BUG, DRUG, TIMING: HOW A REAL-WORLD MULTIDISCIPLINARY FRAMEWORK OPTIMISES ANTIMICROBIAL STEWARDSHIP IN HIP AND KNEE PERIPROSTHETIC JOINT INFECTION

Introduction and aim Periprosthetic joint infection (PJI) remains a challenging complication of hip and knee arthroplasty, with European one-year incidence estimated at 0.5–1.2% for total hip arthroplasty (THA) and ~1–2% for total knee arthroplasty (TKA). Outcomes depend on timely diagnosis, appropriate microbiological sampling, and optimised antimicrobial therapy, increasingly managed by a multidisciplinary team (MDT). This study evaluated real-world antimicrobial stewardship practices, adherence to guideline-based principles, and their association with clinical outcomes within a structured MDT model. Methods Retrospective single-centre study in a Portuguese secondary care hospital. All hip and knee PJI discussed at a MDT meeting (Orthopaedics, Infectious Diseases, Clinical Microbiology and Pharmacy) over two years (2024–2025) were analysed. Results Eighty-five patients were discussed and all but one met EBJIS criteria for confirmed or likely PJI. The cohort was 53% male, median age of 76 years, high comorbidity burden (median five chronic conditions), and functional status impairment in 29%. THA accounted for 66% of discussed cases. Early-onset PJI incidence was 1.7% for THA and 0.8% for TKA. Globally, primary arthroplasties comprised 60% of infections. A total of 124 isolates were identified from deep-tissue sampling (median of seven intraoperative samples per patient), with coagulase-negative staphylococci predominating (58/124). Infections were monomicrobial in 54%, polymicrobial in 38% and culture-negative in 8% (two later yielded Candida spp. on repeat sampling). Multidrug-resistant organisms (MRO) were identified in 13 cases. Empirical antimicrobial therapy was dual in 87% of cases, most commonly piperacillin/tazobactam+vancomycin (65/74), with a median duration of seven days (range 0–46). Targeted therapy almost exclusively involved spectrum narrowing, typically to vancomycin-based regimens (monotherapy in 13 cases and combination therapy in four polymicrobial infections). Two MRO cases required escalation. Early oral switch during this initial phase was uncommon (8%) and median intravenous duration was 14 days. Most cases were managed with debridement, antibiotics, and implant retention (DAIR; 66%), achieving 80% success. Overall treatment failure (defined as persistent infection, the need for suppressive antibiotic therapy, or prosthesis removal) occurred in 25%, with surgical revision in 15%. Mortality was 6% at 30 days and 9% at one year. Conclusions MDT-driven management achieved favourable outcomes, with low TKA PJI incidence and DAIR success exceeding published pooled estimates (~64%). Extensive intraoperative sampling may have contributed to the low culture-negative rate. Higher early mortality likely reflects comorbidity burden. Prolonged intravenous therapy and limited early oral transition highlight opportunities to optimise stewardship, consistent with OVIVA and DATIPO recommendations.