Wound Healing and Infection in Surgery

To clarify the evidence on smoking and postoperative healing complications across surgical specialties and to determine the impact of perioperative smoking cessation intervention. DATA SOURCES: Cohort studies and randomized controlled trials. STUDY SELECTION: Selected studies were identified through electronic databases (CENTRAL, MEDLINE, and EMBASE) and by hand searching. DATA EXTRACTION: Multiple data on study characteristics were extracted. Risk of bias was assessed by means of the Newcastle-Ottawa Scale and Jadad score. Healing outcome was classified as necrosis, healing delay and dehiscence, surgical site infection, wound complications, hernia, and lack of fistula or bone healing. Mantel-Haenszel and inverse variance methods for meta-analysis (fixed- and random-effects models) were used. DATA SYNTHESIS: Smokers and nonsmokers were compared in 140 cohort studies including 479,150 patients. The pooled adjusted odds ratios (95% CI) were 3.60 (2.62-4.93) for necrosis, 2.07 (1.53-2.81) for healing delay and dehiscence, 1.79 (1.57-2.04) for surgical site infection, 2.27 (1.82-2.84) for wound complications, 2.07 (1.23-3.47) for hernia, and 2.44 (1.66-3.58) for lack of fistula or bone healing. Former smokers and patients who never smoked were compared in 24 studies including 47,764 patients, and former smokers and current smokers were compared in 20 studies including 40,629 patients. The pooled unadjusted odds ratios were 1.30 (1.07-1.59) and 0.69 (0.56-0.85), respectively, for healing complications combined. In 4 randomized controlled trials, smoking cessation intervention reduced surgical site infections (odds ratio, 0.43 [95% CI, 0.21-0.85]), but not other healing complications (0.51 [0.22-1.19]).

Postoperative healing complications occur significantly more often in smokers compared with nonsmokers and in former smokers compared with those who never smoked. Perioperative smoking cessation intervention reduces surgical site infections, but not other healing complications.