17.1 Introduction and Background

The unique nature of scoliosis in children with cerebral palsy (CP) presents many challenges for treatment. Instrumented spinal arthrodesis effectively corrects spinal deformity and spinopelvic obliquity in these patients. ^{1 ,​ 2} The decision to pursue surgical correction in children with CP can be difficult. The indications for surgical correction are: (1) scoliosis that exceeds 50 degrees; (2) major impairment (current or predicted) of the patient’s ability to function; or (3) presence of substantial pain. ³ The goal of operative management should be to achieve the greatest improvement of scoliotic deformity and pelvic obliquity possible without compromising safety. Sitting posture can be improved dramatically with simultaneous correction of pelvic obliquity by including the pelvis in the fusion. ⁴ Techniques for posterior spinal fusion have evolved over the past 30 years: early Harrington instrumentation and the later development of Luque rods, modified for the Galveston technique, have been major achievements in the operative treatment of neuromuscular scoliosis. However, the problems associated with each of these techniques led to development of the unit rod and modular constructs.

17.2 Unit Rod Construct

Galveston rod fixation consisted of paired Luque rods bent so they could be anchored in the ilium. ⁵ Differential migration of the two rods was partially solved by adding cross-links and, later, by joining the two rods together at the top as one unit in an inverted “U.” This innovation by Moseley, in 1989, resulted in a one-piece, precontoured system of instrumentation that, when combined with sublaminar wires, followed the principles of the Luque–Galveston technique. ⁶ It allowed for stable, segmental fixation of the spine and pelvis. This produced better correction of spinal and pelvic deformity, as well as restoration of coronal and sagittal trunk balance. ⁶ Once the distal ends had been inserted and seated into the iliac wings, the proximal end of the rod could be cantilevered toward the midline, forcing the pelvis into a horizontal position. The curve of the spine could then be further corrected by transverse approximation of the apex of the curve to the rod with tightening of the sublaminar wires into the fixed vertical plane of the rod (Fig. 17‑1). ⁷

In their review of the literature on degree of correction obtained using the unit rod, Tsirikos et al ² analyzed a series of 287 children and adolescents with severe CP treated with the unit rod technique. They reported an excellent major curve correction of 68% and pelvic obliquity correction of 71%, with good lateral balance of the spine and a low rate of complications. The authors stated that the unit rod was the preferred system for the treatment of patients with CP for several reasons: its relative ease of use, lower cost compared with all–pedicle screw instrumentation, comparable deformity correction, low rate of loss of correction, and low rates of reoperation and complications. An earlier study by Bulman et al ⁷ found similar results, validating the use of the unit rod.

The unit rod produces near-automatic correction of pelvic obliquity. However, several drawbacks are notable. First, it is not optimal for correction of proximal thoracic curves because insertion must start from the pelvis and the cantilever mode of correction becomes less efficient the more proximal the deformity. Second, it is unable to compress the posterior column and correct major thoracic kyphosis for the same reasons. Third, developmental asymmetry of the pelvis in patients with CP may not match the typical pelvic anatomy on which the unit rod relies. Fourth, the many required laminotomies and passage of sublaminar wires contribute to increased bleeding. Fifth, lumbar hyperlordosis virtually necessitates that the unit rod be cut, recontoured, and reconnected with multiple connectors at the thoracolumbar junction. Because of these limitations, other modular systems developed to treat neuromuscular scoliosis have gained popularity.

17.3 Modular Constructs

All–pedicle -screw constructs have been used extensively in patients with adolescent idiopathic scoliosis. The convex rod can be placed first, and alignment is achieved through a cantilever maneuver combined with vertebral compression, translation, and derotation using the screws and the rods. The concave rod is placed second to augment the construct. Or, alternatively, the concave rod can be placed first, the apex translated medially and dorsally, distracted to obtain kyphosis in the thoracic spine, and the convex rod placed to cantilever the apex to the midline and correct the rib prominence. The correction can be individually tailored to the patient. The development of modular constructs using screws, wires, and hooks has been advantageous in treating patients with neuromuscular scoliosis (Fig. 17‑2).

Tsirikos and Mains ⁸ reviewed 45 consecutive patients with severe CP (Gross Motor Function Classification System [GMFCS] level 5) who underwent spinal arthrodesis using pedicle screw/rod instrumentation. They reported mean correction of 74% for scoliosis and 83% for pelvic obliquity, with loss of only 2.5 degrees of correction at a mean 3.5-year follow-up. None of their patients treated by posterior or anteroposterior spinal arthrodesis developed the “crankshaft” phenomenon, possibly because of the three-column fixation provided by segmental pedicle screws. (The crankshaft phenomenon, described by Dubousset et al, ⁹ occurs when posterior spinal fusion stops longitudinal growth in the posterior elements but the vertebral bodies continue to grow anteriorly and result in progressive angulation and rotation of the spine.) They reported no problems related to positioning of the iliac bolts or pelvic fixation of the construct, in contrast to the unit rod technique. Placement of the iliac screws was always performed after exposure of the pelvis and under direct visualization. They reported that, in patients with lumbar hyperlordosis and marked anterolateral pelvic tilt, it is easier and safer to place iliac screws than the pelvic legs of the unit rod, which can cut out from the osteopenic iliac bed. In their study, complications included one deep and five superficial wound infections treated with surgical debridement and antibiotics. There were no detected pseudarthroses and only one reoperation for prominent instrumentation. Their results compared favorably with their previous unit rod study, ² in which 3 nonunions requiring revision surgery and 12 reoperations for prominent implants occurred more than 3 years after combined anteroposterior spine arthrodesis in 45 patients. In their 2012 study, Tsirikos and Mains ⁸ concluded that spinal correction using segmental pedicle screw/rod constructs can be performed safely and with fewer major complications and a lower reoperation rate compared with the traditional unit rod. After 3.5 years of postoperative follow-up, they reported that correction of spinopelvic imbalance was maintained.

In a similar study, Modi et al ¹⁰ reported on 52 patients with CP and various degrees of neurological involvement who underwent scoliosis correction through posterior-only spinal arthrodesis with pedicle screw instrumentation and mean follow-up of 3 years. Mean scoliosis correction was 63%. Overall correction of pelvic obliquity was 56% postoperatively and 43% at 3 years of follow-up. Compared with the study by Tsirikos and Mains, ⁸ in which all patients had quadriplegia with major pelvic obliquity, in the study by Modi et al, 20 patients had diplegia or hemiplegia, which explains the smaller degree of pelvic deformity. ¹⁰ They reported a 33% complication rate, including two perioperative deaths, one neurological deficit caused by screw penetration in the canal, one prominent pelvic screw that required removal, and several respiratory complications but no deep wound infections.