Relation of active, passive, and quitting smoking with incident type 2 diabetes: a systematic review and meta-analysis

Abstract

Background Cigarette smoking remains the leading avoidable cause of disease burden worldwide, and observational studies have linked various smoking behaviors (active and passive smoking and smoking cessation) with risk of type 2 diabetes. We did a meta-analysis of prospective studies assessing the relation of various smoking behaviors and diabetes risk. Methods We systematically searched MEDLINE and EMBASE databases (up to 3 May 2015) with manual searches of the references of retrieved publications and relevant reviews. We included prospective studies that reported risk of type 2 diabetes by baseline smoking status. Relative risks (RRs) and 95% confidence intervals (CIs) were pooled using random-effects models, and subgroup analyses were performed by participant and study characteristics. Findings The search yielded 88 prospective studies with nearly 6 million participants and 295446 incident cases. Compared with never smoking, the pooled RR (95% CI) of type 2 diabetes was 1·37 (1·33-1·42) for current smoking (84 studies), 1·14 (1·10-1·18) for former smoking (47 studies), and 1·22 (1·10-1·35) for passive smoking (7 studies). The associations persisted in all subgroups, and a dose-response relation was found for current smoking and diabetes risk: the RRs (95% CIs) were 1·21 (1·10-1·33), 1·34 (1·27-1·41) and 1·57 (1·47-1·66) for light, moderate, and heavy smokers, respectively, compared with never smokers. We estimated that 10·3% in men and 2·2% in women of type 2 diabetes cases (approximately 25 million) were attributable to cigarette smoking worldwide if smoking is causally related to diabetes. Compared to never smokers, the pooled RR (95% CI) from 10 studies was 1·54 (1·36-1·74) in new quitters (<5 years), and 1·11 (1·02-1·20) in long-term quitters (≥10 years). Interpretation Active and passive smoking are associated with significantly increased risks of type 2 diabetes. The risk of diabetes is elevated in new quitters, but decreases substantially as the time since quitting increases.