Clinical outcomes associated with early in-hospital use of metformin, dipeptidylpeptidase-4 inhibitor or insulin in COVID-19 patients with type 2 diabetes: a territory-wide retrospective observational study

Abstract

Objective: Type 2 diabetes mellitus (T2DM) is associated with worse outcomes for patients with COVID-19. Commonly used glucose-lowering medications, metformin, dipeptidyl peptidase-4 inhibitor (DPP4i), or insulin, were found to impact the level of inflammatory and cytokine, therefore potentially influence the clinical outcomes in SARS-CoV-2 infected patients. Concerns remain on the early in-hospital use of these medications. This territory-wide retrospective observational study evaluated clinical improvement, virality, acidosis, hypoglycemia, hyperinflammation status, composite clinical outcomes associated with use of metformin, DPP4i, or insulin for hospitalized T2DM patients with COVID-19 infection. Methods: Baseline covariate between the two groups were balanced by multiple imputation followed by inverse probability of treatment weighting for propensity score. Multivariable cox proportional hazard models were performed to estimate hazard ratios (HR) and their 95% confidence intervals (CI) of time-to-event outcomes. Exposure was defined as receiving glucose-lowering drug class within 3 days of admission. Results: After propensity-score weighting, baseline characteristics were generally balanced. Metformin users (n = 609) were associated with quicker time to clinical improvement (HR = 1.41, 95%CI 1.19 to 1.65, P < 0.001), hospital discharge (HR = 1.43, 95%CI 1.20 to 1.69, P < 0.001) and recovery (HR = 2.09, 95%CI 1.42 to 3.06, P < 0.001), shorter hospital length of stay among survivors (−6.45 days, 95%CI −8.34 to −4.56, P < 0.001), reduced costs from 30-days onward, lower risk of composite outcome (HR = 0.41, 95%CI 0.23 to 0.74, P = 0.003) and macrophage activation (HR = 0.64, 95%CI 0.47 to 0.87, P < 0.004) when compared to non-metformin users. Shorter hospital length of stay was also observed in DPP4i users (n = 144) compared with non-DPP4i users (−5.32 days, 95%CI to 7.30 to −3.34, P < 0.001). Insulin users (n = 414) have greater risk of hypoglycemia during admission (HR = 2.61, 95%CI 1.23 to 5.53, P = 0.013) when compared to non-insulin users. Conclusion: Metformin and DPP4i have exhibited association with improved clinical outcomes while insulin has proved to be associated with increased risk of hypoglycemia for in hospitalized T2DM patients with COVID-19.