You are here

Article Text

Article menu
Coronary artery disease
Original article
Smoking status and life expectancy after acute myocardial infarction in the elderly
  1. Emily M Bucholz1,
  2. Adam L Beckman2,
  3. Catarina I Kiefe3,
  4. Harlan M Krumholz4,5,6
  1. 1Department of Medicine, Boston Children's Hospital, Boston, Massachusetts, USA
  2. 2Yale School of Medicine and Yale School of Public Health, New Haven, Connecticut, USA
  3. 3Department of Quantitative Health Sciences, University of Massachusetts Medical School, Worcester, Massachusetts, USA
  4. 4Section of Cardiovascular Medicine, Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut, USA
  5. 5Center for Outcomes Research and Evaluation, Yale-New Haven Hospital, New Haven, Connecticut, USA
  6. 6Section of Health Policy and Administration, School of Public Health, Yale University School of Medicine, New Haven, Connecticut, USA
  1. Correspondence to Dr Harlan M Krumholz, Department of Internal Medicine, Yale University School of Medicine, 1 Church St. Suite 200, New Haven, CT 06510, USA; harlan.krumholz{at}yale.edu

Abstract

Objective Smokers have lower short-term mortality after acute myocardial infarction (AMI) than non-smokers; however, little is known about the long-term effects of smoking on life expectancy after AMI. This study aimed to quantify the burden of smoking after AMI using life expectancy and years of life lost.

Methods We analysed data from the Cooperative Cardiovascular Project, a medical record study of 158 349 elderly Medicare patients with AMI and over 17 years of follow-up, to evaluate the age-specific association of smoking with life expectancy and years of life lost after AMI.

Results Our sample included 23 447 (14.8%) current smokers. Current smokers had lower crude mortality up to 5 years, which was largely explained by their younger age at AMI. After adjustment other patient characteristics, smoking was associated with lower 30-day (HR 0.91, 95% CI 0.87 to 0.94) but higher long-term mortality (17-year HR 1.19, 95% CI 1.17 to 1.20) after AMI. Overall, crude life expectancy estimates were lower for current smokers than non-smokers at all ages, which translated into sizeable numbers of life-years lost attributable to smoking. As age at AMI increased, the magnitude of life-years lost due to smoking decreased. After full risk adjustment, the differences in life expectancy between current smokers and non-smokers persisted at all ages.

Conclusions Current smoking is associated with lower life expectancy and large numbers of life-years lost after AMI. Our findings lend additional support to smoking cessation efforts after AMI.

https://doi.org/10.1136/heartjnl-2015-308263

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Smoking increases the risk of coronary heart disease and cardiovascular-related mortality in the general population,1–3 but studies of patients with acute myocardial infarction (AMI) have consistently reported lower crude mortality in smokers compared with non-smokers.4–8 This ‘smoker's paradox’ has been investigated in numerous short-term studies but relatively few long-term studies. In general, short-term studies have reported lower or similar mortality after adjustment in smokers,4–13 whereas long-term studies have found similar or higher adjusted mortality in smokers.14–17 These discrepancies in study findings are likely due to differences in study characteristics and covariate selection. Additionally, short-term and long-term studies available have been limited by small numbers of current smokers,5 ,8 ,11 ,13–15 minimal covariate adjustment8 ,10–12 ,14 ,15 and specific inclusion criteria such as hospital survivors,14 patients undergoing revascularisation5 ,8 ,9 ,13 or non-ST-elevation AMIs,16 which may limit their generalisability to all patients. As a result, it is unclear whether the ‘smoker's paradox’ can be explained by other patient characteristics and whether this short-term phenomenon persists over the long-term.

    Clarifying these issues requires a comprehensive assessment of the relationship between smoking with early and late mortality in patients post AMI. However, such assessments have been difficult to conduct because most datasets lack the size, follow-up, or clinical detail required. Using data from the Cooperative Cardiovascular Project (CCP), we took advantage of a unique opportunity to characterise the association between smoking with short-term and long-term survival after AMI. This large dataset contains over 17 years of follow-up and detailed clinical information. The long follow-up afforded by CCP allows us to estimate the years of life lost or gained after AMI attributable to smoking. This metric quantifies the absolute burden of smoking in patients with AMI by comparing life expectancy in smokers and non-smokers. Although years of life lost has been used extensively in population-based studies to quantify the premature mortality and the burden of behavioural risk factors such as smoking, it is virtually absent from the cardiovascular literature.18 The aims of this study were to determine the association of current smoking with short-term and long-term mortality after AMI and to quantify the burden of smoking after AMI using years of life lost. We hypothesised that the short-term survival advantage observed in current smokers can be explained, in part, by differences in patient characteristics between current smokers and non-smokers, and that over time, this survival advantage reverses and current smokers have higher mortality compared with non-smokers.

    Methods

    Cooperative Cardiovascular Project

    Conducted by the Health Care Financing Administration (now Center for Medicare and Medicaid Services (CMS)), the CCP was a prospective, quality-improvement project designed to improve the care of patients with AMI in the USA.19 ,20 Between January 1994 and February 1996, all fee-for-service Medicare beneficiaries hospitalised with a principal discharge diagnosis of AMI (International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) code 410) were sampled from non-federal acute-care hospitals in the USA for approximately eight consecutive months (n=234 769). When patients had repeat admissions for the same AMI episode (ICD-9-CM code 410.x2), only the initial admission was included. Medical charts were abstracted for patient demographics, cardiovascular risk factors, medical history, clinical presentation, guideline-based admission therapies, inhospital events and discharge disposition. This study was approved by the Yale University institutional review board.

    Study sample

    For this analysis, we included all patients aged ≥65 years with AMI confirmed by medical record, defined as an elevation of creatine kinase-MB level (>5% of total creatine kinase), or an elevation of lactate dehydrogenase enzyme (LDH) level with isoenzyme reversal (LDH1>LDH2), or the presence of at least two of the following: chest pain, twofold elevation in total creatine kinase, and diagnostic ECG changes (ST-segment elevation or new pathological Q-waves). Patients who were <65 years of age (n=17 593) or without clinically confirmed AMI (n=31 186) were excluded. We limited our sample to elderly patients with AMI in order to obtain a nationally representative sample of patients with AMI. Whereas most US citizens are eligible for Medicare upon turning 65 years of age, younger patients are eligible for Medicare only if they qualify for disability or dialysis and thus are not representative of the younger AMI population. For patients with multiple admissions during the study period, we excluded all admissions except the first (n=27 500). We also excluded patients transferred from postambulatory or surgical settings (n=19 379) and patients whose records could not be linked to the Medicare Denominator Files to ascertain vital status (n=2195). Of the 234 769 patients in the CCP, 158 349 met eligibility criteria for this analysis.

    Study variables

    Information on current smoking status at the time of AMI was abstracted from the medical record and recorded as a dichotomous variable. The non-smoker group included never smokers and former smokers. The primary outcome was 17-year survival from admission, which was obtained from the 1994–2012 Medicare Denominator files. These files provide complete follow-up on all patients enrolled in Medicare and thus no patients were lost to follow-up except those that could not be linked to the Medicare Denominator files.

    Statistical analyses

    Baseline characteristics were compared between current smokers and non-smokers using χ2 tests for categorical variables and t-tests for continuous variables. Cox proportional hazards models were constructed to evaluate the HRs for death, with current smoking as the independent variable of interest, at 30 days, 1 year, 5 years and 17 years. For the 30-day, 1-year and 5-year models, patients surviving these time frames were censored at 30 days, 1 year and 5 years, respectively. Proportional hazards assumptions were checked for all models using Schoenfeld residuals, examined graphically and tested formally. We used sequential adjustment to determine which covariates, if any, mediated the relationship between current smoking and mortality. In model 1, we adjusted for patient age only. Model 2 controlled for patient demographics (sex and race), traditional cardiovascular risk factors (diabetes, hypertension, body mass index, history of coronary artery disease) and medical history (congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), cerebrovascular accident, peripheral vascular disease). Model 3 further controlled for clinical presentation (Killip class, AMI location, ST-segment elevation AMI, pulse, systolic blood pressure), and therapies received (percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) within 30 days of admission, fibrinolytic therapy, and among those eligible, aspirin and beta-blockers on admission). Covariates were selected based on prior reports,5 ,6 ,8 ,14–16 face validity and clinical judgement. Patients with missing data on continuous covariates were assigned the median value in the cohort and a binary dummy variable to denote missing. Patients with missing data on categorical covariates were included in the model using a dummy variable for missing data.

    We defined life expectancy, or mean survival, as the area under the survival curve and used a 3-step process to calculate life expectancy for current smokers and non-smokers. These methods have been described previously and are detailed in the online supplementary appendices.21 Briefly, we used Cox proportional hazards models to plot the expected survival curves for current smokers and non-smokers at each age and then extrapolated the curves to age 100 using exponential models. Life expectancy was estimated by summing the areas under the Cox and exponential curves. Years of life lost after AMI attributable to current smoking were calculated by subtracting life expectancy in current smokers from that in non-smokers.

    To determine whether the differences in life expectancy between current smokers and non-smokers were explained by differences in patient characteristics between the two groups, we repeated our calculations of life expectancy adjusting for the full list of covariates listed above. Adjusted years of life lost were again calculated by subtracting adjusted mean survival in current smokers from that in non-smokers. All statistical analyses were performed using SAS V.9.2 (SAS Institute, Cary, North Carolina, USA).

    Results

    Of the 158 349 patients with AMI in our cohort, 23 447 (14.8%) were current smokers. Compared with non-smokers, current smokers were younger and more likely to be male. In addition, they had lower rates of common cardiovascular risk factors including diabetes, hypertension and obesity, but higher rates of COPD and peripheral vascular disease (table 1). They were also more likely to present with ST-elevation AMI and to receive acute reperfusion therapy.

    View this table:
    Table 1

    Baseline characteristics of patients by current smoking status on admission (n=158 349): the Cooperative Cardiovascular Project, 1994–1996*

    Current smokers had lower crude mortality at 30 days, 1 year and 5 years after AMI but slightly higher mortality over 17 years (table 2). By the end of follow-up, 93.6% of current smokers were deceased compared with 92.9% of non-smokers. Unadjusted survival curves showed consistently higher mortality for non-smokers with the curves converging towards the end of follow-up (figure 1); however, adjustment for age attenuated these effects (table 2). Relative to non-smokers, the age distribution of current smokers was concentrated at younger ages and heavily right-skewed (figure 2). After age adjustment, the HR of 30-day mortality in current smokers versus non-smokers was not significantly different from one, but current smokers had significantly higher hazards of mortality at 1, 5 and 17 years. Further adjustment for medical history and clinical presentation again rendered the hazards of 30-day mortality in current smokers lower than for non-smokers (HR 0.91, 95% CI 0.87 to 0.94) and attenuated this difference for 1-year mortality (HR 0.97, 95% CI 0.94 to 1.00). However, the higher hazards of 5-year and 17-year mortality for current smokers versus non-smokers persisted (5 years: 1.07, 95% CI 1.05 to 1.10; 17 years: 1.19, 95% CI 1.17 to 1.20).

    View this table:
    Table 2

    Crude mortality and HRs for mortality in current smokers versus non-smokers by length of follow-up from admission*

    Figure 1
    Figure 1

    Unadjusted survival curves for smokers versus non-smokers calculated from admission. Survival curves were modelled using Cox proportional hazards regression and thus represent expected survival for current smokers and non-smokers.

    Figure 2
    Figure 2

    Age distribution of smokers and non-smokers at the time of acute myocardial infarction hospitalisation.

    Crude life expectancy estimates were lower for current smokers than non-smokers at all ages (figure 3). These differences resulted in sizeable numbers of life-years lost attributable to current smoking; however, the magnitude of life-years lost due to smoking decreased as age at AMI increased (figure 4). Before adjustment, 65-year-old smokers lived 2.90 (95% CI 2.68 to 3.12) fewer years of life after AMI, and 85-year-old smokers lived 0.41 (95% CI 0.30 to 0.52) fewer years. After full risk adjustment, the differences in life expectancy between current smokers and non-smokers persisted but decreased. Current smoking was associated with 1.96 (95% CI 1.76 to 2.16) life-years lost at age 65 and 0.16 (95% CI 0.06 to 0.26) life-years lost at age 85.

    Figure 3
    Figure 3

    Life expectancy estimates for smokers and non-smokers calculated from admission. Estimates are calculated from a Cox proportional hazards model that includes only age, smoking status and their interaction. AMI, acute myocardial infarction.

    Figure 4
    Figure 4

    Unadjusted and adjusted years of potential life lost attributable to smoking at the time of acute myocardial infarction (AMI). Multivariable estimates are adjusted for patient demographics (sex and race), traditional cardiovascular risk factors (diabetes, hypertension, body mass index, history of coronary artery disease), medical history (congestive heart failure, chronic obstructive pulmonary disease, cerebrovascular accident, peripheral vascular disease), clinical presentation (Killip class, AMI location, ST-segment elevation AMI, pulse on admission, systolic blood pressure on admission) and therapies received (percutaneous coronary intervention or coronary artery bypass grafting in the first 30 days after admission, fibrinolytic therapy during hospitalisation, and aspirin and β-blockers on admission).

    Discussion

    Using data from a large, nationally representative cohort of about 150 000 elderly patients with AMI, we found that the lower risk of short-term and long-term mortality in current smokers after AMI was largely explained by their younger age at presentation. After adjustment for age, current smokers had comparable 30-day mortality and higher long-term mortality than non-smokers. However, after adjustment for clinical and treatment characteristics, current smoking was associated with lower 30-day, comparable 1-year, and higher 5-year and 17-year mortality. The higher long-term mortality in current smokers resulted in significantly lower life expectancies after AMI for smokers at all ages and large numbers of life-years lost after AMI attributable to smoking.

    Prior studies evaluating the relationship between smoking and short-term mortality after AMI have been inconsistent. Nearly all studies have reported lower 30-day and 1-year unadjusted mortality for smokers but lower,6 ,7 ,10 ,11 comparable,4 ,5 ,8 ,9 ,13 or even higher12 risk of mortality after adjustment. Inconsistencies between studies are likely due to differences in study samples, length of follow-up, covariate adjustment and the categorisation of smokers.

    Our results may offer insight into why studies of short-term mortality have been mixed. First, we found that the relationship between current smoking status and mortality varies by length of follow-up, even in the short-term. In our study, current smokers had a lower risk of 30-day mortality but comparable risk of 1-year mortality as non-smokers after full adjustment. Accordingly, most studies that have found lower mortality in smokers have evaluated hospital mortality,8 ,10 ,11 whereas studies reporting no differences in mortality by smoking have generally examined 1-year mortality.5 ,8 ,13 Second, we found that the association between current smoking status and short-term mortality depended on covariate selection. Using sequential adjustment, we found no differences in 30-day mortality among current smokers and non-smokers after adjustment for age but lower mortality for current smokers after adjustment for other clinical and treatment characteristics.

    Over the long-term, we found that differences in mortality between current smokers and non-smokers were largely explained by differences in age. Current smokers in our sample were, on average, 4.8 years younger than non-smokers, which may also explain the lower prevalence of cardiovascular risk factors in smokers in our sample. The age difference in smokers and non-smokers is also consistent with observations from the general population suggesting that smoking causes premature AMIs in men by approximately 8.3–8.7 years and in women by 10.8–14.4 years.22 ,23

    Very few studies have evaluated the relationship between smoking and long-term mortality after AMI; however, the few studies available have reported similar or higher mortality for smokers after adjustment.14–17 Using data from the CRUSADE registry, Shen et al16 found that smoking was associated with increased 3-year mortality in patients with non-ST-segment elevation AMI. Likewise, Gerber et al14 found that never smokers and quitters had significantly lower 13-year mortality after AMI than persistent smokers. Our study improves upon prior long-term studies by (1) including large numbers of smokers and non-smokers, (2) extending follow-up to the entire remaining lifespan, (3) adjusting for multiple demographic and clinical characteristics and (4) quantifying the burden of smoking after AMI using years of life lost. This metric has several strengths above traditional mortality estimates.18 ,24 As indices of premature mortality, years of life lost can be used to quantify the effect of smoking after AMI over the entire lifespan and to compare the burden of smoking across multiple ages. In addition, quantifying the years of life lost attributable to smoking may be a more effective way to communicate the harmful effects of smoking to patients as opposed to relative mortality rates.25 To our knowledge, only one other study has used years of life lost to evaluate the relationship between smoking and post-AMI mortality; however, this study estimated the years of life lost attributable to AMI in smokers and non-smokers rather than the years of life lost after AMI attributable to smoking as we have done in our study.26

    Several explanations may account for our finding that smoking is associated with increased mortality after AMI over the long term but not the short term. Whereas mortality in the first 30 days after AMI is largely related to the extent of myocardial damage, comorbid conditions and timely receipt of therapies, mortality over the long term depends largely on the extent and progression of atherothrombotic disease in the coronary arteries as well as other competing risks. We found that current smokers had less comorbidity but higher rates of acute reperfusion therapy after AMI, which may explain their lower mortality in the first 30 days. Although it is unclear why the effect of smoking on 30-day mortality became protective after full adjustment, one explanation may be that current smokers have the opportunity to alter a potent risk factor by quitting smoking and thereby rapidly reduce their risk of short-term death and recurrent AMI. In contrast, non-smokers likely have other risk factors that are more difficult to alter in the short term. Prior studies have shown that up to 74% of smokers may quit, at least temporarily, in the first month after AMI27 and that smoking cessation has immediate benefits after AMI, including reductions in angina and inhospital complications.28 ,29 Over the long term, however, smoking causes endothelial dysfunction by increasing oxidative stress, vascular inflammation and thrombosis, which predisposes patients to recurrent cardiac events.30 Indeed, several studies have shown that smokers are at increased risk of recurrent infarction even in the first year after AMI.13 Over time, this prothrombotic state places smokers at increased risk of cardiovascular mortality. In addition, smoking is associated with other competing risks and higher rates of non-cardiovascular death in the general population from causes such as lung cancer and COPD. Not surprisingly, we found a higher prevalence of COPD in current smokers than non-smokers in our cohort. Thus, the higher risk of long-term mortality in smokers is likely mediated through both cardiac and non-cardiac pathways.

    Interestingly, we observed that the years of life lost attributable to smoking narrowed as age at AMI increased. Although we might expect the gap in life expectancy to be smaller among older patients with lower life expectancies, it is also possible that smokers who reach the oldest ages represent a select group of patients with low frailty and low mortality. This phenomenon has been previously reported in the general population.31 It is hypothesised that older smokers represent a healthier cohort than younger smokers because they have survived the increased risk of premature death due to smoking. Our study is the first to report this phenomenon in a population of patients with established cardiovascular disease.

    This study was subject to several limitations. First, only current smoking status at the index hospitalisation of the patient was recorded. No information on the duration or intensity of smoking or previous habits was collected. Thus, there is likely a fair amount of heterogeneity in risk among both current and non-smokers, and the estimates presented in this study represent averages across smokers and non-smokers with varying smoking histories and intensities. Second, we do not know how many patients quit smoking after AMI. Prior studies have noted that quitters have improved prognosis after AMI relative to persistent smokers.14 ,17 ,27 As a result, including both quitters and persistent smokers together may attenuate our risk estimates. In fact, it is very likely that the life-years lost attributable to smoking differs in quitters and persistent smokers. Third, because the Medicare Denominator files record only all-cause mortality, we were unable to differentiate cardiovascular-related deaths from other causes and thus could not determine whether the higher long-term mortality in current smokers was attributable to cardiovascular disease or other smoking-related deaths. Fourth, our study sample included only elderly patients aged 65 and older and thus may not be generalisable to younger patients. Nevertheless, the greatest differences in life expectancy between current and non-smokers occurred among younger patients suggesting that there may be even larger numbers of years of life lost attributable to smoking among patients <65 years of age. Fifth, in order to calculate life expectancy after AMI, we needed a cohort with long-term follow-up, which only a study like CCP could offer. As a result, the data used for this study were collected in the pre-primary PCI era and thus our life expectancy estimates may differ from those of patients today. It is possible that contemporary treatment with primary-PCI may be associated with better outcomes in both smokers and non-smokers. However, previous literature suggests that smokers may have higher rates of complications after reperfusion therapy such as major bleeding, stroke, or reinfarction.32 ,33 Thus, it is unclear how the magnitude of life-years lost after AMI attributable to smoking differs today. Finally, we adjusted for several clinical and treatment characteristics, some of which may be in the causal pathway between smoking and mortality after AMI. We opted for thorough risk adjustment at baseline in order to determine which variables if any explained the association between current smoking status and long-term mortality after AMI. However, it is possible that we adjustment for certain variables such as COPD or stroke may have resulted in overadjustment of the models.

    In summary, we showed that current smoking was associated with lower 30-day, comparable 1-year, and higher 5-year and 17-year mortality. Although the smoker's paradox persisted in the short term, the association between smoking and long-term mortality was largely explained by the younger age of current smokers at the time of AMI. Using data from the largest and longest follow-up study of elderly patients with AMI in the USA, we demonstrated that current smoking status was associated with lower life expectancy and large numbers of life-years lost after AMI. Our findings lend additional support to efforts that encourage smoking cessation in patients at risk for coronary heart disease by demonstrating poorer prognosis after AMI in current smokers. Years of life lost may be particularly helpful for communicating the harms of smoking to patients with AMI.

    Key messages

    What is already known on this subject?

    • Studies of patients with acute myocardial infarction (AMI) have consistently reported lower short-term crude mortality in smokers compared with non-smokers; however, little is known about the effect of smoking on long-term mortality after AMI and whether the lower mortality in smokers can be explained by other patient characteristics.

    What might this study add?

    • Using data from a large, nationally representative cohort of patients with AMI with detailed clinical information and over 17 years of follow-up, we found that current smokers had significantly shorter life expectancy after AMI than non-smokers. In patients ≤85 years of age, current smoking was associated with between 0.16 and 1.96 years of life lost after AMI relative to non-smokers. Years of life lost attributable to smoking were greatest in younger patients.

    How might this impact on clinical practice?

    • Our findings lend additional support to smoking cessation efforts after AMI. In addition, years of life lost may be particularly helpful for communicating the harms of smoking to patients with coronary artery disease.

    References

    1. ↵
      1. Kannel WB,
      2. Higgins M
      . Smoking and hypertension as predictors of cardiovascular risk in population studies. J Hypertens Suppl 1990;8:S3–8.
      OpenUrlPubMed
    2. ↵
      1. Peto R,
      2. Lopez AD,
      3. Boreham J, et al
      . Mortality from tobacco in developed countries: Indirect estimation from national vital statistics. Lancet 1992;339:1268–78. doi:10.1016/0140-6736(92)91600-D
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Wilhelmsson C,
      2. Vedin JA,
      3. Elmfeldt D, et al
      . Smoking and myocardial infarction. Lancet 1975;305:415–20. doi:10.1016/S0140-6736(75)91488-9
      OpenUrlCrossRef
    4. ↵
      1. Andrikopoulos GK,
      2. Richter DJ,
      3. Dilaveris PE, et al
      . In-hospital mortality of habitual cigarette smokers after acute myocardial infarction; the “smoker's paradox” in a countrywide study. Eur Heart J 2001;22:776–84. doi:10.1053/euhj.2000.2315
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Goto K,
      2. Nikolsky E,
      3. Lansky AJ, et al
      . Impact of smoking on outcomes of patients with ST-segment elevation myocardial infarction (from the HORIZONS-AMI Trial). Am J Cardiol 2011;108:1387–94. doi:10.1016/j.amjcard.2011.06.063
      OpenUrlCrossRefPubMed
    6. ↵
      1. Gourlay SG,
      2. Rundle AC,
      3. Barron HV
      . Smoking and mortality following acute myocardial infarction: Results from the National Registry of Myocardial Infarction 2 (NRMI 2). Nicotine Tob Res 2002;4:101–7. doi:10.1080/14622200110103205
      OpenUrlAbstract
    7. ↵
      1. Kang SH,
      2. Suh JW,
      3. Choi DJ, et al
      . Cigarette smoking is paradoxically associated with low mortality risk after acute myocardial infarction. Nicotine Tob Res 2013;15:1230–8. doi:10.1093/ntr/nts248
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Rakowski T,
      2. Siudak Z,
      3. Dziewierz A, et al
      . Impact of smoking status on outcome in patients with ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention. J Thromb Thrombolysis 2012;34:397–403. doi:10.1007/s11239-012-0764-0
      OpenUrlCrossRefPubMed
    9. ↵
      1. Barbash GI,
      2. Reiner J,
      3. White HD, et al
      . Evaluation of paradoxic beneficial effects of smoking in patients receiving thrombolytic therapy for acute myocardial infarction: Mechanism of the “smoker's paradox” from the GUSTO-I trial, with angiographic insights. Global Utilization of Streptokinase and Tissue-plasminogen Activator for Occluded Coronary Arteries. J Am Coll Cardiol 1995;26:1222–9. doi:10.1016/0735-1097(95)00299-5
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Canto JG,
      2. Kiefe CI,
      3. Rogers WJ, et al
      . Atherosclerotic risk factors and their association with hospital mortality among patients with first myocardial infarction (from the National Registry of myocardial Infarction). Am J Cardiol 2012;110:1256–61. doi:10.1016/j.amjcard.2012.06.025
      OpenUrlCrossRefPubMed
    11. ↵
      1. Chen KY,
      2. Rha SW,
      3. Li YJ, et al
      . ‘Smoker's paradox’ in young patients with acute myocardial infarction. Clin Exp Pharmacol Physiol 2012;39:630–5. doi:10.1111/j.1440-1681.2012.05721.x
      OpenUrlCrossRefPubMed
    12. ↵
      1. Robertson JO,
      2. Ebrahimi R,
      3. Lansky AJ, et al
      . Impact of cigarette smoking on extent of coronary artery disease and prognosis of patients with non-st-segment elevation acute coronary syndromes: An analysis from the ACUITY trial (Acute Catheterization and Urgent Intervention Triage Strategy). JACC Cardiovasc Interv 2014;7:372–9. doi:10.1016/j.jcin.2013.11.017
      OpenUrlCrossRefPubMed
    13. ↵
      1. Weisz G,
      2. Cox DA,
      3. Garcia E, et al
      . Impact of smoking status on outcomes of primary coronary intervention for acute myocardial infarction--the smoker's paradox revisited. Am Heart J 2005;150:358–64. doi:10.1016/j.ahj.2004.01.032
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Gerber Y,
      2. Rosen LJ,
      3. Goldbourt U, et al
      . Smoking status and long-term survival after first acute myocardial infarction a population-based cohort study. J Am Coll Cardiol 2009;54:2382–7. doi:10.1016/j.jacc.2009.09.020
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Grundtvig M,
      2. Hagen TP,
      3. Amrud ES, et al
      . Mortality after myocardial infarction: impact of gender and smoking status. Eur J Epidemiol 2011;26:385–93. doi:10.1007/s10654-011-9557-6
      OpenUrlCrossRefPubMed
    16. ↵
      1. Shen L,
      2. Peterson ED,
      3. Li S, et al
      . The association between smoking and long-term outcomes after non-st-segment elevation myocardial infarction in older patients. Am Heart J 2013;166:1056–62. doi:10.1016/j.ahj.2013.09.011
      OpenUrlCrossRefPubMed
    17. ↵
      1. Kinjo K,
      2. Sato H,
      3. Sakata Y, et al
      . Impact of smoking status on long-term mortality in patients with acute myocardial infarction. Circ J 2005;69:7–12. doi:10.1253/circj.69.7
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Gardner JW,
      2. Sanborn JS
      . Years of potential life lost (YPLL)—what does it measure? Epidemiology 1990;1:322–9. doi:10.1097/00001648-199007000-00012
      OpenUrlCrossRefPubMed
    19. ↵
      1. Marciniak TA,
      2. Ellerbeck EF,
      3. Radford MJ, et al
      . Improving the quality of care for Medicare patients with acute myocardial infarction: Results from the Cooperative Cardiovascular Project. JAMA 1998;279:1351–7. doi:10.1001/jama.279.17.1351
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Ramunno LD,
      2. Dodds TA,
      3. Traven ND
      . Cooperative Cardiovascular Project (CCP) quality improvement in Maine, New Hampshire, and Vermont. Eval Health Prof 1998;21:442–60. doi:10.1177/016327879802100404
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Bucholz EM,
      2. Normand ST,
      3. Wang Y, et al
      . Life expectancy and years of potential life lost after acute myocardial infarction by sex and race: a cohort-based study of Medicare beneficiaries. J Amer Coll Cardiol 2015;66:645–55. doi:10.1016/j.jacc.2015.06.022
      OpenUrlCrossRefPubMed
    22. ↵
      1. Bähler C,
      2. Gutzwiller F,
      3. Erne P, et al
      . Lower age at first myocardial infarction in female compared to male smokers. Eur J Prev Cardiol 2012;19:1184–93. doi:10.1177/1741826711422764
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Grundtvig M,
      2. Hagen TP,
      3. German M, et al
      . Sex-based differences in premature first myocardial infarction caused by smoking: Twice as many years lost by women as by men. Eur J Cardiovasc Prev Rehabil 2009;16:174–9. doi:10.1097/HJR.0b013e328325d7f0
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      Centers for Disease Control (CDC). Premature mortality in the United States: Public health issues in the use of years of potential life lost. MMWR Morb Motal Wkly Rep 1986;35(2 Suppl):1S–11S.
      OpenUrl
    25. ↵
      1. Fontaine KR,
      2. Redden DT,
      3. Wang C, et al
      . Years of life lost due to obesity. JAMA 2003;289:187–93. doi:10.1001/jama.289.2.187
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Grundtvig M,
      2. Hagen TP,
      3. Amrud ES, et al
      . Reduced life expectancy after an incident hospital diagnosis of acute myocardial infarction--effects of smoking in women and men. Int J Cardiol 2013;167:2792–7. doi:10.1016/j.ijcard.2012.07.010
      OpenUrlCrossRefPubMed
    27. ↵
      1. Wilson K,
      2. Gibson N,
      3. Willan A, et al
      . Effect of smoking cessation on mortality after myocardial infarction: meta-analysis of cohort studies. Arch Int Med 2000;160:939–44. doi:10.1001/archinte.160.7.939
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Mallya S,
      2. Calverley PM,
      3. MacFarlane IA, et al
      . Influence of cigarette smoking on the outcome of coronary care unit admissions with chest pain. Postgrad Med J 1992;68:562–5. doi:10.1136/pgmj.68.801.562
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Miwa K,
      2. Fujita M,
      3. Miyagi Y
      . Beneficial effects of smoking cessation on the short-term prognosis for variant angina--validation of the smoking status by urinary cotinine measurements. Int J Cardiol 1994;44:151–6. doi:10.1016/0167-5273(94)90019-1
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Ambrose JA,
      2. Barua RS
      . The pathophysiology of cigarette smoking and cardiovascular disease: An update. J Am Coll Cardiol 2004;43:1731–7. doi:10.1016/j.jacc.2003.12.047
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Hoffman R
      . Socioeconomic differences in old age mortality. Springer Science and Business Media, 2008.
    32. ↵
      1. Barra S,
      2. Providência R,
      3. Caetano F, et al
      . Bleed-myocardial infarction score: Predicting mid-term post-discharge bleeding events. World J Cardiol 2013;5:196–206. doi:10.4330/wjc.v5.i6.196
      OpenUrlCrossRefPubMed
    33. ↵
      1. Honda T,
      2. Fujimoto K,
      3. Miyao Y, et al
      . Current cigarette smoking is an independent risk factor for subacute stent thrombosis in acute myocardial infarction patients. J Cardiol 2014;63:358–64. doi:10.1016/j.jjcc.2013.10.007
      OpenUrlCrossRefPubMed

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors The authors had full access to all of the study data and take responsibility for the accuracy the analyses. All authors were actively involved in manuscript preparation including conception and design (EMB, HMK, CIK), analysis and interpretation of the data (EMB, ALB, CIK, HMK), drafting of the manuscript (EMB, ALB, HMK) and revisions (EMB, ALB, CIK, HMK). All authors have read and approved the manuscript.

    • Funding EMB is supported an F30 Training grant F30HL120498-01A1 from the National Heart, Lung and Blood Institute and by NIGMS Medical Scientist Training Program grant T32GM07205. HMK is supported by grant U01 HL105270-04 (Center for Cardiovascular Outcomes Research at Yale University) from the National Heart, Lung and Blood Institute.

    • Competing interests HMK has received research grants from Medtronic and from Johnson and Johnson through Yale University, to develop methods of clinical trial data sharing. He also works under contract with the Centers for Medicare and Medicaid Services to develop and maintain performance measures and chairs the Cardiac Scientific Advisory Board for UnitedHealth.

    • Ethics approval Yale University Institutional Review Board.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement The data for this study were obtained through an agreement with Qualidigm and the Centers for Medicare and Medicaid Services. Data user agreements can be sought through Qualidigm.

    Linked Articles