Response to Letter Regarding Article, “Aspirin Use and Outcomes in a Community-Based Cohort of 7352 Patients Discharged After First Hospitalization for Heart Failure”
We thank Drs Guazzi and Reina for their interest in our work.1 They believe the 1-year time frame of our study was too short, and they cite their study2 as evidence that a dose-dependent angiotensin-converting enzyme (ACE) inhibitor–aspirin interaction only becomes manifest over a longer time frame. We are not convinced, however, by their finding of a modest interaction that only achieved statistical significance after 3 years (hazard ratio 1.03; 95% confidence interval 1.01 to 1.05) and that was driven by only 35 deaths in the 64 patients treated with high doses of aspirin (≥325 mg daily) in their study.2 In addition, as we understand their study, they classified the 344 study participants on the basis of medication doses at cohort inception.
Thus, to the provisos about observational data raised both in our article1 and the accompanying editorial,3 we would add 2 further cautions. First, although it is well recognized that studies with small sample sizes are susceptible to type 2 error, it should be acknowledged that small studies can also produce type 1 errors (ie, spuriously declaring a difference exists when it does not).4 Second, given the extent to which patient adherence with prescribed therapies decays over time, the validity of attributing differences in long-term outcomes to baseline exposures become increasingly tenuous as the duration of follow-up increases. If there were long-term hazards to the coadministration of aspirin and ACE inhibitors, we would have expected this finding to have emerged from the meta-analysis of long-term ACE inhibitor trials (in which the mean length of follow-up exceeded 36 months); however, ACE inhibitor–treated patients demonstrated a clear reduction in major clinical outcomes whether they were taking aspirin at baseline (odds ratio 0.80, 95% confidence interval 0.73 to 0.88) or were not (odds ratio 0.71, 95% confidence interval 0.62 to 0.81).5
Considering all of the available evidence published thus far, our opinion remains that (1) the use of aspirin is not associated with worse cardiovascular outcomes in patients with heart failure, and (2) there is no strong evidence for a negative aspirin–ACE inhibitor interaction, even in those patients in which such an effect would be most likely to be seen (nonischemic patients, those with renal dysfunction, or patients treated with high doses of aspirin and low doses of an ACE inhibitor).
McAlister FA, Ghali WA, Gong Y, Fang J, Armstrong PW, Tu JV. Aspirin use and outcomes in a community-based cohort of 7352 patients discharged after first hospitalization for heart failure. Circulation. 2006; 113: 2572–2578.
Jhund P, McMurray JJV. Does aspirin reduce the benefit of an angiotensin-converting enzyme inhibitor? Choosing between the Scylla of observational studies and the Charybdis of subgroup analysis. Circulation. 2006; 113: 2566–2568.
Montori VM, Devereaux PJ, Adhikari NKJ, Burns KE, Eggert CH, Briel M, Lacchetti C, Leung TW, Darling E, Bryant DM, Bucher HC, Schunemann HJ, Meade MO, Cook DJ, Erwin PJ, Sood A, Sood R, Lo B, Thompson CA, Zhou Q, Mills E, Guyatt GH. Randomized trials stopped early for benefit: a systematic review. JAMA. 2005; 294: 2203–2209.
Teo KK, Yusuf S, Pfeffer M, Kober L, Hall A, Pogue J, Latini R, Collins R, for the ACE Inhibitors Collaborative Group. Effects of long-term treatment with angiotensin converting enzyme inhibitors in the presence or absence of aspirin: a systematic review. Lancet. 2002; 360: 1037–1043.