SECTION II: ORIGINAL ARTICLES: Spine
The paraspinal approach to the lumbar spine was first described by Watkins in 1959.3 A plane is developed between the lateral border of the sacrospinalis muscles and the quadratus lumborum muscle. Wiltse et al described a modified transmuscular paraspinal approach5,7,10 consisting of a longitudinal separation of the sacrospinalis muscle between its multifidus and longissimus parts (Fig 1). The original description was for spinal fusion, especially regarding lumbosacral spondylolisthesis treatment.9 A one-level or a multilevel fusion can be performed with this approach, leaving the supraspinalis and interspinalis ligaments intact. Its use for removing far lateral disc herniations,6 decompressing a far-out syndrome,8 inserting pedicle screws, and for spinal canal decompression have been described.10 Despite the descriptions by Wiltse et al, the exact location of the sacrospinalis muscle and where it must be split are unclear.
We intended to provide a more precise anatomic description for the transmuscular paraspinal approach and to provide topographic landmarks to facilitate this approach.
MATERIALS AND METHODS
Fifty cadavers of undetermined age were dissected by the same author (RV). There were 27 males and 23 females. There were 33 embalmed cadavers and 17 fresh cadavers (Table 1).
Cadavers were placed in the prone position. For each cadaver, the trunk length was measured from the spinous process of the seventh cervical vertebra to the posterosuperior iliac spine. A midline skin incision was made to the level of muscular fascia. The superficial fascia is in continuity with the fascia of latissimus dorsi muscle. This fascia was opened close to the midline and retracted laterally in an anatomic avascular space between the superficial fascia and the muscular aponeurosis of the sacrospinalis muscle. The muscular aponeurosis of the sacrospinalis muscle contains muscular fibers on its cranial superficial part and only fibrous tissues on its caudal superficial part. The border between muscular and fibrous tissues was noted by two measurements. The d1 distance was the distance between the caudal muscular border and the superior posterior iliac spine. The d2 distance was the distance between the caudal muscular border and the midline. The d1 and the d2 distances were orthogonal (Fig 2). The level of the natural cleavage plane between the multifidus and the longissimus parts of the sacrospinalis muscle was noted, and measurements were made between this level and the midline at the level of the spinous process of L4 (d3 distance) (Fig 3). The approach then was completed through the sacrospinalis muscle to expose cranially the transverse process of L3 and caudally the articular process of S1 (Fig 4). The same procedure was performed on both sides. All measurements were entered into a computer database.
Analysis showed no difference between fresh and embalmed specimens for the d1 distance, for the d2 distance, or the d3 distance; there also were no differences in male and female cadavera for these measures. Therefore, we combined the data for embalmed and unembalmed specimens and males and females. Because of the small number of cadavers, statistical analysis was performed using Wilcoxon's nonparametric test and an unpaired Student's t test.
The mean distance between the spinous process of the seventh cervical vertebra and the posterior superior iliac spine was 48.03 cm (range, 40-58 cm). This value was greater (p < 0.001) in males than in females. In all cases, there was an anatomic avascular space between the superficial muscular fascia and the muscular aponeurosis of the sacrospinalis muscle.
The border between muscular and aponeurotic parts of the sacrospinalis muscle was easily identified in 93 of 100 cases. In seven cases (both sides of three cadavers and one unilateral side), muscular fibers of the sacrospinalis muscle were attached onto the sacrum. The mean distance between the superior posterior iliac spine and the border between muscular and aponeurotic parts of the sacrospinalis muscle (d1 distance) was 4.86 cm (range, 0-8 cm).
The mean distance from the border between muscular and aponeurotic parts of the sacrospinalis muscle to the midline was 4.71 cm (range, 3-7 cm). The closeness of the musculoaponeurotic border to the sacrum correlated (r = 0.262; p < 0.001) to its closeness to the midline.
A natural cleavage plane between the multifidus and the longissimus parts of the sacrospinalis muscle was present in all cadavers. There was a fibrous separation between the two muscular parts in 88 of 100 cases. The fibrous separation was easily identified at the caudal part of the muscular cleft but disappeared gradually above the level of the L4 transverse process. In 12 cases (10 cases were both sides of five cadavers and two were unilateral cases), there was no well-defined fibrous partition inside the sacrospinalis muscle. A natural cleavage plane was seen in other cases. The mean d3 distance between the level of the cleavage plane and the midline was 4.04 cm (range, 2.4-7 cm). Most of the d3 distances ranged from 3.5 to 4.5 cm with a normal distribution (Fig 5). The d3 distance correlated (r = 0.145; p = 0.043) with the trunk length (ie, the d3 distance was longer in the tallest cadavers) (Table 2).
Small arteries and veins were present in all cadavers precisely at the level of the cleavage plane, arising from this anatomic intermuscular space and remaining on the surface of the sacrospinalis muscle (Fig 6). On both sides of 12 cadavers, we identified the emergence of a large posterior ramus of the L3 nerve. It was possible to retract and preserve the nerve cranially and laterally to expose the articular and transverse processes from L3-S1.
Despite descriptions by Wiltse et al, the exact description of the sacrospinalis muscle-splitting approach is unclear. The anatomy of the paraspinal approach is more complicated than the midline approach because the spinous processes are not exposed to provide landmarks. A precise description of the anatomic cleavage plane between the multifidus and the longissimus parts of the sacrospinalis muscle is lacking in the literature. For these reasons, we wanted to provide a more precise anatomic description.
We recognize the limiting factors of such anatomic studies made with anatomic specimens. Our cadavers were old with atrophied muscles. The d3 distance between the midline and the muscular cleft we measured might differ in young and healthy patients, especially teenagers treated for lumbosacral spondylolisthesis. In the description by Wiltse and Spencer, the fascial incision is made 2 cm lateral to the midline and the finger can be plunged inside the sacrospinalis muscle at any point above the L4 level.10 Our results showed that the muscular cleft was 4 cm lateral to the midline. These findings might be explained by the difference between cadaveric and young patients' muscular structures.
Three different paraspinal approaches to the lumbar spine have been described. The surgical approach described by Ray2 and Watkins3,4 are between the sacrospinalis muscle and the quadratus lumborum muscle. The approach described and popularized by Wiltse et al5,7,10 is more medial in the cleft between the multifidus and longissimus parts of the sacrospinalis muscle. These three approaches allow the surgeon to reach the area to be fused or decompressed without incising many of the supporting structures. In our experience, the paraspinal approach is useful to obtain solid posterior fusion in children and adolescents with high-grade lumbosacral spondylolisthesis.1 The L5 and S1 transverse and articular processes are reached easily through this approach, which causes less bleeding than the posterior approach to the lateral parts of the vertebrae.
We propose a specific posterior approach allowing easy identification of the anatomic cleavage plane between the multifidus and longissimus parts of the sacrospinalis muscle. First, the superficial muscular fascia is opened near the midline that exposes the posterior aspect of the sacrospinalis muscle. Then, the location of the muscular cleft can be found by identifying the perforating vessels leaving the anatomic intermuscular space. In all specimens, we identified small arteries and muscular veins at the level of the muscular cleavage plane arising from the anatomic intermuscular space, and on the surface of the sacrospinalis muscle (Fig 7). This cleft provides a good anatomic landmark to open the sacrospinalis fascia (after having cauterized the vessels) and identify the cleavage plane. Except these perforating arteries, we observed no anatomic landmarks allowing differentiation of the muscular cleavage plane.
The cleavage plane seems easier to find caudally, but the approach requires more hemostasis. In 88% of our cadavers, a fibrous partition was identified at the caudal part of the muscular cleft and disappeared gradually above the level of L4 transverse process. This anatomic finding may explain a part of the observations by Wiltse et al, but we did not find specific vessels explaining increased bleeding. It was easy to expose the transverse and articular processes from L3 to the sacrum. In 12 cases, we found a large posterior ramus of the L3 nerve that could have been injured during surgical approach. The dissection must be performed carefully at the cranial part of the surgical approach to avoid damaging this sensitive nerve through the muscular cleavage plane.
Our anatomic study confirms numerous advantages of the paraspinal sacrospinalis splitting approach. The approach is easy and allows exposure of the articular and transverse processes. A one-level or a multilevel fusion can be performed using this approach without widely removing muscle insertions and leaving the supraspinalis and interspinalis ligaments intact.
1. Dubousset J. Treatment of spondylolysis and spondylolisthesis in children and adolescents. Clin Orthop Relat Res
2. Ray CD. The paralateral approach to decompression for lateral stenosis and far lateral lesions of the lumbar spine. In Watkins E, ed. Principles and Techniques in Spine Surgery
. Aspen, CO: Collis; 1987:217-227.
3. Watkins MB. Posterolateral bonegrafting for fusion of the lumbar and lumbosacral spine. J Bone Joint Surg
4. Watkins MB. Posterolateral fusion in pseudarthrosis and posterior element defects of the lumbosacral spine. Clin Orthop Relat Res
5. Wiltse LL. The paraspinal sacrospinalis-splitting approach to the lumbar spine. Clin Orthop Relat Res
6. Wiltse LL. Surgery for intervertebral disk disease of the lumbar spine. Clin Orthop Relat Res
7. Wiltse LL, Bateman JG, Hutchinson RH, Nelson WE. The paraspinal sacrospinalis-splitting approach to the lumbar spine. J Bone Joint Surg
8. Wiltse LL, Guyer RD,Spencer CW. Glenn WV, Porter IS. Alar transverse process impingement of the L5 spinal nerve: the far-out syndrome. Spine
9. Wiltse LL, Hutchinson RH. Surgical treatment of spondylolisthesis. Clin Orthop Relat Res
© 2006 Lippincott Williams & Wilkins, Inc.
10. Wiltse LL, Spencer CW. New uses and refinements of the paraspinal approach to the lumbar spine. Spine