Introduction

The need for revision surgery is an important complication in pediatric spinal deformity surgery, and a significant consideration in the cost and risk of surgery in the adolescent. The need for revision surgery in the pediatric spinal deformity patient is certainly much less common than in the adult deformity population. However, there are a variety of circumstances that require revision surgery (or multiple revision surgeries) in this patient group. The diagnoses requiring revision surgery can be divided into early and late etiologies. Suboptimal correction, wound issues, implant issues, and neural issues typically occur and need treatment in the early postoperative period. Later, issues such as deformity progression in the unfused spine, pseudarthrosis, and junctional pathology can occur. However, wound and implant issues have shown up many years after surgery, and junctional problems have been seen early in some patients. The key is encouraging regular follow-up visits after surgery and carefully evaluating the patient clinically and radiographically. This chapter discusses revision surgery in the pediatric patient with spinal deformity, with a focus on both prevention of occurrence and surgical strategies for revision.

Suboptimal Clinical or Radiographic Outcome

The selection of specific operative treatments for pediatric patients with various spinal deformities is quite variable. Certainly, in the current era of segmental pedicle screw fixation, rigid posterior-only constructs with attendant thorough spinal fusion are the mainstay of treatment that can be applied to any pediatric spinal deformity. However, often these are combined with various posterior-based spinal osteotomies for increased deformity correction. Alternatively, anterior release of the spine, which is now less common, is still performed by some surgeons for severe coronal or sagittal plane deformities, and to limit fusion levels in some type 1 and 5 Lenke curves. The challenges of managing a deformity patient increase as the size and stiffness of the operative curve(s) increase, leading to a higher rate of compli cations and the potential need for revision surgery.

In the adolescent idiopathic scoliosis (AIS) population, the need for revision surgery up to 20 years later varies considerably, ranging from ~ 13% in one series to just under 4% in another series from busy pediatric deformity centers. ¹ , ² The reasons for revision surgery were quite similar, with the most prevalent being problems with instrumentation constructs or fusion, leading to revisions for implant dislodgement, curve progression, truncal imbalance, and pseudarthrosis. However, these series were accrued prior to the common usage of segmental pedicle screw constructs, which have a far lower revision rate than do hook or hybrid constructs in the pediatric deformity surgical population. ³ The most important component to any deformity construct is secure and solid fixation at the most cephalad and caudad ends of the construct. We recommend at least four solid screw anchors at the ends of the construct, and six if the patient or deformity requirements are more substantial. The quality of fixation is as important as the quantity, if not more so. Thus, secure screw purchase must be confirmed intraoperatively following both implant and construct placement. This should be performed by both radiographic and physical means (i.e., in vivo construct security assessment) on every surgical procedure. If a construct is not secure on the operating room table, it will certainly not be secure when the patient is upright and ambulating.

The classification of deformity and the choice of fusion levels are important considerations in limiting the risk of revision surgery. Classifying and then appropriately treating AIS patients remains controversial. Although the Lenke AIS classification system ⁴ is the most widely used, there are known “rule breakers” where the recommended fusion regions may vary from inclusion of the structural curves alone. The most problematic areas are optimizing clinical shoulder balance in main thoracic type 1 and double thoracic type 2 curves. ⁵ There are patients with clinically level shoulders, a large main thoracic curve, and a nonstructural proximal thoracic curve (i.e., a type 1 main thoracic curve) in which the left shoulder will be elevated postoperatively due to the marked correction of the main thoracic curve when the upper instrumented vertebra (UIV) is T4 or T5. This can also become problematic when the UIV is T2 or T3 if the shoulders are not well aligned during the operative proce dure. The production of a “high left shoulder” in AIS surgery is probably the most common suboptimal condition postoperatively, but only a fraction of those patients will require revision surgery for this malalignment.

Another quite controversial area is the decision to perform a selective thoracic fusion (STF) for Lenke 1C, 2C or even 3C and 4C curves. Although it is often advisable to try to leave the nonstructural lumbar curve unfused for functional reasons, postoperative decompensation can occur in the unfused lumbar spine requiring revision surgery. ⁶ The two most common reasons are inappropriate selection of patients to undergo an STF and overcorrecting the thoracic curve beyond which the unfused lumbar curve can accommodate. ⁷ In addition, patients with inadequate, or excessive, sagittal plane restoration of the thoracic or thoracolumbar region and those with subtle connective tissue laxity are also prone to postoperative problems including junctional kyphosis. ⁸ Patients with selective thoracic fusion and postoperative coronal or sagittal plane decompensation that does not spontaneously correct over time may require revision surgery that extends the thoracic instrumentation and fusion to the low lumbar region, either L3 or more commonly L4.

The surgical treatment of skeletally immature patients, whether for AIS or for other diagnoses, is challenging, and may require revision surgery if subsequent growth leads to progressive deformity. The risks of adding on to the upper or lower end of the construct, or progression of the apical segments leading to a crankshaft phenomenon, or both, are important causes of revision surgery in patients with significant spinal growth remaining. ⁹ Revision surgery usually requires extension of the lowest instrumented vertebra (LIV) to L4 or even L5 in some circumstances. Interestingly, at my center we have not seen many problems with lumbar curve progression following successful STF in the immature patient; conversely, we have seen problems with selective lumbar fusions in which thoracic curve progression in the immature patient that is recalcitrant to postoperative bracing attempts require fusion and instrumentation of the thoracic curve. Revision requires the extension of the lumbar construct to the upper thoracic region, with segmental apical screw fixation used to minimize progressive thoracic deformity with continued growth. This is illustrated in Fig. 9.1 in a skeletally immature girl who had a 55-degree Lenke 5CN AIS deformity. After undergoing a posterior selective lumbar instrumentation and fusion from T11 to L4, she had progression of her thoracic curve from 50 degrees at 2 months postoperative to 65 degrees at 2 years. Her lumbar curve increased from 39 to 49 degrees and she had increased her thoracic kyphosis to +69 degrees as well. She underwent a revision posterior procedure consisting of implant removal and fusion exploration (she was solidly fused from T11 to L4), then multiple posterior column osteotomies (PCOs) of the apex of both the thoracic and lumbar curves, with definitive instrumentation and fusion from T3 to L5. As a general rule to follow, fusing the immature spine short will always run the risk of unfused curve progression, and patients and parents need to understand this ahead of time to avoid any surprises postoperatively. ⁹