14 Kyphotic Deformity in the Pediatric Spine
Kyphotic deformity in the pediatric spine may be caused by multiple disorders. The clinical presentation of kyphotic deformity varies depending on the age of the child, the severity of the curve at time of diagnosis, and the underlying cause. In many instances the spinal anomaly is just part of the problem in a globally affected patient, and other systems including neuromuscular disorders and intraspinal abnormalities may occur concordantly. It is important to understand the natural history of specific conditions that cause kyphotic deformity, the dynamics of growth in the developing spine, and axial skeletal biomechanics in order to determine the appropriate treatment. This chapter provides an overview of kyphotic deformity in the pediatric spine, and discusses the etiologies, concurrent pathologies, developmental patterns, and treatment.
Hyperkyphosis, sometimes referred to as round back or hunchback, is defined as a curve with abnormally increased posterior convex angulation. In contrast to scoliotic deformities, kyphotic deformities are more typically confined to one plane (sagittal). In the pediatric population, the etiologies are diverse, encompassing congenital anomalies, neuromuscular conditions, developmental disorders, skeletal dysplasia, infections, trauma, surgical sequela, and, more rarely, neoplastic processes. The most common kyphotic deformity in pediatric patients is Scheuermann kyphosis, which is the focus of this chapter. Other causes, including congenital anomalies and neuromuscular disorders, are also discussed. The risk of pediatric kyphosis progression is dependent on the cause of the deformity, the severity of the deformity, and the growth remaining. Appropriate management of pediatric kyphosis requires an understanding of the natural history and the risk of progression.
Scheuermann disease is the most common cause of structural kyphosis in adolescents. Reports show an incidence of 0.4 to 8.0% in the general population, but the diagnosis is likely underreported, as it often goes unrecognized or is attributed to poor posture. 1 Originally described in 1921 by Holger Scheuermann, 2 the disease was identified as a rigid spinal kyphosis, differentiating it from the more correctable postural roundback. It was further characterized radiographically by Sorenson 3 as vertebral body wedging of greater than 5 degrees in at least three contiguous vertebrae, end plate irregularity, and the presence of Schmorl′s nodes. A definitive distinction between abnormal kyphosis and pathognomonic signs of Scheuermann disease may be difficult because the disease encompasses a broad spectrum of vertebral anomalies. 4
The specific etiology of Scheuermann disease remains unclear. It was initially thought to result from avascular necrosis of the vertebral ring apophyses that leads to premature growth arrest and wedging. 5 Schmorl proposed an inherent weakness of the cartilaginous end plates. Various mechanical theories have also been proposed, implicating the anterior longitudinal ligament and anterior column pressure changes. Bracing has been advocated to relieve pain and reverse vertebral wedging, giving support to a mechanical etiology. 6 Murray et al 1 have questioned the long-term utility of bracing in their long-term follow-up study. No high-quality studies (level 1 or 2) support or refute the evidence for bracing or casting. In addition, Scheuermann disease has been associated with endocrine abnormalities, inflammatory disease, and neuromuscular disorders. Juvenile osteoporosis may also play a role. 7 , 8 Varying degrees of scoliosis are seen radiographically in one third of patients with Scheuermann disease. Isthmic spondylolisthesis has also been reported more commonly in patients with Scheuermann disease and may be a cause of low back pain. 9
Scheuermann disease is commonly detected at puberty. Often it is ignored or dismissed by parents and teachers who attribute it to poor posture. It is rare in patients younger than 10 years of age, with the typical presentation in the late juvenile period, and the more severe form appearing between the ages of 12 and 16 years. 4 Scheuermann disease and postural kyphosis can be differentiated by the rigidity of the kyphotic deformity. Trunk extension will produce deformity correction in those with postural kyphosis, whereas there is little correction in the rigid Scheuermann kyphosis. Patients with Scheuermann disease may also suffer from contractures of the pectoralis muscles, hip flexors, and hamstrings.
The effect of Scheuermann disease on health-related quality of life has been variable. In a study comparing health-related quality of life in patients with Scheuermann kyphosis to patients with adolescent idiopathic scoliosis and normal controls, Lonner et al 10 demon strated that patients with Scheuermann kyphosis had more impairment in every domain of the Scoliosis Research Society Outcomes Questionnaire (SRS-22) including pain, function, mental health, and self image. In contrast, Ristolainen et al 11 evaluated 80 adults with untreated Scheuermann kyphosis, and concluded that affected adults had a higher risk of back pain compared with controls. In addition, the patients with untreated Scheuermann kyphosis reported lower quality of life and poorer general health than did controls. The risk of disabilities affecting the performance of activities of the daily living was high in patients than in controls. However, among the patients there was no correlation between the degree of kyphosis and self-reported quality of life or health status or back pain.
The clinical presentation of the disease may vary from minimal or no symptoms in some patients to significant disability in others. 1 Symptoms, when present, encompass nonradiating pain, physical disability, decreased range of motion, weakness of the back, restrictive lung disease, and hamstring tightness, leading to decreased participation in physical activity in employment and in athletics. 7 Adolescents typically are more concerned with self-image but may report neck pain, back pain, and fatigue as well. In these patients, pain is typically below the apex of the curve in the paraspinal musculature, which may be a result of the increased incidence of lumbosacral spondylolisthesis, spondylosis, disk degeneration, and scoliosis in patients with Scheuermann disease. Spondylolysis in particular has been reported in up to 50% of these patients. 9 Neural symptoms are rare in pediatric patients with Scheuermann disease, although paraparesis has been reported. When present, neural symptoms are often associated with herniated thoracic disks, spinal stenosis, dural cysts, or extreme kyphotic deformity with tenting of the spinal cord over the apex of the deformity. 6 There are reports of acute cord compression due to traumatic disk herniation as well as other reports of intraoperative herniations during kyphosis correction. To summarize, patients with Scheuermann kyphosis may be neurologically intact, although with varying degrees of reserve and significant anterior displacement of the spinal cord within the canal, possibly making them more vulnerable to acute neurologic deterioration from typically benign lesions such as disk bulges, ventral compression, or disruptions of blood flow to the spinal cord. 12
Thoracic kyphosis in unaffected populations is variable, and typically ranges from 20 to 40 degrees using the Cobb method on erect lateral radiograph. 13 Scheuermann disease is best seen by evaluating standing full spine lateral radiographs. 14 Diagnostic criteria based on Sorenson include the presence of > 5 degrees of anterior wedging in at least three consecutive vertebrae at the apex of the kyphotic deformity. Patients typically also have narrowed disk spaces, end-plate irregularities, and Schmorl′s nodes ( Fig. 14.1 ). A radiograph obtained with the patient hyperextended over a bolster dem onstrates the degree of stiffness of the kyphotic curve. Two different curve patterns have been described in Scheuermann disease. 15 The more common pattern is thoracic, with an apex at T8-T9 that is usually quite rigid and balanced. The less common pattern is thoracolumbar, with an apex at T10-T11 that is normally flexible and frequently unbalanced 15 ( Fig. 14.2 ). If neurologic findings are present, additional evaluation by magnetic resonance imaging (MRI) is generally indicated. This imaging may detect myelopathy or cord compression, thoracic disk herniation, and other subtleties that which may alter the treatment approach. MRI may also be helpful in evaluating the neural elements in intact patients and can assist in surgical planning.
The treatment of Scheuermann disease is primarily nonoperative, and includes anti-inflammatory medications, exercise, bracing, and casting. In a skeletally immature patient with a mild deformity, routine evaluation with radiographs every 6 months is recommended. Once the diagnosis has been made, it is important to maintain close observation until skeletal maturity. Adolescents with kyphosis of less than 60 degrees are typically treated with physical therapy and exercise programs until skeletal maturity. Exercise and physical therapy are useful to treat associated back pain and to improve muscle tone and posture, though these modalities have not been shown to alter progression of the deformity.
Bracing and casting have been effectively used in patients with kyphotic deformities and sufficient growth (age < 12 or RIser grade 0–2) remaining. The initial report by Bradford et al 16 demonstrated a 40% decrease in mean thoracic kyphosis, and a 35% decrease in mean lumbar lordosis after 34 months of bracing. Gutowski and Renshaw 17 reported on the use of the Boston and Milwaukee braces in a group of 75 patients. Compliant patients had 35% improvement in the Milwaukee brace compared with 27% in the Boston brace. Milwaukee brace treatment consistently improved kyphosis by ~ 50% during the active phase of treatment. Skeletally immature adolescents with progressive kyphosis > 45 degrees or with curves of up to 65 degrees should be considered for a trial of bracing. Prior to considering brace treatment, a hyperextension X-ray over a bolster is helpful to assess the flexibility of the deformity. Patients with at least 40% passive correction of curves measuring between 50 and 75 degrees often respond well to bracing. Those with > 75 degrees of curvature respond less favorably to bracing, and surgical intervention should be considered for these patients. Murray et al 1 suggest that bracing is not effective in the long term for Scheuermann kyphosis.
Surgery in Scheuermann disease may be indicated in patients with kyphosis > 75 degrees, in patients with kyphosis > 55 degrees and with pain that is unresponsive to nonoperative care, patients with progression of curve despite bracing, those with an unacceptable appearance, and in rare instances of neurologic deficit. Cardiopulmonary compromise is not usually associated with thoracic hyperkyphosis. Surgery is generally not indicated in skeletally immature patients with kyphosis < 75 degrees unless their symptomatic deformity is not responsive to nonoperative care. Other factors to consider are patient age and the location and shape of the kyphosis. The goals of surgical treatment are to prevent curve progression, improve pain, restore sagittal alignment, and improve cosmetic appearance. 18
The surgical techniques for the treatment of Scheuermann kyphosis have evolved over several decades and include posterior-only, combined anterior and posterior, and anterior-alone procedures. Traditionally, anterior release combined with posterior fusion had been used to treat severe rigid deformities that did not correct to < 50 degrees on hyperextension X-rays. Modern pedicle screw instrumentation systems have improved the ability to control and realign the spine, reducing the need for anterior release and making posterior-only approaches sufficient for most cases. Posterior-only procedures enable posterior column shortening via segmental compression, which may be combined with multilevel posterior column or, less frequently, three-column osteotomies for correction of large deformities.
In North America, the terms Smith-Petersen osteotomy and Ponte osteotomy are often used incorrectly to describe the same technique. This can cause confusion. To overcome this difficulty, we will refer to them collectively as a posterior column osteotomy (PCO) throughout the chapter. The Smith-Petersen osteotomy (SPO) was initially described for use in the lumbar spine, and it consisted of a narrow resection of lumbar facet joints with detachment of the ligamentum flavum from the inferior margin of the lamina and inferior articular process. There was no resection of the lamina in the original description, and correction was gained at a single level with osteoclasis of the anterior column and anterior column distraction. 19 When SPOs are used in thoracic defor mity, it is generally to obtain flexibility in stiff curves such as those seen with ankylosing spondylitis. Flexibility and kyphosis correction is then obtained by opening the anterior disk spaces with lengthening of the anterior column.
The Ponte osteotomy consists of complete resection of the thoracic facet joints and wide lamina resection with complete removal of ligamentum flavum. 20 Correction is dependent on the mobility of the intervertebral disk, with distraction of the anterior column and compression of the posterior disk. The absence of anterior column disruption preserves the immediate and long-term load-sharing capacity and the stability of correction. Current recommendations are for bilateral pedicle screw fixation at every instrumented level, with PCOs in the thoracic and lumbar spine. Surgical techniques obtain correction by combining anterior lengthening through the intervertebral disk and posterior shortening by PCO of the spine.
The PCO enables closure and shortening of the posterior elements, resulting in decreased kyphosis. We prefer to first prepare all implant sites prior to the osteotomy. This preserves the anatomy so that proper screw trajectories can be obtained and the risk of neurologic injury is decreased with no dura exposed. In the PCO, the spinous process is removed entirely. A high-speed drill is used to cut a fracture line across the lamina and bilateral pars. Osteotomes and Capener gouge instruments are used to fracture the remaining pars/lamina bone, with this segment generally removed as a solid piece. Any remaining bone of the inferior articulating process is removed with a narrow Leksell rongeur, and the superior articulating process is carefully removed with Kerrison rongeurs, being sure to leave no bone or soft tissue that may be capable of causing foraminal obstruction during compression. Any remaining ligamentum flavum or lamina causing stenosis may then be removed, being careful to maintain a segment of intact lamina for stability and for a posterior fusion surface. This resection of bone can result in 5 to 7 degrees of correction when compressed posteriorly 20 ( Fig. 14.3 ).
Geck et al 20 reviewed 17 consecutive patients undergoing posterior-only pedicle screw instrumentation with PCO, and reported ex cellent correction compared with anterior-posterior technique controls. Lonner et al 21 also compared anterior-posterior and posterior-only procedures. Patients undergoing anterior-posterior were noted to have more overall complications (23.8% versus 5.5%) and an increased rate of junctional kyphosis (32% versus 4%), although they were noted to have decreased loss of correction (3.2 versus 6.4 degrees). The groups had similar reoperation rates (5.1%) due to symptomatic junctional kyphosis.
Treatment of Scheuermann disease should aim to correct the kyphosis to the high normal range of thoracic kyphosis (40 to 50 degrees). Overcorrection of a kyphotic deformity can lead to neurologic complications, postoperative sagittal malalignment, and proximal junctional kyphosis (PJK). Lowe and Kasten 22 recommended that no more than 50% of the preoperative kyphosis be corrected, and that the final kyphosis should never be less than 40 degrees. This is achieved through two techniques: segmentally applied and apically directed com pression forces, and cantilever reduction. With cantilever reduction two 5.5- or 6.0-mm rods are contoured to an anticipated kyphosis, inserted into the pedicle screws proximal to the apex, and cantilevered into the distal implants. This technique has been implicated in junctional kyphosis, with the force concentrated on either end of the construct. Apical compression techniques use multisegment fixation and compression toward the apex to reduce the kyphosis. This has the advantage of spreading out reduction forces to the entire construct rather than being concentrated at junctional levels. A review by Denis et al 23 found the majority of Scheuermann correction was achieved by a combination of cantilever and apical compression techniques. They found no difference in the rate of PJK between any of the techniques used. Both hooks and pedicle screws are safe and efficacious, but most surgeons prefer the mechanical advantage afforded by pedicle screws in deformity correction.
We have found success reducing strain on proximal and distal screw foundations seen in cantilever loading by using temporary apical cantilever rods and sequential reduction of the correction rod. Final full-length rods are then secured when the kyphotic spine is almost fully corrected. We prefer to lock the proximal three or four screws in place with no corrective force, and cantilever them into the distal segments. Although improvement in the instrumentation enables more effective force application to the spine for deformity correction, the ultimate success of the correction relies on the effectiveness of the release.
It is rare for Scheuermann kyphosis to need release beyond PCO, but in some instances a three-column osteotomy may be necessary, which may include a vertebral column resection (VCR) or less commonly a pedicle subtraction osteotomy (PSO). Vertebral column resection is reserved for use in the thoracic or thoracolumbar spine for the treatment of sharp or angular kyphotic deformity, fixed kyphotic segments, or congenital malformations. Advantages include the potential for dramatic correction in all three dimensions and overall shortening of the vertebral column, which help to relieve tension on the anterior neurovascular structures. Correction up to 45 degrees in the sagittal plane has been reported. 24 VCR involves complete resection of all posterior elements at the level of the VCR in addition to complete removal of the vertebral body and adjacent rostral and caudal intervertebral disks. In most cases, an anterior fusion is performed with structural support via an anterior cage that enables preservation of the anterior column height, recreation of the lordosis, and enhanced correction ( Fig. 14.4 ). VCR remains the most powerful method of three-dimensional deformity correction; however, the technique poses the greatest technical challenge and the greatest risk to the patient in terms of possible neurologic injury, operative time, blood loss, and potential morbidity.
Pedicle subtraction osteotomies are seldom used in the thoracic spine because wedging of the thoracic vertebra limit the height of the anterior column, and therefore the potential for correction. However, in the lumbar spine, or if there is sufficient anterior vertebral height, a wedge resection or pedicle subtraction may be a useful technique for deformity correction. The procedure is typically associated with shortening of the posterior column without lengthening the anterior column. A PSO requires removal of all posterior elements at the level of correction, including the pedicles and superior and inferior adjacent facet joints. A posterior wedge of bone is then removed from the vertebral body, including the entire posterior and lateral vertebral body walls, to enable osteotomy closure. This results in bone-on-bone approximation with a high rate of healing. We have found that a PSO can achieve up to 35 degrees of correction ( Fig. 14.5 ). PSOs are again preferred below the level of the conus, but can be performed in the thoracic spine, although with increased risk. 25 With an extended PSO the bony resection extends cranially to include the rostral intervertebral disk. The vertebral end plate is then closed directly onto the cancellous osteotomy wedge or the PSO. An inter-body spacer can be placed in the middle or anterior third of the disk space to be used as a fulcrum, obtaining the same correction with less compromise of the neural structures. Again, PSO is rarely indicated in the thoracic spine but may be considered for fixed thoracic deformity or thoracolumbar congenital malformations.
The selection of the appropriate level for instrumentation and fusion is an important consideration in the treatment of Scheuermann disease. It is important to extend the fusion over the entire length of the kyphotic deformity. Failure to do so may result in PJK or distal junctional kyphosis (DJK). Junctional kyphosis most often is a result of junctional ligamentous disruption, too much deformity correction, and failure to incorporate appropriate vertebra. Most surgeons agree that the upper limit of the fusion must be the proximal end vertebra in the measured kyphosis, or to the level of the first lordotic disk. 22 For high-and midthoracic apex deformities, the upper-instrumented vertebra is typically T2 or T3. Fusions short of the proximal end vertebra and disruption of the junctional ligaments are the main risk factors for PJK. In a study of 40 patients in whom the fusion incorporated the proximal end vertebra, three had disruption of junctional ligaments and all three developed PJK. 23
There has been debate regarding the distal extend of fusion. Some authors advocate extending fusion to the first lordotic vertebra (FLV). Denis et al 23 assessed 67 patients who had correction of Scheuermann kyphosis. Eight patients developed DJK, with seven of them having fusion short of the FLV. Other authors advocate extending the fusion to the second lordotic vertebra. Poolman et al 26 reported on a series of patients with no DJK when the second lordotic vertebra was incorporated into the fusion.
More recently the concepts of sagittal stable vertebra (SSV) and its relationship to the lowest instrumented vertebra (LIV) and FLV have become important considerations. The SSV is the most proximal vertebra touched by the posterior sacral vertical line (PSVL), with the PSVL being a vertical line from the posterior-superior corner of the sacrum on a lateral upright radiograph. When the LIV includes the SSV, the fusion mass will be centered over the sacrum in the sagittal plane, achieving improved global balance. Cho et al 27 reported a series of 24 consecutive patients with no DJK when the SSV was included in the distal fusion. To maintain global sagittal alignment after surgery, both the proximal and distal ends of the fusion should be within the center of gravity. This equates to a fusion incorporating the upper end of kyphosis and SSV distally. The disadvantage of incorporating the SSV is that the fusion level may be one segment longer distal to the FLV, with loss of a mobile segment. Another important consideration is the postoperative lumbopelvic mismatch. Following correction of the thoracic spine, there is simultaneous reduction in lumbar lordosis. Patients with a high pelvic incidence should undergo a smaller thoracic kyphosis reduction to prevent excessive loss of lumbar lordosis. Nasto et al 28 reported a higher incidence of PJK in patients with significant lumbopelvic mismatch. Overall, distal fusion including the FLV but not the SSV is not appropriate in all patients, with distal junctional problems developing more often in patients with fusion to the FLV rather than the SSV. Selection of fusion levels remains an important and challenging aspect of kyphosis surgery, and overcorrection should be avoided.
With Scheuermann disease there are some additional complications to consider. Lonner et al 29 documented a comparison of Scheuermann kyphosis correction with adolescent idiopathic scoliosis (AIS) correction, and found Scheuermann patients were 3.9 times more likely to have a major complication. Scheuermann patients were also significantly more likely to undergo a second operation and expe rience infections. The prolonged operative time, anatomic variation, and lower incidence of operative Scheuermann patients have all been proposed as reasons for increased complications in Scheuermann disease compared with AIS patients.