This is the authors' abstract. We don't add key points for this paper.
Periprosthetic joint infection (PJI) is uncommon after hip and knee arthroplasty but can lead to repeated surgery, prolonged antimicrobial treatment, impaired function, and excess mortality. This critical narrative review examines prediction of a first PJI after an index arthroplasty and separates that task from diagnosis, postoperative surveillance, reinfection, and treatment failure. PubMed and Web of Science searches identified 8,905 records. After deduplication and relevance screening, 500 records formed a prioritized evidence map; full texts were available for 494 records and unavailable for six. A focused appraisal using the Prediction model Risk Of Bias ASsessment Tool (PROBAST) covered 32 model- or analysis-specific assessment units from 11 occurrence-model publications. All 32 units were judged at high overall risk of bias, most often because of analysis limitations, sparse validation events, incomplete handling of missing data, or time-zero concerns. The count ordering of predictor evidence, model development, validation, and clinical translation was unchanged across the top 300, top 500, top 750, and all 1,909 strongly relevant records, although category proportions were threshold-sensitive. Reported discrimination ranged widely, but direct comparisons were generally invalid because populations, landmarks, outcomes, horizons, and validation designs differed. Several models described as preoperative used operative or later information, and independent validations showed moderate discrimination with clinically relevant miscalibration. Future studies should lock the intended task and landmark, validate calibration and net benefit externally, connect thresholds to explicit actions, and test impact prospectively. Reporting was informed by the Scale for the Assessment of Narrative Review Articles (SANRA), and the evidence was synthesized qualitatively rather than by meta-analysis.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Ye et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: