Predictors of and predictive nomogram for cut-out of proximal femur nail anti-rotation device in intertrochanteric fractures.

Abstract

This study determined independent predictors and developed a predictive nomogram for failed correction of intertrochanteric fractures due to cut-out of the proximal femur nail anti-rotation (PFNA) device. Demographic and radiological data of 592 adult patients with intertrochanteric fractures (AO 31A) treated by PFNA were collected retrospectively. Independent predictors of cut-out were obtained through univariate and multivariate analyses, and a predictive nomogram was established. The discrimination, calibration, and clinical utility of the nomogram were based on receiver operating characteristic curve (AUC), concordance index (C-index), calibration curve, and decision curve analysis, respectively. Overall, 18 (3.04%) cases of cut-out occurred. Independent predictors according to the multivariate analysis were body mass index (BMI), poor-to-acceptable quality of reduction, PFNA blade position, and tip-apex distance (TAD). AUC of the nomogram was 0.849, and C-index was 0.849 (95% CI [0.844-0.854]). Bootstrapping yielded a corrected C-index of 0.849. The calibration and decision curves indicated good agreement and clinical benefit of the nomogram. A reliable predictive nomogram was developed for cut-out of the PFNA in intertrochanteric fractures, based on BMI, quality of reduction, blade position, and TAD.