Effects of preoperative smoking cessation on the incidence and risk of intraoperative and postoperative complications in adult smokers: a systematic review

Objectives: To establish the effect of preoperative smoking cessation on the risk of postoperative complications, and to identify the effect of the timing of preoperative cessation. Data sources: The Cochrane Library Database, PsycINFO, EMBASE, Medline, and CINAHL databases were searched, using the terms: smoking'', smoking-cessation'', tobacco-use'', tobacco-abstinence'', cigarett$'', complication$'', postoperative-complication$'', preoperative'', perioperative'' and surg$''. Further articles were obtained from reference lists. The search was limited to articles on adults, written in English and published up to November 2005. Study selection: Prospective cohort designs exploring the effects of preoperative smoking cessation on postoperative complications were included. Two reviewers independently scanned abstracts of relevant articles to determine eligibility. Lack of agreement was resolved through discussion and consensus. Twelve studies met the inclusion criteria. Data extraction: Methodological quality was assessed by both reviewers, exploring validation of smoking status, clear definition of the period of smoking cessation, control for confounding variables and length of follow-up. Data synthesis: Only four of the studies specified the exact period of smoking cessation, with six studies specifying the length of the follow-up period. Five studies revealed a lower risk or incidence of postoperative complications in past smokers than current smokers or reported that there was no significant difference between past smokers and non-smokers. Conclusions: Longer periods of smoking cessation appear to be more effective in reducing the incidence/risk of postoperative complications; there was no increased risk in postoperative complications from short term cessation. An optimal period of preoperative smoking cessation could not be identified from the available evidence.