Endoscopic Intertransverse Lumbar Fusion

Objective:

To assess the feasibility of a new endoscopic technique for intertransverse lumbar fusion using human cadavers.

Summary of Background Data:

Current methods of exposure for posterior lumbar fusion have been shown to cause significant morbidity resulting from iatrogenic muscle denervation, devascularization, ischemia, and increased intramuscular pressure. All of these factors can lead to chronic pain and subsequent disability.

Methods:

Utilizing a percutaneous, posterolateral endoscopic approach via a 16 mm tubular retractor (MED, Sofamor Danek, Memphis, TN), the intertransverse interval was exposed unilaterally at five lumbar levels in two fresh cadavers. The transverse processes and facet complexes were decorticated using an endoscopic, high–speed drill. A standard open Wiltse intertransverse approach was performed on the contralateral side at the same levels. Iliac crest bone graft mixed with a synthetic BMP carrier was delivered to all ten fusion sites. Post–procedure CT scans (1 mm slices) were evaluated for graft placement and host–graft apposition.

Results:

No qualitative or quantitative differences in host bone preparation or graft placement were noted between the two techniques. The endoscopic technique did, however, take longer (15–40 min) than the standard open technique (15–20 min).

Conclusion:

We have developed an endoscopic fusion technique that minimizes surgical muscle trauma, yet provides excellent visualization and access to the pertinent bony anatomy. Utilization of this technique with BMP and a synthetic carrier could minimize patient morbidity and may provide an attractive alternative to current methods of single–level posterior lumbar fusion.

Introduction

The most common type of fusion performed in lumbar surgery is the posterolateral intertransverse fusion. Standard techniques for this procedure entail significant paraspinous muscle trauma. These muscles must be stripped from the transverse processes and retracted for exposure of the underlying bone. This can result in denervation, devascularization, and ischemia of these important structures. Significant spasm in the short–term and atrophy in the long–term contribute to the morbidity and sequelae of fusion. The purpose of this study was to assess the feasibility of a new endoscopic technique for intertransverse lumbar fusion in human cadavers. The primary goal was to demonstrate that this technique was comparable to open techniques in terms of providing posterolateral gutter access, preparation, and adequate graft delivery, while minimizing paraspinous muscle trauma.

Materials & Methods

Figure l. The MED System (Microendoscopic Discectomy, Sofamor Danek Group, Memphis, TN). Illustrated are the sequential dilators, 16 mm tubular retractor, and the 25º endoscope.

Endoscopic Surgical Technique:

The endoscopic technique was performed using the MED system (Sofamor Danek Group, Memphis, TN—see Figure 1). A 16 mm, longitudinal, paramedian incision (4.5 cm from the midline) was centered on the disc space of the level to be fused. With fluoroscopic guidance, the dilators and tubular retractor were directed to the junction of the cephalad pars and transverse process (TP) (Figure 2). The tubular retractor was held in placed by an articulated arm. The endoscope was then positioned within the retractor. The transverse process was directly visualized; proper localization was confirmed by bony palpation and fluoroscopy. The paraspinal muscles were then dissected off the dorsal surface of the transverse process, staying superficial to the intertransverse ligament. Dissection was achieved using standard long curettes, forceps and electrocautery.

Figure 2. Initial dilator located at the right L3 pars–transverse process junction. The transition from the TP to the pars is easily palpated and confirmed by fluoroscopy.

After the plane superficial to the intertransverse ligament was established, complete exposure of the posterolateral aspect of the pars, superior and inferior facet complexes and the caudal transverse process was achieved by sequentially moving (pivoting) the tubular retractor (Figure 3). Once this interval was developed, full access and visualization of the posterolateral gutter was accomplished by sweeping the retractor from TP–to–TP.

Figure 3. Endoscopic view of the right L3 transverse process and the intertransverse ligament.

Decortication of the bony surfaces (Figure 4) was then easily performed under direct visualization with an endoscopic high–speed burr (MedNext, Sofamor Danek Group, Memphis, TN). Graft material was then systematically placed into the prepared posterolateral gutter, with the intertransverse ligament suspending the graft from TP–to–TP (Figure 5). Dilation of the plane between the paraspinal muscles, combined with minimal deep dissection of muscle from the posterolateral bony surfaces, leaves a minimally traumatized, healthy muscular bed that falls back over the exposed space (Figure 6).

Figure 4. A) Endoscopic view of the right L5–S1 posterolateral interval. Due to their close proximity, the L5 transverse process and sacral ala can be visualized and exposed with little or no movement of the tubular retractor. B) Endoscopic view illustrating decortication of the right medial sacral ala using a high–speed burr introduced through the tubular retractor. As demonstrated, the facet complex can also be exposed and decorticated, if desired.

Figure 5. Delivery of graft material can be performed in a piecemeal fashion (autograft mixed with BMP carrier), or via a syringe, depending on its consistency.

Figure 6. As the tubular retractor is removed, the dilated paraspinal muscles are seen falling back into place over the posterolateral gutter.

Discussion

Posterolateral fusion for spinal stenosis, degenerative spondylolisthesis or isthmic spondylolisthesis constitutes a large portion (1/3–1/2) of lumbar fusion procedures. Recent literature suggests that instrumentation does not necessarily improve patient outcome in this specific subset of patients. This would also indicate that the surgical morbidity associated with instrumented fusion may be unnecessary in a large number of patients requiring lumbar fusion. Our current surgical methods for instrumented lumbar fusion create patient morbidity from three main sources: surgical exposure, harvesting of iliac crest bone graft, and instrumentation. The most commonly–used method of exposure for lumbar fusion is a midline approach with extensive paraspinal muscle–stripping out to the lateral aspects of the transverse processes, followed by prolonged retraction of these muscles. This type of exposure has been shown to cause significant muscle morbidity resulting from iatrogenic muscle denervation, devascularization, ischemia, and increased intramuscular pressure. This can lead to subsequent paraspinal muscular atrophy, scarring and decreased extensor strength. The clinical effect of this muscle morbidity has been termed “fusion disease.” It can be a significant source of postoperative pain and functional impairment in the convalescent period, as well as a detriment to long–term paraspinal lumbar muscular function.

The endoscopic technique developed in this study is aimed at minimizing muscle morbidity by dilating the paraspinal muscles and performing only minimal muscle dissection at the bony surface. The thin–walled tubular retractor, which serves as both the endoscopic and working portal, enables minimal muscle retraction (8 mm from the center), thereby minimizing the retraction pressure on the surrounding muscle. As the tube is moved, the muscle is allowed to reapproximate to its natural position.

Furthermore, the “keyhole principle” of endoscopy provided by this procedure allows for a larger field of view at the level of the bone through a much smaller incision. The proposed advantages of endoscopic posterolateral fusion techniques are to decrease postoperative pain, minimize acute and chronic muscular dysfunction, and hence, improve functional recovery. It is hoped that this will result in a significant reduction of in–hospital stay and rehabilitation time.

The MED system provides an ideal, simple method of posterolateral exposure and graft material delivery. When necessary, it also gives the surgeon the ability to perform a decompression via the endoscope or by standard open means. By exposing both sides of the spine simultaneously, our ultimate goal is to be able to perform a single level endoscopic lumbar fusion under epidural or general anesthesia as an overnight stay or outpatient procedure. These proposed benefits of endoscopic posterolateral fusion will need to be further evaluated in a prospective clinical trial.

Figure 7. A) Axial CT image illustrating graft placement at the L3 transverse process. The endoscopic procedure was performed on the right in this cadaver. B) L5 axial CT image illustrating excellent apposition of the graft to host bone on both the open (left) and endoscopic sides. C) Three–dimensional CT reconstruction illustrating graft placement at L3–4 and L5–S1.

Conclusions

Endoscopic posterolateral lumbar fusion is feasible using the MED system. This technique should significantly reduce the muscle morbidity (i.e., “fusion disease”) typically associated with standard open fusion procedures. With the increasing availability of BMP, particularly rhBMP–2, minimally invasive procedures which do not require instrumentation or iliac crest bone graft may provide ideal management alternatives for patients requiring surgery for degenerative spondylolisthesis, spinal stenosis and lower grades of isthmic spondylolisthesis.

Results

The access to and ability to prepare the posterolateral gutter using the endoscopic technique was comparable at all levels to the open technique. The incision size for the open procedure was, on average, 35.2 mm (range: 32–39 mm) versus 16 mm for the endoscopic technique. The paramedian open technique was used to directly compare the same paraspinal intermuscular interval utilized in the endoscopic technique. Use of the more common standard midline approach would have resulted in larger skin incisions and significantly more muscular dissection and stripping to achieve comparable exposure of the posterolateral gutters. The average time to perform the endoscopic procedure was 30 min (range: 14–40 min) compared to 17.2 min (range: 15–20 min) for the open procedure. It must be kept in mind that additional time would be required in–vivo for bleeding control. Qualitative analysis of the post–procedure CT scans revealed indistinguishable bone graft placement and host bone–to–graft apposition between the two techniques at all five levels (Figure 7A,B,C). The mean volume of graft placed was 15.2 cm3 (range: 11.6–20.2 cm3) and 15.4 cm3 (range: 10.9–21.1 cm3) for the endoscopic and open techniques, respectively.

Y. Raja Rampersaud, M.D. and Kevin T. Foley, M.D.
Image–Guided Surgery Research Center; Semmes–Murphey Clinic; Department of Neurosurgery, University of Tennessee, Memphis