In 1975, the Scoliosis Research Society (SRS) Morbidity and Mortality Committee published the first significant analysis of 1885 scoliosis cases, with nearly all patients having undergone posterior spinal fusion with Harrington rod instrumentation. Neurologic complications were recorded in 87 patients (0.72%), with 74 patients having spinal cord injury. Although this study was limited by the heterogeneity of diagnoses, it helped to better understand the effects of spinal instrumentation on neurologic function and stimulate the development of intraoperative neurophysiologic monitoring and the wake-up test. The investigators noted that congenital scoliosis, kyphosis, preexisting neurologic deficits, and high magnitude curves were associated with higher complication rates. Among patients with spinal cord injury, approximately one-third (36%) recovered completely, one-third (32%) had partial recovery, and one-third (32%) had no return of function. Prognosis for recovery was better for incomplete spinal cord deficits than complete and improved spinal cord deficits when instrumentation was removed within 3 hours of neurologic deficit diagnosis.

Coe and colleagues published the most recent analysis of SRS morbidity and mortality data in 2006 that analyzed 58,197 adolescent idiopathic scoliosis procedures performed by SRS members between 2001 and 2003. These investigators noticed that combined anterior and posterior instrumentation and fusion had significantly higher rates of neurologic complications (1.75%) compared with anterior (0.26%) or posterior (0.32%) instrumented spinal fusion alone. Spinal cord injury was recorded in 9 patients treated with combined fusion and 9 patients managed with posterior fusion alone. All spinal cord injuries were incomplete. Complete recovery was noted in 11 patients (7 from the posteriorly fused patients) and incomplete recovery in 6 (2 from the posteriorly fused patients), whereas no neurologic recovery was recorded in 1 patient who underwent combined fusion. Combined fusion was complicated by a dural tear in 0.12% of cases, whereas anterior or posterior fusion alone was associated with dural tear in 0.26% and 0.18% of cases, respectively. Limitations of this study include its retrospective, nonconsecutive character and the fact that it was not limited a priori to diagnosis of adolescent idiopathic scoliosis.

To overcome these limitations, Diab and colleagues reviewed 1301 consecutive surgically treated adolescent idiopathic scoliosis cases using a prospective database. Neurologic complications included 4 spinal cord injuries, 3 thecal penetrations, and 2 nerve root injuries (1 positional femoral neurapraxia). The overall rate of neurologic complications was 0.69%. The complication rate was reduced to 0.38% when dural tears and positional neurapraxia were eliminated. All neurologic injuries resolved completely within 6 months of the index operation. Three neurologic injury cases were associated with apical sublaminar wires placement. The investigators concluded that neural stretch secondary to large reduction and apical sublaminar wires are major risk factors for neurologic injury. In a meta-analysis addressing surgical outcomes after instrumented posterior spinal fusion that analyzed studies with a minimum follow-up of 5 years, neurologic complications were recorded in 2 of 1136 patients (0.17%). Both patients had been treated with Cotrel-Dubousset construct and suffered from unilateral lower extremity paresis immediately after surgery. Complete neurologic recovery was noted in both patients a few months after the operation. In one, hypereflexia of the affected lower extremity was found during the last follow-up.

Factors associated with increased risk for neurologic injury in adolescent idiopathic scoliosis are categorized as surgeon dependent and surgeon independent ( Fig. 1 ). Surgeon-dependent factors include type of procedure (with distraction, overcorrection, kyphosis correction, and osteotomy having the highest rates of neurologic injury), approach (with combined approach having the highest rates of neurologic injury), type of instrumentation (with sublaminar wires having the highest rates of neurologic injury), and excessive hemorrhage/prolonged hypotension resulting in decreased spinal cord perfusion. Surgeon-independent factors include curve magnitude and preexisting neurologic deficit. Most of the neurologic complications are detected intraoperatively as neuromonitoring signal changes and are reversible in a majority of the cases without any long-term sequelae. In cases of intraoperative neurologic change, the authors recommend immediate increase of the mean arterial pressure to 80 mm Hg followed by release of correction, intraoperative imaging to evaluate implant position, steroids administration, and wake-up test if no neurologic improvement is recorded within 30 minutes.

( A ) A 14-year-old girl with progressive adolescent idiopathic scoliosis treated with Harrington instrumentation. During implant insertion removal in another institution, she became completely paralyzed. She had partial recovery of motor function but persistent loss of urinary control. Posteroanterior radiograph shows curve progression after 1 year of bracing. She underwent anterior release and fusion with video-assisted thoracoscopic surgery followed by posterior osteotomies and fusion. ( B ) Intraoperative image of curve correction with Ponte osteotomies at the apex and fixation with a hybrid construct. ( C ) Posteroanterior radiograph 3 years after the last surgery shows good curve correction and coronal alignment. ( D ) Lateral radiograph 3 years after the last surgery reveals good sagittal alignment.

The overall prevalence of non-neurologic complications associated with surgical management of adolescent idiopathic scoliosis ranges from 0% to 15.4%, whereas the overall reoperation rates range from 3.9% to 26%. Non-neurologic complications are categorized as perioperative (intraoperative and postoperative complications occurring during the first postoperative week), early postoperative (occurring between the second and fourth postoperative weeks), and late postoperative complications (occurring after the fourth postoperative week). Perioperative non-neurologic complications include pulmonary, visceral, and urinary complications, excessive hemorrhage, superior mesenteric artery syndrome, and perioperative blindness. Early postoperative complications include ileus, superior mesenteric artery syndrome, wound hematoma, chololithiasis, pancreatitis, syndrome of inappropriate antidiuretic hormone secretion, and wound seroma. Pseudarthrosis, implant failure ( Fig. 2 ), junctional kyphosis, postoperative curve progression, crankshaft phenomenon, implant prominence ( Fig. 3 ), late operative site pain, residual rib prominence, and decompensation are considered late complications.

( A ) A 12-year-old boy with high-grade spondylolisthesis of L5 on S1. ( B ) Sagittal MRI view of the same patient. ( C ) Intraoperative image shows L5-S1 transacral interbody fusion with the Bohlman and Cook technique using a fibula strut graft. ( D ) Immediate postoperative lateral radiograph demonstrates partial reduction of L5 to S1. ( E ) One year after the index procedure, the lateral radiographic view shows failure of the right S1 screw. At that time, the patient remained asymptomatic but was brought to operating room to prevent any further displacement. ( F ) Lateral radiographic view 6 years after the revision surgery reveals no progression of spondylolisthesis and intact implants.

( A ) A 10-year-old boy with infantile scoliosis treated with casting and bracing until age 6 years when he was treated with growing rods. Lateral radiograph of the spine shows distal dislodgment of the inferior hooks. ( B ) Clinical image of the same patient demonstrates significant implant prominence. ( C ) Posteroanterior radiograph of the spine immediately after replacement of the distal anchors with pedicle screws. ( D ) Lateral radiograph after pedicle screw fixation.

According to the only level 1 study evaluating non-neurologic complications after surgery for adolescent idiopathic scoliosis, factors that are responsible for or contribute to non-neurologic complications are prolonged anesthesia time, excessive bleeding, and history of renal disease. No association was found between the prevalence of non-neurologic complications and the number of the levels fused, the type of bone graft, diaphragm detachment, or the approach used (combined anterior and posterior, anterior alone, or posterior alone). Alternatively, the most recent SRS Morbidity and Mortality Committee data revealed that patients who undergo combined anterior and posterior spinal surgery have twice as many complications (10.2%) as patients who have anterior (5.2%) or posterior (5.1%) surgery alone. In regards to the instrumentation type, a recent meta-analysis of midterm to long-term outcomes of surgical management of adolescent idiopathic scoliosis demonstrated that all-pedicle screw fixation is associated with the lower risk of pseudarthrosis, infection, neurologic deficit, and reoperation compared with Harrington rod and Cotrel-Dubousset instrumentation.