Author + information
- Published online October 7, 2019.
- aDeutsches Herzzentrum München, Technische Universität München, Munich, Germany
- bDZHK (German Centre for Cardiovascular Research), partner site Munich Heart Alliance, Munich, Germany
- ↵∗Address for correspondence:
Dr. Michael Joner, Deutsches Herzzentrum München, Technische Universität München, Lazarettstrasse, 36, 80636 Munich, Germany.
Cardiovascular disease and cancer represent the 2 major disease manifestations of smoking, with the former accounting for approximately 48% of smoking-related deaths (1). For this reason, smoking is universally regarded as the strongest behavioral risk factor for the premature onset of atherosclerosis, which often leads to fatal atherothrombotic events, including myocardial infarction (MI) and sudden cardiac death (2). Hydrocarbons in smoke were shown to cause injury to the endothelium, to promote sustained low-grade inflammation, and to increase plasma fibrinogen levels (3), finally inducing platelet aggregation and accelerating atherosclerosis (4). Additional components of cigarette smoke such as nicotine have been associated with augmented catecholamine release (5), which increases heart rate and blood pressure, while predisposing to arrhythmia and vasoconstriction, all adversely affecting cardiovascular health.
The term “smoker’s paradox” was coined and released into scientific discourse more than 25 years ago, after observational studies reported the association between smoking status and improved short-term outcomes in patients with cardiovascular manifestations of atherosclerosis. In the first systematic investigation of smoking and its impact on outcomes after acute MI in 1968, Weinblatt et al. (6) reported a larger proportion of deaths within 1 month in nonsmokers (38%) compared with smokers (27.5%) in a cohort of 881 patients with MI. Some years later, a post hoc analysis from GUSTO-I (Global Utilization of Streptokinase and Tissue-Plasminogen Activator for Occluded Coronary Arteries) reiterated these findings (7). Recently, a systematic review of the “smoker’s paradox” reported mixed results, with improved outcomes in smokers with MIs in 6 studies performed to investigate the impact of thrombolytic therapy in the 1980s and 1990s and defining MI according to the outdated World Health Organization definition of 1979. Among the remaining 11 studies not supporting the existence of the “smoker’s paradox” 5 studies represented patients undergoing more contemporary management of MI (8).
A plausible explanation of the “smoker’s paradox” has been that smokers are significantly younger at the time of their first cardiovascular events, with fewer atherosclerotic risk factors and comorbidities compared with nonsmokers. Thus, part of the association between smoking and improved outcomes might be attributable to these imbalances in cardiovascular risk factors.
In this issue of JACC: Cardiovascular Interventions, Yadav et al. (9) investigate once more the “smoker’s paradox” in a large individual patient data pooled analysis including 18 large-scale, multicenter, prospective, randomized controlled coronary stent trials including a total of 24,354 percutaneous coronary intervention (PCI) patients with available data on smoking status, through a median follow-up period of 5 years. The investigators argue that no prior systematic investigation of this phenomenon until 5-year follow-up had been undertaken in the era of contemporary PCI techniques, with improved stent-based PCI and advanced concomitant pharmacotherapy. The main findings of this study are that smokers had a lower crude (unadjusted) rate of target lesion failure (TLF) (6.0% vs. 7.2%; p = 0.006) compared with nonsmokers, a trend toward less target lesion revascularization (5.8% vs. 6.5%; p = 0.06), but a greater rate of definite or probable stent thrombosis (1.8% vs. 0.8%; p < 0.0001) 1 year after index PCI. There were no differences in the rates of death, cardiac death, and MI between smokers and nonsmokers at 1 year. At 5 years post-PCI, smokers had significantly higher rates of MI (7.8% vs. 5.6%; p < 0.0001) and definite or probable stent thrombosis (3.5% vs. 1.8%; p < 0.0001). There were no significant differences in the rates of death, cardiac death, target lesion revascularization, and TLF. However, after multivariate adjustment for differences, smoking was highly and significantly associated with death, cardiac death, and TLF at 5 years.
First and foremost, the investigators must be congratulated for performing such an analysis by pooling individual patient data from large-scale randomized controlled studies in a timely and scientifically appealing manner. Without any doubt, this study helps clarify the “smoker’s paradox” in an era of primary PCI using cutting-edge drug-eluting stent (DES) technology with adjunct pharmacotherapy in an unprecedented way, outdating most prior analyses on this topic. Furthermore, it deserves mention that independent event committees adjudicated almost all endpoints in the present study, with few missing data in the overall cohort. The strength of the present analysis lies in the fact that, by means of individual patient data, it was possible to show that smoking was associated with increased adjusted risk for death and TLF. We therefore observe with relief that previous hesitation to overcome this paradox has been conquered, while future medical education should clearly focus on establishing effective preventive measures, including dedicated smoking cessation programs for patients with cardiovascular disease to reduce this significant and clinically relevant risk factor.
However, there are important limitations to the present analysis that also deserve discussion. First, although the “smoker’s paradox” was initially observed in studies examining the outcomes of patients with acute MI, the present analysis predominantly included stable and less complex patients, with overall reduced acuity. Indeed, from a total of 18 trials included, 13 originally enrolled patients with stable or unstable angina, in the absence of acute MI. Only 1 trial enrolled patients with ST-segment elevation MI, while 4 additional trials enrolled all-comer populations. Although a sensitivity analysis found no interaction among the adjusted risk for death, smoking status, and clinical presentation, there remains concern on whether the present findings can reliably be applied to patients presenting with acute MI, a clinical subset that deserves further systematic investigation.
Second, smoking status was assessed only at baseline, without differentiation into former and ex-smokers, which limits transferability of current findings to important subgroups at high risk for future cardiovascular events. In addition, smoking was not assumed as a time-varying covariate, prohibiting assessment of smoking crossover and its impact on cardiovascular outcomes over time. In fact, one of the key limitations of prior studies on this topic was the absence of information of smoking cessation and relapse in younger patients after MI, which likely contributed to the paradoxical finding of lower acute mortality in smokers. Unfortunately, the design of the present study did not make it possible to draw definitive conclusions regarding this conundrum. Although this aspect could be perceived as less relevant given the absence of dose-dependent increments in cardiac death after smoke exposure (10), the importance of sequential data capture with regard to smoking status was exemplified in the 5-year follow-up analysis of the SYNTAX (TAXUS Drug-Eluting Stent Versus Coronary Artery Bypass Surgery for the Treatment of Narrowed Arteries) trial, in which smoking status changed in a significant proportion of patients enrolled in the trial (17.9%) over time and was found to be predictive of death, MI, or stroke and major adverse cardiovascular events after percutaneous or surgical revascularization in 1,800 patients (Figure 1) (11).
Finally, this meta-analysis missed the opportunity to specifically address the risk for stent failure associated with smoking status across different DES platforms (namely, earlier and newer generations). Indeed, although the subgroup analysis for definite or probable stent thrombosis revealed a significant interaction between smoking status and DES generation, it is of note that the 5-year risk for definite or probable stent thrombosis in smokers treated with newer DES was as high as 2%, contradicting the relatively lower risk for thrombotic events associated with these improved stent platforms in contemporary stent trials.
In conclusion, Yadav et al. (9) report that smokers are at increased risk for adverse events up to 5 years following index PCI. Any smoke exposure, even that associated with newer tobacco products, carries substantial risk for adverse cardiac events, although data specific to PCI patients have been lacking until now. In addition, in an era of high-technology treatment of MI, the findings of this study also serve to remind us of the central role of effective preventive measures for cardiovascular health.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Joner has received personal fees and grants from Biotronik. Dr. Cassese has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Johnson K.C.
- Messner B.,
- Bernhard D.
- Mizobe F.,
- Livett B.G.
- Yadav M.,
- Mintz G.S.,
- Généreux P.,
- et al.
- Hackshaw A.,
- Morris J.K.,
- Boniface S.,
- Tang J.L.,
- Milenkovic D.
- Zhang Y.J.,
- Iqbal J.,
- van Klaveren D.,
- et al.