Abstract
Methicillin-resistant Staphylococcus aureus (MRSA) poses diagnostic and therapeutic challenges in pediatric bone and joint infections (BJIs). This two-center retrospective study evaluated the admission albumin-to-fibrinogen ratio (AFR) and urea-to-creatinine ratio (UCR) as predictors of MRSA among children with culture-confirmed S. aureus BJIs. Seventy-seven patients (MRSA, n = 23; MSSA, n = 54) admitted from 2015 to 2024 were analyzed. Clinical data and laboratory tests obtained before antibiotic initiation were reviewed. Compared with MSSA, the MRSA group had significantly lower AFR and UCR values. Using Firth's penalized logistic regression, AFR and UCR were incorporated as continuous covariates to construct a composite indicator (AFR + UCR). The model showed acceptable discrimination (AUC = 0.724, 95% CI 0.612-0.854) and good calibration (Hosmer-Lemeshow P = 0.273; Brier score = 0.184), with internal validation by stratified 10-fold cross-validation and 1,000 bootstrap resamples. At a prespecified probability cutoff of 0.293, sensitivity and specificity were 0.826 and 0.611, respectively. Decision curve analysis indicated net clinical benefit across threshold probabilities of approximately 10-60%. In pediatric BJIs, lower AFR and UCR at admission are associated with increased MRSA risk, and their combination may serve as an adjunctive tool for early risk assessment and empirical treatment planning. Findings pertain to culture-positive S. aureus infections; external validation in larger multicenter cohorts is warranted to confirm generalizability and clinical utility.