Preoperative malnutrition and risk of re-revision and mortality following revision arthroplasty.

Abstract

Background: The impact of preoperative nutritional status on long-term outcomes after revision arthroplasty remains poorly defined. This study aimed to (1) determine the association between preoperative malnutrition and the risks of re-revision and all-cause mortality-distinguishing periprosthetic joint infection (PJI)-related from non-PJI failures-and (2) establish clinically relevant thresholds of the Prognostic Nutritional Index (PNI) for high-risk patient stratification.

Methods: In this retrospective cohort study, 1,297 consecutive patients undergoing hip or knee revision arthroplasty between 2008 and 2024 were analyzed. The PNI, calculated from serum albumin and lymphocyte count, was evaluated as both a continuous and categorical variable. Primary outcomes were re-revision (classified as PJI or non-PJI) and long-term all-cause mortality. Analyses included multivariable logistic and Cox regression, restricted cubic spline (RCS) modeling, receiver operating characteristic (ROC) curve analysis, and 1:1 propensity score matching (PSM). Mediation analysis was conducted to assess the inflammatory contribution of C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), while the Geriatric Nutritional Risk Index (GNRI) was used for sensitivity analyses in patients aged ≥60 years.

Results: Each 1-point reduction in PNI was associated with a 15% higher odds of re-revision (odds ratio [OR] = 1.15; 95% confidence interval [CI], 1.10-1.19; P < 0.001) and a 4% higher hazard of long-term mortality (hazard ratio [HR] = 1.04; 95% CI, 1.01-1.09; P = 0.04), with these associations remaining significant after PSM. The increased re-revision risk was predominantly driven by PJI (OR = 1.19; 95% CI, 1.12-1.25; P < 0.001), with no significant association for non-infectious failures. A PNI threshold of 46.63 optimally predicted re-revision (AUC = 0.72; negative predictive value [NPV] = 0.91), while mortality demonstrated a nonlinear, U-shaped relationship with PNI. ESR partially mediated the PNI-PJI association (mediation effect: 4.36%). GNRI analyses corroborated these findings among elderly patients.

Conclusions: Preoperative malnutrition, quantified by PNI, is a robust and independent predictor of re-revision-particularly PJI-related-and long-term mortality following revision arthroplasty. Incorporating PNI-based risk stratification into perioperative care may facilitate targeted nutritional optimization and improve outcomes in this high-risk population.