(B-283) Risk of Hyponatremia Associated With the Use of Antidepressants: a Large-Scale Federated Data Analysis of Multiple Institutions Using Common Data Model

Background: Hyponatremia is a common yet potentially serious adverse event associated with antidepressants. It is often asymptomatic and detected incidentally in laboratory tests, but severe cases can lead to seizure, coma or respiratory failure. Given the clinical benefits of antidepressant therapy, identifying the drug class with the least risk of hyponatremia would help guide clinical decisions and improve patient safety.

Objectives: To assess and compare the risk of hyponatremia among different classes of antidepressants

Methods: This study utilized electronic health records from 15 tertiary hospitals in South Korea (2003–2023), standardized to the Observational Medical Outcomes Partnership Common Data Model (version 5.3). The study population included adults aged 18 years and older newly diagnosed with depression and received antidepressants for at least 30 consecutive days. Antidepressants users were categorized into four mutually exclusive groups (SSRIs, SNRIs, TCAs, others), excluding those prescribed multiple classes at initiation. The primary outcomes were hyponatremia (serum sodium level < 135 mM) and severe hyponatremia ( < 125 mM) within 180 days of initiation. Patients were followed until outcome occurrence, addition of another antidepressant, intensive care unit admission, death or 180 days, whichever was earlier. The risk of hyponatremia in each group was assessed using the remaining groups as comparator. Propensity score-based stratification was used for each set of analysis, and hazard ratios (HRs) with 95% confidence intervals (CIs) using Cox proportional hazards regression models. Results from each site were pooled using fixed effect model.

Results: Among 42,036 new antidepressant users, 17,895 (42%) were prescribed SSRIs, 7,395 (18%) SNRIs, 5,424 (13%) TCAs, and 11,322 (27%) other antidepressants. The incidence of hyponatremia ranged from 43.9 (others) to 49.3 (SSRIs) per 1,000 person-years, with HRs of 1.18 (95% CI 1.01-1.38) for SSRIs, 1.05 (0.87-1.27) for SNRIs, 1.03 (0.84-1.26) for TCAs and 0.90 (0.77-1.06) for others. Notably, SNRIs (HR 1.51 [95%CI 0.68-3.37]) and TCAs (1.58 [0.66-3.79]) showed increased but non-significant risk of severe hyponatremia.

Conclusions: In this large-scale federated analysis of South Korean electronic health records, the overall incidence of hyponatremia was similar across antidepressant classes. However, the potential risk of severe hyponatremia with SNRIs and TCAs should be considered when selecting an initial treatment for depression.