Background: To identify outcome predictors of selective dorsal rhizotomy (SDR) in children with spastic cerebral palsy (CP) using logistic regression models. Methods: A retrospective single-center study was conducted on children with spastic CP who had undergone SDR. Two outcomes were defined, one representing children not improving in motor function and spasticity and one representing children improving in motor function two years after surgery. Baseline variables were chosen based on established regressors and clinical considerations and tested for being predictors using multivariate logistic regression. Results: We included 96 patients (mean age [SD] 6y 9 m [2y 5 m], range 2y to 17y) in the study. Significant predictors of non-improvement in motor function and spasticity two years after SDR were higher age at surgery and higher preoperative Gross Motor Function Measure-88 (GMFM-88) score. Significant predictors of motor function improvement after SDR were lower preoperative Modified Ashworth Scale (MAS) scores, higher preoperative passive range of motion (pROM) and lower age at surgery. Conclusions: For the first time, pROM and preoperative spasticity were identified as predictors of SDR outcome. We confirmed motor function and age as predictors. These variables will be useful for future patient selection. Adjusting for the GMFM-88's ceiling effect, children with higher motor function can still profit significantly from SDR.
