Baseline characteristics of the matched cohort.

BackgroundSodium–glucose cotransporter-2 (SGLT2) inhibitors are increasingly recommended as first-line treatment for type 2 diabetes mellitus (T2DM), but head-to-head data comparing them with metformin, the canonical biguanide, remain sparse in Japan.ObjectiveTo compare the long-term effectiveness and cost of initiating treatment with a biguanide versus an SGLT2 inhibitor, excluding the alternative class for 12 months but permitting other antidiabetic drugs, on a composite of major cardio-cerebrovascular events and all-cause death, and a composite of diabetic complications.MethodsWe emulated a new-user cohort trial using the Shizuoka Kokuho Database (2014–2021). Patients initiating treatment with either a biguanide or an SGLT2 inhibitor, while avoiding the alternative class during the first 12 months but allowing other glucose-lowering agents, were included. Follow-up began at treatment initiation; those who received the comparator drug within 12 months were excluded. After 1:1 propensity-score matching on demographic, clinical, laboratory, and lifestyle variables, cause-specific Cox models estimated hazard ratios (HRs). Daily medication costs were compared.ResultsAfter matching, 1,246 patients (623 per group) were followed for a median of 2.9 years (maximum 7.2 years). Cardio-cerebrovascular composite: 44/623 biguanide users (7.1%) and 35/623 SGLT2 inhibitor users (5.6%) experienced a first event (HR 0.80, 95% CI 0.51–1.24). Diabetic complications: 86/623 (13.8%) vs. 78/623 (12.5%) (HR 0.88, 95% CI 0.70–1.13). Median daily drug cost was 124.7 JPY for biguanides and 184.0 JPY for SGLT2 inhibitors (P ConclusionUsing a large-scale regional database from Japan, we found that among adults with type 2 diabetes without prior major cardiac or renal disease, first-line treatment with an SGLT2 inhibitor did not reduce risks of cardio-cerebrovascular events, mortality, or complications compared with metformin, and cost about 50% more.