Abstract
BACKGROUND: Hip redislocation after reconstructive surgery remains a major challenge in children with cerebral palsy (CP), particularly those with severe motor impairment. The Reimers migration percentage (RMP) has long been used empirically to guide surgical decision-making, but the optimal cutoff for predicting redislocation has not been statistically validated. OBJECTIVE: This study aimed to determine the most accurate RMP threshold for predicting postoperative redislocation using statistical methods in a retrospective cohort of children with CP. METHODS: We retrospectively analyzed 116 hips from 95 children with CP (Gross Motor Function Classification System (GMFCS) Level IV-V) who underwent reconstructive hip surgery. Preoperative RMP was evaluated using receiver-operating characteristic (ROC) analysis, and the Youden index was applied to identify the optimal cutoff. Hierarchical, block-wise logistic regression analysis that incorporated relevant confounders and surgical variables was performed to assess redislocation. RESULTS: To statistically determine the optimal RMP cutoff, we first observed that the median preoperative RMP was significantly higher in the Redislocation Group 72.5 (IQR 50-100) compared to the non-redislocation group 50 (IQR 42.25-70, p=0.002). ROC analysis yielded an area under the curve (AUC) of 0.66, with 70% identified as the optimal threshold. In the fully adjusted model, RMP ≥70% remained independently associated with redislocation (adjusted odds ratio (OR) = 3.59, 95% CI 1.45-8.85, p=0.006). CONCLUSION: This study provides the first statistical validation of the empirically suggested 70% RMP cutoff for predicting redislocation after reconstructive hip surgery in CP. Although discrimination was modest, the threshold offers a practical tool for risk stratification. Clinically, RMP ≥70% should prompt closer postoperative surveillance and may support consideration of earlier or more extensive reconstructive strategies in high-risk patients.