PURPOSE: We evaluated the cardiovascular risk associated with dipeptidyl peptidase-4 inhibitors (DPP-4Is) as monotherapy compared with other antidiabetic drugs in Japan. METHODS: We conducted a nationwide cohort study involving 2 716 000 diabetes patients in Japan. New users of any antidiabetic drug as monotherapy between 1 April 2010 and 31 October 2014 were identified. Occurrences of myocardial infarction (MI), heart failure (HF), and stroke requiring hospitalization associated with DPP-4Is were compared with those associated with biguanides (BGs), sulfonylureas (SUs), or α-glucosidase inhibitors (α-GIs). Adjusted hazard ratios (aHRs) for these outcomes were estimated by Cox proportional hazards model. Propensity score standardization was used to control for confounding. RESULTS: We identified 1 105 103 patients using DPP-4Is, 278 280 patients using BGs, 273 449 patients using SUs, and 217 026 patients using α-GIs. The risks of MI and HF for DPP-4I users were significantly higher than those for BG users (MI: aHR, 1.48 [95%CI, 1.20-1.82], HF: aHR, 1.46 [95%CI, 1.31-1.62]), while significantly lower than those for SU users (MI: aHR, 0.84 [95%CI, 0.72-0.98], HF: aHR, 0.86 [95%CI, 0.81-0.92]). The risk of MI for DPP-4I users was similar to that for α-GI users, while the risk of HF for DPP-4I users was slightly higher than for α-GI users (MI: aHR, 0.98 [95%CI, 0.82-1.17], HF: aHR, 1.12[95%CI, 1.04-1.21]). CONCLUSIONS: Risk of MI and HF requiring hospitalization associated with DPP-4Is as monotherapy was significantly higher than BGs, significantly lower than SUs, and similar to α-GIs.
PURPOSE: We evaluated the cardiovascular risk associated with dipeptidyl peptidase-4 inhibitors (DPP-4Is) as monotherapy compared with other antidiabetic drugs in Japan. METHODS: We conducted a nationwide cohort study involving 2 716 000 diabetespatients in Japan. New users of any antidiabetic drug as monotherapy between 1 April 2010 and 31 October 2014 were identified. Occurrences of myocardial infarction (MI), heart failure (HF), and stroke requiring hospitalization associated with DPP-4Is were compared with those associated with biguanides (BGs), sulfonylureas (SUs), or α-glucosidase inhibitors (α-GIs). Adjusted hazard ratios (aHRs) for these outcomes were estimated by Cox proportional hazards model. Propensity score standardization was used to control for confounding. RESULTS: We identified 1 105 103 patients using DPP-4Is, 278 280 patients using BGs, 273 449 patients using SUs, and 217 026 patients using α-GIs. The risks of MI and HF for DPP-4I users were significantly higher than those for BG users (MI: aHR, 1.48 [95%CI, 1.20-1.82], HF: aHR, 1.46 [95%CI, 1.31-1.62]), while significantly lower than those for SU users (MI: aHR, 0.84 [95%CI, 0.72-0.98], HF: aHR, 0.86 [95%CI, 0.81-0.92]). The risk of MI for DPP-4I users was similar to that for α-GI users, while the risk of HF for DPP-4I users was slightly higher than for α-GI users (MI: aHR, 0.98 [95%CI, 0.82-1.17], HF: aHR, 1.12[95%CI, 1.04-1.21]). CONCLUSIONS: Risk of MI and HF requiring hospitalization associated with DPP-4Is as monotherapy was significantly higher than BGs, significantly lower than SUs, and similar to α-GIs.