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

Summary Background Cigarette smoking remains the leading avoidable cause of disease burden worldwide, and observational studies have linked various smoking behaviours (active smoking, passive smoking, and smoking cessation) with risk of type 2 diabetes. We did a meta-analysis of prospective studies to investigate the associations between various smoking behaviours and diabetes risk. Methods We systematically searched MEDLINE (up to May 3, 2015) and Embase (up to April 16, 2014) for reports of prospective studies, using search terms related to smoking, diabetes mellitus, and studies with a prospective design. We supplemented this strategy with manual searches of the reference lists of retrieved publications and relevant reviews. We included prospective studies that reported risk of type 2 diabetes by baseline smoking status. We calculated pooled relative risks (RRs) with 95% CIs using random-effects models, and did subgroup analyses by participant and study characteristics. Findings We identified 88 eligible prospective studies with 5 898 795 participants and 295 446 incident cases of type 2 diabetes. The pooled RR of type 2 diabetes was 1·37 (95% CI 1·33–1·42) for comparing current smoking with non-smoking (84 studies with 5 853 952 participants), 1·14 (1·10–1·18) for comparing former smoking with never smoking (47 studies with 2 930 391 participants), and 1·22 (1·10–1·35) for comparing never smokers with and without exposure to passive smoke (seven studies with 156 439 participants). The associations persisted in all subgroups, and we identified a dose-response relation for current smoking and diabetes risk: compared with never smokers, the RRs were 1·21 (1·10–1·33) for light smokers, 1·34 (1·27–1·41) for moderate smokers, and 1·57 (1·47–1·66) for heavy smokers. Based on the assumption that the association between smoking and diabetes risk is causal, we estimated that 11·7% of cases of type 2 diabetes in men and 2·4% in women (ie, about 27·8 million cases in total worldwide) were attributable to active smoking. Compared with never smokers, the pooled RR from ten studies with 1 086 608 participants was 1·54 (95% CI 1·36–1·74) for new quitters (