Development of diagnostic and prognostic biomarker models for knee osteoarthritis based on NLRP3 inflammasome...

This study aimed to characterize the expression profiles of the NOD-like receptor pyrin domain-containing protein 3 (NLRP3) inflammasome and its downstream effectors [Interleukin (IL)-1b, IL-18, and Gasdermin-D (GSDMD)] in degenerative knee osteoarthritis (KOA) and to establish an integrated biomarker model for predicting the likelihood of unfavorable rehabilitation outcomes.

We conducted a retrospective study involving 121 KOA patients and 94 age-matched healthy controls. Serum concentrations of NLRP3, IL-1 b, and IL-18 were quantified using ELISA, while GSDMD expression in peripheral blood mononuclear cells was assessed through flow cytometry. Conventional inflammatory markers (CRP ESR, and WBC) and neutrophil-to-lymphocyte ratio (NLR) were measured using automated analyzers. Receiver operating characteristic (ROC) curve analysis and multivariate logistic regression were performed to evaluate the diagnostic and prognostic utility of the integrated biomarker model.

KOA patients exhibited significantly elevated levels of NLRP3, IL-1 b, IL-18, and GSDMD compared to healthy controls (P < 0.05). These biomarkers showed positive correlations with systemic inflammatory markers (CRP ESR) and negative associations with knee joint range of motion (ROM) (P < 0.05). The integrated biomarker model demonstrated robust diagnostic performance for KOA (AUC = 0.928, sensitivity 84.30% , specificity 87.23%). Notably, among treated patients, those with poor recovery (n=37) maintained significantly higher post-treatment levels of NLRP3 pathway components than those with favorable recovery (P < 0.05). The predictive model achieved excellent performance in identifying patients at risk of suboptimal rehabilitation (AUC = 0.911, sensitivity 94.59%, specificity 73.81%).

Our findings highlight the pivotal role of NLRP3 inflammasome activation and GSDMD-dependent pyroptosis in mediating poor rehabilitation outcomes in KOA. The predictive model achieved excellent performance in identifying patients at risk of suboptimal rehabilitation (AUC = 0.911, sensitivity 94.59%, specificity 73.81%).