Smokers have a substantially increased risk of postoperative complications. Preoperative smoking intervention may be effective in decreasing this incidence, and surgery may constitute a unique opportunity for smoking cessation interventions. The objectives of this review are to assess the effect of preoperative smoking intervention on smoking cessation at the time of surgery and 12 months postoperatively, and on the incidence of postoperative complications. We searched the Cochrane Tobacco Addiction Group Specialized Register in January 2014. Randomized controlled trials that recruited people who smoked prior to surgery, offered a smoking cessation intervention, and measured preoperative and long-term abstinence from smoking or the incidence of postoperative complications or both outcomes. The review authors independently assessed studies to determine eligibility, and discussed the results between them. Thirteen trials enrolling 2010 participants met the inclusion criteria. One trial did not report cessation as an outcome. Seven reported some measure of postoperative morbidity. Most studies were judged to be at low risk of bias but the overall quality of evidence was moderate due to the small number of studies contributing to each comparison.Ten trials evaluated the effect of behavioural support on cessation at the time of surgery; nicotine replacement therapy (NRT) was offered or recommended to some or all participants in eight of these. Two trials initiated multisession face-to-face counselling at least four weeks before surgery and were classified as intensive interventions, whilst seven used a brief intervention. One further study provided an intensive intervention to both groups, with the intervention group additionally receiving a computer-based scheduled reduced smoking intervention. One placebo-controlled trial examined the effect of varenicline administered one week preoperatively followed by 11 weeks postoperative treatment, and one placebo-controlled trial examined the effect of nicotine lozenges from the night before surgery as an adjunct to brief counselling at the preoperative evaluation. There was evidence of heterogeneity between the effects of trials using intensive and brief interventions, so we pooled these separately. An effect on cessation at the time of surgery was apparent in both subgroups, but the effect was larger for intensive intervention (pooled risk ratio (RR) 10.76; 95% confidence interval (CI) 4.55 to 25.46, two trials, 210 participants) than for brief interventions (RR 1.30; 95% CI 1.16 to 1.46, 7 trials, 1141 participants). A single trial did not show evidence of benefit of a scheduled reduced smoking intervention. Neither nicotine lozenges nor varenicline were shown to increase cessation at the time of surgery but both had wide confidence intervals (RR 1.34; 95% CI 0.86 to 2.10 (1 trial, 46 participants) and RR 1.49; 95% CI 0.98 to 2.26 (1 trial, 286 participants) respectively). Four of these trials evaluated long-term smoking cessation and only the intensive intervention retained a significant effect (RR 2.96; 95% CI 1.57 to 5.55, 2 trials, 209 participants), whilst there was no evidence of a long-term effect following a brief intervention (RR 1.09; 95% CI 0.68 to 1.75, 2 trials, 341 participants). The trial of varenicline did show a significant effect on long-term smoking cessation (RR 1.45; 95% CI 1.01 to 2.07, 1 trial, 286 participants).Seven trials examined the effect of smoking intervention on postoperative complications. As with smoking outcomes, there was evidence of heterogeneity between intensive and brief behavioural interventions. In subgroup analyses there was a significant effect of intensive intervention on any complications (RR 0.42; 95% CI 0.27 to 0.65, 2 trials, 210 participants) and on wound complications (RR 0.31; 95% CI 0.16 to 0.62, 2 trials, 210 participants). For brief interventions, where the impact on smoking had been smaller, there was no evidence of a reduction in complications (RR 0.92; 95% CI 0.72 to 1.19, 4 trials, 493 participants) for any complication (RR 0.99; 95% CI 0.70 to 1.40, 3 trials, 325 participants) for wound complications. The trial of varenicline did not detect an effect on postoperative complications (RR 0.94; 95% CI 0.52 to 1.72, 1 trial, 286 participants). There is evidence that preoperative smoking interventions providing behavioural support and offering NRT increase short-term smoking cessation and may reduce postoperative morbidity. One trial of varenicline begun shortly before surgery has shown a benefit on long-term cessation but did not detect an effect on early abstinence or on postoperative complications. The optimal preoperative intervention intensity remains unknown. Based on indirect comparisons and evidence from two small trials, interventions that begin four to eight weeks before surgery, include weekly counselling and use NRT are more likely to have an impact on complications and on long-term smoking cessation.