converting enzyme (ACE) inhibitor, and a calcium channel blocker.10 A recent meta-analysis that reports an association between exposure to angiotensin II receptor blockers (ARBs) and increased risk for cancer, especially lung cancer, continues this research.11 Given the widespread use of antihypertensive drugs over long periods for so many patients, it is reasonable to continue surveillance for possible long-term risk that might not appear during the early phases of drug development, availability, and use for approved indications. ARBs with specificity for the angiotensin II type 1 receptor have become valuable, widely used, and effective therapy for hypertension and congestive heart failure (CHF). ARBs reduce activity of the renin-angiotensin system without potentiation of kinins or other mechanisms that account for the well-recognized adverse reaction, cough, that is clearly associated with ACE inhibitors. Well-designed double-blinded trials support the conclusion that ACE inhibitor–associated cough often leads to discontinuation of ACE therapy.12 If continued blockade of the renin-angiotensin system is needed, particularly for treatment of CHF, switch from an ACE inhibitor to an ARB has been beneficial.13 A report from analysis of a large meta-analysis of ARB trials, finding a small but statistically significant association between ARB use and cancer immediately raises concern about the safety of ARBs.14 Does the association represent a causative mechanism or can it be explained by the limitations inherent in retrospective analysis? Could a hidden flaw account for the apparent association between ARB exposure and cancer, especially lung cancer, for which the only significant association was found. It was acknowledged by Sipahi and colleagues11 that their meta-analysis did not include smoking as a covariate due to lack of sufficient baseline information from the studies. This brief review will focus on the issue of smoking as a possible confounder in the ARB-cancer association because cigarette smoking is so highly linked to incidence of lung cancer. We reviewed available outcome trials in which ARBs were compared with either a placebo or an alternate therapy for baseline rates of current smoking or past smoking. The results are given in the Table. The baseline rates for current smoking varied from approximately 10% to 17% in these seven trials. Baseline prevalence rates for past smokers were only available for four trials and varied from approximately 33% to 52%. There is no obvious difference in rates for either current or past smoking in the ARB limbs of these trials compared with the non-ARB limbs. Two trials, the Telmisartan Randomized Assessment Study in ACE-Intolerant Subjects With Cardiovascular Disease (TRANSEND) and the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (In Patients With LV Dysfunction Intolerant to ACE Inhibitors) (CHARM-Alternative) studies, enrolled participants who were considered “ACE inhibitor–intolerant” mostly due to presence of cough (approximately 70%) as the basis for assignment to either an ARB or placebo. The criteria for ACE inhibitor intolerance in these trials was not provided in detail, so baseline assessment for non–ACE inhibitor causes of cough cannot be determined. Likewise the trials provide no detailed information about screening for pulmonary disease such as use of chest x-rays. Thus, lack of stratification for current or previous smoking at baseline could have led to unequal assignment within the ARB trials. It would take only a small extra fraction of patients predisposed to develop lung cancer assigned to ARB cohorts for the trend that was found in the meta-analysis. Cigarette smoking accounts for at least 30% of all cancer deaths, and about 87% of lung cancer deaths are caused by smoking.15 More than half of the sites for the large trials comparing ARB with placebo or other active therapy took place in Northern-Eastern Europe. In one of the largest recent case-control studies carried out in 6 European countries (Sweden among them) to investigate the major risk factors for lung cancer, the results indicate elevated odds ratios: 23.9 among current smoker men and 8.7 among current smoker women and 7.5 among ex-smoker men and 2 among ex-smoker women. All these data highlight the strong association between current and ex-smokers with developing lung cancer. It would be most helpful to have a more definitive history of smoking and smoking-related lung disease in the various studies of ARB therapy as related to incidence of lung cancer during the trials. Aside from cigarette smoke, a wide variety of toxic substances have been identified as potential causes of lung cancer and respiratory impairment in human beings, among them asbestos, polycyclic aromatic carbohydrates, nitrogen oxides, radon, and others. The majority of the population enrolled in those studies lived in large industrialized cities, where exposure to air pollution is common. Although difficult to measure in the past, new technologies have allowed for ecological studies, measuring how air particles can affect human health. In the Danish Cancer Registry, investigators analyzed whether the concentration of nitrogen oxide in the patients’ residences, used as an indicator of air pollution from traffic, was associated with risk for lung cancer. They found that the incidences rates of lung cancer were 1.30 (95% confidence interval, 1.07−1.57; P=.008) for intermediate exposure (30–70 μg/m3), meaning that, although difficult to study, other reasons aside from tobacco exposure could be related to lung cancer.16, 17 These trends reinforce the need for careful stratification at entry to characterize lung cancer risk in clinical trials where rates of lung cancer are examined in relation to baseline characteristics. Experimental studies in animal models for cell systems suggest that exposure to ARBs have either no onchogenic effect or might even be protective in blocking angiotensin II receptor–mediated abnormalities in various mechanisms.18 A retrospective case control study of lung cancer reported that use of either ACE inhibitors or ARBs conferred a life extension of approximately 3 months when compared with lack of exposure to either of these two renin system blockers.19 The small but statistically significant association between ARB exposure and risk of cancer in the meta-analysis by Sipahi and colleagues11 rests on greater rates of lung cancer in the ARB cohorts. Alternate and well-defined causes of lung cancer, ie, current or past smoking, were highly prevalent at baseline in these studies. The Sipahi analysis did not take current or past smoking or pre-existing lung disease into account for calculation of risk ratios. Small differences in assignment to ARB and comparison groups may have resulted in an imbalance, such that a chance result associating ARB exposure to cancer, especially lung cancer, occurred. Additional analyses from all available trials and ongoing surveillance are clearly needed to define cancer risk in patients exposed to ARBs who never smoked, smoked in the past, or are current smokers. Evidence from basic research is somewhat reassuring that ARBs, if anything, may protect against cancer-causing mechanisms.18, 19 Meanwhile, before all the evidence is in, what should clinicians consider when relying on ARBs for effective treatment of hypertension, CHF, or chronic renal disease? We have two suggestions. First, cough as an apparent adverse effect of ACE inhibitors should be evaluated in relation to overall risk for lung cancer in each patient. For current or past smokers, or those with known exposure to lung carcinogens, simply changing to non–ACE inhibitor therapy is insufficient. Thoughtful assessment for alternate causes of cough (including lung cancer) is needed. Second, patients who need ARB therapy should be fully informed about the limitations of the ARB-cancer meta-analysis. Should they refuse to take indicated ARB therapy, but have apparent ACE inhibitor intolerance due to cough, the alternative renin-system blocker aliskiren is available. However, as a recently approved drug, aliskiren’s long-term safety remains to be fully established relative to ARB therapy. What is most important is that hypertension, CHF, and chronic renal disease be effectively treated with the agents that are now available and remarkably safe for these conditions. Disclosure: The authors have nothing to disclose.