INTRODUCTION
Lumbar spine stenosis (LSS) can result in symptomatic compression of the neural elements, requiring surgical treatment if conservative management fails. Decompressive laminectomy is widely used to treat LSS. Although satisfactory surgical outcomes have been reported using this technique, instability following the procedure is one of the greatest concerns amongst surgeons as it may cause deterioration of symptoms33). The following factors are known to be essential for a good clinical outcome : sufficient dural and nerve root decompression, minimum resection of facet joints to maintain spinal stability, early surgical intervention before severe paralysis becomes apparent, postoperative application of hard orthotics for at least 3 months, and postoperative trunk muscle exercises9).
Minimally invasive surgery using microsurgical or endoscopic procedures has come to be more commonly used for the treatment of LSS over the last decade. The goal of these procedures is maximal preservation of structural components such as midline structures, facet joints, and paravertebral muscle to prevent postoperative instability. It has been hypothesized that these techniques yield better clinical outcomes by reducing tissue trauma and preserving spinal architecture, but only limited follow-up data exist to confirm this hypothesis22).
In the current study, we describe the outcomes after at least 3 years of follow-up of patients who underwent bilateral microdecompression by unilateral or bilateral laminotomy (BML) for degenerative LSS and we investigate factors that resulted in poor outcomes.
MATERIALS AND METHODS
Patients
Twenty-one patients were operated on by one author who were followed-up for a minimum of 3-years were included in this study. Twenty-eight patients underwent BML without fusion procedures by one author from 2007 through 2009. Seven patients were not followed-up because of inability to locate new addresses (4 cases), and death due to malignant tumor, heart disease, or stroke (3 cases). The diagnosis of LSS was based on clinical symptoms such as low back pain, leg pain, numbness during standing or walking, and intermittent claudication. In all patients, stenotic spinal lesions were confirmed by magnetic resonance imaging scans. Clinical indications for BML were leg pain and/or leg numbness inducing intermittent claudication (>6 weeks) rather than back pain. Furthermore, this procedure was preferentially considered over fusion surgery in older patients and those with comorbidities, regardless of persistent back pain. Radiologic indications for use of this surgical procedure were LSS without instability and foraminal stenosis excluding degenerative lumbar spondylolisthesis (≥3 mm on the lateral neutral radiograph), and degenerative lumbar scoliosis (Cobb's angle : ≥25°).
Operative technique
BML was modified from the previously reported unilateral approach method for bilateral decompression (ULBD)35,37). Laminotomy was performed on the contralateral side for lateral recess stenosis to preserve as much of the facet joints as possible. If there was bilateral lateral recess stenosis, laminotomy was performed on both sides. A paramedian or midline linear skin incision of approximately 3 cm was made. The latissimus dorsi and the multifidus muscle from one side were stripped from their attachments, the interlaminar space exposed, and a Caspar retractor applied. Under a surgical microscope, the upper and lower laminae were partially removed in the area of the ligamentum flavum insertion. The basal part of the spinous process of the caudal half of the cranial lamina and a small cranial portion of the caudal lamina were removed with a high speed drill. Then the contralateral lamina was undercut with a high-speed air drill leaving the ligamentum flavum in place as protection for the dural sac and the nerve root. Following sufficient resection of the bony segment, the ligamentum flavum was removed en bloc with a curette. With recognition of the inner aspect of the pedicle on the contralateral side, we confirmed adequate decompression of the contralateral side. If the ipsilateral lateral recess needed to be decompressed, we performed an additional contralateral laminotomy for adequate decompression of the ipsilateral side. Concomitant discectomy was not performed except in patients with apparent disc herniation, because preservation of the anterior stabilizer of the spinal column was considered desirable at laminectomy. The average surgical time was 50 minutes for one segmental decompression, and average intraoperative blood loss was less than 50 cc per one level. Patients were generally allowed to walk with a corset brace within 3 days of surgery, and corset brace use was recommended for 4-6 weeks. Rehabilitation was not recommended (Fig. 1).
Clinical evaluation
We used the Japanese Orthopedic Association scoring system for low back pain (JOA score) (Table 1) for clinical evaluation. Preoperative clinical evaluation data and JOA scores were obtained from medical charts, and postoperative final scores were obtained by data collection and telephone interviews. Patients were asked about current status, subjective improvement in symptoms, satisfaction with the surgical procedure and operative results, associated medical conditions, and current therapy. The modified grading system devised by Finneson and Cooper5) was used to assess outcomes (Table 2). Ratings of excellent, good, and fair were classified as successful operative results.
Radiographic evaluation
Radiographic examination was performed by one of the authors without knowledge of the clinical features of 19 patients for whom both pre- and postoperative radiographs were available. Evaluations were also performed independent of clinical assessment. Measurements to determine spondylolisthesis and the sagittal rotation angle were done in operated segments9). Disc heights were measured on a lateral radiograph with identification of the following points : a-anterior disc height, b-middle disc height, c-posterior disc height. Disc height was calculated as the mean of the anterior, middle, and posterior disc heights10).
Statistical analysis
SPSS version 12.0 statistical package (SPSS Inc., Chicago, IL, USA) was utilized for statistical analyses. Data are presented as means±standard deviations. p-values less than 0.05 were considered statistically significant. The paired t-test was used to analyze the significance of differences between the JOA score and radiological measurements pre-operatively and at the latest postoperative follow-up (>3 years), and Fisher's exact test was used to analyze the significance of clinical improvements from the preoperative period to the latest postoperative follow-up (>3 years).
RESULTS
Demographic data
Twenty-one patients (10 men, 11 women) aged 53-82 years (64.1±8.9 years) at index surgery were followed-up for a minimum of 3 years. The average follow-up period was 50.2±10.1 months (range, 36-69 months). Ten patients had primary symptoms of intermittent claudication with bilateral leg numbness and pain. Nine patients had persistent leg pain. Persistent low back pain continuing for more than one month was also found in 10 patients. Manual muscle testing identified apparent motor deficits at less than a fair level in seven patients, including one with complete drop foot (Table 3, 4). Sixteen patients (76.2%) underwent single level surgery, and five patients (23.8%) underwent two-level surgery. The level of surgery was L2-L3 in three patients, L3-L4 in eight patients, L4-L5 in 14 patients, and L5-S1 in one patient (Table 3). Comorbidity was high. Eighteen patients (85.7%) had multiple medical problems, including hypertension (9; 42.9%), diabetes mellitus (9; 42.9%), heart disease (3; 14.3%), chronic renal failure (2; 9.5%), and hypothyroidism (1; 4.8%). Only three patients (14.3%) had no evidence of a concomitant health problem (Table 3).
Clinical outcome
Two of 21 patients underwent reoperation. A 55-year-old male patient had a fusion operation due to foraminal stenosis on the same level 4 months after decompression surgery. A 57-year-old male patient had a fusion operation at the same level due to intractable low back pain 34 months after surgery.
Four patients had excellent outcomes (19.0%), four patients had good outcomes (19.0%), five patients had fair outcomes (24.0%), four patients had marginal outcomes (19.0%), and four patients had poor outcomes (19.0%) at evaluation on the latest follow-up. Thirteen patients (61.9%) had excellent to fair outcomes (Table 2).
The average preoperative JOA score (total points, 15) for all patients was 6.7±1.8 points. It improved to 10.9±1.6 points at 1 year after surgery, and then decreased to 9.2±2.0 points at the latest follow-up (>3 years) (p=0.001) (Fig. 2).
The number of patients who reported persistent low back pain continuing for more than one month increased slightly from 10 to 12 after surgery (p=0.74). In contrast, the number of patients with persistent leg pain decreased from nine to two (p=0.029). The number of patients with neurogenic claudication decreased from 10 before surgery to two after surgery (p=0.013). Seven patients with apparent motor deficits recovered well, but one patient did not show any improvement (p=0.042) (Table 4).
Radiological outcome
Preoperative anterior slippage ranged from -3 to 2 mm, with a mean of 0.6±0.7 mm. The anterior slippage at the latest follow-up ranged from 0 to 5 mm, with a mean of 1.1±1.4 mm. Asymptomatic progression (>3 mm) in anterior slippage after this surgical procedure was observed in one patient. There were no significant differences in progression of anterior slippage among these diseases (p=0.12). The preoperative sagittal angle was 9.6±2.3° and progressed slightly to 10.5±2.8° at the latest follow-up (p=0.23). Preoperative disc height was 11.2±2.5 mm and decreased slightly to 10.4±2.7 mm at the latest follow-up after surgery (p=0.29) (Table 5).
Surgical-related complications
During the follow-up period, two patients underwent subsequent lumbar surgery. Reoperations consisted of spinal fusion to treat foraminal stenosis and secondary lumbar instability inducing persistent low back pain. We observed two surgical-related complications (10.5%); superficial disturbance of wound healing (n=1) and deep wound infection (n=1). Both of these patients recovered without sequelae.
DISCUSSION
Bilateral microdecompression by laminotomy
Unilateral or bilateral laminotomy for decompression of LSS was introduced in 1981 by Getty et al.7) as a less invasive surgical option. Microsurgical bilateral decompression via a unilateral approach was first described by Poletti25). More recently, ULBD initially described by Young et al.37) in 1988, was modified by McCulloch18) and has been successfully used and subsequently modified as a minimally invasive unilateral approach for bilateral decompression17,23,25,28-30,35). Nakanishi et al.21) reported microsurgical lumbar flavectomy that preserves facet joints in cases of lumbar degenerative spondylolisthesis. Comparable results (improvement rates of 59% to 91% after laminotomy) have been reported by other authors for a mean follow-up period of up to 5.5 years1,7,14,15,18,20,28,37).
Laminotomy vs. laminectomy
Laminectomy entails removal of the entire spinous process, the laminae, and the adjacent supra- and interspinous ligaments in addition to the ligamentum flavum and as much as half of each facet complex. Laminotomy involves partial removal of bone from the inferior aspect of one lamina and the superior aspect of the subjacent lamina, excision of the ligamentum flavum, and as much as one-half of the facet, with preservation of the spinous process and supra- and interspinous ligaments31). Only a few series have directly compared laminotomy with laminectomy26,27,31). With success rates of 50% and 58% for laminotomy versus laminectomy, respectively, Thomas et al.31) reported no significant differences in clinical outcome and postoperative instability between these different techniques. However, with 14 and 12 patients in each group and a mean follow-up period of 3.1 years, the patient population was small, and the follow-up period was short. Postacchini et al.26) compared the clinical results and postoperative stability of 41 patients treated by laminectomy with those of 26 patients who underwent bilateral laminotomy after a mean follow-up period of 3.7 years. Spinal instability was more frequent after laminectomy than after laminotomy, whereas the clinical results were comparable (78% vs. 81% good outcomes, respectively). Rompe et al.27) evaluated the results of 117 patients treated either by laminotomy, laminectomy alone, or laminectomy plus instrumented fusion (mean follow-up interval, 8 years). The rate of good outcomes after laminotomy (36%) was better than that after laminectomy alone (31%). These findings indicate that compared with laminectomy, laminotomy can adequately decompress LSS, and that laminectomy may be too aggressive for adequate preservation of spinal stability. Johnsson et al.11) reported postoperative slippage in 18 of 45 patients (40%) who underwent a laminectomy. In their study, 65% of patients with degenerative spondylolisthesis exhibited a high risk of further slippage after operation and 20% of patients with LSS exhibited additional slippage. Mardjetko et al.16) reviewed the incidence of progression of slippage after decompression and reported it to be 31%. In some cases, spinal fusion combined with adequate decompression is therefore required. To sum up these reports, laminotomy showed results in superior than laminectomy in clinical and radiological outcomes to laminectomy.
Laminotomy and postoperative instability
A randomized trial and a study with alternating treatment assignments revealed better outcomes after decompression plus fusion than conventional decompression alone6,8). However, these studies evaluated patients with degenerative spondylolisthesis with LSS. Surgical decompression may lead to increased rates of spondylolisthesis postoperatively. In patients without preoperative spondylolisthesis, the rate of progressive postoperative spondylolisthesis can be as high as 31%, and in patients with preoperative spondylolisthesis, the incidence of progressive slipping after surgery is even higher, ranging from 30 to 100%12,22). Some authors have suggested that patients treated with spinal fusion have a higher likelihood of greater blood loss, a longer operative time, and a higher rate of complications, and thus require more extensive revision surgery, than patients treated by surgical decompression alone33). Thus, decompression surgery without fusion with clinical outcomes equivalent to those of conventional decompression surgery, but less postoperative instability, is desirable33). Recently, Oertel et al.22) demonstrated that in patients treated by ULBD, no postoperative increase of instability should be expected, even in cases with pre-operative fixed spondylolisthesis. Müslüman et al.19) reported that neutral and dynamic slip percentages did not change significantly in patients with Grade I degenerative spondylolisthesis after ULBD. These authors suggested spondylolisthesis in itself is not an indication for fusion. Costa et al.4) reported that 87.9% of 374 patients experienced clinical benefits and only 8% of patients suffered from segmental instability at the treated level after a mean follow-up duration of 30.3 months (range, 16-53 months). In the present study, no patient showed instability in the preoperative radiological study and with the exception of two patients that underwent reoperation, only one patient showed evidence of asymptomatic spondylolisthesis 3 years after surgery. We found little progression of spinal slippage after BML over a minimum of 3-years of follow-up. Less radiographic changes were observed after this procedure than reported for LSS after decompression.
Outcomes
Tsai et al.34) reported clinical results after microscopic laminotomy based on a retrospective study. After a mean follow-up period of 2.25 years, 84% of 50 patients showed good postoperative results. McCulloch18) developed unilateral laminotomy for bilateral ligamentectomy and reported a satisfactory outcome in 76% of 22 patients with degenerative LSS. Weiner et al.35) investigated outcomes of limited osteoplastic laminectomy with spinous process osteotomy preserving the midline osseo-ligamentous structures, and found high rates of satisfaction in 87% of patients at the 9-month follow-up. Thomé et al.32) reported that clinical outcomes after ULBD were equivalent to those obtained with conventional laminectomy after a minimum follow-up of 12 months. Oertel et al.22) reported that 85.3% of 102 patients had excellent to fair surgical results over 4 to 10 years (mean, 5.6 years) of follow-up, with a reoperation rate of 11.8%. Cavuşoğlu et al.3) reported good results in 68% of patients at 4 years and noted that reoperation was not required for recurrent spinal stenosis at the same segments within 4 to 7 years. Toyoda et al.33) evaluated clinical outcome and radiographic changes for LSS including degenerative spondylolisthesis and scoliosis over a minimum 5-year follow-up. They reported that ULBD yielded satisfactory surgical outcomes.
We showed in the present study that 61.9% of 21 patients had excellent to fair results 3 years after surgery, while two patients (9.5%) required reoperation. Persistent low back pain at the latest follow-up in some patients resulted in poor outcome scores, even though there were significant clinical improvements in the other clinical symptoms of persistent leg pain, neurogenic claudication, and apparent motor weakness. In comparison with interbody fusion, a major concern after decompression is the recurrence of low back pain because of secondary instability. However, only a few of our patients showed vertebral hypermobility or a significant increase in spondylolisthesis after surgery. Spine-related adverse events such as osteoporosis, postsurgical clinical instability, and extensive multilevel surgery are known to adversely affect medium- and long-term outcomes24). We demonstrated that persistent back pain was not related to radiological instability. We hypothesize that the causes of persistent low back pain are aging or aggravated osteoporosis and degeneration of disk and facet joints post-operatively. Another possible cause may be that BML was preferentially considered over fusion in older patients and/or those with comorbidities.
Complications
Based on a multicentric study of 270 patients with degenerative lumbar stenosis who were stabilized with or without instrumentation, Weinstein et al.36) reported a dural tear incidence of 10% and a reoperation rate of 12% in the second year as well as a death rate of 1% and a would infection incidence of 3.5%. Oertel et al.22) reported that the incidence of complications related to unilateral laminotomy was 9.8%. These complications included incidental durotomy, disturbance of wound healing, and deep wound infection. In the present study, there were two surgical-related complications (10.5%); superficial disturbance of wound healing and deep wound infection without mortality. Despite the high preoperative comorbidity, we did not experience complications related to the coexisting medical diseases which could be interpreted as an advantage of microdecompression by laminotomy.
In our present series, two patients (9.5%) underwent subsequent lumbar surgery. Reoperations consisted of spinal fusion to treat foraminal stenosis and secondary lumbar instability inducing persistent low back pain. Oertel et al.22) reported a reoperation rate of 11.8% due to complication, restenosis, and spinal instability. Toyoda et al.33) reported a reoperationrate of 0% for LSS, 10% for degenerative spondylolisthesis, and 25% for degenerative scoliosis. The mean rate of reoperation was 7.0%. Katz et al.13) reported that 23% of patients who underwent conventional decompressive surgery had a reoperation over a 7 to 10 year follow-up period. Iguchi et al.9) reported that three of 37 patients (8.1%) who underwent decompression alone with a longer than 10-year follow-up required additional surgery because of disc herniation at segments subjected to laminectomy. Atlas et al.2) found that 23% of patients who underwent decompression alone required at least 1 additional lumbar spine operation within 10 years after their original procedure. Our reoperation rate is consistent with those reported in previous studies.
Study limitations
This study has several limitations. First, this was a retrospective study without any control group. Second, indications for the surgical procedure were limited to patients without degenerative spondylolisthesis and scoliosis. Third, fewer patients were evaluated in this study than in some of the previous studies. Fourth, BML was preferentially considered over fusion surgery in older patients or those with comorbidities, even in patients with persistent back pain. This bias may have affected the outcomes of our study.