Association between smoking, outcomes, and early clopidogrel use in patients with acute coronary syndrome: insights from the Global Registry of Acute Coronary Events

Abstract

BACKGROUND: Smoking induces CYP1A2, thereby enhancing clopidogrel conversion to its active metabolite. We sought to determine the association between clopidogrel use and clinical outcomes in smokers versus nonsmokers with a broad spectrum of acute coronary syndrome (ACS). METHODS: We examined the association between early clopidogrel use in-hospital and 6-month outcomes among 44,426 patients with ACS in relation to smoking status in the Global Registry of Acute Coronary Events. We tested for heterogeneity of clopidogrel effect among smokers versus nonsmokers in separate multivariable models that adjusted for (1) established prognosticators in the Global Registry of Acute Coronary Events risk score and (2) independent predictors of major bleeding. RESULTS: Rates of in-hospital mortality, death/myocardial infarction, and major bleeding were 4.3%, 5.9%, and 2.5%, respectively. Current smokers (n = 12,149) were more likely to be younger men without documented vascular disease; had lower rates of hypertension, hyperlipidemia, and diabetes; and more frequently presented with ST elevation (all P < .0001). Early clopidogrel use (55%) was associated with a reduction in the composite endpoint of mortality and myocardial infarction both in-hospital and at 6 months among current smokers and nonsmokers. There was no interaction between current smoking and clopidogrel use for ischemic endpoints. Major bleeding associated with early clopidogrel use was actually lower among current smokers compared with nonsmokers. CONCLUSIONS: Despite prior observations of smoking-enhanced clopidogrel effects, early clopidogrel use among smokers presenting with ACS compared with nonsmokers was not independently associated with a greater reduction in cardiovascular events. In contrast with nonsmokers, clopidogrel use among smokers was not associated with excess bleeding, perhaps because of unmeasured confounders.