Preoperative RDW-Based composite score for predicting mortality after hip fracture surgery

To develop a simple preoperative score based on age, red cell distribution width–standard deviation (RDW-SD), and the lactate-to-albumin ratio (LAR), and to evaluate its performance in predicting 3-month and 1-year mortality after femoral neck fracture treated with primary total hip arthroplasty. In this single-center retrospective cohort, 666 consecutive patients who underwent hip fracture surgery between January 2020 and June 2024 were analyzed. The primary outcomes were 3-month and 1-year all-cause mortality. Preoperative demographic characteristics and laboratory parameters (RDW-SD, lactate, albumin, and neutrophil-to-lymphocyte ratio) were recorded. Candidate predictors were screened using univariable logistic regression, followed by a multivariable model including age, RDW-SD, and LAR. An integer risk score was derived from regression coefficients. Overall, 245 patients (36.8%) died within 1 year. Non-survivors were older and had higher RDW-SD, lactate, LAR, and neutrophil-to-lymphocyte ratio, and lower albumin levels than survivors (all p < 0.01). In multivariable analysis, age, RDW-SD, and LAR remained independent predictors and were combined into an age–RDW-SD–LAR score ranging from 0 to 12 points. When classified into three risk groups (0–3, 4–7, and 8–12), observed 1-year mortality rates were 10.7%, 35.6%, and 62.8%, respectively, closely matching model-predicted estimates. The score showed good discrimination for 1-year mortality (AUC 0.78; 95% CI 0.75–0.82) with similar performance after bootstrap internal validation. Albumin was expressed in g/L, and the lactate-to-albumin ratio was calculated accordingly. The age–RDW-SD–LAR score, based on routinely available preoperative laboratory measures, provides a simple and clinically applicable tool for estimating early and late mortality after femoral neck fracture treated with primary total hip arthroplasty and may support preoperative risk stratification, pending external validation.