Independent Risk Factors for Impaired Early Outcome after Isolated Ankle Fracture – A Multivariate Analysis and Prognostic Models

Abstract

Abstract Background Ankle fractures are common operative indications in orthopedic surgery. Their incidence is increasing. Objectives To identify independent risk factors and to develop prognostic models for the prediction of prolonged length of hospital stay (LOS) and the onset of postoperative complications. Materials and Methods This is a single-center, retrospective, observational study analyzing data of 154 consecutive, isolated, surgically treated ankle fractures. Multivariate binary logistic regression analysis was applied to identify significant independent risk factors. The validity and clinical applicability of the developed prognostic models was assessed with ROC-curve analysis (ROC: Receiver Operating Characteristic). Internal validation of prognostic models was performed with randomized backwards bootstrapping. Results The median LOS was 7 days. 50 patients (33%) had a longer LOS. 13% of operated patients had a postoperative complication (n = 20). Independent preoperative risk factors for prolonged length of stay were leukocytosis (p = 0.020; OR: 1.211), an increased CRP-level (p = 0.005; OR: 1.901), as well as a bi- (p = 0.002; OR: 15.197) or trimalleolar (p = 0.001; OR: 10.678) fracture type. Immediate operative therapy was an independent beneficial factor (p < 0.001; OR: 0.070). The onset of complications was associated significantly with diabetes mellitus (p = 0.004; OR: 9.903) and an elevated ASA score (p = 0,004; OR: 3.574). The developed prognostic models for the prediction of prolonged LOS (AUROC: 0.736) and postoperative complications (AUROC: 0.724) had a good clinical validity and were internally validated. Conclusion The current data pronounce the importance of preoperative laboratory works. Furthermore, co-morbidities play a major role in the prognosis of outcome. The developed prognostic models are able to reliably predict the outcome and enable the preoperative identification of high-risk patients.