Search
Close
Search
Advanced Search
Search Menu
AI Discovery Assistant

Abstract

OBJECTIVES

We sought to investigate long-term survival of non-smokers undergoing coronary artery bypass surgery (CABG).

METHODS

A prospective database of consecutive patients was retrospectively analysed and cross correlated with the UK strategic tracking service to evaluate survival after primary CABG. Univariate, multivariate and a propensity analyses were performed.

RESULTS

We analysed 13 337 primary CABG procedures. Median follow-up was 7 years. Kaplan–Meier survival curves demonstrate that non-smokers have a significantly improved long-term survival compared with ex- and current smokers, P < 0.0001. Cox regression analysis identified smoking status, age, diabetes, ejection fraction (EF), body mass index, cerebrovascular disease, dialysis, left internal mammary artery (LIMA) non-usage, postoperative creatinine kinase muscle-brain isoenzyme (CKMB), radial artery usage, preoperative rhythm, forced vital capacity (FVC) and logistic EuroSCORE as significant risk factors determining long-term survival. Propensity matching resulted in 3575 non-smokers being matched 1:1, with ex-smokers. After matching, univariate analysis demonstrated the significantly worse long-term survival of ex-smokers compared with non-smokers, P < 0.0001. Cox regression analysis identified smoking status, age, postoperative CKMB, cerebrovascular disease, dialysis, diabetes, EF, FVC, LIMA non-usage, radial artery used, sinus rhythm and logistic EuroSCORE as significant risk factors determining long-term survival. Survival by smoking status plotted at the mean of the covariates, prepropensity matching, demonstrated that non-smokers had a significantly better long-term survival than ex-smokers, P < 0.0001; however, after propensity matching, non-smokers under 65 years of age had a significantly worse long-term survival compared with ex-smokers, P < 0.0001.

CONCLUSIONS

Non-smokers under the age of 65 years of age have significantly worse long-term survival compared with ex-smokers after risk factor adjustment. We speculate that this is because ex-smokers have had the causative factor, smoking, removed, but non-smokers have not.

Coronary, Survival, Smoking

INTRODUCTION

Smoking status is a risk factor for in-hospital morbidity, but not mortality [1, 2]. The detrimental impact of smoking status on long-term survival post-coronary artery bypass surgery (CABG) is widely accepted [3–6].

Smoking causes cardiovascular and respiratory compromises, which are associated with an increased risk of death post-surgery and reduced long-term survival.

We hypothesized that non-smokers have poorer long-term survival after CABG than ex-smokers. Non-smokers have an intrinsic risk factor for coronary/vascular disease, which does not alter after surgery, unlike ex-smokers whose main risk factor is usually smoking, but as they are ex-smokers this risk factor is no longer present. As non-smokers are less likely to have peripheral vascular disease, cerebrovascular disease or chronic obstructive airway disease, a propensity analysis to allow adjustment for these factors will be needed.

METHODS

Local institutional review board approval was granted for this study.

Database

A prospective single-institutional cardiac surgery database validated by the hospital data analysis department and accredited by the Society of Cardiothoracic Surgeons (SCTS) of Great Britain and Ireland was utilized in conjunction with the national strategic tracing service for long-term follow-up that exists in the UK, as has been described previously [7, 8].

All patients have their forced expiratory volume in 1 s (FEV1), and forced vital capacity (FVC) measured prior to surgery, by a single dedicated respiratory department.

The presence of peripheral vascular disease was based on clinical examination, or prior vascular assessment by ankle-brachial blood pressure index, computer tomography, magnetic resonance imaging, angiography, prior angioplasty/stenting or peripheral bypass surgery.

Smoking status was broken down into three categories: current smoker, ex-smoker and never smoker: current smoker is someone who currently smokes cigarettes regularly, ex-smoker is someone who used to smoke cigarettes regularly until 8 weeks prior to the date of surgery and non-smoker is someone who has never smoked cigarettes regularly [3].

Patients

Patients who had undergone isolated CABG only were included, n = 13 337, between 1 April 1997 and 31 March 2010.

Benchmarking

We benchmarked our in-hospital mortality figures against the UK national results (http://www.scts.org).

Methodology

The characteristics of the patients in this study are given in Table 1.

Table 1:
Open in new tab

Pre-, peri- and postoperative characteristic of patients in study group

Data, n = 13 337
Preoperative
 Age (years)64.911 (65.6), [58.9–71.6]
 Female, % (n)19.6 (2614)
 Body mass index (kg/m2)28.4, (27.9), [25.4–30.9]
 Heart failure, % (n)0.14 (19)
 Critical preoperative state, % (n)0.13 (17)
 Previous cerebral vascular accident, % (n)1 (133)
 Diabetes, % (n)36.7 (4895)
 Preoperative dialysis, % (n)0.4 (53)
 Previous myocardial infarction, % (n)48.9 (6522)
 Peripheral vascular disease, % (n)13.6 (1814)
 Hypertension, % (n)60.1 (802)
 Chronic obstructive pulmonary disease, % (n)
  None79.5 (10 603)
  Mild7.7 (1027)
  Moderate10.9 (1454)
  Severe1.9 (253)
 Type of lung disease, % (n)
  None85.0 (11 336)
  Obstructive13.2 (1757)
  Restrictive1.0 (133)
  Mixed0.8 (111)
 FEV1, %90, (93.0), [79.2–100]
 Preoperative IABP, % (n)1.3 (173)
 Ejection fraction
  Good, % (n)59.7 (7962)
  Moderate, % (n)31.4 (4188)
  Poor, % (n)8.9 (1187)
 Previous PCI, % (n)6.4 (854)
 Status
  Elective, % (n)80.5 (10 736)
  Urgent, % (n)18.2 (2428)
  Emergency, % (n)1.3 (173)
 Log EuroSCORE4.7, (2.6), [1.5–5.0]
Operative
 Left internal mammary artery, % (n)91.9 (12 257)
 Number of grafts3.4, (3.0), [3–4]
 Cardiopulmonary bypass time (min)109, (106), [88–127]
Postoperative
 Creatinine kinase myocardial isoenzyme30.3, (15), [4–34]
 Intensive care length of stay (days)2.2, (1), [1–2]
 Hospital length of stay (days)9.5 (7.0) [6–9]
 Mortality, % (n)2.1 (280)
 Median follow-up (years)7.02
Data, n = 13 337
Preoperative
 Age (years)64.911 (65.6), [58.9–71.6]
 Female, % (n)19.6 (2614)
 Body mass index (kg/m2)28.4, (27.9), [25.4–30.9]
 Heart failure, % (n)0.14 (19)
 Critical preoperative state, % (n)0.13 (17)
 Previous cerebral vascular accident, % (n)1 (133)
 Diabetes, % (n)36.7 (4895)
 Preoperative dialysis, % (n)0.4 (53)
 Previous myocardial infarction, % (n)48.9 (6522)
 Peripheral vascular disease, % (n)13.6 (1814)
 Hypertension, % (n)60.1 (802)
 Chronic obstructive pulmonary disease, % (n)
  None79.5 (10 603)
  Mild7.7 (1027)
  Moderate10.9 (1454)
  Severe1.9 (253)
 Type of lung disease, % (n)
  None85.0 (11 336)
  Obstructive13.2 (1757)
  Restrictive1.0 (133)
  Mixed0.8 (111)
 FEV1, %90, (93.0), [79.2–100]
 Preoperative IABP, % (n)1.3 (173)
 Ejection fraction
  Good, % (n)59.7 (7962)
  Moderate, % (n)31.4 (4188)
  Poor, % (n)8.9 (1187)
 Previous PCI, % (n)6.4 (854)
 Status
  Elective, % (n)80.5 (10 736)
  Urgent, % (n)18.2 (2428)
  Emergency, % (n)1.3 (173)
 Log EuroSCORE4.7, (2.6), [1.5–5.0]
Operative
 Left internal mammary artery, % (n)91.9 (12 257)
 Number of grafts3.4, (3.0), [3–4]
 Cardiopulmonary bypass time (min)109, (106), [88–127]
Postoperative
 Creatinine kinase myocardial isoenzyme30.3, (15), [4–34]
 Intensive care length of stay (days)2.2, (1), [1–2]
 Hospital length of stay (days)9.5 (7.0) [6–9]
 Mortality, % (n)2.1 (280)
 Median follow-up (years)7.02

Continuous variables are shown as mean, (median) [25th–75th interquartile range]. Categorical variables are shown as a percentage (n).

IABP: intra-aortic balloon pump; PCI: percutaneous coronary intervention.

Table 1:
Open in new tab

Pre-, peri- and postoperative characteristic of patients in study group

Data, n = 13 337
Preoperative
 Age (years)64.911 (65.6), [58.9–71.6]
 Female, % (n)19.6 (2614)
 Body mass index (kg/m2)28.4, (27.9), [25.4–30.9]
 Heart failure, % (n)0.14 (19)
 Critical preoperative state, % (n)0.13 (17)
 Previous cerebral vascular accident, % (n)1 (133)
 Diabetes, % (n)36.7 (4895)
 Preoperative dialysis, % (n)0.4 (53)
 Previous myocardial infarction, % (n)48.9 (6522)
 Peripheral vascular disease, % (n)13.6 (1814)
 Hypertension, % (n)60.1 (802)
 Chronic obstructive pulmonary disease, % (n)
  None79.5 (10 603)
  Mild7.7 (1027)
  Moderate10.9 (1454)
  Severe1.9 (253)
 Type of lung disease, % (n)
  None85.0 (11 336)
  Obstructive13.2 (1757)
  Restrictive1.0 (133)
  Mixed0.8 (111)
 FEV1, %90, (93.0), [79.2–100]
 Preoperative IABP, % (n)1.3 (173)
 Ejection fraction
  Good, % (n)59.7 (7962)
  Moderate, % (n)31.4 (4188)
  Poor, % (n)8.9 (1187)
 Previous PCI, % (n)6.4 (854)
 Status
  Elective, % (n)80.5 (10 736)
  Urgent, % (n)18.2 (2428)
  Emergency, % (n)1.3 (173)
 Log EuroSCORE4.7, (2.6), [1.5–5.0]
Operative
 Left internal mammary artery, % (n)91.9 (12 257)
 Number of grafts3.4, (3.0), [3–4]
 Cardiopulmonary bypass time (min)109, (106), [88–127]
Postoperative
 Creatinine kinase myocardial isoenzyme30.3, (15), [4–34]
 Intensive care length of stay (days)2.2, (1), [1–2]
 Hospital length of stay (days)9.5 (7.0) [6–9]
 Mortality, % (n)2.1 (280)
 Median follow-up (years)7.02
Data, n = 13 337
Preoperative
 Age (years)64.911 (65.6), [58.9–71.6]
 Female, % (n)19.6 (2614)
 Body mass index (kg/m2)28.4, (27.9), [25.4–30.9]
 Heart failure, % (n)0.14 (19)
 Critical preoperative state, % (n)0.13 (17)
 Previous cerebral vascular accident, % (n)1 (133)
 Diabetes, % (n)36.7 (4895)
 Preoperative dialysis, % (n)0.4 (53)
 Previous myocardial infarction, % (n)48.9 (6522)
 Peripheral vascular disease, % (n)13.6 (1814)
 Hypertension, % (n)60.1 (802)
 Chronic obstructive pulmonary disease, % (n)
  None79.5 (10 603)
  Mild7.7 (1027)
  Moderate10.9 (1454)
  Severe1.9 (253)
 Type of lung disease, % (n)
  None85.0 (11 336)
  Obstructive13.2 (1757)
  Restrictive1.0 (133)
  Mixed0.8 (111)
 FEV1, %90, (93.0), [79.2–100]
 Preoperative IABP, % (n)1.3 (173)
 Ejection fraction
  Good, % (n)59.7 (7962)
  Moderate, % (n)31.4 (4188)
  Poor, % (n)8.9 (1187)
 Previous PCI, % (n)6.4 (854)
 Status
  Elective, % (n)80.5 (10 736)
  Urgent, % (n)18.2 (2428)
  Emergency, % (n)1.3 (173)
 Log EuroSCORE4.7, (2.6), [1.5–5.0]
Operative
 Left internal mammary artery, % (n)91.9 (12 257)
 Number of grafts3.4, (3.0), [3–4]
 Cardiopulmonary bypass time (min)109, (106), [88–127]
Postoperative
 Creatinine kinase myocardial isoenzyme30.3, (15), [4–34]
 Intensive care length of stay (days)2.2, (1), [1–2]
 Hospital length of stay (days)9.5 (7.0) [6–9]
 Mortality, % (n)2.1 (280)
 Median follow-up (years)7.02

Continuous variables are shown as mean, (median) [25th–75th interquartile range]. Categorical variables are shown as a percentage (n).

IABP: intra-aortic balloon pump; PCI: percutaneous coronary intervention.

In-hospital mortality

Stepwise logistic regression was utilized to identify potential significant determining factors with regard to in-hospital mortality. The entry criterion was P < 0.05, and the removal criterion was P > 0.1 (Table 2).

Table 2:
Open in new tab

Logistic regression analysis of in-hospital mortality

VariableOR95% CI (OR)P-value
Non-smoker0.700.51–0.960.03
EuroSCORE1.051.04–1.06<0.0001
CKMB1.011.01–1.01<0.0001
Diabetes2.081.38–3.140.001
Age1.061.04–1.08<0.0001
Ejection fraction
 Moderate1.431.06–1.930.02
 Poor1.971.33–2.920.001
Cerebral vascular disease0.630.41–0.980.04
Dialysis5.102.00–13.00.001
Forced vital capacity0.400.22–0.730.003
VariableOR95% CI (OR)P-value
Non-smoker0.700.51–0.960.03
EuroSCORE1.051.04–1.06<0.0001
CKMB1.011.01–1.01<0.0001
Diabetes2.081.38–3.140.001
Age1.061.04–1.08<0.0001
Ejection fraction
 Moderate1.431.06–1.930.02
 Poor1.971.33–2.920.001
Cerebral vascular disease0.630.41–0.980.04
Dialysis5.102.00–13.00.001
Forced vital capacity0.400.22–0.730.003

CKMB: creatinine kinase muscle-brain isoenzyme.

Table 2:
Open in new tab

Logistic regression analysis of in-hospital mortality

VariableOR95% CI (OR)P-value
Non-smoker0.700.51–0.960.03
EuroSCORE1.051.04–1.06<0.0001
CKMB1.011.01–1.01<0.0001
Diabetes2.081.38–3.140.001
Age1.061.04–1.08<0.0001
Ejection fraction
 Moderate1.431.06–1.930.02
 Poor1.971.33–2.920.001
Cerebral vascular disease0.630.41–0.980.04
Dialysis5.102.00–13.00.001
Forced vital capacity0.400.22–0.730.003
VariableOR95% CI (OR)P-value
Non-smoker0.700.51–0.960.03
EuroSCORE1.051.04–1.06<0.0001
CKMB1.011.01–1.01<0.0001
Diabetes2.081.38–3.140.001
Age1.061.04–1.08<0.0001
Ejection fraction
 Moderate1.431.06–1.930.02
 Poor1.971.33–2.920.001
Cerebral vascular disease0.630.41–0.980.04
Dialysis5.102.00–13.00.001
Forced vital capacity0.400.22–0.730.003

CKMB: creatinine kinase muscle-brain isoenzyme.

Long-term survival

Univariate analysis

Kaplan–Meier survival curves were constructed, for patients classified as non-, ex- or current smokers (Fig. 1).

Kaplan–Meier survival curves for non-, ex-, and current smokers, P < 0.0001.
Figure 1:

Kaplan–Meier survival curves for non-, ex-, and current smokers, P < 0.0001.

Open in new tabDownload slide

Multivariate analysis

Stepwise Cox proportional hazards regression analysis was utilized to identify potential significant determining factors with regard to long-term survival. The entry criterion was P < 0.05, and the removal criterion was P > 0.1 (Table 3). The effect of smoking status was plotted at the mean of the covariates (Fig. 2).

Table 3:
Open in new tab

Cox regression analysis of survival

CovariateHR95% CI (HR)P-value
Study group
 Non-smoker0.740.66–0.83<0.0001
 Smoker1.341.18–1.52<0.0001
 Age1.061.05–1.07<0.0001
 Diabetes oral1.471.30–1.67<0.0001
 Diabetes insulin1.771.50–2.08<0.0001
 Ejection fraction
  Moderate1.341.22–1.48<0.0001
  Poor2.171.90–2.49<0.0001
 Body mass index0.990.98–1.000.02
 Cerebrovascular disease1.201.04–1.380.01
 Dialysis5.813.95–8.54<0.0001
 LIMA non-usage1.331.17–1.51<0.0001
 EuroSCORE1.021.02–1.03<0.0001
 CKMB1.001.00–1.00<0.0001
 Radial usage0.880.79–0.980.02
 Atrial fibrillation1.551.29–1.85<0.0001
 Forced vital capacity0.450.36–0.55<0.0001
After propensity matching
 Non-smoker0.710.63–0.81<0.0001
 Age1.071.06–1.08<0.0001
 CKMB1.001.00–1.00<0.0001
 Cerebrovascular disease1.261.03–1.540.03
 Dialysis7.134.64–10.9<0.0001
 Diabetes1.561.36–1.80<0.0001
 Ejection fraction
  Moderate1.381.21–1.58<0.0001
  Poor2.141.73–2.67<0.0001
 EuroSCORE1.021.01–1.020.0002
 Forced vital capacity0.470.35–0.64<0.0001
 LIMA non-usage1.471.23–1.77<0.0001
 Radial usage0.830.71–0.970.02
 Atrial fibrillation1.491.15–1.940.003
CovariateHR95% CI (HR)P-value
Study group
 Non-smoker0.740.66–0.83<0.0001
 Smoker1.341.18–1.52<0.0001
 Age1.061.05–1.07<0.0001
 Diabetes oral1.471.30–1.67<0.0001
 Diabetes insulin1.771.50–2.08<0.0001
 Ejection fraction
  Moderate1.341.22–1.48<0.0001
  Poor2.171.90–2.49<0.0001
 Body mass index0.990.98–1.000.02
 Cerebrovascular disease1.201.04–1.380.01
 Dialysis5.813.95–8.54<0.0001
 LIMA non-usage1.331.17–1.51<0.0001
 EuroSCORE1.021.02–1.03<0.0001
 CKMB1.001.00–1.00<0.0001
 Radial usage0.880.79–0.980.02
 Atrial fibrillation1.551.29–1.85<0.0001
 Forced vital capacity0.450.36–0.55<0.0001
After propensity matching
 Non-smoker0.710.63–0.81<0.0001
 Age1.071.06–1.08<0.0001
 CKMB1.001.00–1.00<0.0001
 Cerebrovascular disease1.261.03–1.540.03
 Dialysis7.134.64–10.9<0.0001
 Diabetes1.561.36–1.80<0.0001
 Ejection fraction
  Moderate1.381.21–1.58<0.0001
  Poor2.141.73–2.67<0.0001
 EuroSCORE1.021.01–1.020.0002
 Forced vital capacity0.470.35–0.64<0.0001
 LIMA non-usage1.471.23–1.77<0.0001
 Radial usage0.830.71–0.970.02
 Atrial fibrillation1.491.15–1.940.003

LIMA: left internal mammary artery; CKMB: creatinine kinase muscle-brain isoenzyme.

Table 3:
Open in new tab

Cox regression analysis of survival

CovariateHR95% CI (HR)P-value
Study group
 Non-smoker0.740.66–0.83<0.0001
 Smoker1.341.18–1.52<0.0001
 Age1.061.05–1.07<0.0001
 Diabetes oral1.471.30–1.67<0.0001
 Diabetes insulin1.771.50–2.08<0.0001
 Ejection fraction
  Moderate1.341.22–1.48<0.0001
  Poor2.171.90–2.49<0.0001
 Body mass index0.990.98–1.000.02
 Cerebrovascular disease1.201.04–1.380.01
 Dialysis5.813.95–8.54<0.0001
 LIMA non-usage1.331.17–1.51<0.0001
 EuroSCORE1.021.02–1.03<0.0001
 CKMB1.001.00–1.00<0.0001
 Radial usage0.880.79–0.980.02
 Atrial fibrillation1.551.29–1.85<0.0001
 Forced vital capacity0.450.36–0.55<0.0001
After propensity matching
 Non-smoker0.710.63–0.81<0.0001
 Age1.071.06–1.08<0.0001
 CKMB1.001.00–1.00<0.0001
 Cerebrovascular disease1.261.03–1.540.03
 Dialysis7.134.64–10.9<0.0001
 Diabetes1.561.36–1.80<0.0001
 Ejection fraction
  Moderate1.381.21–1.58<0.0001
  Poor2.141.73–2.67<0.0001
 EuroSCORE1.021.01–1.020.0002
 Forced vital capacity0.470.35–0.64<0.0001
 LIMA non-usage1.471.23–1.77<0.0001
 Radial usage0.830.71–0.970.02
 Atrial fibrillation1.491.15–1.940.003
CovariateHR95% CI (HR)P-value
Study group
 Non-smoker0.740.66–0.83<0.0001
 Smoker1.341.18–1.52<0.0001
 Age1.061.05–1.07<0.0001
 Diabetes oral1.471.30–1.67<0.0001
 Diabetes insulin1.771.50–2.08<0.0001
 Ejection fraction
  Moderate1.341.22–1.48<0.0001
  Poor2.171.90–2.49<0.0001
 Body mass index0.990.98–1.000.02
 Cerebrovascular disease1.201.04–1.380.01
 Dialysis5.813.95–8.54<0.0001
 LIMA non-usage1.331.17–1.51<0.0001
 EuroSCORE1.021.02–1.03<0.0001
 CKMB1.001.00–1.00<0.0001
 Radial usage0.880.79–0.980.02
 Atrial fibrillation1.551.29–1.85<0.0001
 Forced vital capacity0.450.36–0.55<0.0001
After propensity matching
 Non-smoker0.710.63–0.81<0.0001
 Age1.071.06–1.08<0.0001
 CKMB1.001.00–1.00<0.0001
 Cerebrovascular disease1.261.03–1.540.03
 Dialysis7.134.64–10.9<0.0001
 Diabetes1.561.36–1.80<0.0001
 Ejection fraction
  Moderate1.381.21–1.58<0.0001
  Poor2.141.73–2.67<0.0001
 EuroSCORE1.021.01–1.020.0002
 Forced vital capacity0.470.35–0.64<0.0001
 LIMA non-usage1.471.23–1.77<0.0001
 Radial usage0.830.71–0.970.02
 Atrial fibrillation1.491.15–1.940.003

LIMA: left internal mammary artery; CKMB: creatinine kinase muscle-brain isoenzyme.

The effect of smoking status on long-term survival plotted at the mean of the covariates. (A) Whole study cohort, (B) patients <65 years old and (C) patients over 65 years old.
Figure 2:

The effect of smoking status on long-term survival plotted at the mean of the covariates. (A) Whole study cohort, (B) patients <65 years old and (C) patients over 65 years old.

Open in new tabDownload slide

Propensity analysis

Logistic regression for group membership of smoking status was used to calculate the propensity score for 1:1 matching. Nearest neighbour matching without replacement with a caliper of 0.2 was utilized. Current smokers were excluded from analysis.

Variables used in the propensity match included: age, ejection fraction (EF), dialysis, EuroSCORE, postoperative creatinine kinase muscle-brain isoenzyme (CKMB), left internal mammary artery (LIMA) usage, radial artery usage, previous cerebrovascular disease, sinus rhythm, diabetes, preoperative FEV1 and FVC.

A dotplot of standardized mean differences (Cohen's d) for all covariates before and after matching was produced for patients who only received a statin. Kaplan–Meier and Cox survival curves post-propensity matching were created. Full propensity analysis data are supplied in Supplementary data.

Statistical software

All statistical analyses other than the propensity matching were performed with MedCalc for Windows (version 12.1.4, MedCalc Software, Mariakerke, Belgium). The propensity matching was performed with SPSS (version 20.0 for Windows, SPSS, Inc., Chicago, IL, USA), SPSS Statistics Integration Plug-In for R, and R 2.12.2.

RESULTS

100% long-term follow-up via the national strategic tracing service was achieved. Benchmarking of our institutional mortality rates compared with the UK did not reveal any differences (part of the continuous UK cardiac surgery quality assessment programme by the SCTS). The logistic EuroSCORE ranged from 0.881 to 80.411. The long-term follow-up was from 6 months to 12 years, mean 6.6 years and median 7 years.

In-hospital mortality

Logistic regression (ROC 0.83, Hosmer–Lemeshow 0.13), identified logistic EuroSCORE (odds ratio [OR] 1.05, 95% confidence interval [CI] 1.04–1.06), P < 0.0001, age (OR 1.06, 95% CI 1.04–1.08), P < 0.0001, dialysis (OR 5.1, 95% CI 2.0–13.0), P = 0.001, FVC (OR 0.40, 95% CI 0.22–0.73), P = 0.003, cerebrovascular disease (OR 0.64, 95% CI 0.41–0.98), P = 0.04, diabetes (OR 2.1, 95% CI 1.4–3.1), P = 0.001, moderate left ventricular EF (OR 1.4, 95% CI 1.1–1.9), P = 0.02, poor EF (OR 2.0, 95% CI 1.3–2.9), P = 0.001, CKMB release (OR 1.01, 95% CI 1.01–1.01), P < 0.0001 and non-smoker (OR 0.70, 95% CI 0.51–0.96), P = 0.03, as significant risk factors determining in-hospital mortality, Table 2. LIMA usage and FEV1 were excluded as significant risk factors.

Long-term survival

Univariate

Kaplan–Meier survival curves demonstrate that non-smokers have a significantly improved long-term survival compared with ex- and current smokers, P < 0.0001 (Fig. 1). With regard to survival, 433 events were recorded in non-smokers, 1331 events were recorded in ex-smokers and 332 events were recorded in current smokers.

Cox regression analysis

Study group

Cox regression analysis (c-index 0.75) identified smoking status, age, diabetes, EF, body mass index (BMI), cerebrovascular disease, dialysis, LIMA non-usage, postoperative CKMB, radial artery usage, preoperative rhythm, FVC and logistic EuroSCORE as significant risk factors determining long-term survival, Table 3 (study group). Female sex, hypertension, graft number, off-pump surgery and FEV1 were excluded by the model as being significant factors.

Cox regression analysis demonstrated that non-smokers had a hazard ratio (HR) <1, indicating that they have better survival than ex-smokers, as demonstrated when survival by smoking status was plotted at the mean of the covariates (Fig. 2A). When survival was broken down by age of surgery, a break point of 65 years old was found. Below 65 years of age non-smokers had significantly improved long-term survival compared with ex-smokers, Fig. 2B. However, above 65 years of age, non-smokers and ex-smokers had identical long-term survival (Fig. 2C).

Propensity-matched group

Propensity matching resulted in 3575 non-smokers being matched 1:1, with ex-smokers. Current smokers were excluded from propensity matching due to the methodological nature of matching. With regard to survival, 433 events were recorded in non-smokers (identical to prepropensity matching as 1:1 matching occurred), and 593 events were recorded in ex-smokers.

The mean logistic EuroSCORE before matching was 4.7 (25th–75th percentile 1.5–5.0) compared with a mean risk of 2.6 (25th–75th percentile of 1.4–4.8) post-matching, P = 0.01 (Mann–Whitney U-test—independent samples).

A dotplot of standardized mean differences (Cohen's d) for all covariates before and after matching for statin therapy is shown in Fig. 3, and overall χ2 balance test was not significant, χ2 (12) = 5.0, P = 0.96. Full details of the propensity match are enclosed in the Supplementary data.

Dotplot of standardized mean differences (Cohen's d—x axis) for all covariates before and after matching for statin therapy. Overall χ2 balance test was not significant, χ2 (12) = 5.0, P = 0.96. EF: ejection fraction, CKMB: postoperative creatinine kinase muscle-brain isoenzyme, LIMA: left internal mammary artery, CVASC: previous cerebrovascular event, SR: sinus rhythm preoperatively, DM: diabetes (oral, medication or insulin controlled).
Figure 3:

Dotplot of standardized mean differences (Cohen's d—x axis) for all covariates before and after matching for statin therapy. Overall χ2 balance test was not significant, χ2 (12) = 5.0, P = 0.96. EF: ejection fraction, CKMB: postoperative creatinine kinase muscle-brain isoenzyme, LIMA: left internal mammary artery, CVASC: previous cerebrovascular event, SR: sinus rhythm preoperatively, DM: diabetes (oral, medication or insulin controlled).

Open in new tabDownload slide

Univariate analysis demonstrated the significantly worse long-term survival of ex-smokers compared with non-smokers, Fig. 4A, P < 0.0001.

Post-propensity matching, n = 3575. (A) Kaplan–Meier survival curves for non- and ex-smokers, P < 0.0001 and (B) the effect of smoking status on long-term survival plotted at the mean of the covariates.
Figure 4:

Post-propensity matching, n = 3575. (A) Kaplan–Meier survival curves for non- and ex-smokers, P < 0.0001 and (B) the effect of smoking status on long-term survival plotted at the mean of the covariates.

Open in new tabDownload slide

Cox regression analysis (c-index 0.73) identified smoking status, age, postoperative CKMB, cerebrovascular disease, dialysis, diabetes, EF, FVC, LIMA non-usage, radial artery used, sinus rhythm and logistic EuroSCORE as significant risk factors determining long-term survival, Table 3 (after propensity matching). FEV1 was excluded by the model as being significant factors.

Non-smokers had a HR of <1, indicating that they have better survival than ex-smokers. Survival post propensity matching identified smoking status when plotted at the mean of the covariates, had minimal clinical effect on survival, Fig. 4B. When survival was broken down by age group, patients <65 years old undergoing CABG had a significantly worse long-term survival, Fig. 5A, compared with those >65 years old, Fig. 5B.

The effect of smoking status post-propensity matching on long-term survival plotted at the mean of the covariates. (A) Patients <65 years old and (B) patients >65 years old.
Figure 5:

The effect of smoking status post-propensity matching on long-term survival plotted at the mean of the covariates. (A) Patients <65 years old and (B) patients >65 years old.

Open in new tabDownload slide

Two additional propensity matches were performed in patients below and above 65 years of age to eliminate any bias due to subgroup analysis. The same conclusions with regard to age, smoking status and long-term survival were found (data not shown).

DISCUSSION

Non-smokers as a group have significantly improved long-term survival compared with ex- and current smokers. However, after risk adjustment, non-smokers under 65 years of age actually have a worse long-term survival compared with ex-smokers. As neither of these groups now smoke, we postulate that the difference is secondary to the removal of the inciting risk factor, smoking, in the ex-smokers.

The explanation of the finding of an age cut-off of 65 years with regard to the effect of smoking status and survival is speculative. We postulate that non-smoking individuals that develop coronary artery disease prior to 65 years of age are genetically susceptible and are thus different to their smoking counterparts; however we provide no evidence to support this.

Smoking causes cardiovascular and respiratory diseases, which are associated with reduced long-term survival. Cessation of smoking removes the causative risk factor but the risk secondary to previous organ damage remains. After cessation of smoking, the long-term survival slowly reduces back to baseline. In non-smokers the underlying risk factor(s) if not hypertension, hypercholesterolaemia, thyroid dysfunction, renal failure or diabetes remain unchanged. The above risk factors are treated in the same manner regardless of smoking status. We hypothesized that non-smokers will have a worse long-term survival compared with ex-smokers as they have something genetically different in their constitution. The incidence of this ‘genetic difference’ will be the same as the general population in the ex-smoker cohort. We thus speculate that non-smokers are an enriched cohort of ‘bad genetic risk’ individuals.

COPD has been shown to be a significant factor determining long-term survival in patients with ischaemic heart disease patients and post-CABG [3, 9]. Numerous publications frequently do not include pulmonary disease as a potential risk factor for long term survival, or use qualitative criteria, such as clinical opinion. We recommend that FEV1, FVC and preferably if available pulmonary diffusion should be utilized to classify patients pulmonary disease by the Global Initiative for Chronic Obstructive Lung Disease criteria, and restrictive or obstructive [10]. Failure to quantify pulmonary dysfunction makes comparison between different publications difficult. Obstructive and restive lung diseases represent different pathological processes. Analysis of the restrictive and obstructive components separately identified that restrictive lung disease implies a worse long-term survival due to a higher HR on Cox analysis, confirming previous work [3].

Arterial grafting is beneficial to long-term survival [11–13]. The use of the internal mammary artery is known to be associated with increased survival post-CABG; however, it has not been included in a number of previous long-term prediction studies. Concerns about LIMA usage in patients who smoke with COPD exist due to the concerns about pulmonary dysfunction, but are unfounded [3]. Multivariate analysis revealed that LIMA usage is not a risk factor with regard to in-hospital mortality, but is a significant factor determining long-term survival regardless of pulmonary function. The number of patients with bilateral internal mammary arteries was too small in our dataset to analyse. We feel that our data demonstrating radial artery superiority with regard to long-term survival are subject to bias as this conduit is only used selectively in ‘fitter’ patients in our institution. This finding is supported by the observation that uncontrolled studies demonstrate a superiority of the radial artery, but randomized trails do not [14, 15].

Intraoperative myocardial damage is the main cause of cardiac dysfunction post-cardiac surgery, be it myocardial stunning, inadequate cardioplegia or coronary graft dysfunction. The finding that CKMB was a significant factor affecting long-term outcome demonstrates the importance of intraoperative myocardial protection and coronary anastomosis technique and has been previously identified as a significant factor effecting long-term survival [8, 16]. None of the previous studies of smoking status have adjusted for this intraoperative risk factor.

Debate continues with regard to the use of off-pump techniques. Multivariate analysis demonstrated that off-pump is not associated with a significant difference in hospital mortality, or long-term survival in our cohort. Identifying BMI and the absence of sinus rhythm as significant factors determining long-term survival confirms previous findings [17–19].

The value of multivariate analysis and propensity matching was demonstrated by the difference in outcomes by smoking status. Univariate analysis pre- and post-propensity matching demonstrated the superior survival of non-smokers; however, after risk factor adjustment (propensity matching), non-smokers under 65 years of age actually had a significantly worse long-term survival, a finding missed by Cox regression of the raw dataset. Performing a multivariate analysis after a propensity match is a previously described methodology [20]. The concept of survival plotted at the mean of the covariates, a well-described Cox output, is key to understanding the results of this manuscript [21].

Non-smokers anecdotally rarely start smoking after cardiac surgery; however ex-smokers frequently restart or have occasional cigarettes. As the ex-smoker group had better long-term survival than the non-smokers after risk factor adjustment, this potential confounding error will not have influenced our results.

Non-smokers under the age of 65 years have significantly worse long-term survival compared with ex-smokers after risk factor adjustment. We speculate that this is because ex-smokers have had the causative factor, smoking, removed, but non-smokers have not.

Limitations

We do not have cause of death of the patients, a frequent limitation of retrospective analysis; however, our series is large, prospective and consecutive and features 100% long-term follow-up. Very few patients received bilateral internal mammary grafts (BIMA), so this factor could not be analysed without introducing a statistical error to a lack of power. We are unable to account for patient selection bias, a limitation of all retrospective analyses. We made no attempt to breakdown the follow-up period into time intervals as have been done previously. The 14-year period that the patients in this study were recruited from may introduce a time factor error. This is a common confounding factor of any large study over a prolonged period. Cox modelling is a mathematical approximation; however, a c-index above 0.7, for both analyses, indicates that our model is reasonably accurate.

Smoking status was not recorded postoperatively; however, we speculate that non-smokers are unlikely to start smoking after undergoing CABG. In addition, the number of ex-smokers (defined as having stopped completely for a period of 6 weeks prior to surgery) who restart smoking after undergoing CABG is likely to be small.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Funding

None.

Conflict of interest: none declared.

REFERENCES

1
Al Sarraf
N
Thalib
L
Hughes
A
Tolan
M
Young
V
McGovern
E
,
Effect of smoking on short-term outcome of patients undergoing coronary artery bypass surgery
, 
Ann Thorac Surg
,
2008
, vol.
86
(pg.
517
-
23
)

Google Scholar

Crossref
Search ADS
PubMed

 
2
Ashraf
MN
Mortasawi
A
Grayson
AD
Oo
AY
,
Effect of smoking status on mortality and morbidity following coronary artery bypass surgery
, 
Thorac Cardiovasc Surg
,
2004
, vol.
52
(pg.
268
-
73
)

Google Scholar

Crossref
Search ADS
PubMed

 
3
O'Boyle
F
Mediratta
N
Chalmers
J
Al Rawi
O
Mohan
K
Shaw
M
et al. ,
Long-term survival of patients with pulmonary disease undergoing coronary artery bypass surgery
, 
Eur J Cardiothorac Surg
,
2013
, vol.
43
(pg.
697
-
703
)

Google Scholar

Crossref
Search ADS
PubMed

 
4
Saxena
A
Shan
L
Reid
C
Dinh
DT
Smith
JA
Shardey
GC
et al. ,
Impact of smoking status on early and late outcomes after isolated coronary artery bypass graft surgery
, 
J Cardiol
,
2013
, vol.
61
(pg.
336
-
41
)

Google Scholar

Crossref
Search ADS
PubMed

 
5
Cavender
JB
Rogers
WJ
Fisher
LD
Gersh
BJ
Coggin
CJ
Myers
WO
,
Effects of smoking on survival and morbidity in patients randomized to medical or surgical therapy in the Coronary Artery Surgery Study (CASS): 10-year follow-up. CASS investigators
, 
J Am Coll Cardiol
,
1992
, vol.
20
(pg.
287
-
94
)

Google Scholar

Crossref
Search ADS
PubMed

 
6
van Domburg
RT
Meeter
K
van Berkel
DF
Veldkamp
RF
van Herwerden
LA
Bogers
AJ
,
Smoking cessation reduces mortality after coronary artery bypass surgery: a 20-year follow-up study
, 
J Am Coll Cardiol
,
2000
, vol.
36
(pg.
878
-
83
)

Google Scholar

Crossref
Search ADS
PubMed

 
7
Chalmers
J
Mediratta
N
McShane
J
Shaw
M
Pullan
M
Poullis
M
,
The long-term effects of developing renal failure post-coronary artery bypass surgery, in patients with normal preoperative renal function
, 
Eur J Cardiothorac Surg
,
2013
, vol.
43
(pg.
555
-
9
)

Google Scholar

Crossref
Search ADS
PubMed

 
8
Mediratta
N
Chalmers
J
Pullan
M
McShane
J
Shaw
M
Poullis
M
,
In-hospital mortality and long-term survival after coronary artery bypass surgery in young patients
, 
Eur J Cardiothorac Surg
,
2013
, vol.
43
(pg.
1014
-
21
)

Google Scholar

Crossref
Search ADS
PubMed

 
9
Nishiyama
K
Morimoto
T
Furukawa
Y
Nakagawa
Y
Ehara
N
Taniguchi
R
et al. ,
Chronic obstructive pulmonary disease—an independent risk factor for long-term cardiac and cardiovascular mortality in patients with ischemic heart disease
, 
Int J Cardiol
,
2010
, vol.
143
(pg.
178
-
83
)

Google Scholar

Crossref
Search ADS
PubMed

 
10
Vestbo
J
Hurd
SS
Rodriguez-Roisin
R
,
The 2011 revision of the global strategy for the diagnosis, management and prevention of COPD (GOLD)—why and what?
, 
Clin Respir J
,
2012
, vol.
6
(pg.
208
-
14
)

Google Scholar

Crossref
Search ADS
PubMed

 
11
Boylan
MJ
Lytle
BW
Loop
FD
Taylor
PC
Borsh
JA
Goormastic
M
et al. ,
Surgical treatment of isolated left anterior descending coronary stenosis. Comparison of left internal mammary artery and venous autograft at 18 to 20 years of follow-up
, 
J Thorac Cardiovasc Surg
,
1994
, vol.
107
(pg.
657
-
62
)

Google Scholar

PubMed
OpenURL Placeholder Text

 
12
Cameron
A
Davis
KB
Green
G
Schaff
HV
,
Coronary bypass surgery with internal-thoracic-artery grafts—effects on survival over a 15-year period
, 
N Engl J Med
,
1996
, vol.
334
(pg.
216
-
9
)

Google Scholar

Crossref
Search ADS
PubMed

 
13
Lytle
BW
Blackstone
EH
Loop
FD
Houghtaling
PL
Arnold
JH
Akhrass
R
et al. ,
Two internal thoracic artery grafts are better than one
, 
J Thorac Cardiovasc Surg
,
1999
, vol.
117
(pg.
855
-
72
)

Google Scholar

Crossref
Search ADS
PubMed

 
14
Hayward
PA
Gordon
IR
Hare
DL
Matalanis
G
Horrigan
ML
Rosalion
A
et al. ,
Comparable patencies of the radial artery and right internal thoracic artery or saphenous vein beyond 5 years: results from the Radial Artery Patency and Clinical Outcomes trial
, 
J Thorac Cardiovasc Surg
,
2010
, vol.
139
(pg.
60
-
5
)

Google Scholar

Crossref
Search ADS
PubMed

 
15
Zacharias
A
Habib
RH
Schwann
TA
Riordan
CJ
Durham
SJ
Shah
A
,
Improved survival with radial artery versus vein conduits in coronary bypass surgery with left internal thoracic artery to left anterior descending artery grafting
, 
Circulation
,
2004
, vol.
109
(pg.
1489
-
96
)

Google Scholar

Crossref
Search ADS
PubMed

 
16
Domanski
MJ
Mahaffey
K
Hasselblad
V
Brener
SJ
Smith
PK
Hillis
G
et al. ,
Association of myocardial enzyme elevation and survival following coronary artery bypass graft surgery
, 
JAMA
,
2011
, vol.
305
(pg.
585
-
91
)

Google Scholar

Crossref
Search ADS
PubMed

 
17
Kalavrouziotis
D
Buth
KJ
Vyas
T
Ali
IS
,
Preoperative atrial fibrillation decreases event-free survival following cardiac surgery
, 
Eur J Cardiothorac Surg
,
2009
, vol.
36
(pg.
293
-
9
)

Google Scholar

Crossref
Search ADS
PubMed

 
18
Kuduvalli
M
Grayson
AD
Oo
AY
Fabri
BM
Rashid
A
,
Risk of morbidity and in-hospital mortality in obese patients undergoing coronary artery bypass surgery
, 
Eur J Cardiothorac Surg
,
2002
, vol.
22
(pg.
787
-
93
)

Google Scholar

Crossref
Search ADS
PubMed

 
19
Warwick
R
Mediratta
N
Chalmers
J
Pullan
M
Shaw
M
McShane
J
et al. ,
Is single-unit blood transfusion bad post-coronary artery bypass surgery?
, 
Interact CardioVasc Thorac Surg
,
2013
, vol.
16
(pg.
765
-
71
)

Google Scholar

Crossref
Search ADS
PubMed

 
20
Angeloni
E
Melina
G
Benedetto
U
Refice
S
Bianchi
P
Quarto
C
et al. ,
Statins improve outcome in isolated heart valve operations: a propensity score analysis of 3217 patients
, 
Ann Thorac Surg
,
2011
, vol.
92
(pg.
68
-
73
)

Google Scholar

Crossref
Search ADS
PubMed

 
21
Kleinbaum
D
Klein
M
, 
Survival Analysis: A Self-Learning Text (Statistics for Biology and Health)
,
2012
New York, NY
Springer
(pg.
97
-
160
)
 
© The Author 2013. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
Topic:
Issue Section:
ADULT CARDIAC
Download all slides
  • Supplementary data

    • Supplementary data

    Supplementary Data - zip file