Author + information
- Received April 18, 2013
- Revision received November 7, 2013
- Accepted November 21, 2013
- Published online April 1, 2014.
- Jason O. Robertson, MD, MS∗,
- Ramin Ebrahimi, MD†,
- Alexandra J. Lansky, MD‡,
- Roxana Mehran, MD§,‖,
- Gregg W. Stone, MD‖,¶ and
- A. Michael Lincoff, MD∗∗ ()
- ∗Department of Cardiovascular Medicine, Cleveland Clinic, Cleveland, Ohio
- †University of California Los Angeles and the Greater Los Angeles VA Medical Center, Los Angeles, California
- ‡Yale University School of Medicine, New Haven, Connecticut
- §Icahn School of Medicine at Mount Sinai, New York, New York
- ‖Cardiovascular Research Foundation, New York, New York
- ¶Columbia University Medical Center, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. A. Michael Lincoff, Department of Cardiovascular Medicine/F25, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, Ohio 44195.
Objectives This study sought to evaluate the short- and long-term outcomes for smokers with non–ST-segment elevation acute coronary syndromes (NSTE-ACS).
Background Smoking has been associated with the “paradox” of reduced mortality after acute myocardial infarction (MI). This is thought to be due to favorable baseline characteristics and less diffuse coronary artery disease (CAD) among smokers.
Methods In the ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy) trial, 13,819 patients (29.1% smokers) with moderate- to high-risk NSTE-ACS underwent angiography and, if indicated, revascularization.
Results Smokers were significantly younger and had fewer comorbidities than nonsmokers. Incidence of death and MI were comparable at 30 days, although smokers had significantly reduced risks of 30-day major bleeding (hazard ratio [HR]: 0.80, 95% confidence interval [CI]: 0.67 to 0.96; p = 0.016) and 1-year mortality (HR: 0.797, 95% CI: 0.65 to 0.97; p = 0.027). After correction for baseline and clinical differences, smoking was no longer predictive of major bleeding (odds ratio: 1.06, 95% CI: 0.86 to 1.32; p = 0.56) and was associated with higher 1-year mortality (HR: 1.37, 95% CI: 1.07 to 1.7; p = 0.013). This pattern of reversed risk after multivariable correction held true for those smokers requiring percutaneous coronary intervention. Core laboratory angiographic analysis showed that smokers and nonsmokers were comparable in terms of the extent of CAD, Thrombolysis In Myocardial Infarction flow, myocardial blush, and the presence of thrombi.
Conclusions In contrast to the paradox previously described in ST-segment elevation MI, our analysis finds smoking to be an independent predictor of higher 1-year mortality in patients presenting with NSTE-ACS, and our angiographic study demonstrates CAD in smokers that is comparable to that in nonsmokers but evident ∼1 decade earlier. (Acute Catheterization and Urgent Intervention Triage Strategy [ACUITY]; NCT00093158)
ST-segment elevation myocardial infarction (STEMI) and non–ST-segment elevation acute coronary syndromes (NSTE-ACS) result from similar pathophysiological processes but comprise a spectrum of disease severity (1). The primary difference is that diagnosis of STEMI usually indicates the presence of an occlusive thrombus, with the paradigm that immediate myocardial reperfusion by thrombolysis or mechanical revascularization will improve patient outcome (2). Despite this, recent studies show that in-hospital and 1-year mortality are similar between the 2 groups (1).
The U.S. Department of Health and Human Services estimates that as many as 30% of all coronary heart disease deaths may be attributed to cigarette smoking (3). Due in part to the prothrombotic effects of smoking (4), cigarette smokers are more likely to present with STEMI than with NSTE-ACS (1,5). Several studies have demonstrated a higher incidence of acute myocardial infarction (MI) but improved survival after reperfusion among smokers (6–9). This phenomenon, termed the smoker's paradox, has been postulated to be due to the fact that smokers present at a younger age, typically with fewer comorbidities (6–8), and with a higher incidence of thrombo-occlusive disease that is optimally treatable with intravenous thrombolysis (10). Angiographic evidence suggests that smokers presenting with acute MI also have less extensive coronary artery disease (CAD) than their nonsmoking counterparts (11–13). Nevertheless, smoking has been associated in a dose-dependent fashion with the development of coronary artery and aortic atherosclerosis in several autopsy studies (14,15), and current smoking was recently confirmed as an independent predictor of greater plaque burden as assessed by intravascular ultrasound imaging (16).
Smokers account for ∼20% to 30% of patients with NSTE-ACS (1,5,17), but few data exist regarding the relationship between smoking and prognosis in NSTE-ACS. Given that thrombolysis is not an effective treatment strategy for NSTE-ACS (18) and that patients with non–ST-segment elevation myocardial infarction (NSTEMI) typically present at an older age with significantly more cardiovascular comorbidities than those with STEMI (1), the effect of smoking on the extent of CAD and prognosis may vary between populations of patients with different types of ACS. The ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy) trial (19) enrolled 13,819 patients with unstable angina and NSTEMI. In the current study, we evaluated short- and long-term outcomes for smokers with NSTE-ACS for both the overall study population of the ACUITY trial and for the subset requiring percutaneous coronary intervention (PCI). Furthermore, we assessed angiographic data to correlate outcomes with the magnitude of vascular disease burden as a function of smoking status.
Study design and protocol
The detailed methodology for the ACUITY trial was previously published (19). In short, the study prospectively randomized 13,819 patients with moderate- to high-risk NSTE-ACS (unstable angina and NSTEMI) from ∼600 different centers to receive enoxaparin or unfractionated heparin plus glycoprotein IIb/IIIa inhibitors, bivalirudin plus glycoprotein IIb/IIIa inhibitors, or bivalirudin monotherapy. Inclusion criteria were at least 10 min of cardiac chest pain within 24 h of presentation in addition to 1 of the following: 1) troponin or creatine kinase-myocardial band (CK-MB) elevation; 2) dynamic electrocardiogram changes; or 3) documented previous CAD. Alternatively, patients with chest pain could be included if they met all 4 variables for predicting Thrombolysis In Myocardial Infarction (TIMI) risk scores for unstable angina (20): 1) 65 years of age and older, 2) aspirin taken within the past 7 days; 3) ≥2 episodes of angina in the past 24 h; and 4) a history of ≥3 of the following risk factors: hypertension, hypercholesterolemia, family history of CAD, diabetes, or current cigarette smoking.
Coronary angiography was performed using a standardized acquisition protocol and analyzed by an independent angiographic core laboratory (Cardiovascular Research Foundation, New York, New York). Complete 3-vessel and lesion-specific qualitative and quantitative angiographic analysis and quantitative left ventriculography were performed using standardized criteria by reviewers blinded to randomization assignment and clinical outcome and reviewed for accuracy by a qualified physician. Vessel disease was defined as quantitative coronary angiography diameter stenosis ≥30%. Epicardial coronary flow was assessed using TIMI flow grade (21) and corrected TIMI frame count (22). Myocardial blush grade was assessed before and after PCI based on standard methodology (23). By protocol, the first 7,000 consecutive patients enrolled in the United States were analyzed by the core angiographic laboratory.
Endpoint and other definitions
The primary endpoints for the present analysis of outcomes among smokers in the ACUITY trial were 1-year mortality and 30-day major bleeding (unassociated with coronary artery bypass surgery). Secondary endpoints were a 30-day composite ischemia endpoint of death, MI, and urgent target vessel revascularization, and individual components of that composite. These variables have been defined elsewhere (19,24). Smokers were identified as patients who reported having smoked cigarettes within 30 days of randomization.
Baseline characteristics and outcomes are summarized as means and SDs for continuous variables and as frequencies and percentages for categorical variables. Missing/unknown data values were not imputed, and, in the presentation of categorical variables, denominators excluded unknown or missing values. The nonparametric Wilcoxon rank sum test was used for comparison of continuous variables. Categorical variables were compared using chi-square statistics. A p value <0.05 was considered statistically significant.
Univariate analyses were completed for 30-day major bleeding and 1-year mortality using time-to-event methodology, and the corresponding hazard ratios (HRs) and 95% confidence intervals (CIs) were determined. A multivariate Cox proportional hazards regression analysis, which was selected to take into account time to event data, was then performed on these variables. Covariates were selected using a forward stepwise procedure from a large number of clinical and demographic candidate variables: 65 years of age and older, smoking status, anemia, study arm, baseline troponin/CK-MB elevation, baseline creatinine clearance <60 ml/min, diabetes, electrocardiogram changes, sex, hyperlipidemia (on medication), hypertension (on medication), procedure performed, PCI versus medical management, previous MI, previous percutaneous transluminal coronary angioplasty, previous CABG, and previous angioplasty/PCI. A significance level of 0.15 was required to allow a variable into the models, and a significance level of 0.10 was required to allow a variable to stay in the models. For overall 1-year mortality, the final model included the following variables: 65 years of age and older, anemia, baseline CK-MB/troponin elevation, baseline creatinine clearance <60 ml/min, procedure performed, smoking status, diabetes, sex, electrocardiogram changes, hypertension (on medication), previous CABG, and previous MI. All of these variables made it into the final model for 1-year mortality among PCI, excluding procedure performed, sex, and hypertension (on medication). The final bleeding model included the following additional variables: previous percutaneous transluminal coronary angioplasty, hyperlipidemia, and study arm randomization code. All statistical analyses were performed by SAS version 8.2 (SAS Institute Inc., Cary, North Carolina).
Outcomes based on smoking status for both the overall population and the subset of patients requiring PCI
Table 1 lists the characteristics of patients based on smoking status. Smokers were on average a decade younger at presentation, more commonly men, and more likely to have received heparin before randomization compared with nonsmokers. Smokers also had fewer comorbidities: they less commonly had diabetes, hypertension, hyperlipidemia, anemia, renal insufficiency, previous MI, and previous coronary intervention. Baseline characteristics for smokers and nonsmokers undergoing PCI were similar to those for the overall population.
Table 2 compares 30-day and 1-year outcomes for smokers and nonsmokers. Smokers had significantly lower incidences of 30-day major bleeding and 1-year mortality. When outcomes were compared among the subset of patients requiring PCI, smokers maintained a significantly lower 1-year mortality, but exhibited only a trend toward reduced 30-day major bleeding (Table 3). These mortality differences are illustrated as Kaplan-Meier curves for both the overall population (3.4% vs. 4.4%) (Fig. 1A) and those patients undergoing PCI (2.6% vs. 4.0%) (Fig. 1B).
Figure 2 shows the HR and respective CIs for the effect of smoking on 1-year mortality and 30-day major bleeding that were derived from both the univariate and multivariate models. In the univariate model for 1-year mortality, smoking was associated with a 20% decreased risk in the overall population (HR: 0.80, 95% CI: 0.65 to 0.98) and a 31% decreased risk among PCI patients (HR: 0.69, 95% CI: 0.51 to 0.92). After accounting for all significant covariables, however, smoking was associated with a significant 37% increased risk for mortality at 1 year among the overall population (HR: 1.37, 95% CI: 1.07 to 1.75) and a 28% increased risk of 1-year mortality among patients undergoing PCI (HR: 1.28, 95% CI: 0.89 to 1.84). Interaction effects were examined for the covariates in the model, and although diabetes and previous MI were statistically significant for death, the fact that there were very few deaths in each subgroup makes inferences regarding these data difficult.
Similar trends were seen in the analysis for 30-day major bleeding (Fig. 2). Univariate models demonstrated 20% and 16% lower risk for smokers in the overall (HR: 0.80, 95% CI: 0.67 to 0.96) and PCI (HR: 0.84, 95% CI: 0.68 to 1.03) populations, respectively. After multivariate correction, there were no significant differences in bleeding risks related to smoking in the overall population (HR: 1.06, 95% CI: 0.86 to 1.32) or among PCI patients (HR: 1.08, 95% CI: 0.86 to 1.37). No interaction effects were noted for the covariates with respect to 30-day major bleeding.
PCI was required in 7.5% and 7.1% of smokers and nonsmokers over 1-year follow-up, respectively (HR: 1.067, 95% CI: 0.93 to 1.22; p = 0.35). Similarly, there was no difference between the 2 groups in the 1-year incidences of coronary artery bypass graft surgery (1.4% vs. 1.5%, HR: 0.91, 95% CI: 0.67 to 1.25; p = 0.56).
Angiographic comparison of smokers and nonsmokers
As seen in Table 4, smokers and nonsmokers were largely comparable in terms of the number of vessels with CAD, the extent of disease per vessel, TIMI flow, myocardial blush, and the presence of thrombi. The specific coronary arteries with disease are delineated for each population. Smokers had significantly less calcification and fewer collaterals present.
Smokers more commonly had low TIMI flow (0/1) and low myocardial blush grade (0/1) pre-PCI (Table 5); however, angiographic outcomes for smokers requiring PCI were similar to those for nonsmokers.
In this study, we have shown in a large, contemporary cohort of patients with NSTE-ACS that smoking is an independent predictor of higher 1-year mortality. Although univariate analysis suggested a protective effect of smoking that is similar to that observed among populations with STEMI, this mortality advantage of smokers did not withstand multivariate analysis accounting for differences in baseline characteristics. This finding provides compelling evidence that the smoker's paradox among NSTE-ACS patients may be explained in similar fashion to that seen in patients with acute MI (12). Additionally, we demonstrated that smokers presenting with NSTE-ACS have an extent and a complexity of CAD that are largely comparable to those seen in nonsmokers, despite disparities in age and other risk factors. These results vary from what has been previously observed of the relationship between smoking and mortality and CAD in patients with STEMI.
The demographic and clinical profiles of the smokers enrolled in the ACUITY trial are comparable to those of smokers in other large studies of patients with ACS (5,12,25), although, to our knowledge we are the first to demonstrate that smoking is independently predictive of higher 1-year mortality after multivariate analysis among patients with NSTE-ACS. When multivariate regression was performed on outcomes from the GRACE (Global Registry of Acute Coronary Events) database, consistent with our data, smoking within 30 days of admission was not a predictor of in-hospital mortality for patients with either unstable angina or NSTEMI (5), but late mortality was not examined. In a smaller retrospective study by Castela et al. (26) consisting of patients with both NSTE-ACS and STEMI, it was shown that smokers have higher rates of vascular complications but similar mortality at 1 year. Differences between the results of that study and ours may be attributable to smaller sample size, inclusion of STEMI patients, and a different definition of what constitutes a smoker (i.e., any smoking within the past year) in the Castela et al. (26) analysis. Other studies have demonstrated that smoking is associated with an increased relative risk of cardiovascular death (27) and higher mortality among younger patients who present with acute MI (28), but these results are derived from many years of follow-up data rather than the relatively early 1-year results that we present, and they are not specifically representative of patients with NSTE-ACS. At 1 year, smoking was, in fact, not predictive of mortality from STEMI in the CADILLAC (Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications) trial (25).
It has been shown that major bleeding has a large impact on early mortality among patients with NSTE-ACS (29). In our study, the reduced rates of 30-day major bleeding did not withstand multivariate analysis. However, neither was an excess risk of bleeding observed among smokers, suggesting that the influence of smoking on mortality may be largely explained by age and clinical comorbidities.
Atherosclerotic disease of the coronary arteries is known to have an impact on the development and severity of acute coronary syndromes. Angiographic CAD burden has been shown to increase with the number of documented risk factors in NSTEMI patients (30). Given that smokers with NSTE-ACS, similar to those with STEMI (12,25), presented an average of a decade earlier with fewer risk factors than nonsmokers, one would expect to observe less multivessel disease, higher TIMI flow and myocardial blush grades, less calcification, and the development of fewer collateral vessels in smokers compared with nonsmokers with NSTE-ACS. Indeed, angiographic substudies of the GUSTO-I (Global Utilization of Streptokinase and Tissue-Plasminogen Activator for Occluded Coronary Arteries), TIMI, and CADILLAC trials demonstrated significantly lower incidences of triple-vessel disease and significantly higher incidences of TIMI 3 flow among smokers with acute STEMI (11,12). These differences were not noted in the population of smokers from the ACUITY trial.
Our study demonstrates significantly more multivessel disease overall than was observed among patients with STEMI in the GUSTO-I, TIMI, and CADILLAC trials. In those trials, smokers and nonsmokers with acute MI had incidences of triple-vessel disease ranging from 12% to 17% and 18% to 22%, respectively, compared with incidences of 42% and 45% for the corresponding groups of patients with NSTE-ACS in the ACUITY trial (11,12,25). Incidences of multivessel disease in our study are similar to those observed in a smaller angiographic analysis of 521 patients with ACS in whom unstable angina accounted for 43% and 60% of patients in the smoker and nonsmoker groups, respectively (26). However, variant definitions of what constitutes vessel disease could explain some of these differences, so further studies are required to determine whether the extent of CAD is greater in smokers who present with NSTE-ACS compared with acute MI.
We did observe anticipated trends toward less coronary artery calcification and collateral vessel formation among smokers. This was expected because calcific deposits are more commonly seen in advanced lesions and with increasing age, but the correlation is not absolute, and the prognostic significance of calcification is controversial (31). As observed in our data, the degree of calcification does not necessarily correlate with the degree of stenosis, as the progression of CAD is neither linear nor predictable (32,33). Rapidly progressing lesions, which are thought to form from plaque disruption and subsequent thrombus formation of pre-existing lipid-rich lesions (34,35), may account for up to two-thirds of patients in whom ACS develops (33,36). The major predictor of collateral vessel development, on the other hand, is severity of coronary stenosis (37), but we observed only modest differences in collateral formation in smokers versus nonsmokers. Thus, these differences, although significant, are likely of less clinical importance than the measures of extent and severity of disease discussed previously.
The similarities in the CAD burden between smokers and nonsmokers are also evident in the proportions of patients requiring coronary artery bypass graft or PCI. We observed no difference between groups, but previous studies of ACS patients that included large percentages of STEMI patients have demonstrated that smokers are less likely to require coronary artery bypass graft due to favorable coronary anatomy (5,26) and more likely to require PCI (5). The differences observed in our study likely relate to the similar extent of CAD noted in smokers and nonsmokers with NSTE-ACS. When the subpopulation of patients who required PCI was considered, a trend was noted for slightly more diseased vessels among smokers, although they responded equally well as nonsmokers to intervention.
First, former smokers were not identified by our analysis, and no data were available regarding the number of cigarettes smoked or smoking status after PCI. Second, we only had data regarding a subset of the angiograms obtained. Third, the number of variables that we are able to include in our multivariate analysis is limited by the event rate within the study. The large changes in the HR estimates of smoking suggests that many covariates influence this estimate, and there may be additional confounders that are not controlled for within our model. Nevertheless, this is the largest core laboratory analysis to date comparing smokers and nonsmokers in any patient population, both in terms of outcomes and angiographic data, and we believe that we have accounted for the most clinically-relevant variables in our model.
This study has demonstrated a clear and important impact of smoking on mortality and development of CAD in patients with NSTE-ACS. Smoking is associated with the development of CAD and the presentation with ACS nearly a decade earlier than in nonsmokers and portends a significant increase in subsequent mortality. NSTEMI patients are less likely than STEMI patients to be discharged with appropriate pharmacological therapies after hospitalization (1), which could additionally affect disease progression and prognosis. Because a significant population of NSTE-ACS patients are smokers, our findings highlight the importance of continued efforts at smoking cessation and use of appropriate secondary prevention measures in this population.
Dr. Ebrahimi is a speaker and consultant for sanofi-aventis, AstraZeneca, Boehringer Ingelheim, Janssen, Abbott, The Medicines Company, Bristol-Myers Squibb, Pfizer, and Amarin. Dr. Mehran has received institutional research grant support from The Medicines Company, Bristol-Myers Squibb/Sanofi, and Eli Lilly/Daiichi-Sankyo Covidien, and Maya Medical; and is a consultant to Abbott Vascular, AstraZeneca, Janssen Pharmaceuticals, Regado Biosciences, The Medicines Company, Bristol-Myers Squibb/Sanofi, and Merck. Dr. Stone has served as a consultant to Abbott Vascular, Boston Scientific, Medtronic, and The Medicines Company. Dr. Lincoff receives grant support from AstraZeneca and Daiichi-Sankyo/Eli Lilly. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- creatine kinase-myocardial band
- hazard ratio
- myocardial infarction
- non–ST-segment elevation acute coronary syndromes
- non–ST-segment elevation myocardial infarction
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Thrombolysis In Myocardial Infarction
- Received April 18, 2013.
- Revision received November 7, 2013.
- Accepted November 21, 2013.
- American College of Cardiology Foundation
- Montalescot G.,
- Dallongeville J.,
- Van Belle E.,
- et al.
- Antman E.M.,
- Hand M.,
- Armstrong P.W.,
- et al.
- United States Department of Health and Human Services
- Grines C.L.,
- Topol E.J.,
- O'Neill W.W.,
- et al.
- Lee K.L.,
- Woodlief L.H.,
- Topol E.J.,
- et al.,
- GUSTO-I Investigators
- de Chillou C.,
- Riff P.,
- Sadoul N.,
- et al.
- Barbash G.I.,
- Reiner J.,
- White H.D.,
- et al.
- Nicholls S.J.,
- Tuzcu E.M.,
- Crowe T.,
- et al.
- Ambrose J.A.,
- Almeida O.D.,
- Sharma S.K.,
- et al.
- Gibson C.M.,
- Cannon C.P.,
- Daley W.L.,
- et al.
- van 't Hof A.W.,
- Liem A.,
- Suryapranata H.,
- Hoorntje J.C.,
- de Boer M.J.,
- Zijlstra F.
- Manoukian S.V.,
- Feit F.,
- Mehran R.,
- et al.
- Wexler L.,
- Brundage B.,
- Crouse J.,
- et al.
- Falk E.,
- Shah P.K.,
- Fuster V.
- Fuster V.
- Davies M.J.,
- Richardson P.D.,
- Woolf N.,
- Katz D.R.,
- Mann J.
- Falk E.
- Steinberg D.
- Seiler C.