AIMS: It is uncertain whether instrumented spinal fixation in nonambulatory children with neuromuscular scoliosis should finish at L5 or be extended to the pelvis. Pelvic fixation has been shown to be associated with up to 30% complication rates, but is regarded by some as the standard for correction of deformity in these conditions. The incidence of failure when comparing the most caudal level of instrumentation, either L5 or the pelvis, using all-pedicle screw instrumentation has not previously been reported. In this retrospective study, we compared nonambulatory patients undergoing surgery at two centres: one that routinely instrumented to L5 and the other to the pelvis.

METHODS: In all, 91 nonambulatory patients with neuromuscular scoliosis were included. All underwent surgery using bilateral, segmental, pedicle screw instrumentation. A total of 40 patients underwent fusion to L5 and 51 had their fixation extended to the pelvis. The two groups were assessed for differences in terms of clinical and radiological findings, as well as complications.

RESULTS: The main curve (MC) was a mean of 90° (40° to 141°) preoperatively and 46° (15° to 82°) at two-year follow-up in the L5 group, and 82° (33° to 116°) and 19° (1° to 60°) in the pelvic group (p < 0.001 at follow-up). Correction of MC and pelvic obliquity (POB) were statistically greater in the pelvic group (p < 0.001). There was no statistically significant difference in the operating time, blood loss, or complications. Loss of MC correction (> 10°) was more common in patients fixated to the pelvis (23% vs 3%; p = 0.032), while loss of pelvic obliquity correction was more frequent in the L5 group (25% vs 0%; p = 0.007). Risk factors for loss of correction (either POB or MC) included preoperative coronal imbalance (> 50 mm, odds ratio (OR) 11.5, 95%confidence interval (CI) 2.0 to 65; p = 0.006) and postoperative sagittal imbalance (> 25 mm, OR 11.0, 95% CI1.9 to 65; p = 0.008).

CONCLUSION: We found that patients undergoing pelvic fixation had a greater correction of MC and POB. The rate of complications was not different. Preoperative coronal and postoperative sagittal imbalance were associated with increased risks of loss of correction, regardless of extent of fixation. Therefore, we recommend pelvic fixation in all nonambulatory children with neuromuscular scoliosis where coronal or sagittal imbalance are present preoperatively. Cite this article: Bone Joint J 2020;102-B(2):261-267.