Author + information
- Received August 21, 2014
- Revision received November 26, 2014
- Accepted January 5, 2015
- Published online March 24, 2015.
- Yao-Jun Zhang, PhD∗,†,
- Javaid Iqbal, PhD‡,
- David van Klaveren, MSc§,
- Carlos M. Campos, MD∗,
- David R. Holmes, MD‖,
- Arie Pieter Kappetein, MD∗,
- Marie-Claude Morice, MD¶,
- Adrian P. Banning, MD#,
- Ever D. Grech, MD‡,
- Christos V. Bourantas, PhD∗,
- Yoshinobu Onuma, PhD∗,
- Hector M. Garcia-Garcia, PhD∗,
- Michael J. Mack, MD∗∗,
- Antonio Colombo, MD††,
- Friedrich W. Mohr, MD‡‡,
- Ewout W. Steyerberg, PhD§ and
- Patrick W. Serruys, MD, PhD∗,‖‖∗ ()
- ∗Thoraxcenter, Erasmus Medical Centre, Rotterdam, the Netherlands
- †Department of Cardiology, Nanjing First Hospital, Nanjing Medical University, Nanjing, China
- ‡Sheffield Teaching Hospitals and the University of Sheffield, Sheffield, United Kingdom
- §Department of Public Health, Erasmus Medical Centre, Rotterdam, the Netherlands
- ‖Mayo Clinic, Rochester, Minnesota
- ¶Institut Jacques Cartier, Massy, France
- #Oxford University Hospitals, Oxford, United Kingdom
- ∗∗The Heart Hospital, Plano, Texas
- ††San Raffaele Scientific Institute, Milan, Italy
- ‡‡Herzzentrum Universität Leipzig, Leipzig, Germany
- ‖‖International Centre for Circulatory Health, Imperial College London, London, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Patrick W. Serruys, Department of Interventional Cardiology, Erasmus University Medical Centre, Thoraxcenter, ‘s-Gravendijkwal 230, P.O. Box 2125, 3000 CC Rotterdam, the Netherlands.
Background Cigarette smoking is a well-known risk factor for development of coronary artery disease (CAD). However, some studies have suggested a “smoker’s paradox,” meaning neutral or favorable outcomes in smokers who have developed CAD, especially myocardial infarction (MI).
Objectives The study aimed to examine the association of smoking status with clinical outcomes in the randomized controlled SYNTAX (SYNergy Between PCI With TAXUS and Cardiac Surgery) trial at 5-year follow-up.
Methods Detailed smoking history was collected at baseline, 6-month, 1-year, 3-year, and 5-year follow-up. The composite endpoints included death/MI/stroke (primary endpoint) plus major adverse cardiac and cerebrovascular events (MACCE) (combination of death/MI/stroke and target lesion revascularization) according to patient smoking status. The comparison of 5-year clinical outcomes between the groups according to smoking status was performed with Cox regression using smoking status at baseline or smoking as a time-dependent covariate.
Results A sizeable proportion (n = 322, 17.9%) of patients had changing smoking status during 5-year follow-up. One in 5 patients with complex CAD was smoking at baseline. However, 60% stopped after revascularization while others continued to smoke. Smokers had worse clinical outcomes due to a higher incidence of recurrent MI in both revascularization arms. Smoking was an independent predictor of the composite endpoint of death/MI/stroke (hazard ratio [HR]: 1.8; 95% confidence interval [CI]: 1.3 to 2.5; p = 0.001) and MACCE (HR: 1.4; 95% CI: 1.1 to 1.7; p = 0.02).
Conclusions Smoking is associated with poor clinical outcomes after revascularization in patients with complex CAD. This places further emphasis on efforts at smoking cessation to improve revascularization benefits. (SYNTAX Study: TAXUS Drug-Eluting Stent Versus Coronary Artery Bypass Surgery for the Treatment of Narrowed Arteries; NCT00114972)
Cigarette smoking is a globally well-recognized risk factor for coronary artery disease (CAD) (1–5). Smokers with ischemic heart disease also have higher incidence of death and other adverse events (5,6). However, some studies have suggested existence of a “smoker’s paradox,” meaning the outcomes in smokers who have developed CAD, especially myocardial infarction (MI), may be neutral or better than nonsmokers (7–9). It is likely that the observed paradox is largely due to differences in the baseline characteristics of smokers and nonsmokers in these studies (7,8), or there may be an interaction between smoking and efficacy of antiplatelet drugs (10,11). However, such reports can have a negative public health impact. Indeed, more than one-third of smokers believe that the dangers of smoking are greatly exaggerated (12). It is, therefore, important to study the effect of smoking on outcomes in patients with established CAD to better understand the effects of smoking on outcomes.
The deleterious effects of smoking after coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) have been shown previously (13–17), although other studies suggested a smoker’s paradox in this context as well (18,19). However, the smoking status in most of these studies is usually taken at baseline and not collected again at regular intervals. Conversely, smokers who learn they have established CAD may stop smoking permanently or intermittently. Therefore, assessing the impact of smoking status at baseline for long-term outcomes may be unreliable. Additionally, most studies were performed in the era of PCI without stents or with bare-metal stents, providing little insight into the impact of smoking in patients undergoing revascularization with drug-eluting stents. Furthermore, the effect of smoking on clinical events in patients with complex CAD undergoing PCI or CABG remains undefined.
This study aimed to examine the smoking status in patients with complex coronary artery disease undergoing PCI or CABG and its prognostic significance in the SYNTAX (SYNergy Between PCI With TAXUS and Cardiac Surgery) trial at the final 5-year follow-up.
The SYNTAX trial (NCT00114972) was a prospective, multicenter, randomized trial that compared PCI with CABG for patients with complex CAD (left main stem and/or 3-vessel disease) (20). Eligible patients (n = 1,800) were randomized on a 1:1 ratio to CABG (n = 897) or PCI (n = 903) with TAXUS Express paclitaxel-eluting stents (Boston Scientific Corporation, Natick, Massachusetts) and followed for 5 years. The study complied with the Declaration of Helsinki and was approved by the ethical review board of the institution involved.
Smoking status was checked for all patients at baseline, 6 months, and 1, 3, and 5 years. The study endpoint was the impact of smoking on a composite endpoint (death/MI/stroke) at 5-year follow-up. We also evaluated impact of smoking on other clinical endpoints including major cardiac and cerebrovascular events (MACCE) (combination of death/MI/stroke/target vessel revascularization). Definitions of various clinical endpoints (e.g., MI, stroke) in the SYNTAX trial have been previously reported (20). All events were adjudicated by an independent clinical event committee comprising interventional cardiologists, cardiac surgeons, and a neurologist (21).
Continuous variables are expressed as mean ± SD and categorical variables are shown as counts and percentages of the total. Comparison of 5-year clinical outcomes between the groups according to the smoking status was performed with Cox regression using smoking status at baseline or smoking as a time-dependent covariate. Hazard ratios (HR) with 95% confidence intervals (CI) are shown. Chi-square test was used to assess interaction p values for treatment arms. A probability value of <0.05 was considered statistically significant. All analyses were performed using SPSS version 21.0 (IBM Corp., Armonk, New York) and STATA version 12.0 (Stata Corp. LP, College Station, Texas).
In the randomized SYNTAX trial (n = 1,800), information on smoking status was available for 1,793 (99.6%) patients. The proportion of patients smoking at different time points is shown in Figure 1. While 20.2% patients with extensive CAD were still smoking at baseline, a significant number stopped smoking after revascularization (8.6% were smokers at 6 months and 8.7% at 1 year). However, after 1 year, there was a modest increase in number of smokers, especially in the CABG arm (10.8% were smokers in the CABG arm vs. 8.7% in the PCI group at 5-year follow-up). Overall, 321 (17.9%) patients had a change in smoking status during follow-up. While nonsmokers remained nonsmokers, current and ex-smokers had considerable change in their smoking status across the 5 years (Table 1).
Comparison of baseline characteristics according to smoking status—stratified as smokers, ex-smokers, and nonsmokers (Online Table 1A) and smoking or nonsmoking at baseline (Online Table 1B)—revealed that smokers were younger and predominantly male and had extensive CAD despite lower prevalence of other cardiovascular risk factors including diabetes mellitus, hypertension, or dyslipidemia. Conversely, smokers had more prevalent chronic obstructive pulmonary disease (COPD), peripheral vascular disease (PVD), and poor left ventricular ejection fraction.
Effect of smoking on clinical outcomes
Baseline smoking status appeared to have no impact on death/MI/stroke and MACCE at 5-year follow-up (Figure 2A). Analyzing the data for those smoking or not smoking at baseline also yielded similar results (Figure 2B). However, considering that the smoking status changed during 5-year follow-up, the analysis with smoking as a time-dependent variable revealed that smoking was associated with increased risk of death/MI/stroke (HR: 1.38; 95% CI: 1.02 to 1.86; p = 0.035) and MACCE (HR: 1.28; 95% CI: 1.01 to 1.61; p = 0.041) at 5-year follow-up (Figure 2C). In total, 1,374 patients never smoked and 98 patients always smoked during 5-year follow-up. The patients who always smoked had significantly higher risk of MI and stent thrombosis/graft occlusion, but not all-cause revascularization, compared with those who never smoked (Online Figure 1).
The baseline smoking status (current smoker vs. ex-smoker vs. nonsmoker) was associated with higher risk of MI during 5-year follow-up (Online Figure 2A). Analyzing the data for those smoking or not smoking at baseline yielded similar results (Online Figure 2B). The predominant effect of smoking as a time-dependent covariate was on subsequent MI (unadjusted HR: 1.86; 95% CI: 1.21 to 2.86; p = 0.005; adjusted HR: 2.08; 95% CI: 1.30 to 3.32; p = 0.002) (Figure 3, Table 2), whereas no statistically significant effect was observed on mortality or repeat revascularization at 5-year follow-up (Figure 3). The interaction p values for these outcomes between CABG and PCI arms were not significant, suggesting smoking was associated with poor outcomes independent of revascularization strategy (Figure 3).
On multivariable Cox regression analysis using smoking status as a time-dependent covariate, smoking was an independent predictor of death/MI/stroke as well as MACCE (Table 3). Subgroup analysis showed that smoking was an independent predictor of poor outcomes in CABG patients (HR: 1.52; 95% CI: 1.02 to 2.25; p = 0.038) as well as the PCI arm (HR: 1.26; 95% CI: 0.90 to 1.75; p = 0.177). After adjusting for other independent predictors, baseline smoking status also was associated with poor outcomes (Central Illustration).
The most important finding in this paper is that among patients with complex CAD undergoing revascularization with either PCI or CABG, those who are smokers have a higher incidence of adverse outcomes including MI than those who never smoked or stopped smoking. This finding is seen irrespective of the specific revascularization strategy.
The SYNTAX trial enrolled patients with complex CAD requiring coronary revascularization. It is noteworthy that 1 in 5 patients with extensive coronary disease were still smoking at the time of trial enrollment. These data are consistent with other studies showing that a sizeable proportion (10% to 30%) of patients with established CAD continue to smoke (18,22–24). The number of smokers was halved after coronary revascularization, suggesting that most of the patients took the smoking cessation advice at the time of revascularization seriously (24). The time of revascularization, therefore, presents a good opportunity to reinforce smoking cessation advice and to offer practical help.
Smoking is a modifiable risk factor, and not all current smokers remain smokers after a coronary event or procedure. Conversely, patients may start smoking again at any time during follow-up. Thus, using baseline smoking status to predict long-term outcomes is a potentially flawed approach. We have, therefore, used smoking status as a time-dependent variable. Indeed, our results confirm that this approach, not unadjusted baseline smoking status, is able to identify the link between smoking and poor outcomes after revascularization. Another study looking at the effect of smoking in patients undergoing CABG has suggested that smoking status at baseline (pre-surgery) was not associated with adverse outcomes but smoking status at follow-up (post-surgery) predicted poor outcomes (13).
Impact of smoking on outcomes: no paradox exists
Smoking was associated with adverse outcomes, especially recurrent MI in our study. This was seen after adjusting for confounding variables or using smoking as a time-dependent covariate. Smoking is a well-established risk factor for acute MI; quitting reduces the risk of subsequent myocardial infarctions (25). It has also been shown in multiple studies that a ban on smoking substantially reduces the incidence of MI, even at the population level (26–28). Adherence to behavioral advice, including smoking cessation, after an acute coronary syndrome is associated with a substantially lower risk of recurrent cardiovascular events (29). There was no significant effect of smoking on repeat revascularization, consistent with previous reports (18,30), although some studies have reported opposing results (31). It is possible that drug-eluting stents are associated with a lower incidence of target lesion revascularization even in those who continue to smoke (30). Our study did not show any statistically significant influence of smoking on mortality despite a 17% higher relative risk of mortality in smokers, possibly reflecting lack of adequate power to demonstrate an effect on mortality. Other studies with larger sample sizes have shown higher mortality in smokers undergoing CABG or PCI (15,23). It has also been shown that smokers demonstrate less improvement in quality of life after coronary revascularization (32). Therefore, it is prudent to highlight that smoker’s paradox does not exist for patients undergoing coronary revascularization and that smoking cessation before such revascularization is strongly advised to improve outcomes.
Smoking is associated with poor outcomes after coronary revascularization with PCI or CABG. This message needs to be disseminated to patients, primary and secondary care physicians, and the general population. It is important to inquire about smoking status at each clinical encounter, and appropriate advice should be offered to help patients stop smoking. Dedicated smoking cessation programs may help patients achieve smoking cessation (33). However, these patients remain at risk of relapse. Patients who had bypass surgery are more prone to return to smoking than those who had an MI (34). Primary care and community-based physicians and nurses may play an important role improving smoking cessation rates.
Study strengths and limitations
The main strength of our study is that smoking status was checked at serial time points in this all-comers clinical trial in patients with complex CAD. However, this is a post-hoc analysis and has limitations inherent to any such analysis. The smoking status presented here is self-reported; there was no biochemical testing (e.g., cotinine measurement) or cross validation of smoking status. Number of cigarettes smoked by current or ex-smokers was not recorded. However, it could be argued that for cardiovascular outcomes, including MI and stroke, smoking status is more important than actual number of cigarettes smoked; in the CADILLAC (Controlled Abciximab and Device Investigation to Lower Late Angioplasty Complications) trial, 1-year mortality rates were similar among those smoking less than a half pack per day, half to 1 pack per day, or more than 1 pack per day (19). Smoking may have an acute effect on platelets and endothelial function, leading to arterial thrombosis and MI (35,36).
Another limitation is that the smoking status was collected at pre-defined intervals, and exact date of change was not available. For example, if a patient was smoking at 1 year and not smoking at 3 years, we considered the patient a smoker at 1 year and ex-smoker at 3 years, although the patient could have stopped smoking at any time between 1 and 3 years. However, we also performed a sensitivity analysis in a completely opposite manner with the assumption that patient stopped smoking just after the last visit (data not shown), and the results remained unchanged.
Finally, the association does not confer a cause-effect relationship. It is plausible that those who continue to smoke may have additional lifestyle risk factors (e.g., lack of exercise, unhealthy diet) that could be contributing to the adverse outcomes observed.
Baseline smoking status may not provide adequate information to predict the longer-term effects of smoking on clinical outcomes due to changes in smoking pattern during follow-up. Smoking is associated with adverse clinical outcomes, especially MI, after revascularization for complex CAD. Abstinence from smoking may improve the outcomes achieved with coronary revascularization, and all patients undergoing PCI or CABG should be encouraged to stop smoking indefinitely before revascularization. Smoking cessation programs may help improve the benefits achieved with revascularization.
COMPETENCY IN MEDICAL KNOWLEDGE: Cigarette smoking is an established risk factor for development of coronary artery disease, and studies suggesting that smokers with coronary disease might enjoy paradoxically better outcomes than nonsmokers have not sufficiently considered baseline differences in clinical risk factors or changes in smoking status over time. After PCI or CABG, smokers with complex coronary disease face a higher incidence of myocardial infarction, and abstinence from smoking improves outcomes.
COMPETENCY IN PATIENT CARE: All patients undergoing PCI or CABG should be encouraged to completely and permanently stop smoking.
TRANSLATIONAL OUTLOOK: Additional work is needed to develop more effective and well-tolerated methods to facilitate cessation and sustained abstinence from cigarette smoking.
The authors express their gratitude to all study centers and participants in the SYNTAX trial, whose work made this study possible.
For supplemental tables and figures, please see the online version of this article.
Dr. Banning has received research grant support from Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Zhang and Iqbal contributed equally to this work. John Ambrose, MD, served as the Guest Editor for this paper.
- Abbreviations and Acronyms
- coronary artery bypass grafting
- confidence interval
- hazard ratio
- major adverse cardiac and cerebrovascular event(s)
- myocardial infarction
- percutaneous coronary intervention
- Received August 21, 2014.
- Revision received November 26, 2014.
- Accepted January 5, 2015.
- American College of Cardiology Foundation
- Sakata R.,
- McGale P.,
- Grant E.J.,
- et al.
- Ambrose J.A.,
- Barua R.S.
- Ezzati M.,
- Henley S.J.,
- Thun M.J.,
- et al.
- Barbash G.I.,
- Reiner J.,
- White H.D.,
- et al.
- Andrikopoulos G.K.,
- Richter D.J.,
- Dilaveris P.E.,
- et al.
- Violaris A.G.,
- Thury A.,
- Regar E.,
- et al.
- Gagne J.J.,
- Bykov K.,
- Choudhry N.K.,
- et al.
- Gurbel P.A.,
- Bliden K.P.,
- Logan D.K.,
- et al.
- ↵Department of Health (United Kingdom). Hidden Dangers Highlighted in New Smoking Campaign. 2012. Available at: https://www.gov.uk/government/news/hidden-dangers-highlighted-in-new-smoking-campaign. Accessed January 11, 2015.
- van Domburg R.T.,
- Meeter K.,
- van Berkel D.F.,
- et al.
- Cohen D.J.,
- Doucet M.,
- Cutlip D.E.,
- et al.
- Prugger C.,
- Wellmann J.,
- Heidrich J.,
- et al.
- Scholte op Reimer W.,
- de Swart E.,
- De Bacquer D.,
- et al.
- Bartecchi C.,
- Alsever R.N.,
- Nevin-Woods C.,
- et al.
- Meyers D.G.,
- Neuberger J.S.,
- He J.
- Chow C.K.,
- Jolly S.,
- Rao-Melacini P.,
- et al.
- Taira D.A.,
- Seto T.B.,
- Ho K.K.,
- et al.
- Quist-Paulsen P.,
- Gallefoss F.
- Hajek P.,
- Taylor T.Z.,
- Mills P.
- Levine P.H.