BackgroundOrthopaedic infections are difficult to eradicate because biofilm and poor local vascularity limit antibiotic exposure. Continuous local antibiotic perfusion (CLAP) delivers sustained, titratable antibiotics directly into infected compartments. We used harmonised individual participant data (IPD) to quantify early effectiveness, longer-term control, safety, and patient-level modifiers.MethodsWe performed an IPD review of observational reports using CLAP as primary or adjunctive therapy (January-May 2025). The primary outcome was 30-days early response (C-reactive protein ≤3 mg/L or earliest sustained clinical/wound improvement). Secondary outcomes were durable infection control at ≥6 and ≥12 months using evaluable denominators with best-worst bounds, infection-free days) and safety. One-stage analyses used mixed-effects logistic regression; Restricted Mean Survival Time (RMST) was preferred when proportional hazards were violated. Multiple imputation supported inferences.ResultsEighty-one studies (n = 256) were included; 164 patients had observed time-to-response. Fifty-nine percent achieved a 30-days response; median time-to-response was 26 days. Implant involvement was associated with lower odds of 30-days response; trajectories were slower with implants and higher organism burden (polymicrobial ≥3), while osteomyelitis responded faster than fracture-related infection. RMST (30) showed delays with implants (+4.43 days) and polymicrobial infection (+6.74 days), and faster response for osteomyelitis versus fracture-related infection (-9.06 days). Durable control among evaluable patients was 88.4% at ≥6 months and 90.2% at ≥12 months, with best-worst bounds of 89.2-82.2% and 90.9-83.5%, respectively. Infection-free-day RMST supported substantial time free of recurrent infection within the first year. Adverse events were uncommon; renal events were generally reversible.ConclusionsCLAP achieved encouraging early response and high durability among evaluable patients, with slower trajectories when implants were retained or pathogen burden was high and faster responses in osteomyelitis. Safety appeared acceptable with monitoring. Prospective comparative studies using standardised endpoints, with RMST for non-proportional hazards, are warranted.
Ng et al. (Thu,) studied this question.