Kyphectomy in myelomeningocele revisited: risk factors for failure.

purposeLumbar kyphosis occurs in approximately 8-20% of patients with myelomeningocele (MMC). The purpose of this article is to analyze the risks and benefits of vertebrectomy and spinal stabilization in MMC children with severe lumbar kyphosis and to establish treatment guidelines.

methodsThis is an IRB-approved retrospective analysis of 59 patients with MMC who underwent kyphectomy and posterior instrumentation in three centers. Average age at surgery was 7.9 years (2 weeks-17 years). Sitting trunk position, skin status, kyphosis angle, and thoracic lordosis were analyzed preoperatively, postoperatively, and at an average follow-up of 8.2 years (range 2.5-16). The correction was maintained by applying a short posterior instrumentation in 6 patients, and extending to the pelvis in 53 cases. Pelvic fixation was achieved using the Warner and Fackler technique in 24 patients, the Dunn-McCarthy in 8, Luque-Galveston in 8, sacral screws in 2, and ilio-sacral screws in 11.

resultsSitting position improved postoperatively in 47 of the 53 patients who underwent pelvic fixation and only in one patient with short instrumentation. All 6 patients with long instrumentation and poor postoperative sitting balance were in the Dunn-McCarthy fixation group. Skin sores at the apex of the deformity disappeared postoperatively in all patients but recurred in two patients with short instrumentations. Kyphosis angle improved from 109° (45°-170°) preoperatively to 10° (0°-45°) postoperatively and 21° (0°-55°) at last follow-up. The best results were seen in cases where a cross-k-wire fixation of the kyphectomy site was used, augmented with a long thoraco-pelvic instrumentation consisting of Luque sublaminar wires in the thoracic region and a Warner-Fackler type of pelvic fixation. Good results were also found with the bipolar technique and ilio-sacral screw fixation. Six over 24 patients with the Warner and Fackler technique showed gradual dislodgment or hardware failure, with subsequent nonunion of the kyphectomy site in four. Infection, with or without wound dehiscence and/or hardware exposure, occurred in 17 cases, necessitating hardware removal in 9 patients.

conclusionLumbar kyphosis in MMC children is best managed by resection of enough vertebrae from the apex to produce a flat lumbar spine, with perfect bone-to-bone contact and long thoraco-pelvic instrumentation using the Warner and Fackler technique through the S1 foramina or the bipolar technique with ilio-sacral screw fixation. Additional local fixation of the osteotomy site using cross-wires with or without cerclage increases the stability of the construct. The majority of complications occurred in patients with short instrumentations or where residual kyphosis persisted postoperatively regardless of the type of pelvic fixation or hardware density. The Dunn-McCarthy technique for pelvic fixation following kyphectomy in MMC was less successful in producing stable pelvic fixation and should not be considered in this patient category.