Association between Metformin Use and Psychiatric Disorders in Patients with Diabetes

Background: Laboratory evidence has suggested better control of glucose levels in patients with diabetes may be associated with a lower incidence of psychiatric disorders. However, limited data was available on the association between metformin use and the risk of psychiatric conditions in patients with diabetes. Additionally, when conducting an observational study to investigate the association between metformin use and the risk of psychiatric disorders in patients with diabetes, selecting an appropriate comparator would be challenging.

Objectives: To investigate the association between metformin use and the risk of psychiatric disorders in patients with diabetes.

Methods: We conducted a population-based cohort study using a target trial emulation framework from the National Health Insurance Research Database (NHIRD). Patients from diagnosed with diabetes between 2002 and 2013. Patients previously diagnosed with psychiatric disorders within the past year were excluded to avoid misclassification of the psychiatry events. The outcomes were the composite psychiatric disorders, defined as bipolar disorder, depression, or schizophrenia). Specifically, we employed the clone-censor-weight approach to address the confounding by indication. Patients were assigned to an exposure group during the 6-month grace period from their diabetes diagnosis. We subsequently censored patients when they deviated from the assigned strategy and adjusted the probability of censoring via an inverse weighting approach. Cox proportional hazard models and non-parametric bootstrap were applied to estimate the effect of metformin on psychiatry disorders by approximate hazard ratios (HRs) and 95% confidence intervals (95% CIs).

Results: A total of 127,886 patients with diabetes were included in the analysis. The mean age was 58.0 (SD 15.7) years, and 50.4% (n=64,469) were male. Under the clone-censor-weight approach, of the included patients, 42,926 (33.6%) initiated metformin, while 84,960 (66.4%) did not, over a 6-month grace period following the diabetes diagnosis. The approximated hazard ratio was 0.78 (95% CI: 0.66–0.94), suggesting metformin users were associated with a reduced risk of psychiatric disorders. The individual outcomes including results of bipolar disorder (HR: 0.89; 95% CI: 0.84-0.94), depression (HR: 0.91; 95% CI: 0.88-0.95), or schizophrenia (HR: 0.84; 95% CI: 0.70-1.00) were consistent with the primary analysis.

Conclusions: The study showed that controlling the diabetes with metformin may even reduce the risk of the psychiatric disorders. The finding provides a foundation to observational studies when no comparable comparator is available.