Introduction

Degenerative lumbar scoliosis (DLS) is lumbar scoliosis that is secondary to degenerative lumbar disc, bone, and joint changes, which can develop after a person reaches full skeletal maturity [1]. DLS is a common disease among the elderly over the age of 50 [1]. Back pain and nerve compression symptoms are the main clinical complaints. Different from idiopathic scoliosis, congenital scoliosis, and neuromuscular scoliosis, DLS is often accompanied by lateral lumbar displacement and rotatory dislocation, as well as complications of other degenerative lumbar diseases. Nonsurgical treatment is generally recommended and effective for relieving pain and restoring normal activity in DLS patients. However, the treatment of severe DLS is relatively more complex than that of general degenerative lumbar diseases. Surgery is mainly indicated for severe back pain and/or progressive neurological symptoms refractory to nonsurgical treatment [1, 2].

Decompression surgery is essential for relieving the symptoms of neurogenic claudication [3]. Decompression can be used alone (isolated decompression), but most surgeons recommend decompression at the time of spinal fusion, including short fusion and long fusion [4, 5]. Spinal fusion is the most widely performed surgery for a variety of degenerative spine disorders including scoliosis, spondylolisthesis, spinal instability, deformity, and spinal stenosis [6]. During spinal fusion, bone is grafted to the vertebrae to form a solid bone mass during the healing process, and then the vertebral column becomes rigid. Thus, spinal fusion prevents worsening of the curve, but reduces some spinal movement [7]. The intercorporal implantation of cage devices has been used to avoid the high rate of pseudarthrosis produced during traditional spinal fusion [8]. The cage can be loaded with bone graft materials. Spinal fusion can be performed from the anterior aspect of the spine by entering the thoracic or abdominal cavities, or more commonly, performed from the posterior. A combination of anterior and posterior procedure is used in more severe cases. However, anterior surgery has been associated with complications, such as vascular complications, retrograde ejaculation, postoperative colonic obstruction, and injury to the sympathetic chain [9]. Posterior surgery has been associated with dural tears, paraspinal muscle denervation, and neural complications [10]. The lateral fusion has been described as a minimally invasive procedure [11] and has the advantage of reducing complications associated with anterior and posterior surgery [12].

In 1962, Paul Harrington introduced a metal spinal system of instrumentation to straighten the spine and to hold it rigid during spinal fusion [13]. Modern spinal instrumentations attempt to address sagittal imbalance and rotational defects unresolved by the Harrington rod system. Recently, all-screw systems have become the gold-standard technique for adolescent idiopathic scoliosis. Pedicle screws have achieved better fixation of the vertebral column and exhibit better biomechanical properties [14]. The anterior instrumentation system has been demonstrated to be successful in restoring immediate post-operative stability and correcting post-traumatic deformities [15]. Posterior instrumentation alone indicated 76 % less axial stiffness compared to the intact spine [6]. Combined use of anterior and posterior instrumentations can increase stability and reduce mobility.

There are a variety of available surgical procedures for the treatment of patients with DLS. The aim of this study was to review the surgical procedures for DLS adopted during the last three decades with the main focus on the outcome and complications.

Methods

Identification of relevant studies

Electronic searching of PubMed and EMbase databases was conducted in February 2014 using the keywords, lumbar degenerative scoliosis and degenerative lumbar scoliosis. Then, the abstracts or full texts, if available, were reviewed one-by-one for the exclusion criteria. Reviews as well as research articles that were not in English, did not involve humans, and non-surgical studies were excluded. We only included studies with surgical treatment for degenerative lumbar scoliosis, number of cases greater than 1, and where surgical procedure can be accurately identified. Surgical studies for other types of degenerative scoliosis were also excluded (Fig. 1).

Fig. 1
figure 1

A flowchart of literature search strategy. A flowchart of the included and excluded studies in the current meta analysis

Full size image

Data extraction

From each publication, we extracted data on the number of cases, surgical procedure, follow-up, outcomes, and complications. The surgical procedures were classified into isolated decompression, short fusion—if the fusion can be clearly identified in 1–2 segments or if the mean number of segments fused was less than three, long fusion—if the fusion can be clearly identified in more than two segments or if the mean number of segments fused was greater than or equal to three, and mix fusion—if a study contained short and long fusion in one group without providing a mean number of segments fused, or other surgical procedure if neither decompression nor fusion was used. If a study included more than one surgical procedure to compare the differences between two or more surgery procedures, each individual procedure in the study was classified in this review and the number of cases was extracted.

Data analysis

The duration of follow-up, preoperative ODI (Oswestry Disability Index), ODI at final follow-up, change in ODI, Cobb angles, curve reduction, curve reduction as a percentage of the original curve, and the incidence of complications were calculated. A pooled analysis of the data recorded in the individual studies was undertaken. The incidence of complications was calculated as the total number of complications divided by the total number of patients in the reported studies.

Results

The literature search identified 365 articles from PubMed and 752 articles from EMbase using the keywords degenerative lumbar scoliosis. A total of 364 articles were identified after removing duplicates. After excluding 46 articles not written in English, 56 reviews or comments, 77 articles of non-surgical studies, 121 articles of non-DLS or non-surgery studies for DLS, and 15 others including two studies reporting previous results, and two studies that reported one case, 45 full texts were included in this study (Fig. 1). The 45 studies were classified into four main surgical categories for analysis: isolated decompression, short fusion, long fusion, and mixed fusion or other surgical procedure. If multiple surgical types were used, the article will be included in each category for analysis. In this review, nine studies utilized isolated decompression with [4, 16, 17] or without [5, 18–22] instrumentation, 12 used short interbody fusion [3, 5, 21, 23–31], 17 used long interbody fusion [4, 5, 30, 32–45], 11 contained samples with mixed short and long interbody fusion [46–56], and one study used neither isolated decompression nor fusion [57] (Table 1). The interbody infusion techniques are composed of anterior lumbar interbody fusion (ALIF) [3, 47], posterior lumbar interbody fusion (PLIF) [33, 48, 49], direct lateral interbody fusion (DLIF) [50], lateral lumbar interbody fusion (LLIF) [24, 25], transforaminal lumbar interbody fusion (TLIF) [26, 27, 34, 35, 51], and extreme lateral interbody fusion (XLIF) [36]. The 12 articles for short fusion included ten articles using instrumentation [3, 5, 21, 23, 26–31] and two articles not using instrumentation [24, 25]. The 17 articles for long fusion included 16 articles using instrumentation and one study where instrumentation usage could not be determined [32]. The ten articles for mixed short and long fusion included nine articles using instrumentation and one study where instrumentation usage could not be determined [47] (Table 1).

Table 1 Surgical treatments of degenerative lumbar scoliosis
Full size table

Finally, nine articles for isolated decompression (352 patients), 12 for short fusion (614 patients), 17 for long fusion (517), and 10 for mixed short and long fusion studies, as well as one study for other surgical procedure (356 patients) were included for clinical outcome analysis (Table 2). The mean age varied from 64 to 77 years for isolated decompression, 53 to 70 years for short fusion, 54 to 66 years for long fusion, and 43 to 66 years for mixed short and long fusion or other surgical procedure. The follow-up duration ranged from 1 to 10 years, and the mean follow-up in most studies was 2 years. The data quality of the included studies was relatively low according to their MINORS scores. The study by Ploumis et al. [54] was the one with the highest quality. The ODI was used as a measure of functional improvement in 18 studies. At final follow-up, patients that underwent isolated decompression surgery showed 20–57 % improvement in mean ODI score, those that underwent short fusion showed 0–44 % improvement in mean ODI score, those that underwent long fusion showed 0–55.5 % improvement in mean ODI score, and those that underwent mixed short/long fusion or other surgical procedure showed 0–56.6 % improvement in mean ODI score. Pre- and post-operative Cobb angles were reported in 29 studies. Patients that underwent isolated decompression showed −7.9 % worsening to 79.9 % improvement in mean Cobb angle, those that underwent short fusion showed 0–52 % improvement in mean Cobb angle, those that underwent long fusion showed 0–72.3 % improvement in mean Cobb angle, and those that underwent short or long fusion showed 0–67.5 % improvement in mean Cobb angle. Surgical complication was reported alone or together with other outcomes in 26 studies with a range of 10–18 % in isolated decompression, 6.5–40 % in short fusion, 18–58.3 % in long fusion, and 20–32.6 % in mixed samples with short or long fusion. Pain release was evaluated as an outcome in ten studies with a range of 38.4–77.1 % decrease in VAS score in nine studies and 65.1 % increase in one study [49].

Table 2 Clinical outcome of surgical treatment of degenerative lumbar scoliosis
Full size table

Discussion

Although nonsurgical treatment is effective for relieving symptoms of DLS, surgery is commonly thought of as a crucial treatment option for severe DLS. A wide spectrum of surgical procedures has been developed with varying treatment efficacy and rate of complications. This meta analysis demonstrated that surgery is effective in treating DLS regardless of its high rate of complications and repeat procedures. Surgical treatment significantly improved the function of the spine (based on ODI), corrected deformities (based on Cobb angle), and released pain.

This review demonstrated that isolated decompression, short interbody fusion, and long interbody fusion are the three main surgical procedures for DLS treatment. Instrumentation is used to increase stability and reduce mobility of spine. In this review, the instrumentation is commonly used together with both short and long fusion (35 in 39 studies for fusion). Instrumentation was also used, but less commonly, for isolated decompression (3 in 9 studies). Isolated decompression (9 in 45 studies) is less common than interbody fusion (39 in 45 studies). This review demonstrated that isolated decompression without instrumentation was not a common choice of surgery for DLS (6 in 45 studies). In contrast, decompression is a basic procedure for fusion surgery [5].

The ODI is a valid and rigorous functional index for assessing spinal disorders [58] and 18 of the included 45 studies used ODI as an outcome of surgical treatment. However, the mean percentage decrease in ODI varied between studies, and no obvious differences was observed between studies with isolated decompression, short fusion, long fusion, and mixed fusion. Although the length of follow-up may be a critical factor that affects readout of the final ODI value, all surgical procedures significantly improved spine function. This suggests that surgery is an effective and reasonable treatment option for recovering the spine function of DLS patients. Deformity correction, as measured by the Cobb angle, is another critical index to evaluate the efficacy of surgery for DLS. Patients treated with long interbody fusion gained up to 72 % curve correction after surgery and patients treated with short fusion can gain up to 52 % curve correction. In contrast, there is a very large range in the percent improvement in curve correction in patients who received isolated decompression treatment with a range from 79.9 % correction to 7.9 % worsening. However, most studies utilizing these three surgical procedures reported effective correction of deformities. Moreover, pain release was evaluated as a single outcome or one of the outcomes in ten studies. Surgery decreased VAS score in nine studies with a range of 38.4–77.1 % release, although one study showed that pain is a complication of surgery. Decompression was not as effective in restoring spine function and correcting the Cobb angle compared to interbody fusion. Other than the one study that showed a 79.9 % reduction in Cobb angle [19], the three remaining studies showed an increase in Cobb angle [20–22]. The ODI data were only available in two studies with isolated decompression and showed a 20 % and 32 % increase, respectively.

Complications are an important concern in surgical treatment of DLS and a high incidence of surgical complications has been reported. The observed complications were classified into major complications and minor complications, such as repeated surgery, back pain, sensory deficit, vascular complications, global thoracic kyphosis, and stenosis. The percentage of major complications was 10–18 % in patients treated with isolated decompression, 6.5–40 % in those treated with short fusion, 18–58.3 % in those treated with long fusion, and 20–32.6 % in studies containing patients treated with mixed short and long fusion. However, the data quality assay yielded low MINORS scores [59] for these studies and complications were not fully observed and recorded. Therefore, comparison between surgical procedures becomes unreliable. Overall, the rate of complications was generally high. There is an urgent need to establish a standard to systemically observe and record the complications related to different surgical procedures. Most importantly, the complications related to surgery must be taken into account when deciding whether or not to proceed with surgery.

In general, most patients with LDS only needed conservative treatment. Surgery was only used for sacral nerve roots and cauda equina nerve decompression, restoring or rebuilding the mechanical balance of the lumbar spine, preventing the progression of the deformity, and in particular restoring physiological lumbar lordosis [60]. The surgical indications in the clinic include: (1) severe lower back pain and/or sciatica associated with lumbar scoliosis, which increases gradually and impacts daily life, and was not relieved with non-surgical treatment; (2) lumbar spinal stenosis and nerve dysfunction with progressive intermittent claudication; (3) progression of scoliosis with complication of acute and chronic cauda equina syndrome loss, as well as sagittal and coronal imbalance [60]. Among spinal fusion procedures, ALIF is a minimally invasive surgical approach with advantages of direct visualization of the intervertebral space, small incision size, and reduced tissue invasiveness [61], but with disadvantages of increased risk of intra-abdominal vascular and visceral injury [62]. In contrast, PLIF has advantages in avoiding intra-abdominal vascular and visceral injury, but is associated with an increased risk of damage to neural structures and epidural vein injury [63]. LLIF reduces the risk of surgical collateral damage associated with ALIF and PLIF approaches and brings increased post-operative vertebral column stability and improved alignment, but LLIF has a risk of neurological adverse sequelae [64]. TLIF is another minimally invasive approach to achieve circumferential fusion, which could reduce the extent of nerve root retraction associated with the PLIF procedure [65, 66]. DLIF is a minimally invasive lateral approach for placement of an interbody fusion device with advantages of not requiring nerve root retraction, reducing damage to the midline back muscles, and improving alignment of the spinal bones, but its use is limited to the lumbosacral (L5–S1) junction [11]. XLIF is a minimally invasive lateral transpsoas approach similar to DLIF, but cannot be used for the lumbosacral (L5–S1) junction [36]. In this review, TLIF was found to be the most commonly used interbody fusion and most effective in restoring spine function, and induced a 53–68 % improvement in correction of Cobb angle, although the complication rates were up to 49–60 %.

This meta-analysis made no recommendation for which specific type of surgery is the best and which surgical technique should be selected for different patients because the circumstances surrounding each patient are highly complex. This meta analysis included a study that found no significant differences in Roland–Morris score, Oswestry score, and patients’ satisfaction between patients who underwent isolated decompression, short fusion, and long fusion surgery [5]. One study compared the clinical outcome, recurrent leg pain, and complications between isolated decompression and decompression plus limited fusion, and revealed that recurrent leg pain occurred significantly more often in patients within 6 months post isolated decompression [21]. One study compared the clinical outcome, radiologic data, and complications between patients who underwent decompression with dynamic stabilization and patients who underwent long instrumented fusion [4]. Although decompression plus dynamic stabilization was less invasive than long instrumented fusion, no difference was found between the two techniques in terms of functional clinical outcomes [4]. It would be misleading to conclude that there were no significant differences between surgical techniques based on only 3 studies. Also, it is meaningless to make a simplistic conclusion on which surgical technique is best for DLS by using studies from different centers. This review demonstrated that the indications for surgery, surgical procedures, and outcome measures varied among the studies. In addition, the data quality of the included studies was not high.

However, despite a high rate of complications, this systematic review demonstrates that surgery is an effective and reasonable treatment intervention for DLS, providing improvements in spine function indicated by ODI and deformity correction as measured using the Cobb angle. A detailed evaluation of indications for surgery, standardization of surgical procedures, systemic evaluation of spinal function and deformity correction, and detailed records on complications should enhance the comparability and validity of future studies on this subject. Large scale, high quality studies with long term follow-up are needed to provide more reliable evidence for future evaluation.