Secondary Logo

Journal Logo

Ambulatory Anesthesiology: Research Report

Long-Term Quit Rates After a Perioperative Smoking Cessation Randomized Controlled Trial

Lee, Susan M. MD, FRCPC; Landry, Jennifer MD, FRCPC; Jones, Philip M. MD, FRCPC, MSc (Clinical Trials); Buhrmann, Ozzie BScPhm, RPh; Morley-Forster, Patricia MD, FRCPC

Author Information
doi: 10.1213/ANE.0000000000000555


Patients who smoke experience increased perioperative complications, particularly wound and pulmonary complications.1,2 Large cohort studies have even shown smoking to increase mortality after elective surgery.3–5 Undergoing surgery can serve as a “teachable moment” that may motivate patients to engage in permanent smoking cessation.6,7 A few studies have found that in addition to the short-term benefits of smoking reduction on postoperative complications,8,9 smoking cessation interventions initiated in the perioperative period may increase the likelihood of long-term cessation.10–12 However, a meta-analysis showed that only intensive interventions, compared with brief interventions, resulted in long-term cessation.13

The aim of this study was to determine whether a perioperative smoking cessation intervention designed to minimize nursing and physician time in a busy preadmission clinic would be successful in reducing smoking rates, including long-term cessation. Another aim of this study was to explore preoperative factors that might be associated with successful long-term abstinence. Short-term results were previously reported.14 We now report our 1-year follow-up outcomes.


Detailed methods are previously described.14 This randomized controlled trial was conducted at St. Joseph’s Hospital, an ambulatory and short-stay hospital (with anticipated surgical inpatient stays <3 days) affiliated with the University of Western Ontario in London, Canada. The research protocol was approved by the local research ethics board, and written informed consent was obtained from all study participants. This trial was registered at (NCT01260233).

Adult daily smokers of 2 or more cigarettes per day were identified in the preadmission clinic at least 3 weeks preoperatively. Patients were ineligible if they were pregnant, breastfeeding, poorly proficient in the English language, or unable to consent. Randomization was computer generated in a 1:1 ratio in randomly permuted blocks of sizes 2, 4, and 6. Allocation was concealed by consecutively numbered sealed opaque envelopes. The control group received usual care. The intervention group received (1) brief counseling by the preadmission nurse, (2) smoking cessation brochures, (3) referral to the Canadian Cancer Society’s free Smokers’ Helpline, which proactively telephoned patients to provide ongoing counseling as agreed on by the patient, and (4) a free 6-week supply of transdermal nicotine replacement therapy. All health care providers on the operative day were blinded. Blinded observers collected self-reported smoking status of 7-day point prevalence abstinence by telephone interview 12 months postoperatively. For patients who had their original surgical date postponed or cancelled, follow-up calls were made 12 months after the original preadmission encounter.

The study was powered for the primary outcome of smoking cessation on the day of surgery, anticipating a baseline quit rate of 20% and an intervention group quit rate of 40% based on previous studies.15,16 Accepting a 2-tailed α error of 5% and a β error of 20%, 158 patients (79 per group) were needed, and an additional 5 patients per group were recruited to account for losses to follow-up.

This trial was analyzed by the intention to treat. Baseline characteristics of patients remaining at 1-year follow-up were analyzed by the Fisher exact test for categorical variables (gender, surgery type, current diseases). Histograms were generated to assess for normality of continuous variables and if normally distributed (age, height, weight, body mass index, number of years smoking, Fagerström score, exhaled carbon monoxide) analyzed by t test. Nonparametric continuous variables (cigarettes per day) were analyzed by Wilcoxon rank-sum test. The 1-year outcome of smoking cessation was analyzed with the Fisher exact test. The comparison was repeated assuming all patients with missing data continued to smoke (i.e., worst-case scenario analysis). Confidence intervals (CI) for numbers needed-to-treat (NNT) were calculated using the method described by Bender.17

Multivariable logistic regression modeling was used to study baseline patient characteristics that could affect the likelihood of abstinence at 1 year. Because the overall rate of smoking cessation was low, an exact logistic regression model was used.18 Prespecified predictors were selected on the basis of the likely relationship between each potential explanatory variable and the primary outcome. The predictor variables were as follows: randomization group, age ≥55 years, gender, ASA physical status (class ≤2), obesity, comorbid diabetes, hypertension, heart disease, chronic obstructive pulmonary disease (COPD) or asthma, number of pack-years of smoking ≥30, and the Fagerström score for nicotine dependency <4. Univariable analyses were performed on each predictor variable and then included in the multivariable model if the P value of the univariable analysis resulted in P < 0.1. A P value of 0.1 rather than 0.05 was chosen as the marker to include in the multivariable analysis to avoid exclusion of potentially important predictors that were negatively confounded before adjusted analysis. Continuous predictor variables were dichotomized at their median values, rounded to the nearest clinically meaningful value. Analyses were repeated with cut points 1 standard deviation above and below (25th and 75th percentiles for the nonparametric predictor pack-years) to assess the sensitivity of the resulting models to changes in cut points. The Hosmer-Lemeshow goodness-of-fit test (using 10 groups) was used to test model fit, and the c-statistic (the area under the receiver operating characteristic curve) was used to test model discrimination. Poisson regression using robust standard errors was performed to produce more interpretable relative risks in the final model.19 A 2-tailed P value of <0.05 was considered significant for all analyses. Stata version 13.0 (StataCorp LP, College Station, TX) was used for all analyses.


Between October 2010 and April 2012, 168 patients were randomized. Results for smoking status on the day of surgery and at 30 days postoperatively are previously reported.14 At 1 year, 127 patients (76%) were available for follow-up telephone interview. The telephone interview occurred a median of 369 (interquartile range [IQR], 366–378) days after surgery. As shown in Table 1, baseline characteristics were similarly balanced at baseline and for those that remained at 1-year follow-up. There were more patients with baseline diabetes (P = 0.040) and hypertension (P = 0.052) in the intervention group remaining at 1 year. However, these were the 2 characteristics that appeared unbalanced at baseline due to chance, suggesting that losses to follow-up were noninformative. Details of losses to follow-up are shown in the Consolidated Standards of Reporting Trials (CONSORT) flow chart in Figure 1. As shown in Table 2, smoking cessation occurred in 5 of 60 (8%) control patients compared with 17 of 67 (25%) patients in the intervention group (relative risk, 3.0; 95% CI, 1.2–7.8; P = 0.018). The NNT to achieve smoking cessation for 1 patient at 1 year postoperatively was 5.9 (95% CI, 3.4–25.9). Among those who did not quit, the number of cigarettes smoked per day did not differ significantly between groups (P = 0.23), with the control group smoking an average of 14.5 (IQR, 7.5–20) cigarettes per day compared with the intervention group that smoked an average of 12.2 (IQR, 5–20) cigarettes per day.

Table 1
Table 1:
Baseline Characteristics of All Study Participants and Those Remaining at 1-Year Follow-Up
Table 2
Table 2:
Smoking Cessation and Reduction at 1 Year
Figure 1
Figure 1:
Consolidated Standards of Reporting Trials (CONSORT) flow chart. Details of excluded patients: (a) Scheduling problems included patients missing their preadmission appointment, surgical date or location moved, or having no time to be assessed during the appointment; and (b) of the 36 ineligible patients, 15 smoked <2 cigarettes per day, 10 smoked something other than cigarettes, 2 were under age 18 years, 5 were already in the study or another smoking cessation study, and 1 had a previous allergic reaction to transdermal nicotine. *Abstinence confirmed by preoperative exhaled carbon monoxide≤10 ppm.

Continuous variables were dichotomized for logistic regression analyses. Age was dichotomized at 50 years and was not predictive of smoking cessation by univariable analysis (P = 1.0), which was consistent with cut points of 40 (P = 1.0) and 60 (P = 0.30). There were few patients with American Society of Anesthesiologists class 1 or 4, so ASA class was dichotomized to ASA 1 and 2 versus ASA 3 and 4. Pack-years of smoking were dichotomized at 20 pack-years and were not predictive of smoking cessation by univariable analysis at this cut point (P = 0.20), although this was somewhat sensitive to varying cut points (P = 0.53 for 10 pack-years, P = 0.086 for 30 pack-years). By univariable analysis, the Fagerström score was predictive of long-term cessation at cut points of 4 (P < 0.001) and 6 (P = 0.033) but not at 2 (P = 0.42).

The association between baseline risk factors and successful abstinence at 1 year postoperatively using exact logistic regression is shown in Table 3. On the basis of univariable analysis, the following predictors were used for the multivariable model: randomization group, history of COPD, and Fagerström score. Because of the sensitivity of univariable models to varying cut points for pack-years of smoking, the multivariable model was repeated including varying cut points. Pack-years was not a significant predictor at any cut point in the adjusted models (P = 0.95, 0.97, and 0.69 for cut points of 10, 20, and 30 pack-years). Pack-years were therefore not included in the final model.

Table 3
Table 3:
Baseline Characteristics Associated with Abstinence at 1 year

As shown in Table 3, in addition to the intervention (adjusted odds ratio [OR], 3.5; 95% CI, 1.02–13.9; P = 0.046), a lower level of nicotine dependency at baseline (as determined by Fagerström20 score <4) was predictive of success at smoking cessation at 1 year (adjusted OR, 6.3; 95% CI, 1.9–24.8; P = 0.001). Although none of the 22 patients with a history of COPD achieved long-term cessation, it was not a statistically significant predictor in the multivariable exact logistic regression model (adjusted OR, 0.22; 95% CI, 0–1.51; P = 0.14). A final model using the intervention group and the Fagerström score as predictors in an ordinary logistic regression model is shown in Table 4. The model performed well, with a high c-statistic of 0.79 indicating good discrimination and a Hosmer-Lemeshow test indicating good fit (P = 0.99). Finally, a Poisson regression, also shown in Table 4, was performed to produce more easily interpreted relative risks and showed that adjusted for the Fagerström score, those randomized to the intervention group were 2.7 times (95% CI, 1.1–6.7, P = 0.028) more likely to achieve long-term cessation than those in the control group. Adjusted for randomization group, a low level of nicotine dependency resulted in a relative risk of quitting of 5.1 (95% CI, 2.0–12.8, P = 0.001). Anonymized raw data and all statistical analyses are available as online supplemental content (Supplemental Digital Contents 1–3,.

Table 4
Table 4:
Baseline Characteristics Associated with Abstinence at 1 Year by Ordinary Logistic Regression and Poisson Regression


This study demonstrates that a smoking cessation intervention started preoperatively is successful at achieving smoking cessation at least as long as 12 months after surgery. The strengths of this study include the ease of implementation of the intervention and the long duration of follow-up. This trial design intentionally minimized the time spent in clinic and did not involve any additional visits beyond the regularly scheduled preadmission appointment, which should simplify clinical implementation of similar programs. Furthermore, the finding of successful self-reported smoking cessation 1 year after surgery suggests a public health benefit beyond the immediate perioperative period.

A Cochrane review suggested that long-term cessation occurs after intensive perioperative interventions, requiring weekly counseling sessions for 4 to 8 weeks but not after brief single-encounter interventions.13 Thus, the design used in this study might offer a compromise that is brief in terms of minimizing nursing or physician time, yet still effective at long-term cessation. As found in previous studies, in addition to the smoking cessation intervention, the level of nicotine dependency at baseline was predictive of smoking status at 1-year follow-up.10,12 However, this study may have been limited by small sample size in determining other predictors of long-term cessation. Further investigation into a wider array of predictors will be useful in tailoring smoking cessation interventions perioperatively to have the most long-term benefit.

It is unclear which specific component of the intervention used in this study (brief counseling, brochures, telephone quitline, or nicotine replacement) was most responsible for the outcome because it is common to combine strategies to maximize outcome.1 However, given that a previous study of a telephone counseling and newsletter program (without nicotine replacement), initiated at the time of surgical or diagnostic outpatient procedure, did not show a reduction in smoking at 1 year,21 we suspect that nicotine replacement therapy is a vital component of a successful perioperative smoking cessation program. The findings of our study, with its NNT of only 6, may serve as a call to action for governments and health insurers to take advantage of the teachable moment6 and support more widespread funding of drugs for smoking cessation therapy around the time of surgery.

The loss to follow-up may limit the validity of the results. However, the results are preserved if one assumes that all lost to follow-up continued to smoke. As with several previous long-term follow-up studies after perioperative smoking cessation interventions, smoking status determination was limited to self-report rather than biochemical verification.10,12 Self-reported smoking cessation has varying accuracy when compared with biochemical validation22 and is dependent on the type of test and the population under study. Encouragingly, another Canadian perioperative smoking cessation study did use biochemical validation with urine cotinine at 12 months postoperatively and found good correlation (0.91–0.95) to self-reported smoking status.11 Furthermore, discrepancies between self-reported abstinence and exhaled carbon monoxide on the day of surgery in our original study were infrequent (6–7%) and did not differ between groups (P = 1.0).14

Our study design used 3 weeks preoperatively as the minimum time to be eligible for inclusion to the trial based on prior literature that has shown that 2 weeks may not be adequate to reduce postoperative complications,16 while 4 weeks is.23 The need to see patients 3 weeks preoperatively hindered patient recruitment because many of the patients were referred too late to be included in the trial. However, given that long-term cessation was achieved with higher success in the intervention group in this study, future research could focus on shorter preoperative cessation intervals because there would likely be a long-term public health impact even if a reduction of postoperative complications could not be shown.

This study demonstrated that an intervention designed to work within existing infrastructure in a preadmission clinic results in decreased smoking rates not only around the time of surgery but also at 1 year. Anesthesiologists and perioperative providers have a unique opportunity to help patients achieve both short-term and long-term smoking cessation.


Name: Susan M. Lee, MD, FRCPC.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Susan M. Lee has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Jennifer Landry, MD, FRCPC.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Jennifer Landry has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Philip M. Jones, MD, FRCPC, MSc (Clinical Trials).

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Philip M. Jones has seen the original study data, reviewed the analysis of the data, approved the final manuscript, and is the author responsible for archiving the study files.

Name: Ozzie Buhrmann, BScPhm, RPh.

Contribution: This author helped design the study and conduct the study.

Attestation: Ozzie Buhrmann has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

Name: Patricia Morley-Forster, MD, FRCPC.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

Attestation: Patricia Morley-Forster has seen the original study data, reviewed the analysis of the data, and approved the final manuscript.

This manuscript was handled by: Peter S. A. Glass, MB ChB, FFA.


1. Warner DO. Helping surgical patients quit smoking: why, when, and how. Anesth Analg. 2005;101:481–7
2. Myles PS, Iacono GA, Hunt JO, Fletcher H, Morris J, McIlroy D, Fritschi L. Risk of respiratory complications and wound infection in patients undergoing ambulatory surgery: smokers versus nonsmokers. Anesthesiology. 2002;97:842–7
3. Singh JA, Hawn M, Campagna EJ, Henderson WG, Richman J, Houston TK. Mediation of smoking-associated postoperative mortality by perioperative complications in veterans undergoing elective surgery: data from Veterans Affairs Surgical Quality Improvement Program (VASQIP)—a cohort study. BMJ Open. 2013;3:e002157
4. Hawn M, Houston T, Campagna E, Graham L, Singh J, Bishop M, Henderson H. The attributable risk of smoking on surgical complications. Ann Surg. 2011;254:914–20
5. Musallam KM, Rosendaal FR, Zaatari G, Soweid A, Hoballah JJ, Sfeir PM, Zeineldine S, Tamim HM, Richards T, Spahn DR, Lotta LA, Peyvandi F, Jamali FR. Smoking and the risk of mortality and vascular and respiratory events in patients undergoing major surgery. JAMA Surg. 2013;148:755–62
6. Shi Y, Warner DO. Surgery as a teachable moment for smoking cessation. Anesthesiology. 2010;112:102–7
7. McBride CM, Emmons KM, Lipkus IM. Understanding the potential of teachable moments: the case of smoking cessation. Health Educ Res. 2003;18:156–70
8. Møller AM, Villebro N, Pedersen T, Tønnesen H. Effect of preoperative smoking intervention on postoperative complications: a randomised clinical trial. Lancet. 2002;359:114–7
9. Lindström D, Sadr Azodi O, Wladis A, Tønnesen H, Linder S, Nåsell H, Ponzer S, Adami J. Effects of a perioperative smoking cessation intervention on postoperative complications: a randomized trial. Ann Surg. 2008;248:739–45
10. Sadr Azodi O, Lindström D, Adami J, Tønnesen H, Nåsell H, Gilljam H, Wladis A. The efficacy of a smoking cessation programme in patients undergoing elective surgery: a randomised clinical trial. Anaesthesia. 2009;64:259–65
11. Wong J, Abrishami A, Yang Y, Zaki A, Friedman Z, Selby P, Chapman KR, Chung F. A perioperative smoking cessation intervention with varenicline: a double-blind, randomized, placebo-controlled trial. Anesthesiology. 2012;117:755–64
12. Villebro NM, Pedersen T, Møller AM, Tønnesen H. Long-term effects of a preoperative smoking cessation programme. Clin Respir J. 2008;2:175–82
13. Thomsen T, Villebro N, Moller AM. Interventions for preoperative smoking cessation. Cochrane Database Syst Rev. 2010:CD002294
14. Lee SM, Landry J, Jones PM, Buhrmann O, Morley-Forster P. The effectiveness of a perioperative smoking cessation program: a randomized clinical trial. Anesth Analg. 2013;117:605–13
15. Møller AM, Villebro N, Pedersen T, Tønnesen H. Effect of preoperative smoking intervention on postoperative complications: a randomised clinical trial. Lancet. 2002;359:114–7
16. Sørensen LT, Jørgensen T. Short-term pre-operative smoking cessation intervention does not affect postoperative complications in colorectal surgery: a randomized clinical trial. Colorectal Dis. 2003;5:347–52
17. Bender R. Calculating confidence intervals for the number needed to treat. Control Clin Trials. 2001;22:102–10
18. Fernandez NP, Mulla ZD. Avoiding sparse data bias: an example from gynecologic oncology. J Registry Manag. 2012;39:167–71
19. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159:702–6
20. Fagerstrom KO, Schneider NG. Measuring nicotine dependence: a review of the Fagerstrom Tolerance Questionnaire. J Behav Med. 1989;12:159–82
21. Glasgow RE, Gaglio B, Estabrooks PA, Marcus AC, Ritzwoller DP, Smith TL, Levinson AH, Sukhanova A, O’Donnell C, Ferro EF, France EK. Long-term results of a smoking reduction program. Med Care. 2009;47:115–20
22. Sarah Connor G, Sean S-H, Jill H, Geneviève L, Mark T. The accuracy of self-reported smoking: a systematic review of the relationship between self-reported and cotinine-assessed smoking status. Nicotine Tob Res. 2009;11:12–24
23. Lindström D, Sadr Azodi O, Wladis A, Tønnesen H, Linder S, Nåsell H, Ponzer S, Adami J. Effects of a perioperative smoking cessation intervention on postoperative complications: a randomized trial. Ann Surg. 2008;248:739–45
© 2015 International Anesthesia Research Society