Materials and Methods

Institutional review board approval was obtained for all aspects of the study. Patients were identified by conducting a comprehensive search of hospital medical and surgical databases, and those undergoing spinous process–splitting laminoplasty for lumbar spinal stenosis performed by the senior author (A.N.N.) at the authors' institution were included. Between 2010 and 2012, 46 patients underwent lumbar spinous process–splitting laminoplasty at the authors' institution. The medical records for all potential patients were reviewed. Thirty-seven (80%) patients were followed by the senior author at the outpatient spinal clinic for a minimum of 1 year, and these patients composed the study group.

Surgical Technique

Templating, which is performed preoperatively, consists of identifying the anteroposterior axis of the spinous process and measuring its shortest depth to understand the spine anatomy for each patient and avoid overpenetration of the osteotomy into the spinal canal. The maximum anteroposterior depth of the spinous process above the spinous process junction with the dorsal lamina is chosen as the depth of the osteotomy. Using fluoroscopy, the spinous process to be osteotomized is identified and the skin is marked (Figure 1). The spinous process to be split is always superior to the interspace to be decompressed (eg, for an L3–4 decompression, the L3 spinous process would be split).

Figure 1: Needle localization and marking of the spinous process. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

A midline skin incision is made, and the subcutaneous tissue is dissected until the thoracolumbar fascia is reached (Figure 2). A radiograph in the anteroposterior and lateral position is performed as a secondary pause with a spinal needle in the spinous process(es) to verify the level(s) to be split. The dorsal aspect of the cortical bone of the spinous process is then decorticated using a 2-mm round burr, and a #15 blade is used to incise the surpraspinous and interspinous ligaments in line with the midsubstance of their fibers above and below the level to be split (Figure 3). The spinous process is then divided in 2 halves with a thin osteotome and detached from the spinous process lamina junction at the predetermined depth (Figure 4). All muscular attachments remain intact and attached to the lateral walls of each spinous process. A working space is made available by inserting a self-retaining retractor between the split halves of the spinous process until the medial edges of the facet joints are identified laterally (Figure 5).

Figure 2: Subcutaneous dissection and identification of the thoracodorsal fascia. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Figure 3: Splitting of the spinous process using a burr and sectioning of the supra and interspinous ligaments. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Figure 4: Osteotomy of the base of the spinous process using a curved osteotome. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Figure 5: Insertion of a self-retaining retractor to create the working window and planned bony resection. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Decompression is then performed under microscope or loupe magnification by removing the distal aspect of the superior lamina as well as the proximal aspect of the inferior lamina. The ligamentum flavum is resected, and lateral recess decompression is performed by undercutting the facet on both sides through a traditional midline approach (Figure 6). Bleeding is controlled with bipolar coagulation and hemostatic agents. Once decompression is achieved, a deep drain is placed, and the split spinous processes are approximated using transosseous restorable sutures (Figure 7). The supraspinous and interspinous ligaments are repaired with interrupted sutures.

Figure 6: Preoperative and postoperative appearance after central and lateral recess decompression. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Figure 7: Transosseous sutures to repair the spinous process split and repair of the supra and infraspinous ligaments. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Postoperatively, all patients are placed into a soft abdominal binder for comfort and activities are restricted to walking without trunk flexion/extension for 6 weeks. The procedure can be performed at multiple segments using a similar approach (Figure 8). In cases of fusion, fixation can be achieved through the same or paramedian incisions using the interval between the multifidus and the other paraspinal muscles.

Figure 8: Technique extended to a multilevel decompression. (Used with permission of the Mayo Foundation for Medical Education and Research, all rights reserved.)

Results

At a minimum 1-year follow-up (mean, 1.3 years; range, 1–2.4 years), patients experienced substantial improvements in pain and overall health and function after lumbar decompression with the use of the spinous process–splitting laminoplasty technique. There were statistically significant improvements in all collected patient-reported outcome scores. Mean VAS for back pain significantly improved from 6.8±2.3 points to 3±3 points at most recent follow-up (P<.0001) (n=37). Likewise, mean VAS for leg pain significantly improved from 6.4±2.6 points to 3±3 points at most recent follow-up (P<.0001) (n=37). In addition, mean ODI significantly decreased by 17.5±19.1 points (P<.0001) (n=35). At most recent follow-up, the ODI was graded as minimal disability (0–20 points) for 17 (46%) patients, moderate disability (21–40 points) for 14 (38%) patients, and crippling or severe disability (>40 points) for the remaining 6 (16%) patients. Mean SF-36 showed statistically significant improvement by 29±30.4 points (P=.0017) for the physical component and 21.8±25.6 points (P=.0062) for the mental health component (n=17) (Table).

Table: Operative and Clinical Outcomes

Overall, 4 (10.8%) patients had a dural tear requiring repair, 3 (8%) developed neurological symptoms resulting from a compressive epidural hematoma requiring evacuation during the index hospital stay, 1 (2.7%) underwent irrigation and debridement for infection, and 2 (5%) underwent an additional decompression procedure for symptom recurrence secondary to instability. Intraoperatively, 3 patients had dural tears, which were repaired without incident during the same operative setting. Postoperatively, 5 (14%) patients underwent 1 or more reoperations, including 2 (5%) requiring further decompression. Of these 5 patients, 2 had hematoma evacuation during index hospital stay and 1 had late pseudomeningocele repair. The remaining 2 patients were reoperated twice. The first patient who underwent L4 spinous process–splitting laminoplasty for canal stenosis and spondylolisthesis had a hematoma evacuation during the index hospital stay, and then required a revision decompression of L3–S1 with posterior instrumentation and transforaminal interbody fusion of L4–L5 for recurrent stenosis and worsening of preexisting instability at that segment with foraminal stenosis (after 0.77 years). The second patient who underwent L4 spinous process–splitting laminoplasty for canal stenosis underwent irrigation and debridement followed by a revision decompression of L4–L5 with posterior instrumentation and transforaminal interbody fusion of L4–L5 for recurrent stenosis with spondylolisthesis and radiculopathy symptoms secondary to foraminal stenosis (after 1.82 years).

Discussion

Spinous process–splitting laminoplasty is a recently reported technique to decompress the lumbar spine in a minimally invasive fashion. The rationale is to preserve the paraspinal muscles with the assumption that this will result in quicker recovery and better long-term function in these patients (Figure 9). The current study cannot address all of these questions; however, it demonstrates that this procedure can be used safely to decompress the spine in a North American cohort of patients with spinal stenosis. In this study, patients demonstrated improvements in all key outcomes for pain and health-related quality of life measures at a minimum 1-year follow-up with complications that are comparable with those reported with conventional and other minimally invasive spine surgery techniques.^26–29

Figure 9: Preoperative axial T2-weighted magnetic resonance image of a 45-year-old man with a severe spinal canal stenosis at L2–L3 showing severe spinal canal stenosis (A). Postoperative axial T2-weighted magnetic resonance image showing adequate decompression with healed spinous process (arrow) and preserved paraspinal muscle bulk and signal without evidence of atrophy (B).

In the current study, 4 (10.8%) patients had a dural tear with repair, 3 (8%) had a hematoma evacuation, 1 (2.7%) underwent irrigation and debridement for infection, and 2 (5%) required additional decompression with posterior fusion for persistent symptoms secondary to worsening instability. The reported number of dural tears is consistent with previous studies.^27–29 Wang et al²⁹ reported approximately 14% of patients with dural tears in a cohort of 641 patients who underwent lumbar spine surgery. Tsutsumimoto et al²⁸ reported that approximately 5% of patients sustained dural tears in a cohort of 555 patients who underwent a microscopic lumbar decompressive procedure. Although the development of symptomatic epidural hematoma is infrequent after lumbar decompression surgery,²⁶ the rate of postoperative hematoma requiring evacuation was high in the current study cohort (8%). The authors theorize that this is due to some continued bleeding postoperatively from the cut edges of the spinous process. This was a complication encountered early, before routine drain placement, and therefore it is likely preventable. Subsequent to routine drain placement, no epidural hematomas were encountered.

A few studies have been published on lumbar spinous process–splitting laminoplasty or slight variations of this technique. These have included a brief description of the technique and reports on the clinical outcomes in patients with lumbar canal stenosis.^14–25 Watanabe et al¹⁴ compared the clinical outcomes at a minimum of 2 years between 18 patients treated with spinous process–splitting laminectomy and 20 patients treated with conventional laminectomy for lumbar canal stenosis. They concluded that comparable clinical results can be achieved for both techniques based on the Japanese Orthopaedic Association (JOA) score. They also demonstrated significantly less muscular atrophy and fatty infiltration of the paraspinal muscles in patients undergoing spinous process–splitting laminectomy compared with traditional decompression based on postoperative MRI.¹⁴

Cho et al¹⁶ compared the results of 40 patients undergoing a spinous process–splitting laminoplasty technique (Marmot technique) with 30 patients undergoing conventional laminectomy. They reported significant clinical improvements at 1 year favoring patients undergoing spinous process–splitting laminoplasty. They used the JOA score for low back disorders, the VAS for back pain, and the Prolo scale for function and employment status. They also noted that patients undergoing the spinous process–splitting laminoplasty had significantly lower creatine phosphokinase–muscular-type isoenzyme levels during the early postoperative period, indicating less muscle trauma with this technique.¹⁶

Watanabe et al¹⁵ noted a significant improvement in wound pain during the early recovery period with no difference in activities of daily living in patients undergoing spinous process–splitting laminoplasty compared with those undergoing conventional laminectomy. They also noted that patients undergoing spinous process–splitting laminoplasty had significantly lower rates of atrophy of the paraspinal muscles on axial MRI 1 month postoperatively.¹⁵

Rajasekaran et al²² conducted a randomized, controlled trail and compared outcomes at a minimum of 1 year (mean, 14.2 months) between 28 patients undergoing lumbar spinous process–splitting laminoplasty and 23 patients undergoing conventional decompression. They found no difference between the groups in number of decompressed levels, operative time, intraoperative blood loss, and length of hospital stay. They also reported no difference in clinical outcomes between patients based on JOA score, neurogenic claudication outcome score, VAS for back pain, and VAS for neurogenic claudication. Furthermore, there was no difference between patients in C-reactive protein and creatine phosphokinase–muscular-type isoenzyme levels on postoperative days 1 and 3. Thus, they could not establish the advantage of lumbar spinous process–splitting laminoplasty vs conventional decompression.²²

Although many studies focusing on the potential benefits of minimally invasive spine surgery to preserve the posterior musculature have failed to show long-term benefit over conventional techniques, several studies demonstrate some potential advantages to these techniques, including saving the paraspinal muscles, avoiding the detachment of tendons, and minimizing tissue retraction.^5,8–11,30 Gejo et al³⁰ assessed postoperative back muscle injury in rats and reported that 2-hour retraction resulted in a high signal intensity for the multifidus muscles 21 days postoperatively. Histologically, these were also associated with small-diameter and large extracellular fluid space for the regenerating muscle fibers. Kim et al⁸ evaluated trunk extensor muscle strength following open vs minimally invasive spinal fusion and demonstrated a significantly better extension strength in the minimally invasive spinal fusion group. Kim et al¹¹ studied serum skeletal muscle tissue injury markers and inflammatory cytokines in patients undergoing conventional posterior lumbar decompression and fusion vs patients undergoing a mini-open approach using tubular retractors. They found that patients who underwent the conventional technique had significantly higher levels of serum creatinine kinase, aldolase, proinflammatory cytokines (interleukin [IL]-6, IL-8), and anti-inflammatory cytokines (IL-10, IL-1 receptor antagonist) during the early postoperative period.¹¹ Nevertheless, more studies are required to evaluate the potential benefits of this procedure on clinical outcomes and functional recovery in the long term.

The current study has some limitations, including its retrospective nature and the limited number of patients who underwent the proposed technique; thus, further analysis of the factors associated with the outcomes in these patients was difficult. Several patients were lost to follow-up, possibly due to a poor outcome, and therefore the complication rate may be underrepresented.

Despite these limitations, a strength of this study is that all patients were treated with the same technique performed by a single surgeon. This cohort also included all cases since adopting the technique, including early cases during the learning curve, which presumably may have a higher complication rate. To the authors' knowledge, this series is the first to report the use of spinous process–splitting laminoplasty in a North American cohort of patients, demonstrating that this technique may be suited to this population.

Conclusion

Spinous process–splitting laminoplasty for the treatment of lumbar canal stenosis has favorable clinical outcomes with the potential advantage of avoiding extensive damage of the posterior spinal musculature. Future research using different study designs and comparing the spinous process–splitting laminoplasty with other traditional surgical treatments at long-term follow-up is required.