Renal and Hepatic Biochemical Trajectories During Lithium and Valproate Treatment in Bipolar Disorder

Purpose: To compare within-person renal and hepatic biochemical trajectories during lithium and valproate treatment in bipolar disorder and determine whether longitudinal slopes differed between treatment groups.Methods:This single-center retrospective cohort included 96 patients with bipolar disorder treated with lithium (n=35) or valproate (n=61), contributing 282 laboratory records. Random-intercept linear mixed-effects models for creatinine, blood urea nitrogen (BUN), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) included age at the first available laboratory record, sex, time, treatment group, and the time × treatment interaction. Outcome-specific adjusted models comprised 275-278 observations from 95 patients. For creatinine, a treatment-specific piecewise model with a 6-month knot was supplemented by data-quality and random-slope sensitivity analyses.Results: Mean follow-up was 312.22 ± 199.21 days. The time × treatment interaction was not statistically significant for creatinine (B=-0.0019, p=.303), BUN (B=-0.0853, p=.160), AST (B=-0.0070, p=.285), or ALT (B=-0.0059, p=.442). Male sex was associated with higher creatinine, AST, and ALT levels, whereas older baseline age was associated with higher BUN. In the piecewise model, neither early nor late slopes differed by treatment (joint Wald χ²(2)=0.544, p=.762). All sensitivity analyses supported the primary finding.Conclusion: No statistically detectable difference was observed between lithium and valproate groups in the longitudinal slopes of the evaluated markers. The findings do not establish equivalence or long-term safety; average short- to mediumterm trajectories did not diverge under regular, individualized laboratory monitoring.