Abstract
BACKGROUND: Intraoperative periprosthetic femoral fracture is an uncommon but clinically relevant complication of total hip arthroplasty (THA). While several patient-related risk factors have been described, the role of surgical approach remains incompletely understood. METHODS: A retrospective cohort study was conducted, including 1,384 patients who underwent primary THA with an uncemented femoral component between January 2019 and March 2025 at a tertiary-care hospital. Surgical approaches were categorized as anterior, anterolateral, direct lateral, and posterolateral. The primary outcome was the occurrence of intraoperative femoral periprosthetic fracture. Fractures were classified using the Vancouver intraoperative classification, the Unified Classification System for Periprosthetic Fractures (UCPF), and the Mallory classification. Multivariable logistic regression analysis was performed to evaluate the association between surgical approach and fracture risk. RESULTS: A total of 50 intraoperative periprosthetic femoral fractures were identified, corresponding to an incidence of 3.6%. The anterolateral (odds ratio (OR) 2.46, 95% CI 1.07-5.66, p = 0.03) and direct lateral (OR 2.47, 95% CI 1.14-5.34, p = 0.02) approaches were associated with a higher likelihood of fracture compared with the posterolateral approach. The anterior approach was not significantly associated with fracture risk (OR 1.32, 95% CI 0.58-3.01, p = 0.50). Most fractures were stable metaphyseal patterns, predominantly classified as Vancouver A2 (39/50, 78%), UCPF B1 (35/50, 70%), and Mallory type I (21/40, 52.5%). The majority occurred during femoral canal preparation or final stem implantation and were primarily managed with cerclage fixation. CONCLUSIONS: Surgical approach was associated with differences in the occurrence of intraoperative periprosthetic femoral fracture in primary THA. The anterolateral and direct lateral approaches demonstrated higher fracture rates compared with the posterolateral approach. However, given the retrospective observational design and the potential for residual confounding, these findings should be interpreted with caution and do not establish a causal relationship.