Asymmetric Spondylolisthesis as the Cause of Childhood Lumbar Scoliosis—Can New Imaging Modalities Help Clarify the Relationship?

Abstract

The etiology of idiopathic scoliosis is likely genetic. Research is proceeding to identify the responsible genes. Although genetics accounts for the majority of idiopathic scoliosis, a subset of curves occur secondary to mechanical “foundation”issues at the lumbosacral junction. The most common mechanical “foundation”issue at the lumbosacral junction is spondylolisthesis. A relationship between lumbar scoliosis and spondylolisthesis has been well documented. Modern imaging studies are providing an opportunity to cast new light on this inter-relationship. First, computerized tomography (CT) studies, and now 3-D CT studies of the lumbosacral area have been performed in an attempt to further elucidate this matter. The purpose of this paper is to provide an introduction to the topic and to present images that suggest an etiologic relationship between lumbar scoliosis and spondylolisthesis with mild asymmetric spondylolisthesis proposed as the cause of the lumbar curve.

INTRODUCTION

The precise etiology of idiopathic scoliosis remains unknown. A complex mixture of inherited biochemical abnormalities of the discs, ligaments, and/or bone of the spinal column likely predisposes individuals to development of spinal curves. Extensive research is proceeding to identify the responsible genes, with some success.¹ This new molecular information predicts an era in which genetic testing will be utilized to distinguish true scoliosis from curves secondary to anatomic disorders.

Thoracic and thoracolumbar curves likely develop secondary to the aforementioned genetic abnormalities during periods of rapid growth. Lumbar scoliosis may be explained by similar mechanisms in some cases, but mechanical “foundation” issues also need to be considered because of the lumbar spine's proximity to the sacrum and pelvis.“Foundation” issues are especially pertinent in cases of lumbar scoliosis in which there is no family history of spinal asymmetry.

In some cases, curvature of the lumbar spine is associated with structural abnormalities of the lumbosacral junction, ranging from asymmetry of the sacrum to L5 lateral masses (Bertolotti syndrome)² to spina bifida occulta.³ The most common known structural abnormalities at the lumbosacral junction are spondylolysis at the L5 level and associated spondylolisthesis with the L5 vertebra slipping anteriorly on the sacrum.

Approximately 6% of children will develop spondylolysis by adulthood and 74%of those children will develop spondylolisthesis.³ Additionally, children with spondylolisthesis commonly develop spinal asymmetry and/or scoliosis. This increased incidence of spinal curvature in patients with spondylolisthesis has been well documented.^4,5,6,7,8,9 Studies have shown that up to 48% of children with spondylolisthesis develop at least five degrees of lumbar spinal asymmetry.⁸ Although most experts would agree that patients with spondylolisthesis and scoliosis of the upper spine likely have two separate and unrelated conditions, asymmetry in the lumbar spine presents a different possibility.

Modern imaging studies (CT and 3-D CT of the lumbosacral area) are providing an opportunity to cast new light on this inter-relationship. Review of multiple cases suggests that L5 spondylolisthesis commonly does not displace symmetrically as evidenced by a greater space in the pars defect on one side versus the other, leading to rotation of the L5 vertebrae as it relates to the sacrum. This asymmetric slip is a probable cause of lumbar scoliosis in some children.

We suggest an etiologic relationship between lumbar scoliosis and spondylolisthesis with mild asymmetric spondylolisthesis proposed as the cause of the lumbar curve. This report previews a larger retrospective study being conducted at our institution.

CASE REPORTS

Case 1

A ten-year-old female with a history of possible mild Ehlers-Danlos syndrome presented with lower back pain. A postero-anterior standing spine film showed mild lumbar scoliosis (Figure 1A) and a lateral spine film showed grade 1-2 spondylolisthesis (Figure 1B). A CT study of her lumbar spine was obtained to evaluate possible pars interarticularis defects and spondylolisthesis.

Figure 1A.

(left)(Case 1) Postero-anterior scoliosis radiograph of a 10-year-old female.

Figure 1B.

(above)(Case 1) A lateral radiograph demonstrates grade 1-2 spondylolisthesis at the L5 level.

The CT study (Figure 2) showed bilateral pars interarticularis defects with associated grade 1-2 spondylolisthesis of L5 on S1. Of note, the olisthetic L5 vertebrae had an asymmetric rotational slip with the pars defect greater on the left than on the right. We refer to this pattern of CT imaging as an “asymmetric ring”. The anatomy of the asymmetric spondylolisthesis was then visualized using a 3-D CT study (Figures 3A, 3B).

Figure 2.

(Case 1) A CT study at the L5 level demonstrates asymmetric spondylolisthesis. We refer to this pattern as an“asymmetric ring.” The pars displacement on the left (arrow) is greater than on the right.

Figure 3A.

(Case 1) A 3D-CT study (right-sided oblique view) shows a smaller L5 pars defect on the right as compared to the left (see Figure 3B).

Figure 3B.

(Case 1) A 3D-CT study (left-sided oblique view) shows a larger pars defect on the left at L5 (arrow).

Case 2

An otherwise-healthy twelve-year-old male presented to our clinic with a complaint of right hip pain. Initial x-rays were obtained and the patient was found to have a sclerotic lesion of the right pedicle of L5 consistent with osteoid osteoma versus a healing spondylolytic lesion (Figures 4A, 4B). A CT study was ordered to further investigate the lesion and demonstrated bilateral pars interarticularis defects at the L5 level and grade 1 spondylolisthesis with sclerosis along the margins of the right-sided spondylolytic lesion (Figure 5). In this case the pars lesions appeared to be symmetric and there was no associated lumbar scoliosis.

Figure 4A.

(Case 2) This postero-anterior scoliosis radiograph in a 12-year-old boy shows no evidence of scoliosis. (Radiograph taken as scoliosis film – re: R vs. L.) The L5 pedicle shows greater density.

Figure 4B.

(Case 2) A lateral radiograph demonstrating grade 1 spondylolisthesis.

Figure 5.

(Case 2) A CT study at L5 shows symmetric pars defects associated with grade 1 spondylolisthesis. The bilateral pars defects are of equal size and the L5 ring is symmetric. Lumbar scoliosis may be less likely with this symmetric pattern.

Case 3

An eight-year-old female with a previous diagnosis of L5-S1 grade 1 spondylolisthesis and both lumbar and thoracic juvenile idiopathic scoliosis presented to clinic with a complaint of worsening spinal deformity. Radiographs confirmed the presence of spondylolisthesis and a worsening scoliosis (Figures 6A, 6B). A CT study was obtained to investigate the progression of the spondylolisthesis. Imaging showed an asymmetric grade 1 spondylolisthesis (Figure 7) with an “asymmetric ring” at the L5 level.

Figure 6a.

(Case 3) Postero-anterior scoliosis radiograph of an 8-year-old female showing lumbar asymmetry. The patient also had a 20-degree secondary thoracic scoliosis on the right.

Figure 6b.

(Case 3) Lateral radiograph demonstrating grade 1 spondylolisthesis.

Figure 7.

(Case 3) A CT study shows bilateral pars defects with an asymmetric ring at the L5 level.

The patient underwent in situ posterior spinal fusion of L5-S1 via the Wiltse technique, was placed in a single-leg spica cast for one month and was then transitioned to a Boston-type scoliosis brace. The brace was worn for 23 hours per day the first 3 months and 16 hours a day thereafter. Despite the successful L5-S1 fusion with stabilization of her spondylolisthesis, she continued to have progression of her scoliosis (Figure 8).

Figure 8.

(Case 3) Postero-anterior radiograph taken eight months after in situ L5-S1 fusion. The lumbar curve has progressed moderately.

DISCUSSION

Both idiopathic scoliosis and spondylolysis are very common childhood conditions. The incidence of scoliosis in this age group is estimated to be 0.5-3% of the population^10,11,¹²,¹³,¹⁴,¹⁵,¹⁶ while spondylolysis is far more common at 6%.³ Patients with true scoliosis often have a strong family history for curve progression due to its complex genetic etiology. Spondylolysis and spondylolis-thesis also have a familial predisposition,^17,18 however, only a very small percent of the population are aware that they have the condition. When spondylolisthesis and adjacent scoliosis are seen together in a child with no family history of scoliosis one might reasonably query as to whether the spinal curvature was in some way caused by the adjacent structural abnormality.

A high incidence of lumbar scoliosis associated with spondylolisthesis has been shown in several series.^4–9 In fact, studies have shown that up to 48% of children with spondylolisthesis develop at least five degrees of scoliosis.⁸

Spinal curvatures occurring concomitantly with spondylolisthesis have been divided into three main categories.⁸ First, the curvature can simply be idiopathic scoliosis of the upper spine, likely unrelated to the olisthetic defect. Second, the curvature can be of the “sciatic” type in which irritation associated with the olisthetic defect induces scoliosis via muscle spasm. Third, the curvature can be the result of an asymmetric olisthetic defect as first described by Tojner.⁹ In this case, the displaced vertebra is translated in both the sagittal and coronal planes and additionally rotated around the vertical axis thereby creating an asymmetric foundation leading to a rotatory deformity of the spine above.

In most prior studies, scoliosis in association with spondylolisthesis has been thought to be of the sciatic type.^6,7,19,20 Only one series, in which olisthetic scoliosis was found in 30% of patients with symptomatic spondylolysis, supported the asymmetric foundation view.⁹ A better understanding of the relationship between scoliosis and spondylolisthesis using modern imaging techniques should help clarify the issue.

The three cases presented are meant to provide an overview of the possible relationships between spondylolisthesis and scoliosis.

Case 1 demonstrates a patient in which lumbar scoliosis is associated with obviously asymmetric spondylolis-thesis. It is important to note that there was no family history of scoliosis over several generations. The current view that idiopathic scoliosis is a genetic, inherited condition makes the appearance of scoliosis in a patient with no family history for the disorder somewhat problematic and adds suggestive evidence that the spinal asymmetry associated with spondylolisthesis may be due to the olisthetic disorder rather than a true scoliosis.

This case illustrates the key issue of this paper and focus of our study: Asymmetric spondylolisthesis may prove to be of value in predicting progression of lumbar scoliosis associated with spondylolisthesis. Two mechanisms may be responsible for this association. First, asymmetric olisthesis may be more likely to trigger muscle spasm via tissue irritation inducing sciatic scoliosis. In fact, prior studies have shown that lumbar curves with rotatory olistheses are more likely to be associated with radicular pain.²¹ Second, the asymmetric olisthesis may create an asymmetric foundation which causes the vertebrae above the slip to rotate into a torsional lumbar scoliosis.

It is interesting to note that previous investigators have found that the incidence of scoliosis associated with spondylolisthesis is much greater when the pars defect is at the L4-L5 level versus the L5-S1 level.⁷ These investigators found that eight out of nine patients with L4-L5 spondylolisthesis also had scoliosis. They attributed the higher incidence of scoliosis in these patients to the absence of stabilizing ligaments, such as the iliolumbar ligament, at the L4-L5 level. It's likely that the olisthetic vertebrae in patients with ligamentous laxity, such as the girl in case 1 with an Ehlers-Danlos variant syndrome, would have more freedom to rotate and slip in an asymmetric fashion. Perhaps children with ligamentous laxity are more prone to develop lumbar scoliosis if they have spondylolisthesis.

Case 2 demonstrates a patient with grade 1 symmetric spondylolisthesis without an associated scoliosis and is meant to serve as a contrast to case 1. Without significant vertebral rotation there is no asymmetric foundation for the proximal vertebrae and there is less tissue irritation to induce muscle spasm. Hence, the mechanisms for the olisthetic induction of scoliosis are absent.

Case 3 demonstrates a patient with lumbar scoliosis and a secondary thoracic curve. She was also noted to have grade 1 spondylolisthesis. She was initially treated with physical therapy but she had both an increase in her scoliosis and greater lumbosacral pain. A CT scan demonstrated an “asymmetric ring” at L5. She was treated with a Wiltse-type L5-S1 fusion which relieved her back pain. Unfortunately her scoliosis continued to progress despite the successful fusion.

There are two explanations for the progression of her scoliosis. First, the vertebrae may have been fused in situ into a position of permanent asymmetry. Hence, the suspected force driving the lumbar scoliosis may have remained post-operatively. Of note, Seitsalo reported that spondylolisthesis fusions often fail to correct the scoliosis when a significant rotatory component was present.⁸ Perhaps de-rotational correction of the ring asymmetry will be considered in the future when these relationships become better established. Alternatively, the child could have true genetic scoliosis unrelated to the spondylolisthesis.

Important questions arising from our study will focus on treatment options. Should asymmetric spondylolisthesis be treated aggressively with reduction and fusion? Should asymmetric spondylolisthesis be treated in the absence of symptoms to prevent the development of scoliosis?

Previous studies have shown that fusion of the lumbosacral area does not effectively treat thoracic or thoracolumbar curves (as in our case 3), but lumbar curves typically disappear post fusion.⁸ The resolution of lumbar curves is presumed to be the result of relieving muscle spasm caused by the spondylolisthesis. However, lumbar curves with a significant torsional component are sometimes, but not always, relieved by an in situ lumbosacral fusion. These curves have been described as “fixed structural scoliosis”and attributed to long-term muscle contraction.²⁰ Perhaps these curves are caused by asymmetric spondylolisthesis, and reduction of these slips at the time of a single-level fusion may be necessary to prevent curve progression and even potentially reverse the scoliosis.

Although the majority of children with asymptomatic spondylolisthesis do not require treatment and can have a normal childhood, certain children may benefit from early fusion of asymmetric spondylolisthesis to avoid development of scoliosis and associated morbidity. If significant rotatory change at the L5 level is noted (i.e.,an “asymmetric ring” is present on CT imaging), pedicle screw rotational reduction plus fusion could be considered rather than the currently favored in-situ fusion. Such an approach would seem radical at this time, however future analysis (and similar surgical methods) may make such an approach reasonable.

The relationship between scoliosis and spondylolisthesis in children is well established, but the actual etiologic relationship requires further research. Modern imaging modalities, specifically CT and 3D-CT studies, should help to elucidate the relationship.