Residual lumbar hyperlordosis is associated with worsened hip status 5 years after scoliosis correction in non-ambulant patients with cerebral palsy

Background: Cerebral palsy (CP) is a static encephalopathy with progressive musculoskeletal pathology. Non-ambulant children (GMFCS IV and V) with CP have high rates of spastic hip disease and neuromuscular scoliosis. The effect of spinal fusion and spinal deformity on hip dislocation following total hip arthroplasty has been well studied, however in CP this remains largely unknown. This study aimed to identify factors associated with worsening postoperative hip status (WHS) following corrective spinal fusion in children with GMFCS IV and V CP. Methods: Retrospective review of GMFSC IV and V CP patients in a prospective multicenter database undergoing spinal fusion, with 5 years follow-up. WHS was determined by permutations of baseline (BL), 1 year, 2 years, and 5 years hip status and defined by a change from an enlocated hip at BL that became subluxated, dislocated or resected post-op, or a subluxated hip that became dislocated or resected. Hip status was analyzed against patient demographics, hip position, surgical variables, and coronal and sagittal spinal alignment parameters. Cutoff values for parameters at which the relationship with hip status was significant was determined using receiver operating characteristic curves. Logistic regression determined odds ratios for predictors of WHS. Results: Eighty four patients were included. 37 (44%) had WHS postoperatively. ROC analysis and logistic regression demonstrated that the only spinopelvic alignment parameter that significantly correlated with WHS was lumbar hyperlordosis (T12-L5) > 60° (p = 0.028), OR = 2.77 (CI 1.10–6.94). All patients showed an increase in pre-to-postop LL. Change in LL pre-to-postop was no different between groups (p = 0.318), however the WHS group was more lordotic at BL and postop (pre44°/post58° vs pre32°/post51° in the no change group). Age, sex, Risser, hip position, levels fused, coronal parameters, global sagittal alignment (SVA), thoracic kyphosis, and reoperation were not associated with WHS. Conclusion: Postoperative hyperlordosis(> 60°) is a risk factor for WHS at 5 years after spinal fusion in non-ambulant CP patients. WHS likely relates to anterior pelvic tilt and functional acetabular retroversion due to hyperlordosis, as well as loss of protective lumbopelvic motion causing anterior femoracetabular impingement. Level of evidence: III.