Prediction of postoperative mortality in elderly patient with hip fractures: a single-centre, retrospective cohort study.

Niessen, Romain [UCL] Bihin, Benoît [UCL] Gourdin, Maximilien [UCL] Yombi, Jean Cyr [UCL] Cornu, Olivier [UCL] Forget, Patrice

Elderly patients are at high risk for postoperative complications and increased mortality after hip fracture (HF) surgery due to frailty and co-morbidities. The prediction of postoperative outcome could be used for clinical decision making. A reliable score to predict postoperative mortality after HF surgery in this sub-population remains unavailable. A single-centre retrospective cohort study was performed in 782 patients who were operated on for HF. Receiver Operating Characteristic (ROC)-curves were used to analyse the performance of gender, age, neutrophil-to-lymphocyte ratio (NLR) and C-reactive protein (CRP) at admission (D) as prognostic factors, alone or combined with the PreOperative Score to predict PostOperative Mortality (POSPOM) in univariate and multivariate linear regression models. No correlation between gender, age, NLR D or CRP D and postoperative, intra-hospital mortality was found. The Area Under the ROC-curve (AUC) for age, male gender, NLR and CRP were 0.61 [95% confidence interval (CI) = 0.45-0.61], 0.56 [95% CI = 0.42-0.56], 0.47 [95% CI = 0.29-0.47] and 0.49 [95% CI = 0.31-0.49] respectively. Combination with the POSPOM score did not increase its discriminative capacity as neither age (AUC = 0.69, 95% CI = 0.54-0.69), gender (AUC = 0.72, 95% CI = 0.58-0.72), NLR D (AUC = 0.71, 95% CI = 0.56-0.71), nor the CRP D (AUC = 0.71, 95% CI = 0.58-0.71) improved the POSPOM performance. Neither age, gender, NLR D nor CRP D are suitable parameters to predict postoperative, intra-hospital mortality in elderly patients undergoing surgery for HF.