Objective <p>Olanzapine and sodium valproate combination (OLZ/VPA) and OLZ and samidorphan combination (OLZ/SAM) are commonly used for treating bipolar disorder type I (BD-I). However, direct comparative evidence regarding their efficacy and metabolic impacts is lacking. This study evaluates the efficacy and metabolic safety of OLZ/VPA versus OLZ/SAM in BD-I patients and explores risk factors associated with metabolic syndrome (MetS).</p> Methods <p>In this single-center retrospective study, 180 BD-I patients were divided into OLZ/VPA and OLZ/SAM groups (<i>n</i> = 90/group). Bech-Rafaelsen Mania Scale (BRMS) and Hamilton Depression Rating Scale (HAMD) scores and metabolic parameters were compared before and after treatment. Logistic regression was utilized for identifying MetS risk factors, with predictive performance assessed by ROC curve analysis.</p> Results <p>Both groups showed comparable efficacy with similar &lt; mark&gt; improvements in BRMS (ΔBRMS, <i>P</i> = 0.705) and HAMD scores (ΔHAMD, <i>P</i> = 0.329). However, the OLZ/VPA group had significantly higher MetS incidence (<i>P</i> = 0.002) and greater worsening of fasting plasma glucose (FPG; ΔFPG, <i>P</i> &lt; 0.001), triglycerides (ΔTG, <i>P</i> = 0.033), and blood pressure (ΔSBP, <i>P</i> = 0.065; ΔDBP, <i>P</i> = 0.052), with the changes in blood pressure showing a strong trend towards significance. Elevated baseline FPG (OR = 5.693, 95% CI: 2.026–15.997, <i>P</i> &lt; 0.01) and diastolic blood pressure (DBP) (OR = 1.133, 95% CI: 1.015–1.265, <i>P</i> = 0.026) were independent MetS risk factors. Their combined prediction model showed optimal performance (AUC = 0.7938, 95% CI: 0.7287–0.8588).</p> Conclusion <p>OLZ/SAM may demonstrate metabolic safety advantages over OLZ/VPA while maintaining comparable efficacy in BD-I treatment. Higher baseline FPG and DBP are independent MetS risk factors, and their combined assessment may facilitate early identification of high-risk patients.</p> Clinical trial number <p>Not applicable.</p>