Metabolic risk stratification and psychotic symptoms in first-episode drug-naïve major depressive disorder: a principal component-based cross-sectional study

Abstract

Objective To examine the association between glucose-lipid metabolism and psychotic symptoms in first-episode drug-naïve (FEDN) major depressive disorder (MDD) patients. Methods A single-center cross-sectional study enrolled 1718 FEDN MDD patients (171 with, 1547 without psychotic symptoms). Fasting blood glucose, lipid profiles, body mass index, and HAMD, HAMA, PANSS scores were collected. Principal component analysis (PCA) reduced collinearity, serial association analysis explored association patterns, K-means clustering stratified metabolic risk, and multivariate logistic regression confirmed independent associations. Results Two metabolic components were extracted by PCA. The cholesterol-glucose-related first component (PC1) was most closely linked to psychotic symptoms. Cross-sectional path analysis showed that the statistical covariance between PC1 and psychotic symptoms was predominantly accounted for by the severity gradients of depressive and anxiety symptoms, with total indirect association 0.0440 (95% CI: 0.0361 - 0.0533) and no significant direct association. Participants were classified into three metabolic risk subgroups by PC1 scores. Prevalence of psychotic symptoms increased progressively across the low, medium, and high risk groups (0.61%, 3.44%, and 32.23%, respectively). Multivariate logistic regression validated metabolic risk as an independent predictor of psychotic symptoms, with adjusted ORs 0.076 (medium-risk) and 0.013 (low-risk) (all P < 0.001). Receiver operating characteristic analysis also showed its discriminative ability. Conclusion Cholesterol-glucose metabolic alterations are associated with psychotic symptoms via sequential depressive and anxiety symptoms in FEDN MDD patients.