Validation of the European Bone & Joint Infection Society recommendations for debridement, antibiotics, and implant retention in hip and knee periprosthetic joint infection

Aims The aim of this study was to externally validate the European Bone & Joint Infection Society (EBJIS) recommendations for the use of debridement, antibiotics, and implant retention (DAIR) in the treatment of periprosthetic joint infection (PJI), examining the independent associations of the selection of patients and the operative technique with failure after DAIR. Methods This was a multicentre retrospective cohort study involving seven UK hospitals and including adults with hip or knee PJI treated with DAIR. Patients were classified by EBJIS recommendations (DAIR-indicated, risk factors, or not recommended) and operative technique (exchange of components vs no exchange). The primary endpoint was the failure of treatment as defined by the need for further surgery for infection, long-term suppressive antibiotics, or death within 90 days, amputation, or arthrodesis. Sensitivity analyses were stratified by joint and repeated, excluding long-term suppressive antibiotics from the composite endpoint. Results A total of 411 patients (165 hips; 246 knees) were analyzed: 82 (20.0%) DAIR-indicated, 103 (25.1%) with risk factors, and 226 (55.0%) not recommended. Modular components were exchanged in 213 (51.8%). Over a median 2.1 years of follow-up (IQR 0.58 to 5.5), failure occurred in 193 patients (47.0%). Compared with the DAIR-indicated group, the DAIR not recommended group had a significantly higher risk of failure (hazard ratio (HR) 1.64 (95% CI 1.14 to 2.35); p = 0.007), whereas the risk factor group had a similar risk (HR 1.14 (95% CI 0.60 to 2.15); p = 0.689). DAIR without exchange of components was strongly significantly associated with failure of treatment (HR 2.40 (95% CI 1.89 to 3.06); p < 0.001). Joint-stratified and sensitivity analyses showed a similar direction of effect. Treatment was successful in 78.8% (26/33) in the most favourable scenario (DAIR-indicated with exchange of components) compared with 28.0% (26/93) in the least favourable scenario (DAIR not recommended without exchange of components). Conclusion In this external validation, both EBJIS patient selection and operative technique were associated with the outcomes after DAIR, with lack of exchange of components showing the strongest association with failure of treatment. The EBJIS framework is pragmatic and generalizable across specialist and non-specialist centres, and can support shared decision-making and the standardization of surgery in these patients. Cite this article: Bone Joint J 2026;108-B(10):1255–1264.