Several surgical options are available for patients with degenerative lumbar scoliosis (DLS).^2,12,15,16 The recommended surgery for DLS with a coronal curve of less than 30° is posterior decompression and short fusion without correction of the curve.¹⁵ For a coronal curve of more than 30° that requires correction of global coronal balance, a long corrective fusion should be performed.¹⁶ The classification of adult spinal deformity (Schwab adult deformity classification) is well established, easy to use, and consistent with the treatment of adult scoliosis.¹² Some surgeons recommend performing fusion with instrumentation to prevent curve collapse.^4,18 However, the surgical treatment of DLS with stenosis in elderly patients is often challenging because these patients frequently have medical comorbidities, which are associated with higher complication rates.¹² DLS includes a diversity of symptoms, such as back pain, claudication, and neurological deficits. The management of DLS is characterized by significant variability because of the absence of a clear and evidence-based approach to treatment. The appropriate indications for the optimal treatment of DLS have yet to be defined. We hypothesized that the outcome of decompression surgery for mild DLS, based on our definition by radiological measurements of instability, would be as good as decompression with limited fusion. We prospectively analyzed the clinical outcomes of surgical treatment for patients with mild DLS. All patients underwent decompression surgery with or without limited segmental fusion, depending on radiological parameters for segmental spinal instability determined by preoperative coronal and lateral radiographs.

The purpose of this study was to compare the clinical and radiological results of decompression alone versus decompression with short fusion for mild DLS and to assess the usefulness of radiological parameters for surgical decision-making in these patients.

Methods

This study was approved by the ethics research board of the Yokohama City University School of Medicine. All patients provided informed consent for their participation in the study.

From a total of 298 patients in whom we prospectively planned surgical treatment for degenerative lumbar disease between September 2005 and March 2013 in our hospital, we chose patients with lumbar canal stenosis and mild DLS for this study. We excluded 223 patients without scoliosis or whose Cobb angle was less than 10°, as well as 11 patients whose Cobb angle was more than 25°. Sixty-four patients whose Cobb angle ranged from 10° to 25° were enrolled in this study (Fig. 1), and we were able to monitor 57 patients (89%) for a minimum of 2 years postoperatively. The surgical modality was determined based on the degree of spinal instability in the radiological findings. Radiological evaluation was performed based on functional lateral radiographs in a lateral decubitus position and anteroposterior lumbar radiographs in a standing position. A lateral view was obtained in the lateral decubitus position because maximal-effort flexion-extension radiographs were obtained in the patient without pelvic restraint.¹

Patient sampling and selection. Flowchart illustrating the enrollment of patients into the study. Patients with degenerative lumbar scoliosis (Cobb angle 10°–25°) with a minimum follow-up of 2 years are included and divided into 2 surgical groups. The patients in the decompression group (n = 25) and the patients in the fusion group (n = 32) are then analyzed.

The radiological parameters were measured manually on preoperative radiographs using digital calipers (Fig. 2). The dynamic parameters included segmental lordosis during lumbar extension, anterior translation during flexion, posterior translation during extension, and disc wedging. Our definition of radiological intervertebral instability was based on the report by Aota et al.,¹ which established 5 criteria for intervertebral instability: 1) 15° or greater segmental range of motion, 2) 3 mm or greater anterior translation, 3) 3 mm or greater posterior translation in the sagittal plane, 4) 5° or greater disc wedging, and 5) 3 mm or greater lateral translation in the coronal plane (Fig. 2).¹

Illustrations representative of radiographs used for evaluating spinal instability. A: Functional lateral radiographs in flexion and extension are obtained in the lateral decubitus position; measurements of segmental range of motion in the sagittal plane (a), anterior translation with flexion (b), and posterior translation with extension (c). B: Anteroposterior radiographs in a standing position; measurements of disc wedging (d) and lateral translation (e).

All patients underwent neurological decompression. Patients who were 75 years or older whose spinal column met 3 or more of the 5 parameters and patients younger than 75 years whose spinal column met 2 or more of the parameters also underwent limited segmental fusion. Using these criteria, 25 patients underwent decompression alone and 32 patients underwent decompression and fusion.

Results

Radiological intervertebral instability was measured for 85 segments in 57 patients. In the coronal plane, disc wedging was most common and seen in 61 segments (72%), followed by lateral translation in 52 segments (61%). In the sagittal plane, rotational mobility was seen in 34 segments (40%), followed by anterior translation in 28 (33%), and posterior translation in 7 (8%).

The mean age of the patients who underwent decompression alone was 73 ± 7.1 years (Table 2). There were 14 men and 11 women in the decompression group. The mean ± SD age of the patients who underwent decompression and fusion was 68 ± 9.0 years. There were 12 men and 20 women in the fusion group.

The mean preoperative JOA scores for the decompression group and fusion group were 5.9 ± 1.6 and 7.2 ± 2.0, respectively (Fig. 3). The JOA scores improved significantly to 10 ± 2.8 and 11.3 ± 2.8, respectively, at final follow-up. The net change in both groups after surgery did not differ significantly. The recovery rates were 45% and 53%, respectively, and there was no statistically significant difference between the groups.

Mean JOA scores in the decompression and fusion groups (preoperatively and at final follow-up). The mean JOA scores show significant improvement postoperatively in both groups. Data are presented as the mean ± SD. *p < 0.05. Pts = points.

The preoperative Cobb angles of the decompression group and fusion group were 14° ± 2.9° and 14.8° ± 4.0°, respectively (Fig. 4). The postoperative Cobb angles were 13.2° ± 4.7° and 8.5° ± 6.1°, respectively, and the Cobb angles at the final follow-up were 14.3° ± 6.4° and 10.0° ± 8.5°, respectively. The postoperative Cobb angle of the decompression group did not progress significantly during the follow-up period. Complications, such as nerve injury, infection, or hematoma, were not observed. Two patients in the decompression group underwent a second operation. One patient developed anterior translation and rotatory instability of the decompressed level, the L3–4 level on radiographic evaluation, and complained of lumbar back pain and claudication 6 months after decompression surgery. The patient underwent TLIF of the affected level 1 year after the decompression surgery. Another patient developed intervertebral disc protrusion at the decompressed level without radiological instability. The patient underwent discectomy 1 year after the decompression surgery.

Mean Cobb angle in the decompression and fusion groups (preoperatively, postoperatively, and at final follow-up). This graph shows that the postoperative Cobb angle in the decompression group does not progress significantly during the follow-up period. The Cobb angle in the fusion group decreases significantly immediately after the surgery and does not progress significantly during the follow-up period. There is no significant postoperative progression of the Cobb angle in either group. Data are presented as the mean ± SD. *p < 0.05; **not significant.

The Cobb angle in the fusion group decreased significantly immediately after surgery and did not progress until final follow-up. The Cobb angle in the decompression group did not change immediately after surgery and did not progress during the follow-up period. Lumbar lordosis in both groups did not change significantly after surgery (Fig. 5).

Mean lumbar lordosis in the decompression and fusion groups (preoperatively, postoperatively, and at final follow-up). There is no significant change in lumbar lordosis in either group after surgery, nor is there a statistical difference in lordosis between the groups. Data are presented as the mean ± SD. **not significant.

In the fusion group, 2 patients developed a deep infection in the early postoperative phase and underwent surgical debridement. One patient who underwent TLIF from L3 to L5 developed lower adjacent-segment degeneration and complained of low-back pain and claudication. The patient underwent additional fusion surgery at the L5–S1 level 1 year after the first operation. One patient who underwent TLIF from L3 to L5 developed upper adjacent-segment degeneration and complained of low-back pain. The patient underwent additional fusion surgery at the L2–3 level 1 year after the first operation.

Discussion

A diagnosis of DLS with lumbar spinal stenosis makes surgical decision-making difficult because DLS can exhibit various neurological symptoms and complex coronal and sagittal findings on radiographs. Surgical options for mild DLS include decompression alone or decompression and limited fusion within the deformity. Surgical realignment of the deformity is necessary for severe DLS. Several factors, including instability, must be considered when choosing the optimal surgical method. In this study, we focused on coronal curve progression after surgical treatment of mild DLS using our radiological instability criteria.

Simmons suggested that the neural elements as well as the stability and alignment of the vertebral column are important in deciding surgical treatment in patients with spinal stenosis associated with DLS.¹⁶ Neglect of either of these components will lead to inadequate results and less than ideal clinical outcomes. However, the optimal surgical treatment is unclear. Postacchini and Gupta reported that decompression surgery alone can be performed in patients with minor curves and no back pain.^7,14 Some authors have suggested that simple decompression is sufficient, provided that there is no significant kyphosis, olisthesis, or back pain because surgery can lead to further collapse, instability, continued back pain, and exacerbation of neural symptoms.^16,18 Daubs et al. reported that decompression with limited fusion for mild DLS prevents early return of stenotic symptoms compared with decompression alone and recommended limited fusion for mild DLS.⁵ Liu et al. reported that it is preferable to perform fixation and fusion in conjunction with decompressive laminectomy when 2 or more segments are involved, and that, although simple decompression is indicated for a small number of patients, posterior fixation and fusion is necessary for most.¹¹ Frazier et al. demonstrated a clear association between preoperative spinal deformity and poor outcomes following decompression surgery, regardless of whether concurrent fusion was performed.⁶ Many authors have suggested that short or long fusion is necessary for DLS after decompression. On the other hand, in a study by Hosogane et al., in which patients with DLS were retrospectively divided into 2 surgical decompression groups depending on postoperative curve progression, there were no significant differences in revision rates between the groups regardless of curve progression, and none of the patients required arthrodesis due to curve progression.⁹

Chen et al. reported that decompression alone is appropriate for patients with moderate stenosis affecting up to 2 spinal levels and smaller curves, provided that there is no evidence of progression or sagittal imbalance, while decompression with limited fusion is appropriate when patients have moderate or severe stenosis affecting 3 or more levels, smaller curves, no progression of deformity, no imbalance, and no more than moderate risk factors for suboptimal outcomes.³ We hypothesized that further collapse, instability, and curve progression would not occur in patients in the decompression group, whose surgical treatment was determined based on our radiological parameters. In this study, the average lumbar curve progression in the decompression group was 1.2°. No patient in this group demonstrated Cobb angle progression greater than 10°, and the Cobb angle in the decompression group did not progress significantly after surgery.

In previous reports, reoperation rates after decompression surgery for spinal stenosis have varied from 5% to 23%. Hansraj et al. reported that only 5% of patients required reoperation during a 5-year follow-up period, and Katz et al. reported that 23% of patients required revision surgery 7 to 10 years after decompression for lumbar canal stenosis.^8,10 In the decompression group, 2 patients underwent a second surgery, although only 1 patient (4%) underwent a second surgery with instrumentation due to instability of the previously decompressed level. Most patients did not show further collapse or instability. The patients who underwent a lumbar spinous process–splitting technique versus open fenestration showed the same net change in JOA score and an equally maintained Cobb angle after surgery. Early-phase complications did not occur in the patients in this group. In the fusion group, 2 patients developed surgical site infections and underwent surgical debridement in the early phase. Two patients exhibited adjacent-segment degeneration and underwent additional surgeries. The complication rate in the decompression group was lower than that in the fusion group. The high infection rate and adjacent-segment degeneration in the fusion group might be due to the long operative time and use of instrumentation.

Transfeldt et al. compared the results of different surgical methods in patients who underwent decompression alone, decompression with limited fusion, and decompression with full curve fusion.¹⁷ They reported that the group receiving decompression alone had the lowest rate of complications and the best Oswestry Disability Index results, whereas the satisfaction was lowest in the decompression-only group.

In this study, the JOA score improved in both groups, and the recovery rate based on the JOA score showed no difference between the groups. The patients assigned to the decompression group using our radiological parameters did well after surgery from the standpoint of radiological and clinical progression. The results of this study indicate that the evaluation of spinal instability by radiological parameters is appropriate for surgical decision-making.

One limitation is that the study does not evaluate whether the curve progressed, or whether the clinical outcomes and risk of complications of decompression surgery alone were appropriate for patients with positive radiological instability criteria and mild DLS. Furthermore, there is limited information on complications due to the small sample size, and further studies are needed to discuss whether fusion is required or decompression alone is appropriate for radiologically unstable segments.

Conclusions

This prospective study of patients with mild DLS treated by different surgical methods showed good results for both methods. The results of this study indicate that the radiological evaluation of spinal instability is appropriate for surgical decision-making.

Acknowledgments

We are grateful to Editage for the English language review.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Masuda. Acquisition of data: Masuda. Analysis and interpretation of data: Masuda. Drafting the article: Masuda. Reviewed submitted version of manuscript: all authors.