Development and validation of baseline, perioperative and at-discharge predictive models for postsurgical prosthetic joint infection. (February 2019)

• Record Type: Journal Article
• Authors: del Toro, M.D.; Peñas, C.; Conde-Albarracín, A.; Palomino, J.; Brun, F.; Sánchez, S.; Rodríguez-Baño, J.
• Is Part Of: Clinical microbiology and infection. Volume 25:Number 2(2019)
• Page Start: 196
• Page End: 202
• Publication Date: 2019-02
• Subjects: Predictive model -- Prosthetic joint infection -- Risk factor -- Risk score -- Surgical infection; Medical microbiology -- Periodicals; Diagnostic microbiology -- Periodicals; Communicable diseases -- Periodicals; Infection -- Periodicals; 616.01
• DOI: 10.1016/j.cmi.2018.04.023
• Languages: English
• ISSNs: 1198-743X

Abstract

Abstract: Objectives: To develop and validate baseline, perioperative and at-discharge risk-scoring systems for postsurgical prosthetic joint infection (PJI) in patients undergoing arthroplasty. Methods: A multicentre prospective cohort study of patients undergoing hip and knee arthroplasty was performed. Patients were randomly assigned (2:1) to a derivation cohort (DC) or a validation cohort (VC). Multivariable predictive models of PJI were constructed at baseline (preoperative period), perioperative (adding perioperative variables) and at-discharge (adding wound state at discharge). The predictive ability of the models and scores was evaluated by area under the receiving operating characteristic curves (AUROC). Results: The DC and VC included 2324 and 1245 patients, respectively. Baseline model included total hip arthroplasty (THA), revision arthroplasty (RA), Charlson index and obesity. The AUROC for the score was 0.75 and 0.78 in the DC and VC, respectively. Perioperative model included THA, RA, obesity, National Nosocomial Infections Surveillance (NNIS) index ≥2, significant wound bleeding and superficial surgical site infection; the AUROC was 0.81 and 0.77 in the DC and VC, respectively. The at-discharge model included THA, RA, obesity, NNIS index ≥2, superficial surgical site infection and high-risk wound; the AUROC was 0.82 and 0.84 in the DC and VC, respectively. A score ≥8 points provided 99% negative predictive values for all models. Conclusions: Simple scores for predicting PJI at three different moments of care in patients undergoing arthroplasty were developed and validated. The scores allow early and accurate identification of high-risk individuals in whom enhanced preventive measures and follow-up may be needed. Further external validation is needed.