Is the Effect of Angiotensin Receptor Blockade in Patients with Heart Failure Modified by Treatment with Aspirin? The Answer is Not so Clear!

This editorial refers to ‘Efficacy and safety of angiotensin receptor blockade are not modified by aspirin in patients with chronic heart failure: a cohort study from the candesartan in heart failure: Assessment of reduction in mortality and morbidity (CHARM) programme’, by S.M. Chang et al., published in this issue on page 738. In this issue of the European Journal of Heart Failure, Chang et al.,1 examine the question of whether the effect of candesartan is modified by concomitant administration of aspirin in the large and heterogeneous population of heart failure patients enrolled in the Candesartan in Heart Failure—Assessment of Reduction in Mortality and Morbidity (CHARM) Program.2 To appreciate the relevance of this question, some background is required. There is a substantial literature suggesting that non-steroidal anti-inflammatory drugs, including aspirin, may diminish the beneficial effects of angiotensin-converting enzyme-inhibitors (ACE-I), particularly in patients being treated for chronic heart failure (CHF),3,4 but also in post-myocardial infarction patients and in patients at high risk for vascular events.5,6 Angiotensin-converting enzyme is a multifunctional protease that not only converts angiotensin I to angiotensin II, but also degrades kinins. Angiotensin-converting enzyme-inhibitors both reduce the formation of angiotensin II and increase concentrations of kinins. Kinins are important regulators of prostaglandin production and NO synthesis, both of which appear to play a beneficial counter-regulatory role in CHF.7 In a group of patients with severe heart failure, Dzau et al.8 found elevated levels of both angiotensin II and metabolites of prostaglandin I2 and E2, particularly in patients with hyponatraemia. Treatment with indomethacin resulted in significant declines in cardiac index and increases in pulmonary capillary wedge pressure, supporting a counter-regulatory role for prostaglandins. Aspirin has also been shown to counteract the vasodilator effects of enalapril and adversely affect renal function in patients with severe heart failure.9 Subsequently, a considerable body of mechanistic, observational, and prospective trial data have suggested that patients who are taking aspirin on a chronic basis, either for vascular protection or inhibition of platelet function, may obtain less benefit from ACE-I given for a variety of indications.10 An adverse interaction with aspirin was first noted in the Studies Of Left Ventricular Dysfunction (SOLVD) treatment and prevention trials, with enalapril reducing all-cause mortality by 23% [hazard ratio 0.77, 95% confidence interval (CI) 0.67–0.87] in patients not taking antiplatelet agents (predominantly aspirin) and a trend toward higher mortality in those who received antiplatelet agents (hazard ratio 1.10, 95% CI 0.93–1.30).4 Warfarin treatment, in contrast, was associated with a lower mortality rate. An additional combined analysis of three long-term post-MI ACE-I trials, the Heart Outcomes Prevention Evaluation (HOPE), and the SOLVD prevention trial also suggested lesser benefit from ACE-inhibitors in patients receiving aspirin, although the interaction did not reach statistical significance.5 Since these analyses were post hoc and the use of aspirin was not randomized, they are best viewed as hypothesis generating, albeit concerning. On the basis of these data and the plausible mechanism, two prospective randomized trials, WASH (Warfarin-Aspirin Study in Heart Failure) and WATCH (Warfarin and Aspirin Trial in Chronic Heart Failure),11–13 were conducted to rigorously test the hypothesis that aspirin interferes with the therapeutic effect of ACE-inhibitors in patients with CHF. WASH was designed as a pilot study, randomizing 279 patients with heart failure and left ventricular systolic dysfunction to warfarin with a target INR of 2.5, aspirin 300 mg daily or no-antithrombotic therapy administered in an open-label manner with blinded endpoint adjudication.11 No significant differences were observed for the primary endpoint of death, non-fatal myocardial infarction, or non-fatal stroke, but more patients assigned to aspirin were hospitalized for cardiovascular reasons (primarily heart failure), with a nominal P = 0.044. The WATCH trial was substantially larger, randomizing 1587 patients with left ventricular ejection fractions (LVEFs) ≤40% to aspirin 162 mg or clopidogrel 75 mg in a double-blind manner or unblinded warfarin with a target INR of 2.5–3.0 with a total of 3073 patient-years of follow-up.13 No significant differences were found for the primary endpoint of death, non-fatal myocardial infarction or stroke. However, as in WASH, more patients in the aspirin group than in the warfarin group were hospitalized for heart failure, the pre-specified secondary endpoint (22.2 vs. 16.5%, respectively, P = 0.019) and a trend toward more frequent heart failure hospitalizations was also found when the aspirin and clopidogrel patients were compared (P = 0.14). The paper by Chang and colleagues provides further insight into the possible interaction of aspirin with another class of renin–angiotensin system blockade in heart failure patients. The CHARM Program enrolled 7599 patients with clinical heart failure into three separate but simultaneously conducted trials: CHARM-Alternative (candesartan vs. placebo in ACE-I intolerant patients with LVEF ≤ 40%), CHARM-Added [candesartan vs. placebo in patients with ongoing ACE-I therapy (enalapril) with LVEF ≤ 40%), and CHARM-Preserved (candesartan vs. placebo in patients with EF > 40%).2 The outcomes of patients were analyzed for interactions based on the use or non-use of aspirin as prescribed by the treating physicians. Unfortunately, the dosages of aspirin employed were not available. The primary analysis was to assess whether there was an interaction between aspirin use at baseline on the effect of candesartan on the primary outcome of cardiovascular death or heart failure hospitalization in the overall three trial programme. These findings are not surprising, given that angiotensin receptor blockers (ARBs) would not be expected to stimulate prostaglandin production, at least not to the same extent as ACE-I, and therefore any modification of the effect of candesartan by aspirin would be expected to be lesser in magnitude or absent. However, the three component trials enrolled very different patient groups, characterized by different background therapy (ACE-I or no ACE-I) and different mechanism of heart failure (LVEF ≤ 40% or >40%) and therefore, it is critical to examine each trial separately, as the authors have done. In the CHARM-Added trial, all patients were receiving an ACE-I.14 If one hypothesizes that aspirin would be expected to have a greater magnitude of effect on the action of the ACE-I than an ARB, this would likely dilute the power to discern a further modification of the effect of candesartan even if such an effect were present. The findings of the CHARM-Preserved trial are also difficult to interpret in this context.15 Since much less is known about the mechanism of this form of heart failure, and there is little information on prostaglandin levels and their physiological effects in this group, as well as no convincing evidence of efficacy for either ACE-I or ARBs, it is difficult to conclude that there is no modification of the candesartan effect in this population. Therefore, in evaluating whether aspirin use modifies the effect of an ARB, the most informative trial is CHARM-Alternative.16 If there is an interaction between candesartan and aspirin, this is the trial in which one would expect it to be apparent. No interaction was found for the primary endpoint of cardiovascular death or non-fatal MI or stroke, but differences were found for both all-cause mortality and cardiovascular mortality. In patients taking aspirin, the proportions of patients who died from any cause were similar (27.3 and 26.4% in the candesartan and placebo groups, respectively); however, candesartan treated patients not taking aspirin had a mortality rate that was 25.1% lower in the candesartan group compared with the placebo group (24.6 vs. 33.1%). These differences in candesartan effect in patients who were and were not taking aspirin resulted in a highly significant interaction (P < 0.01). A lesser effect of candesartan in patients taking aspirin was also found for cardiovascular deaths. In patients taking aspirin, again there was little difference between candesartan and placebo (22.0 vs. 22.3%, respectively, died of cardiovascular causes). In the patients not receiving aspirin, cardiovascular mortality was 28% lower with candesartan treatment (20.7 vs. 28.8%), again a highly significant interaction (P = 0.02; Figure Figure 1). Surprisingly, then, these results do suggest a possible modification of the effect of candesartan on survival by concomitant aspirin therapy in the CHARM-Alternative trial, which enrolled a patient population (low LVEF heart failure) known to benefit from both ACE-I and ARBs and not receiving background ACE-I. However, these findings should be interpreted with caution, since they were not present for the primary endpoint of the trial and were not present for heart failure hospitalizations, for which the aspirin–ACE-I interaction was most evident in SOLVD, WATCH, and WASH. Furthermore, a plausible mechanism for an interaction between aspirin and ARBs is not readily apparent. Nonetheless, further thought and investigation are warranted, particularly since there is no convincing evidence that aspirin has a beneficial effect on outcomes in heart failure patients, and a strong argument can be made for avoiding aspirin in most heart failure patients.17