Background: To report on early failures of pelvic fixation in posterior spinal fusions for neuromuscular scoliosis.
Methods: A retrospective review of posterior-only spinal instrumentation and fusion to the pelvis with iliac screws was performed. Forty-one patients with a mean age of 14 years and mean 16 levels fused met the inclusion criteria. Diagnoses include cerebral palsy (22), Duchenne muscular dystrophy (7), other neuromuscular (10), and spina bifida (2). Cox proportional hazards regression modeling was used to compare rates of failure.
Results: The mean preoperative primary Cobb angle was 82 degrees (range, 21 to 144 degrees). The pelvic obliquity correction was 76%. The fixation in the pelvis failed in 12/41 patients (29%). Failures include: screw head disengaged from screw shaft (5), iliac screw disengaged from rod (2), iliac connector disengaged from rod (2), iliac connector disengaged from iliac screw (4), and iliac screw loosened from bone (3). No failures occurred if there were at least 6 screws in L5, S1, and pelvis (0/7 patients). The failure rate with <6 screws in L5, S1, and pelvis was higher at 35% (12/34 patients) (P=0.16). When using traditional iliac screws with connectors to rods, all constructs had <6 screws in L5, S1, and pelvis. Only 1 failure occurred when S2-iliac screws were used, but was without clinical consequence. The mean time from surgery to failure was 18 months (range, 1 to 49 mo).
Conclusion: Not placing bilateral pedicle screws at L5 and S1, in addition to 2 iliac screws, was associated with a 35% early failure rate of pelvic fixation.
Level of Evidence: Level IV.