There were 140 cohort studies (479,150 participants) and four RCTs (477 participants). There was considerable clinical heterogeneity in the cohort studies. Various methodological flaws including retrospective data collection, no reporting of missing data, inadequate outcome definition, inadequate post discharge follow-up reporting and inadequate confounder control. All RCTs had a low risk of bias and achieved a maximum Jadad score.
RCTs: In four trials, smoking cessation intervention significantly reduced surgical site infections (OR 0.40, 95% CI 0.20 to 0.83) but no other healing complications.
Cohort studies: Compared to nonsmokers, smokers had significantly increased odds of necrosis of wounds and tissue flaps (OR 3.60, 95% CI 2.62 to 4.93; nine studies), healing delay and dehiscence (OR 2.07, 95% CI 1.53 to 2.81; 12 studies), surgical site infection (OR 1.79, 95% CI 1.57 to 2.04; 32 studies), unspecified wound complications (OR 2.27, 95% CI 1.82 to 2.84; 17 studies), hernia (OR 2.07, 95% CI 1.23 to 3.47; seven studies) and lack of fistula or bone healing (OR 2.44, 95% CI 1.66 to 3.58; four studies).
For all healing complications combined, former smokers had increased odds of a complication compared with never smokers (OR 1.31, 95% CI 1.10 to 1.56; 15 studies) and reduced odds compared with current smokers (OR 0.28, 95% CI 0.12 to 0.72; two studies).
Sensitivity analyses of only studies with a maximum quality score and more than 1,000 patients showed similar results for smokers versus nonsmokers but a non-significant result for former versus never smokers.
There was some evidence of publication bias for the outcomes of surgical site infection and wound complications in the cohort studies.