Surgical treatment for lumbar spinal stenosis (LSS) is widely performed, and its effectiveness has been demonstrated in prospective randomized studies [12]. Expansive laminectomy is the most commonly used surgical procedure for LSS. The potential risks of conventional laminectomy include persistent back pain due to the extensive detachment of the paraspinal muscle, trunk extensor weakness, and iatrogenic spondylolisthesis [34567]. Less invasive approaches and methods have been developed to preserve the posterior elements, including the lamina, spinous processes, interspinous ligaments, facet joint, and to decrease paraspinal muscle damage. Lumbar spinous process-splitting laminectomy (LSPSL) was introduced in 2005 as a less invasive technique that preserves the paravertebral muscles and posterior elements and reduces muscle atrophy and postoperative wound pain for treatment of LSS [8]. This method can easily allow a wide field of view of the central canal and lateral recess while preserving the facet joint. The LSPSL technique has been performed since 2008 at our institution for the treatment of LSS, mild degenerative spondylolisthesis (DS), and mild degenerative lumbar scoliosis (DLS). This procedure may avoid iatrogenic lumbar instability more than conventional laminectomy. However, there have been few reports regarding the clinical outcomes and radiographic evaluations after LSPSL. The objective of this study was to assess the clinical and radiographic outcomes and identify the predictive factors associated with poor clinical outcomes after LSPSL.

A total of 52 consecutive patients (32 males, 20 females) met the inclusion criteria and had an average follow-up of 2.6±0.8 years (range, 2-4.5 years). The characteristics of the patients are shown in Table 1. The average age at the time of surgery was 72.1±7.1 years (range, 61-87 years). The average number of decompression levels was 2.5±0.1 (range, 1-5). The preoperative diagnosis consisted of LSS in 19 patients, DS in 19 patients, and DLS in 14 patients. The mean JOA score significantly increased from 14.6±3.7 to 23.2±3.4 (p<0.001) after the operation at the last followup (Table 2). The mean recovery rate was 60.1%±22.1% (range, 0%-100%). Intraoperative and postoperative complications associated with LSPSL included a dural tear in two patients. Three cases required reoperation caused by progression of slippage in one case, far-out syndrome in one case, and a synovial cyst in one case. A progression of slippage ≥5 mm was found in 8 of 24 patients (33%) in the DS group. A progression of curvature ≥5° was found in 5 of 14 patients (36%) in the DLS group. Among the 52 patients, 13 patients (25%) were classified into the poor clinical outcome group (recovery rate <50%). None of the patient factors, including age, gender, preoperative JOA score, preoperative distance of IC, motor weakness, urinary dysfunction, duration of symptoms, and multilevel decompression were independent predictors of poor clinical outcome (Table 3). There was no statistically significant difference in the follow-up periods between the two groups. We compared postoperative leg pain and back pain using subscales of the JOA score between the good and poor outcome groups. The postoperative back pain was significantly worse in the poor group than in the good group (1.7±0.1 vs. 2.3±0.9, respectively; p=0.195). In contrast, the postoperative leg pain score was similar between the poor and good outcome groups (2.5±0.1 vs. 2.7±0.8, respectively; p=0.001). Regarding the radiographic parameters, there was a trend towards smaller preoperative and postoperative LL in the poor outcome group relative to that of the good outcome group, although the difference was not significant (preoperative, 31.7° vs. 35.5°, p=0.249; postoperative, 26.6° vs. 32.8°, p=0.099, respectively). Higher rates of DLS patients were observed in the poor clinical outcome group than in the good outcome group (15% vs. 62%, p=0.003, respectively). The preoperative and postoperative Cobb angles were greater in the poor outcome group than in the good outcome group (preoperative: 4.8° vs. 10.7°, p=0.001; postoperative: 6.4° vs. 13.7°, p=0.002, respectively). However, there was no significant difference in the change in the Cobb angle (1.5° vs. 3.0°, p=0.135). These results indicate that pre-existing DLS was associated with poor clinical outcome, whereas progression of the Cobb angle did not have an effect on the clinical outcome. There was no difference in the percentage of DS patients between the good and poor outcome groups (46% vs. 46%, p=1.000). The preoperative percentage of slip was similar in both groups. However, there was a trend towards a progression of slip in the poor outcome group, although the difference was not significant. A progression of slippage was found in 8 of the 24 patients (33%) with DS. A progression of scoliosis was found in 5 of the 14 patients (36%) with DLS. The recovery rate was significantly lower in the DLS group than in the LSS and DS groups (DLS group, 46.9%; LSS group, 62.8%; DS group, 67.1%) (Fig. 3). There was no significant difference in the three groups (LSS, DS, DLS) with respect to age (70.6±7.3, 71.6±7.2, 75.0±6.4 years, respectively; p=0.204), sex (20%, 40%, 40%, respectively; p=0.100), number of decompression levels (2.5±0.9, 2.4±0.7, 2.6±0.9, respectively; p=0.642), and preoperative JOA score (14.0±3.7, 15.3±3.5, 14.6±3.7, respectively; p=0.527). These results suggested that only pre-existing DLS was significantly associated with poor clinical outcome.

Degenerative LSS is the most common cause of lumbar spine surgery in adults [10]. Recent prospective randomized studies with long-term follow-up have demonstrated that surgical treatment is superior to nonsurgical treatment for degenerative LSS [2]. In the current study, with a mean follow-up of 2.6 years, the overall mean JOA score significantly improved, and the recovery rate was 60% after LSPSL. The results of this study indicate that the clinical outcomes of LSPSL for LSS were favorable. Conventional laminectomy has been reported as having a potential risk of iatrogenic instability and back muscle damage [345]. Johnsson et al. [3] reported that 40% of patients showed progressive slippage postoperatively with conventional laminectomy. Less invasive decompression procedures, including microscopic or microendoscopic lumbar decompression via a unilateral approach, have been reported as giving satisfactory clinical outcomes for the treatment of LSS [1112131415]. However, these procedures are technically demanding, and the working space is limited, particularly in cases of upper lumbar lesions. LSPSL has been reported to reduce muscle damage and postoperative wound pain in the treatment of LSS [8161718]. The advantages of this procedure include obtaining a wider surgical working space with less invasive muscle damage and the possible prevention of excessive removal of the facet joint. In our study, the radiographic outcomes after LSPSL in the LSS and DS groups were equivalent to those of conventional laminectomy [4]. The progression of spillage after LSPSL did not influence clinical outcomes and reoperation. These results, which are based on an average follow-up of 2.6 years, suggest that LSPSL is an effective less invasive surgical treatment for patients with LSS or DS. LSPSL has been increasingly used as an alternative to lumbar laminectomy with some modifications [192021].

We found that patients with pre-existing DLS were associated with worse clinical outcomes after surgical treatment. Kelleher et al. [22] compared the clinical outcomes of minimally invasive decompression for LSS in patients with and without deformities. They concluded that DLS patients, particularly those with lateral listhesis, had a significantly higher rate of revision surgery (25%) than LSS patients (0%) and DS patients (4%). Toyoda et al. [14] reported 5-year and longer clinical outcomes in patients treated with microsurgical decompression for LSS, DS, and DLS, finding no significant differences in the postoperative JOA scores at the final follow-up. However, a higher rate of reoperation was found in the DLS group (25%) than in the LSS group (0%) and DS group (10%) [14]. Decompression with fusion surgery for patients with DLS is still controversial. A previous study suggested that decompression and fusion surgery prevent revision surgery better than decompression alone in the setting of mild degenerative scoliosis [2324]. A higher rate of recurrence of symptoms was observed in patients who underwent decompression alone (75%) than in those who underwent decompression and fusion (34%) during a 5-year follow-up [23]. Progression of scoliosis was found in 36% of patients with DLS, a percentage similar to that in all patients with DLS in Japan [24].

There were several limitations in this study, which include the relatively small sample size and retrospective data collection. In addition, the postoperative follow-up period may not have been long enough to assess the potential risk of iatrogenic instability. Recently, a prospective randomized study demonstrated that clinical outcomes and back pain were similar between LSPSL and conventional techniques [25]. A prospective randomized study with a larger sample size and a longer follow-up period would be required to thoroughly assess the clinical outcomes after LSPSL. Furthermore, we did not evaluate the preoperative global spinal alignment for most of these patients. In recent literature, sagittal global alignment and pelvic parameters were shown to correlate with healthrelated quality of life scores [26]. In the current study, there was a trend towards smaller preoperative and postoperative LL in the poor outcome group. There is some possibility that the global spinal alignment was worse in the poor outcome group than in the good outcome group and may have influenced the clinical outcomes.

The results of this study showed that the clinical and radiographic outcomes of LSPSL for LSS and DS were favorable. Progression of slippage and scoliosis were not shown to affect the clinical outcomes. However, progression of slippage was found in 33% of the DS group with average 2.6-year follow-up. A long-term follow-up study for DS after LSPSL is needed. Pre-existing DLS was a significant risk factor associated with poor clinical outcomes. Appropriate patient selection with consideration to decompression surgery may be important for DLS patients.