BACKGROUND: In patients with type 2 diabetes, sodium-glucose cotransporter-2 (SGLT2) inhibitors are known to reduce glucose concentrations, blood pressure, and weight, but to increase LDL cholesterol and the incidence of urogenital infections. Protection against cardiovascular events has also been reported, as have possible increased risks of adverse outcomes such as ketoacidosis and bone fracture. We aimed to establish the effects of SGLT2 inhibitors on cardiovascular events, death, and safety outcomes in adults with type 2 diabetes, both overall and separately for individual drugs. METHODS: In this systematic review and meta-analysis, we searched MEDLINE, Embase, the Cochrane Library, and websites of US, European, and Japanese regulatory authorities from Jan 1, 1950, to Sept 30, 2015, for data from prospective randomised controlled trials assessing the effects of SGLT2 treatment compared with controls. We excluded duplicate reports, trials of compound drugs, trials that lasted 7 days or fewer, trials that did not report on outcomes of interest, and articles that presented pooled trial data for which the individual trials could not be identified. We extracted data in duplicate using a standardised approach. The primary outcome was major adverse cardiovascular events. Secondary outcomes were cardiovascular death, non-fatal myocardial infarction, non-fatal stroke, admission to hospital for unstable angina, heart failure, and all-cause mortality. We estimated summary relative risks with fixed-effects meta-analysis, with the I(2) statistic used to estimate heterogeneity of results beyond chance. FINDINGS: The analyses included data from six regulatory submissions (37 525 participants) and 57 published trials (33 385 participants), which provided data for seven different SGLT2 inhibitors. SGLT2 inhibitors protected against the risk of major adverse cardiovascular events (relative risk 0·84 [95% CI 0·75-0·95]; p=0·006), cardiovascular death (0·63 [0·51-0·77]; p<0·0001), heart failure (0·65 [0·50-0·85]; p=0·002), and death from any cause (0·71 [0·61-0·83]; p<0·0001). No clear effect was apparent for non-fatal myocardial infarction (0·88 [0·72-1·07]; p=0·18) or angina (0·95 [0·73-1·23]; p=0·70), but we noted an adverse effect for non-fatal stroke (1·30 [1·00-1·68]; p=0·049). We noted no clear evidence that the individual drugs had different effects on cardiovascular outcomes or death (all I(2)<43%). Safety analyses showed consistent increased risks of genital infections (regulatory submissions 4·75 [4·00-5·63]; scientific reports 2·88 [2·48-3·34]), but findings for some safety outcomes varied depending on whether anlayses were based on data extracted from regulatory submissions or trials reported in the scientific literature. INTERPRETATION: These data suggest net protection of SGLT2 inhibitors against cardiovascular outcomes and death. The efficacy results were driven by findings for empagliflozin (the only SGLT2 inhibitor for which data from a dedicated long-term cardiovascular safety trial have been reported), although results for the other drugs in the class were not clearly different. Adverse events were more difficult to quantify than was efficacy, with the effects of individual drugs in the class seeming to differ for some safety outcomes. Results from ongoing studies will be crucial to substantiate these findings across the drug class, but the available data provide a strong rationale to expect benefit from use of SGLT2 inhibitors in patients with type 2 diabetes at high risk of cardiovascular events. FUNDING: National Health and Medical Research Council of Australia.