Individualized Medication Selection and Outcome Prediction for Patients With Dementia and BPSD Based on Systemic Immune-Inflammation Index (SII): A Retrospective Case-Control Study

Abstract

Background: Behavioural and psychological symptoms of dementia (BPSD) are often treated with atypical antipsychotics, but heterogeneous treatment responses remain a clinical challenge. A systemic inflammatory state may contribute to this heterogeneity, but it is unclear whether the baseline systemic immune-inflammation index (SII) is associated with differential short-term responses among BPSD patients treated with risperidone, olanzapine, and quetiapine. Therefore, this study aimed to investigate whether baseline SII was associated with differential short-term responses among antipsychotic treatment groups. Methods: This was a retrospective cohort study. We reviewed the electronic medical records of hospitalised patients with dementia and BPSD from January 2020 to December 2024 and divided the patients into a risperidone group, an olanzapine group and a quetiapine group according to the antipsychotic treatment in the medical records. Using the risperidone group as the reference, 1:1 nearest-neighbour matching without replacement was performed to obtain three balanced groups of 77 patients each (n = 231 in total). Normally distributed continuous data were analysed using t-tests, analysis of variance (ANOVA) and paired t-tests; non-normally distributed continuous data were analyzed using Wilcoxon rank-sum tests, Kruskal–Wallis tests and Wilcoxon signed-rank tests; categorical data were analysed using chi-square tests. Correlation analysis, multivariate linear regression and restricted cubic splines (RCS) were used for additional analyses. Results: Spearman’s rank correlation test revealed divergent therapeutic trajectories. Baseline SII was positively correlated with the reduction rate of the Scale for the Assessment of Positive Symptoms (SAPS) in the risperidone group (r = 0.863, p < 0.001), and negatively correlated with the reduction rate of the SAPS in the olanzapine group (r = –0.513, p < 0.001) and the quetiapine group (r = –0.368, p < 0.001). RCS analysis further revealed a non-linear relationship between SII and SAPS reduction rate (non-linear test p = 0.001). A significant interaction between medication type and SII was confirmed by the multivariable regression model (interaction p < 0.001). Conclusions: Baseline systemic inflammatory status is differentially associated with short-term antipsychotic response in hospitalised patients with dementia and BPSD. These findings suggest that inflammatory burden may be clinically relevant to the heterogeneous short-term response patterns across antipsychotic groups. However, prospective validation is required before any implications for treatment stratification can be drawn.