Abstract
BACKGROUND: Inflammation is pivotal in psychiatric pathophysiology, yet evidence linking inflammatory biomarkers to postpartum anxiety (PPA) remains lacking. METHODS: We included 14,419 eligible postpartum women and calculated composite inflammatory biomarkers during second and third trimesters and the postpartum period, including platelet-neutrophil product (PPN), systemic inflammatory response index (SIRI) and pan-immune-inflammation value (PIV). PPA symptoms were only assessed by Self-Rating Anxiety Scale (SAS) at 5-8 weeks after childbirth. Propensity score matching (PSM) was performed to minimize confounding bias. Multivariable logistic regression analyses were used to examine the associations of inflammatory biomarkers with risk of PPA symptoms. Model discrimination, calibration, and clinical utility were assessed via the receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA). Machine learning was applied to strengthen the robustness. RESULTS: After PSM, the final analysis comprised 113 women with PPA symptoms (SAS ≥50) and 113 asymptomatic women (SAS < 50). After confounder adjustment, elevated second-trimester PPN, SII and PIV levels were associated with a reduced risk of PPA symptoms, with adjusted ORs of 0.21 (95% CI: 0.10, 0.45), 0.34 (95% CI: 0.16, 0.72), 0.44 (95% CI: 0.26, 0.75), respectively. Consistent results were observed during third trimester and the postpartum period (all P (FDR) < 0.05). The areas under the ROC curves (AUCs) of PPN at the three time points for PPA symptom were 0.67, 0.65, and 0.66, respectively. PPN had acceptable calibration and favorable net clinical benefit. The SVM-Radial model achieved robust performance, with AUCs ranging from 0.73 to 0.86 across all time points. CONCLUSIONS: Our findings demonstrated that inflammatory biomarkers were associated with PPA symptoms, emphasizing the potential role of PPN in the pathogenesis and application values of PPA symptoms.