The clinical features and surgical treatment of degenerative lumbar scoliosis: A review of 112 patients

Abstract

Objective:  To investigate the clinical features, radiological characteristics and surgical results of degenerative lumbar scoliosis (DLS).

Methods:  One hundred and twelve cases of DLS treated surgically from June 2001 to February 2006 were retrospectively reviewed for clinical features, characteristics of nerve root compression and imaging presentations. According to the preoperative clinical manifestations and imaging findings, different surgical modalities were performed, including simple nerve decompression and decompression with short or long posterior fusion (less or more than three segments, respectively).

Results:  The mean age of 47 male and 65 female patients was 54.7 years. Clinical manifestations included lower back pain (76.8%), radiculopathy (79.5%) and claudication (48.2%). Plain lumbar radiograph showed right scoliosis in 87 and left scoliosis in the other 25 cases; the Cobb angle was 10°–46°; the apex of scoliosis mostly located at L3 (48.2%); L3 and L4 nerve roots were usually compressed on the concave side and L5 and S1 nerve roots on the convex side. The Cobb angle and physiologic lordosis angle of patients who underwent multi‐segment (>3 segments) fusion improved to a greater extent than did that of patients who had simple decompression without fusion. A mean 5.7‐year follow‐up showed that the average improvement in Oswestry disability index (ODI) scores was 32.6, 26.3 and 13.5 for long segment fusion, short segment fusion and simple decompression without fusion, respectively.

Conclusion:  Decompression surgery with or without fusion, the main purpose of which is to relieve nerve root compression and stabilize the spinal column, is an effective treatment for chronic DLS. The treatment should be individualized according to the patient's age, general and economic factors, severity of deformity and other coexisting lumbar degenerative disorders.

Introduction

Degenerative lumbar scoliosis (DLS) is a condition secondary to degeneration of the lumbar spine in adults ¹ . The Cobb angle in DLS usually does not exceed 40°, but as this condition is often associated with rotatory subluxation and lateral listhesis, and complicated with other lumbar degenerative disorders, treatment of DLS is complex and relatively difficult.

The pathogenesis of DLS is believed to be mainly related to degeneration of the lumbar intervertebral discs and zygapophysial joints, and changes in canal morphology and volume ² , ³ , ⁴ , ⁵ . Some researchers believe that DLS may also be secondary to compressive fracture and asymmetric intervertebral space collapse due to osteoporosis, but most researchers maintain that compressive fracture caused by osteoporosis is a contributing rather a causative factor for DLS ⁶ . When intervertebral collapse or asymmetry of bilateral zygapophysial joints occur, the intervertebral discs and zygapophysial joints lose their function of maintaining the normal conformation of the spine. The vertebrae then change their alignment and abnormal activities lead to instability and asymmetry of the spinal column and eventually to DLS ¹ , ⁷ . DLS progresses slowly. Some researchers have found that the Cobb angle increases about 1.8° annually, but it may also progress quickly in some patients ⁸ . The purpose of the present study was to analyze clinical features, and evaluate surgical outcomes of DLS, by retrospective review of 112 surgically treated patients with DLS.

Materials and methods

From all patients who received surgical treatment for DLS from June 2001 to February 2006 in Changzheng Hospital (Shanghai, China), 112 patients whose Cobb angle was greater than 10° and who were without congenital abnormality of the lumbar spine, tuberculosis, tumors or traumatic wounds were selected for this study. They included 47 men and 65 women ranging in age from 43 to 67 years with a mean of 54.7 years. When the patients sought medical treatment in our hospital for the first time, the duration of disease ranged from 6 months to 15 years with a mean of 3.4 years. Main clinical manifestations were low back pain, radiculopathy and claudication. Preoperative assessment included routine lumbar posterior‐anterior and flexion‐extension position radiographs, lumbar computed tomography (CT) and magnetic resonance imaging (MRI) in all patients; and lumbar myelography in some patients to ascertain whether there was complicated lumbar intervertebral disc protrusion, canal stenosis or spondylolisthesis.

Most patients (91%) underwent conservative treatment which included bed rest, administration of non‐steroidal anti‐inflammatory drugs, avoidance of lifting heavy objects and other changes in lifestyle for about 6 months before surgery. When all these treatments failed to work effectively and symptoms deteriorated, surgical intervention was performed. In patients (9%) who had severe symptoms of canal stenosis or nerve root compression, especially cauda equina syndrome, when they first visited the clinic, surgical treatment was the initial form of management.

Surgical modalities were dependent on the clinical symptoms and imaging findings. In our series, simple nerve root decompression was performed in 15 cases whose main clinical presentation was nerve root symptoms without lower back pain, intermittent claudication or significant canal stenosis; decompression by posterior laminectomy plus bone graft fusion of short (≤3 segments) (Fig. 1) or long (>3 segments) (Fig. 2) fixation with pedicle screw instrumentation was performed in the other 97 patients. The latter surgical modality was used mainly for DLS complicated by intervertebral disc protrusion, canal stenosis, and instability or spondylolisthesis of the lumbar vertebrae. Short fusion and fixation was mainly used in patients with local/sectional intraspinal nerve compression and without obvious symptoms of instability and nerve compression in the upper and lower vertebral segments, especially for those in poor condition who could not tolerate significant trauma. Long fusion and fixation was mainly used in patients with multi‐segmental neural compression and/or spinal instability. Intervertebral fusion, posterolateral fusion or interlaminar fusion was performed after fixation.

Figure 1.

Imaging findings in a 70‐year‐old man with DLS. This patient, who was in very poor general condition, complained of low back pain, severe claudication and radiculopathy. (a) Anteroposterior and (b) lateral radiographs show degenerative lumbar scoliosis (DLS) with segmental instability and (c) CT scan shows spinal stenosis in the L3–L4 and L4–L5 segments. A nerve decompression with short fusion was performed and produced a good result after 3.5 years follow‐up (d) anteroposterior radiograph, (e) lateral radiograph.

Figure 2.

Imaging findings in a 59‐year‐old man with DLS. This patient complained of low back pain and claudication. Standing lumbar radiographs show (a) a lordosis angle of 22° and (b) a Cobb angle of 24°. (c) With myelography severe central canal stenosis was observed below the L2 vertebra. (d) The lordosis angle was significantly improved after broad decompression and long fusion surgery to 40° and (e) the Cobb angle to 3°.

All patients were followed up clinically at 2‐month intervals for the first year, and then yearly. Average follow‐up was 5.7 years (range, 2.5–7 years). The Oswestry disability index (ODI) ⁹ was used to evaluate surgical outcomes. We also evaluated the results as follows: excellent = normal working capacity, normal walking distance, no or occasional residual pain; good = normal or mildly reduced working capacity, mild restriction of walking distance, mild residual back or radicular pain; fair = incapable of working, walking restricted to short distances, low back or radicular pain slightly improved compared to that experienced pre‐operatively; poor = incapable of working, massive restriction of walking distance, pain unchanged or worse ¹⁰ .

We analyzed for statistically significant correlations by the χ² test, Dunnett's t‐test and repeated measures ANOVA test with SAS 8.2 statistical software (SAS Institute, Cary, NC, USA). All differences were regarded as significant only if P < 0.05.

Results

Clinical features

The main clinical manifestations for all patients were low back pain (86/112, 76.8%), radiculopathy (89/112, 79.5%), and claudication (54/112, 48.2%). Most patients had more than one complaint, although some patients had a single symptom, for example, lower back pain only in 10 patients (8.9%), radicular pain only in 21(18.8%), and claudication only in 5 (4.5%). There were three cases (2.7%) complicated by cauda equina syndrome (CES) (Table 1).

Table 1.

Clinical features in 112 patients with DLS

	Number	Percentage (%)
Low back pain with	86	76.8
Radicular pain	26	23.2
Claudication	6	5.4
Radicular pain and claudication	41	36.6
Radicular pain and cauda equina syndrome	1	0.9
Claudication and cauda equina syndrome	2	1.8
Simple low back pain	10	8.9
Simple radicular pain	21	18.8
Simple claudication	5	4.5

Characteristics of nerve root compression

Excluding patients with significant intervertebral disc protrusion, lumbar spondylolisthesis, CES, claudication only and lower back pain only, 37 patients in our series had nerve root symptoms, including those of the concave side at L3 in 23 patients (62.2%), at L4 in 27(73%), at L5 in 7(18.9%) and at S1 in 5(13.5%); and those of the convex side at L3 in 11 patients (29.7%), at L4 in 8(21.6%), at L5 in 16(43.2%), and at S1 in 12(32.4%). The above results demonstrate that L3 and L4 nerve roots were usually compressed on the concave side while L5 and S1 nerve root were more often compressed on the convex side (χ² test, P < 0.001). While most patients had involvement of multiple or bilateral nerve roots, clinically pain, numbness and motor dysfunction were localized in most patients to the innervation area of 1–2 roots.

Imaging evaluation

Preoperative imaging showed that in 43 of the 112 patients the DLS was complicated by severe intervertebral disc protrusion, and in 20 cases by degenerative lumbar spine instability or spondylolisthesis. Preoperative MRI or lumbar myelography showed that all 112 patients had central canal stenosis or lateral recess stenosis of varying degrees.

Posterior‐anterior X‐ray examination showed that 87 patients had right lumbar scoliosis and the remaining 25 left lumbar scoliosis, and that the apex of scoliosis was mostly seen at the L3 vertebra. In our series, the apex of scoliosis was seen at the L1–2 intervertebral disc in two cases, at the L2 vertebra in five, at the L2–3 intervertebral disc in ten, at the L3 vertebra in 54, at the L3–4 intervertebral disc in 23, and at the L4 vertebra in 18 cases. Apical vertebrae or their superior vertebrae were the sites where rotatory subluxation was most likely to occur. Rotatory subluxation occurred in about 70% (15/21) patients of our series whose Cobb angle was larger than 30° (Table 2).

Table 2.

Radiological characteristics in 112 patients with DLS

	Number	Percentage (%)
Direction of lumbar scoliosis
Left	25	22.3
Right	87	77.6
Degree of lumbar scoliosis
10°–20°	53	47.4
21°–30°	38	33.9
>30°	21	18.7
Degree of lumbar lordosis
10°–30°	75	67
31°–50°	33	29.5
>50°	4	3.5
Apex of lumbar scoliosis
L1–2 intervertebral disc	2	1.8
L2 vertebra	5	4.5
L2–3 intervertebral disc	10	8.9
L3 vertebra	54	48.2
L3–4 intervertebral disc	23	20.5
L4 vertebra	18	16.1

Surgical results

The preoperative Cobb angle was 10°–46°, including 10°–20° in 53 cases, 21°–30° in 38 and >30° in 21. The degree of lumbar scoliosis and lordosis improved more markedly in patients receiving multi‐segment (>3 segments) fixation and fusion than in those patients receiving short fusion (≤3 segments) or simple decompression without fusion (Dunnett's t‐test, P < 0.05). In 15 patients who did not receive fusion, the mean Cobb angle increased (repeated measures ANOVA test, P= 0.0135) and the lordosis angle decreased during the follow‐up period (repeated measures ANOVA test, P < 0.001) (Table 3).

Table 3.

Comparisons among three groups of patients using different surgical methods

	Fusion less than 3 segments (short fusion) (n= 34)	Fusion more than 3 segments (long fusion) (n= 63)	Decompression without fusion (n= 15)
Lumbar scoliosis (Cobb angle, o)
Preoperative	17.6 ± 2.8	24.3 ± 4.5	15.3 ± 3.7
Postoperative †	11.3 ± 2.4*	12.8 ± 3.9*	14.9 ± 2.8
Final evaluation ‡	15.5 ± 3.7**	14.6 ± 5.1*	19.9 ± 6.1*
Lumbar lordosis (lordosis angle, o)
Preoperative	30.6 ± 8.1	21.7 ± 7.5	28.7 ± 4.2
Postoperative †	31.7 ± 9.6	29.5 ± 10.1*	29.3 ± 5.3
Final evaluation ‡	27.3 ± 7.2**	28.2 ± 8.4*	24.4 ± 3.4*
Blood loss (ml)	809 ± 95	1627 ± 63	205 ± 72
Operative time (min)	131 ± 53	184 ± 49	78 ± 26
Secondary operation	1 (3%)	5 (8%)	5 (33.3%)
Adjacent segment disease (No.)	16 (47.1%)	23 (36.5%)
Clinical outcomes (ODI) (No.)
Preoperative	50.5 ± 6.5	53.3 ± 5.8	45.3 ± 7.7
Postoperative †	17.3 ± 4.9*	15.8 ± 6.9*	15.9 ± 5.4*
Final evaluation ‡	25.1 ± 5.2*	21.6 ± 5.7*	32.4 ± 8.1*

^* P < 0.01,

^** **P < 0.05, compared with preoperatively;

^{† †} one year after surgery;

^{‡ ‡} average follow‐up of 5.7 years.

Complications

Cerebrospinal fluid leakage, which occurred postoperatively in six cases, was cured by prolonged drainage and pressure dressing. Intracanal hematoma and mild paralysis of both extremities occurred in one patient, the symptoms were relieved by emergency removal of the hematoma. Poor bone grafting fusion and pseudoarthrosis occurred in nine out of 97 patients who received posterior fixation and fusion. Re‐operation was done in six patients because of loosening or breaking of a pedicle screw. Five of 15 patients who received simple interlaminar decompression underwent a second operation in which posterior bone grafting fixation and fusion was performed (three for intervertebral disc re‐protrusion and two for instability of the spinal column).

Adjacent segment disease (ASD) after lumbar fusion was also analyzed. ASD in our series included degeneration, protrusion, retrolisthesis and anterolisthesis of intervertebral discs of the superior and inferior adjacent segments, and instability of the spinal column on dynamic X‐ray films. Follow‐up showed that ASD occurred in 16 (47.1%) patients from the short fusion group and 23 (36.5%) patients from the long fusion group. Only 21 (53.8%) ASD patients complained of significant clinical symptoms.

Clinical outcomes

Preoperative symptoms were relieved within one year of surgery in most patients, these were allowed to resume normal daily life. The clinical outcome was excellent in 73 patients (65.2%), good in 21(18.7%), fair in 15 (13.4%) and poor in 3(2.68%). There was no significant difference in ODI among the three groups of patients one year after operation (Dunnett's t‐test, P > 0.05). After a mean of 5.7 years follow‐up, the clinical outcome was excellent in 62 patients (55.3%), good in 15(13.4%), fair in 22(19.6%), and poor in 13(11.6%). Group pairwise comparison showed that there was a significant difference between the long segment fixation/fusion, short segment fixation/fusion and simple decompression groups in ODI (Dunnett's t‐test, P < 0.05). The average improvement in ODI scores was 32.6, 26.3 and 13.5, respectively. The fusion rate was about 90% in patients with lumbar fixation and fusion after an average of 5.7 years follow‐up.

Discussion

DLS mostly occurs after 40–50 years of age ¹¹ . The mean age at time of diagnosis in our series was 54.7 years, and the male/female ratio was 1:1.4. DLS rarely occurs alone, and is often associated with lumbar canal stenosis, intervertebral disc protrusion, degenerative lumbar spine instability and other lumbar degenerative disorders. Most patients in our series had multiple co‐existing lumbar degenerative disorders. The Cobb angle in DLS is usually <40°, and may be as large as 60° in rare cases ¹² . It may include 4–5 vertebral segments, mainly involving lumbar vertebrae, plus inferior thoracic vertebrae in some patients as compensatory scoliosis. Imaging presentation of DLS may also include vertebral rotation, semi‐dislocation and lateral displacement. The top vertebra is most likely to rotate, usually by 1–2 degrees. Rotatory subluxation may occur in any segment of DLS, though it is mostly seen in the top vertebra or its immediately superior vertebra.

CT presentation of DLS mainly includes vertebral rotation, osseous hyperplasia of the vertebral edge on the concave side, hyperplastic degeneration of the zygapophysial joints on the convex side, and deformity of the neck of the vertebral arch, which is mainly due to compression on the concave side and traction of the ligament on the convex side. Central canal stenosis, lateral recess stenosis or foraminal stenosis of varying degrees were seen on MRI or myelography in almost all scoliotic patients. Zygapophysial joints and ligamentum flavum hyperplasia, lamina thickening, vertebral arch neck deformity, hyperostosis, and intervertebral disc bulge or protrusion are the main causes of canal stenosis or lateral recess stenosis, though the canal volume is also related to congenital development.

When DLS is complicated by intervertebral disc protrusion or spondylolisthesis, nerve root symptoms of the involved segments are often present. Apart from nerve root compression for these specific reasons, root compression in DLS is characteristically present in any case. In their study of 22 DLS patients with simple nerve root pain, Liu et al. found that L3 and L4 roots were mostly compressed on the concave side while L5 and S1 roots were most often compressed on the convex side ¹³ , which is consistent with our findings. There are multiple causes for nerve root compression, the main ones being degenerative central canal or foraminal stenosis, hyperostosis outside the intervertebral foramen, rotatory subluxation and lateral listhesis. Nerve root involvement on the convex side may be related to foraminal stenosis or extra foraminal stenosis, while root compression on the convex side may be related to canal deformity, lateral recess stenosis or nerve root traction.

Conservative treatment may be effective in some DLS patients, but when the symptoms are uncontrollable and exacerbate progressively, surgical intervention should be considered ⁴ , ¹⁴ . The deformed spinal column in middle and old aged patients with DLS is often rigid; in these patients the main purpose of surgery is to decompress the cauda equina and lumbosacral roots so as to stabilize the spinal column. Modalities of surgical treatment include nerve root decompression, laminectomy decompression and fusion with pedicle fixation, and combined anterior‐posterior decompression with orthopaedic fixation.

Simple nerve root decompression is rarely used, and is reserved mainly for patients with simple nerve root pain without lower back pain, intermittent claudication, cauda equina syndrome or significant lumbar dynamic instability. Assessment of the degree of nerve root decompression depends mainly on clinical symptoms and myelographic findings. In cases of nerve root compression complicated by segmental canal stenosis, hemi or total laminectomy decompression should be considered. However multi‐segment laminectomy may cause lumbar intervertebral instability and kyphosis. For this reason, it is preferable to perform fixation and fusion in laminectomy decompression involving two or more segments. In most cases of decompression, it is necessary to perform fixation and fusion with pedicle screws. Fusion is mainly indicated for patients with complicated lumbar intervertebral instability and spondylolisthesis who have to undergo multi‐segment laminectomy, and those with lumbar kyphosis, severe lumbar scoliosis or complicated rotatory subluxation and lateral listhesis.

Gupta believes that a combination of anterior‐posterior approaches offers a better surgical outcome for patients with severe coronary and sagittal imbalance and significant lumbar kyphosis ¹⁵ . But there are also researchers who think that there is no significant difference in results between a combination of anterior‐posterior approaches and the simple posterior approach ¹⁶ . As a combination of anterior‐posterior approaches is more traumatic, it should be used prudently in aged patients who have a lower tolerance for surgery.

While simple decompression is indicated for a small number of patients, posterior fixation and fusion is necessary for most. The following factors should be considered in selecting the extent of surgery: (1) segments of nerve root compression; (2) segments of lumbar canal stenosis, intervertebral disc protrusion, spondylolisthesis, and instable vertebral arrangement; (3) characteristics of lumbar scoliosis, including the presence or absence of complicated rotatory subluxation and lateral listhesis; and (4) other factors, including patient's age, general condition, the presence or absence of other complicated problems, and economic status.

Bridwell pointed out that fusion should be performed between the neutral and stable vertebrae, and that the fusion area should include the related rotatory subluxation and lateral listhesis vertebrae ¹⁷ . In most cases of DLS, the lower end of the deformity is usually below the L3 or L4 segment and therefore distal fusion should not be limited to L3 or L4, but this does not mean that fusion should involve the sacrum. In patients with mild degeneration of L5S1 intervertebral discs without L5S1 spondylolisthesis or degenerative declination, the L5S1 segment should be stable and therefore fusion can be terminated at the L5 segment. In patients with significant declination of the L5 vertebra (angle with the horizontal line >20°), L5S1 intervertebral disc protrusion, L5S1 spondylolisthesis or lateral listhesis greater than 5 mm, fusion should reach the sacrum. The scope of fusion should first of all include the segment that needs laminectomy for decompression, and then the unstable segment(s). However in some cases the segment that requires laminectomy for decompression and the unstable segments are not contiguous, and therefore fusion of the segments should be justified according to individual circumstances.

Whether DLS requires orthopaedic therapy remains controversial. Simmons maintained that correction of deformity should be achieved by surgical treatment of DLS for the sake of restoring the normal configuration of the spine ¹⁸ . But most researchers believe that the main purpose of surgical treatment for DLS is to stabilize the spinal column rather than correct the scoliosis, so they do not favor excess correction on the coronary surface ¹⁹ . In our series, the occurrence of ASD in patients receiving short segment fixation was higher than that in those receiving long segment fixation. In addition, the Cobb angle was smaller in the former group. If short fusion were used in patients with greater Cobb angle, the occurrence of ASD would be even higher. But this does not mean that long fusion is a must in patients with DLS. The formation of DLS is a long‐term process.

Excess correction of the scoliotic spinal column would lead to new biodynamic disturbance of the lumbar spine, especially in patients whose Cobb angle is large and who have significant zygapophysial joints hyperplasia with anterior vertebral osteophyte formation. As a result, the internal fixer would bear greater stress causing pseudoarthrosis or long‐term rupture of the pedicle crews. Marchesi and Aebi reported that the occurrence of pseudoarthrosis was greater than 4% if correction of the scoliotic degree was greater than 50% overall ²⁰ . In our series, scoliotic correction was performed in early fixation, and late stage orthopaedic treatment for the sake of correcting the deformed appearance was performed only in patients whose Cobb angle was greater than 30°. Such orthopaedic treatment was not carried out in patients whose Cobb angle was smaller than 30°. The overall clinical outcome proved to be satisfactory. Especially for patients in very poor condition, we prefer to significantly relieve the symptoms through limited decompression and fusion surgery rather than performing multi‐segmental, greatly traumatic surgery. This approach is helpful in reducing surgical risk and promoting postoperative recovery. Although there may be some residual clinical symptoms in these patients and the lumbar scoliosis Cobb and physiologic lordosis angles are not greatly improved, short fusion with limited decompression might still be a better option for these patients than multi‐segmental orthopaedic surgery.

Conclusions

DLS is often complicated by other lumbar degenerative disorders, presenting with certain distinctive imaging and clinical characteristics. The main goal of surgical treatment for DLS is to relieve nerve root compression and stabilize the spinal column. Correction of the deformed appearance should be a secondary consideration in elderly patients. Although long segment fixation is beneficial for deformity correction, it is more traumatic and more costly. Excess correction may increase the risk of long‐term pseudoarthrosis and internal fixation failure, while short fusion may increase the long‐term occurrence of ASD. Simple decompression may relieve short‐term symptoms but is unable to prevent DLS from progressing, so that some patients may need a secondary operation. To achieve the best therapeutic outcome, surgical treatment should be individualized according to the patient's age, general condition, economic status, severity of deformity, clinical symptoms, characteristics of other co‐existing lumbar degenerative disorders, and personal requirements.