A case-control study of patients with type 2 diabetes admitted with acute infection

Introduction Despite the long-term renoprotective effects of Metformin, a recent study on data from the U.S. Food and Drug Administration reported a possible nephrotoxic effect, contributing to the development of acute kidney injury (AKI). We investigated the association between metformin and AKI in patients admitted with the AKI-prone condition of acute infection and compared results with corresponding results of other antidiabetics. Methods In a nationwide register-based case-control study, we identified Danish patients with type 2 diabetes hospitalized with acute infection between 2008 and 2018. Cases of AKI had an increase in plasma creatinine >= x 1.5 during admission, controls did not. Antidiabetics were identified up to 6 months before admission. Odds ratio (OR) of each antidiabetic was computed in separate multiple logistic regression models adjusted for relevant medication and comorbidities and results compared. Results We included 46,811 patients, hereof 9454 AKIs (20%) and 2186 (4.7%) severe AKIs. Overall, 56% were males, median age (IQR) was 73 (65-81). Sixty percent received metformin, 13% sulfonylurea, 31% insulin and 8% dipeptidyl peptidase-4 inhibitors (DPP-4i), with equal distribution between cases and controls. Metformin was associated with increased OR (CI) for AKI, 1.07 (1.02-1.12), equally to sulfonylurea, 1.10 (1.03-1.18) and DPP-4i, 1.11 (1.02-1.20), but not insulin, 0.99 (0.93-1.05). In severe AKI, results for metformin were 1.27 (1.25-1.40) but increased equivalently to other antidiabetics. Conclusions In patients with type 2 diabetes hospitalized with acute infection, metformin was not independently associated with AKI, since other antidiabetics were also significantly associated, indicating confounding by indication.