J Neurosurg 120:811–819, 2014 ©AANS, 2014
Smoking and postoperative outcomes in elective cranial surgery Clinical article Nima Alan, B.S.,1 Andreea Seicean, Ph.D., M.P.H.,1,2 Sinziana Seicean, M.D., M.P.H., Ph.D., 3,4 Nicholas K. Schiltz, Ph.D., 2 Duncan Neuhauser, Ph.D., 2 and Robert J. Weil, M.D. 5 Case Western Reserve University School of Medicine, Cleveland; 2Department of Epidemiology and Biostatistics, Case Western Reserve University; 3Departments of Pulmonary, Critical Care, and Sleep Medicine, University Hospitals, Cleveland; 4Heart and Vascular Institute, Cleveland Clinic; and 5The Rose Ella Burkhardt Brain Tumor & Neuro-Oncology Center, and Department of Neurosurgery, the Neurological Institute, Cleveland Clinic, Cleveland, Ohio 1
Object. The goal in this study was to assess whether a current or prior history of smoking and the number of smoking pack years affect the risk for adverse outcomes in the 30-day postoperative period in patients who undergo elective cranial surgery. Methods. Data from the 2006–2011 American College of Surgeons’ National Surgical Quality Improvement Project were used in this study. The authors identified 8296 patients who underwent elective cranial surgery, of whom 1718 were current smokers, 854 were prior smokers, and 5724 were never smokers. Using propensity scores and age, the authors matched current and prior smokers to never smokers. Odds ratios for adverse postoperative outcomes were predicted with logistic regression. The relationship between number of pack years and poor outcomes was also examined. Results. In unadjusted analyses, prior and current smokers did not differ from never smokers for having poor outcomes postoperatively. Similarly, in matched analyses, no association was found between smoking and adverse outcomes. Number of pack years in propensity-matched analyses did not predict worse outcomes in prior or current smokers versus never smokers. Conclusions. The authors did not find smoking to be associated with 30-day postoperative morbidity or mortality. Although smoking cessation is beneficial for overall health, it may not improve the short-term (≤ 30 days) outcome of elective cranial surgery. Thus postponement of elective cranial cases only for smoking cessation may not be necessary. (http://thejns.org/doi/abs/10.3171/2014.1.JNS131852)
Key Words • smoking • cranial surgery • morbidity • mortality • health services research
A
to a 2011 report from the Centers for Disease Control and Prevention, more than 19% of adults in the US smoke cigarettes.5 It is well known that smoking is associated with significant morbidity and mortality.22 Evidence for the benefits of smoking cessation are significant; recently Jha et al.8 demonstrated that abstention from tobacco smoking at any age between the ages of 25 and 54 years reduces risk of death associated with smoking. ccording
Abbreviations used in this paper: ACS = American College of Surgeons; BMI = body mass index; BUN = blood urea nitrogen; CPT = Current Procedure Terminology; IQR = interquartile range; LOS = length of stay; NSQIP = National Surgical Quality Improvement Program; OR = operating room; SAH = subarachnoid hemorrhage.
J Neurosurg / Volume 120 / April 2014
In patients undergoing a variety of nonneurosurgical operations, smoking has been identified as an independent risk factor for poor perioperative outcomes.7,13,23 In spine surgery there is equivocal evidence of an association between smoking and postoperative outcomes.18,20,24 In cranial neurosurgery the impact of smoking on outcomes has been studied in only limited clinical settings. Smoking is a known risk factor for formation of intracranial aneurysm,3 subarachnoid hemorrhage (SAH),9,10 and symptomatic vasospasm.14 In patients undergoing craniotomy for tumor resection, Lau et al.15 reported that smoking increased the risk for morbidity and the 1-year mortality rate. However, the effect of cigarette smoking as a potentially modifiable preoperative risk factor that influences outcomes in patients undergoing elective cranial surgeries is unknown. 811
N. Alan et al. We used a large, national, prospectively collected database from the National Surgical Quality Improvement Program (NSQIP) to examine whether perioperative smoking in patients undergoing elective cranial surgery influenced 30-day outcomes postoperatively.
Methods Data Source
We identified patients who underwent elective cranial surgery by using the American College of Surgeons (ACS)–NSQIP database for records obtained between 2006 and 2011. Detailed descriptions of the ACS-NSQIP database can be found elsewhere.1,12 The study was approved by the Cleveland Clinic Institutional Review Board. Patient Population
We used the Current Procedure Terminology (CPT) codes to identify 10,651 patients who were 18 years of age and older, and who underwent cranial surgery between 2006 and 2011. We excluded 2122 emergency cases because smoking cessation intervention would not be possible. We also excluded 233 patients with unknown smoking status. Our final study sample consisted of 8296 patients with known smoking status, who underwent elective cranial surgery.
Smoking Status
We grouped patients according to their smoking status: current, prior, and never smoker.7,20 The NSQIP database contains 2 smoking variables: the first is “current smoker,” defined as smoking within 1 year prior to surgery and smoking duration as determined by total number of pack years. A current smoker was defined as a patient with “yes” for the NSQIP current-smoker variables and having either a missing value or a value > 0 for pack years. The second is a “prior smoker,” defined as having a “no” for the current-smoker variable and a value > 0 for pack years. We defined a “never smoker” as someone having a “no” for the current-smoker variable and a value of 0 or missing for the pack-year variable. Covariates
Available perioperative factors that may affect postoperative outcome were analyzed (Table 1).
Postoperative Outcomes
Short-term postoperative complications, defined as occurring within 30 days after surgery, were assessed for each patient group. Prolonged length of stay (LOS) was defined as LOS longer than that of the third quartile of the mean LOS; thus, prolonged LOS was considered to be ≥ 8 days. Minor complications consisted of ≥ 1 of the following: superficial surgical site infection, urinary tract infection, deep venous thrombosis, or thrombophlebitis. Major complications were ≥ 1 of the following: deep incision surgical site infection, organ or space surgical site infection, wound disruption, pneumonia, unplanned in-
812
tubation, pulmonary embolism, > 48 hour postoperative ventilator-assisted respiration, progressive renal insufficiency, acute renal failure, cardiovascular accident with neurological deficit, coma of > 24 hours, peripheral nerve injury, cardiac arrest requiring cardiopulmonary resuscitation, myocardial infarction, graft, prosthesis or flap failure, sepsis, septic shock, and/or return within 30 days to the operating room (OR). Any complication was defined as ≥ 1 minor or major complication. Thirty-day return to the OR, defined as any unplanned return to the OR for a surgical procedure, and 30-day mortality were also assessed. Statistical Analyses
Preoperative and intraoperative and 30-day outcomes were compared between smokers (current and prior) and never smokers by using Pearson chi-square tests for categorical variables and ANOVA for continuous variables. Propensity scores, including all variables from Table 1, were generated to obtain an approximately unbiased measure of the association between smoking and adverse outcomes.19 A 1:1 greedy matching technique4 was used first to match prior smokers with a unique never smoker on both propensity score and age, and similarly, current smokers with a unique never smoker. Restrictions were set on the matching criteria to within 0.2 SDs of the propensity score, to allow for increased accuracy in matching. Covariates were compared between smoking groups in the matched sample. Logistic regression analysis was used to test whether different smoking categories were independently associated with adverse outcomes. Covariates that remained unbalanced after matching based on propensity scores were included in the final models. The relationship between pack years and adverse outcomes was also tested in the entire sample, pooling current and prior smokers and, after matching based on propensity score was performed, separately comparing current with never smokers and prior with never smokers. A p value of < 0.05 was considered statistically significant; SAS (version 9.2, SAS Institute) was used.
Results
Compared with never smokers, current smokers were younger by an average of 4 years, and were more likely to be male and African American (Table 1). Daily consumption of > 2 alcoholic beverages was more prevalent among current smokers versus never smokers (4.8% vs 1.3%, p < 0.001), and the former had a slightly lower body mass index ([BMI] 27.6 ± 6.4 vs 28.5 ± 6.5, p < 0.001). Furthermore, current smokers were more likely to have pulmonary (10.7% vs 3.1%, p < 0.001) and CNS (19.1% vs 13.1%, p < 0.001) comorbidities than never smokers. Disseminated cancer was present in 16.8% of current smokers versus 10.0% of never smokers (p < 0.001). However, hypertension (34.9% vs 39.9%, p < 0.001) and diabetes (9.8% vs 11.6%, p = 0.04) were more common among never smokers. Current smokers, compared with never smokers, were more likely to have abnormal albumin (10.9% vs 8.3%, p < 0.001), serum glutamic oxaloacetic J Neurosurg / Volume 120 / April 2014
Smoking and outcomes in cranial surgery TABLE 1: Demographic data, comorbidities, preoperative laboratory values, and intraoperative factors according to smoking status in 8296 patients* Smoking Status Baseline Characteristic
Current
Prior
Never
no. of patients (%) age in yrs, mean ± SD female sex African American admitted from home >2 drinks per day partially or fully dependent functional status ASA classification 1+2 3+4 BMI in kg/m2, mean ± SD pulmonary comorbidities cardiovascular comorbidities hypertension requiring meds dyspnea CNS comorbidities diabetes mellitus disseminated cancer weight loss >10% in 6 mos prior to surgery steroid use for chronic condition preop chemo &/or RT bleeding disorder prior op w/in 30 days abnormal albumin abnormal alkaline phosphatase abnormal BUN abnormal creatinine abnormal hematocrit abnormal platelet count abnormal SGOT abnormal sodium abnormal total bilirubin abnormal WBC count level of residency supervision in OR attending alone attending & resident wound class clean clean-contaminated contaminated or infected intraop or postop transfusions
1718 (20.7) 51 ± 14 51.9 9.7 88.0 4.8 10.8
854 (10.3) 61 ± 13 46.5 6.6 92.4 3.9 12.4
5724 (69) 55 ± 17 54.9 6.7 89.5 1.3 10.8
27.6 72.1 27.6 ± 6.4 10.7 5.1 34.9 10.3 19.1 9.8 16.8 3.7 13.1 2.2 2.4 3.1 10.9 2.7 21.1 3.7 44.8 7.6 6.1 9.6 1.9 29.8
25.4 73.8 29.5 ± 6.8 9.7 12.8 52.2 10.9 21.4 16.7 14.6 3.3 15.0 6.0 3.9 5.3 8.2 3.4 27.2 7.1 43.6 9.7 6.8 8.2 2.1 23.2
35.2 64.4 28.5 ± 6.5 3.1 4.7 39.9 4.8 13.1 11.6 10.0 1.8 12.7 2.1 2.9 2.8 8.3 3.0 24.7 4.7 44.3 8.1 5.7 8.4 2.5 24.8
27.2 49.0
33.1 66.5
25.8 44.6
92.9 4.3 2.8 5.0
91.2 6.3 2.5 6.2
92.8 4.8 2.4 5.5
p Value†
p Value‡
Smoking and postoperative outcomes in elective cranial surgery Clinical article Nima Alan, B.S.,1 Andreea Seicean, Ph.D., M.P.H.,1,2 Sinziana Seicean, M.D., M.P.H., Ph.D., 3,4 Nicholas K. Schiltz, Ph.D., 2 Duncan Neuhauser, Ph.D., 2 and Robert J. Weil, M.D. 5 Case Western Reserve University School of Medicine, Cleveland; 2Department of Epidemiology and Biostatistics, Case Western Reserve University; 3Departments of Pulmonary, Critical Care, and Sleep Medicine, University Hospitals, Cleveland; 4Heart and Vascular Institute, Cleveland Clinic; and 5The Rose Ella Burkhardt Brain Tumor & Neuro-Oncology Center, and Department of Neurosurgery, the Neurological Institute, Cleveland Clinic, Cleveland, Ohio 1
Object. The goal in this study was to assess whether a current or prior history of smoking and the number of smoking pack years affect the risk for adverse outcomes in the 30-day postoperative period in patients who undergo elective cranial surgery. Methods. Data from the 2006–2011 American College of Surgeons’ National Surgical Quality Improvement Project were used in this study. The authors identified 8296 patients who underwent elective cranial surgery, of whom 1718 were current smokers, 854 were prior smokers, and 5724 were never smokers. Using propensity scores and age, the authors matched current and prior smokers to never smokers. Odds ratios for adverse postoperative outcomes were predicted with logistic regression. The relationship between number of pack years and poor outcomes was also examined. Results. In unadjusted analyses, prior and current smokers did not differ from never smokers for having poor outcomes postoperatively. Similarly, in matched analyses, no association was found between smoking and adverse outcomes. Number of pack years in propensity-matched analyses did not predict worse outcomes in prior or current smokers versus never smokers. Conclusions. The authors did not find smoking to be associated with 30-day postoperative morbidity or mortality. Although smoking cessation is beneficial for overall health, it may not improve the short-term (≤ 30 days) outcome of elective cranial surgery. Thus postponement of elective cranial cases only for smoking cessation may not be necessary. (http://thejns.org/doi/abs/10.3171/2014.1.JNS131852)
Key Words • smoking • cranial surgery • morbidity • mortality • health services research
A
to a 2011 report from the Centers for Disease Control and Prevention, more than 19% of adults in the US smoke cigarettes.5 It is well known that smoking is associated with significant morbidity and mortality.22 Evidence for the benefits of smoking cessation are significant; recently Jha et al.8 demonstrated that abstention from tobacco smoking at any age between the ages of 25 and 54 years reduces risk of death associated with smoking. ccording
Abbreviations used in this paper: ACS = American College of Surgeons; BMI = body mass index; BUN = blood urea nitrogen; CPT = Current Procedure Terminology; IQR = interquartile range; LOS = length of stay; NSQIP = National Surgical Quality Improvement Program; OR = operating room; SAH = subarachnoid hemorrhage.
J Neurosurg / Volume 120 / April 2014
In patients undergoing a variety of nonneurosurgical operations, smoking has been identified as an independent risk factor for poor perioperative outcomes.7,13,23 In spine surgery there is equivocal evidence of an association between smoking and postoperative outcomes.18,20,24 In cranial neurosurgery the impact of smoking on outcomes has been studied in only limited clinical settings. Smoking is a known risk factor for formation of intracranial aneurysm,3 subarachnoid hemorrhage (SAH),9,10 and symptomatic vasospasm.14 In patients undergoing craniotomy for tumor resection, Lau et al.15 reported that smoking increased the risk for morbidity and the 1-year mortality rate. However, the effect of cigarette smoking as a potentially modifiable preoperative risk factor that influences outcomes in patients undergoing elective cranial surgeries is unknown. 811
N. Alan et al. We used a large, national, prospectively collected database from the National Surgical Quality Improvement Program (NSQIP) to examine whether perioperative smoking in patients undergoing elective cranial surgery influenced 30-day outcomes postoperatively.
Methods Data Source
We identified patients who underwent elective cranial surgery by using the American College of Surgeons (ACS)–NSQIP database for records obtained between 2006 and 2011. Detailed descriptions of the ACS-NSQIP database can be found elsewhere.1,12 The study was approved by the Cleveland Clinic Institutional Review Board. Patient Population
We used the Current Procedure Terminology (CPT) codes to identify 10,651 patients who were 18 years of age and older, and who underwent cranial surgery between 2006 and 2011. We excluded 2122 emergency cases because smoking cessation intervention would not be possible. We also excluded 233 patients with unknown smoking status. Our final study sample consisted of 8296 patients with known smoking status, who underwent elective cranial surgery.
Smoking Status
We grouped patients according to their smoking status: current, prior, and never smoker.7,20 The NSQIP database contains 2 smoking variables: the first is “current smoker,” defined as smoking within 1 year prior to surgery and smoking duration as determined by total number of pack years. A current smoker was defined as a patient with “yes” for the NSQIP current-smoker variables and having either a missing value or a value > 0 for pack years. The second is a “prior smoker,” defined as having a “no” for the current-smoker variable and a value > 0 for pack years. We defined a “never smoker” as someone having a “no” for the current-smoker variable and a value of 0 or missing for the pack-year variable. Covariates
Available perioperative factors that may affect postoperative outcome were analyzed (Table 1).
Postoperative Outcomes
Short-term postoperative complications, defined as occurring within 30 days after surgery, were assessed for each patient group. Prolonged length of stay (LOS) was defined as LOS longer than that of the third quartile of the mean LOS; thus, prolonged LOS was considered to be ≥ 8 days. Minor complications consisted of ≥ 1 of the following: superficial surgical site infection, urinary tract infection, deep venous thrombosis, or thrombophlebitis. Major complications were ≥ 1 of the following: deep incision surgical site infection, organ or space surgical site infection, wound disruption, pneumonia, unplanned in-
812
tubation, pulmonary embolism, > 48 hour postoperative ventilator-assisted respiration, progressive renal insufficiency, acute renal failure, cardiovascular accident with neurological deficit, coma of > 24 hours, peripheral nerve injury, cardiac arrest requiring cardiopulmonary resuscitation, myocardial infarction, graft, prosthesis or flap failure, sepsis, septic shock, and/or return within 30 days to the operating room (OR). Any complication was defined as ≥ 1 minor or major complication. Thirty-day return to the OR, defined as any unplanned return to the OR for a surgical procedure, and 30-day mortality were also assessed. Statistical Analyses
Preoperative and intraoperative and 30-day outcomes were compared between smokers (current and prior) and never smokers by using Pearson chi-square tests for categorical variables and ANOVA for continuous variables. Propensity scores, including all variables from Table 1, were generated to obtain an approximately unbiased measure of the association between smoking and adverse outcomes.19 A 1:1 greedy matching technique4 was used first to match prior smokers with a unique never smoker on both propensity score and age, and similarly, current smokers with a unique never smoker. Restrictions were set on the matching criteria to within 0.2 SDs of the propensity score, to allow for increased accuracy in matching. Covariates were compared between smoking groups in the matched sample. Logistic regression analysis was used to test whether different smoking categories were independently associated with adverse outcomes. Covariates that remained unbalanced after matching based on propensity scores were included in the final models. The relationship between pack years and adverse outcomes was also tested in the entire sample, pooling current and prior smokers and, after matching based on propensity score was performed, separately comparing current with never smokers and prior with never smokers. A p value of < 0.05 was considered statistically significant; SAS (version 9.2, SAS Institute) was used.
Results
Compared with never smokers, current smokers were younger by an average of 4 years, and were more likely to be male and African American (Table 1). Daily consumption of > 2 alcoholic beverages was more prevalent among current smokers versus never smokers (4.8% vs 1.3%, p < 0.001), and the former had a slightly lower body mass index ([BMI] 27.6 ± 6.4 vs 28.5 ± 6.5, p < 0.001). Furthermore, current smokers were more likely to have pulmonary (10.7% vs 3.1%, p < 0.001) and CNS (19.1% vs 13.1%, p < 0.001) comorbidities than never smokers. Disseminated cancer was present in 16.8% of current smokers versus 10.0% of never smokers (p < 0.001). However, hypertension (34.9% vs 39.9%, p < 0.001) and diabetes (9.8% vs 11.6%, p = 0.04) were more common among never smokers. Current smokers, compared with never smokers, were more likely to have abnormal albumin (10.9% vs 8.3%, p < 0.001), serum glutamic oxaloacetic J Neurosurg / Volume 120 / April 2014
Smoking and outcomes in cranial surgery TABLE 1: Demographic data, comorbidities, preoperative laboratory values, and intraoperative factors according to smoking status in 8296 patients* Smoking Status Baseline Characteristic
Current
Prior
Never
no. of patients (%) age in yrs, mean ± SD female sex African American admitted from home >2 drinks per day partially or fully dependent functional status ASA classification 1+2 3+4 BMI in kg/m2, mean ± SD pulmonary comorbidities cardiovascular comorbidities hypertension requiring meds dyspnea CNS comorbidities diabetes mellitus disseminated cancer weight loss >10% in 6 mos prior to surgery steroid use for chronic condition preop chemo &/or RT bleeding disorder prior op w/in 30 days abnormal albumin abnormal alkaline phosphatase abnormal BUN abnormal creatinine abnormal hematocrit abnormal platelet count abnormal SGOT abnormal sodium abnormal total bilirubin abnormal WBC count level of residency supervision in OR attending alone attending & resident wound class clean clean-contaminated contaminated or infected intraop or postop transfusions
1718 (20.7) 51 ± 14 51.9 9.7 88.0 4.8 10.8
854 (10.3) 61 ± 13 46.5 6.6 92.4 3.9 12.4
5724 (69) 55 ± 17 54.9 6.7 89.5 1.3 10.8
27.6 72.1 27.6 ± 6.4 10.7 5.1 34.9 10.3 19.1 9.8 16.8 3.7 13.1 2.2 2.4 3.1 10.9 2.7 21.1 3.7 44.8 7.6 6.1 9.6 1.9 29.8
25.4 73.8 29.5 ± 6.8 9.7 12.8 52.2 10.9 21.4 16.7 14.6 3.3 15.0 6.0 3.9 5.3 8.2 3.4 27.2 7.1 43.6 9.7 6.8 8.2 2.1 23.2
35.2 64.4 28.5 ± 6.5 3.1 4.7 39.9 4.8 13.1 11.6 10.0 1.8 12.7 2.1 2.9 2.8 8.3 3.0 24.7 4.7 44.3 8.1 5.7 8.4 2.5 24.8
27.2 49.0
33.1 66.5
25.8 44.6
92.9 4.3 2.8 5.0
91.2 6.3 2.5 6.2
92.8 4.8 2.4 5.5
p Value†
p Value‡
