BACKGROUND:Cardiovascular outcome trials have shown cardiovascular benefit with sodium-glucose co-transporter-2 (SGLT2) inhibitors in patients with type 2 diabetes, whereas dipeptidyl peptidase-4 (DPP-4) inhibitors have not shown an effect. We aimed to address knowledge gaps regarding the comparative effectiveness of SGLT2 inhibitor use in clinical practice (with DPP-4 inhibitor use as an active comparator) across a range of cardiovascular risks and in diverse geographical settings. METHODS:In this comparative cohort study, we used data from clinical practice from 13 countries in the Asia-Pacific, Middle East, European, and North American regions to assess the risk of cardiovascular events and death in adult patients with type 2 diabetes newly initiated on SGLT2 inhibitors compared with those newly initiated on DPP-4 inhibitors. De-identified health records were used to select patients who were initiated on these drug classes between Dec 1, 2012, and May 1, 2016, with follow-up until Dec 31, 2014, to Nov 30, 2017 (full range; dates varied by country). Non-parsimonious propensity scores for SGLT2 inhibitor initiation were developed for each country and patients who were initiated on an SGLT2 inhibitor were matched with those who were initiated on a DPP-4 inhibitor in a 1:1 ratio. Outcomes assessed were hospitalisation for heart failure, all-cause death, myocardial infarction, and stroke. Hazard ratios (HRs) were estimated by country and then pooled in a weighted meta-analysis. FINDINGS:Following propensity score matching, 193 124 new users of SGLT2 inhibitors and 193 124 new users of DPP-4 inhibitors were included in the study population. Participants had a mean age of 58 years (SD 12·2), 170 335 (44·1%) of 386 248 were women, and 111 933 (30·1%) of 372 262 had established cardiovascular disease. Initiation of an SGLT2 inhibitor versus a DPP-4 inhibitor was associated with substantially lower risks of hospitalisation for heart failure (HR 0·69, 95% CI 0·61-0·77; p<0·0001), all-cause death (0·59, 0·52-0·67; p<0·0001), and the composite of hospitalisation for heart failure or all-cause death (0·64, 0·57-0·72; p<0·0001). Risks of myocardial infarction (HR 0·88, 0·80-0·98; p=0·020) and stroke (0·85 0·77-0·93; p=0·0004) were significantly but modestly lower with SGLT2 inhibitors versus DPP-4 inhibitors. INTERPRETATION:In this large, international, observational study, initiation of SGLT2 inhibitors versus DPP-4 inhibitors was associated with lower risks of heart failure, death, myocardial infarction, and stroke, providing further support for the cardiovascular benefits associated with use of SGLT2 inhibitors in patients with type 2 diabetes. FUNDING:AstraZeneca.