Fusions of the lumbosacral spine have been and continue to be the bane of the spine surgeon. The complex local anatomy, unique biomechanical forces, and poor bone quality of the sacrum are just a few of the many reasons why fusions of the lumbosacral spine have been notoriously difficult to perform with various complications. In this chapter, we hope to familiarize the reader with the complicated anatomy, specific entities that involve this region, and the biomechanical forces that lead to high pseudarthrosis rates. We also help to outline an algorithm of treatment options, both conservative and surgical, to treat the specific problem of pseudarthrosis in this region.

The sacrum consists of five fused vertebras with transverse processes that merge to form the lateral masses. It is, for the most part, a very cancellous structure with a thin cortical rim. The areas of increased bone density are the sacral promontory and the sacral ala. Geometrically, the sacrum is wedge-shaped and its anteroposterior diameter rapidly decreases from 45 to 50 mm at S1 to 20 to 30 mm at S3 (1). Numerous other important anatomic structures are in very close proximity to the sacrum. The internal iliac artery and vein, middle sacral artery and vein, sympathetic chain, lumbosacral trunk, and the sigmoid colon lie anterior to the sacrum (2). All these structures, at some point in their course, lie directly on the sacrum. This increases the chance of perforation of these structures during instrumented fusions performed either anteriorly or posteriorly (2).

The sacrum is connected to each ilium by the sacroiliac (SI) joint. This joint is a fibrocartilaginous joint with interdigitating contours on sacrum and iliac bones (3,4). It is, for the most part, an immobile joint and functions mainly as a shock absorber to transfer loads to hip joint. The SI joints are stabilized by interosseous, dorsal, and ventral ligaments, with the dorsal ligaments being the strongest (3,4).

There are numerous pathologic entities that involve the lumbosacral junction. These may include scoliosis of any type, instability secondary to spondylolysis or spondylolisthesis, degenerative disk disease, trauma, tumors, and infection. Each pathology presents with its own unique set of challenges for achieving lumbosacral fusion. In this chapter, however, we will focus on the complication of lumbosacral and spinopelvic pseudarthrosis and potential treatment options.

Aside from the difficulties of the local anatomy and poor bone quality, large lumbosacral loads are placed on implants at this junction (4,5,6). This includes large flexural bending moments and concomitant cantilevered forces on posteriorly placed implants. Also, being the terminal extension of the mobile spinal column, large axial, translational, and rotational loads are present at the lumbosacral junction. This is partly due to the unique anatomical attributes of this region. The L5-S1 disk space has the largest flexion-extension range of motion than any other level in the lumbar spine. It also has the steepest slope with the highest translational shear forces. This leads to the large bending moment forces resulting in implant failure, pseudarthrosis, and implant pull-out (4).

Historically, fusions in the lumbosacral junction were attempted with in situ fusion. This had the high rates of pseudarthrosis, reportedly 46% to 100% (6,7,8,9). With the advent of instrumentation, pseudarthrosis rate has decreased (10,11,12,13,14). The first advancement was with the Harrington instrumentation for scoliosis developed in the 1960s (15). The system consisted of hooks and compression and distraction rods. Later, sacral rods were added for cases requiring fusion to the sacrum. However, high rates of pseudarthrosis, as high as 40%, were reported in fusions to the sacrum due to poor control of flexion-extension moments at the lumbosacral junction. This eventually led to instrumentation failure and migration of sacral bar and dislodgment of hooks (6,16,17). Other cases resulted in flat-back syndrome with its subsequent associated problems (18). In the 1970s, Dr. Edoardo Luque developed the concept of segmental spinal instrumentation (19,20). This entailed the use of multiple points of fixation with wires passed beneath the lamina and attached to Harrington rods. This allowed for restoration of frontal and sagittal alignment and eliminated flat-back syndrome. Yet despite these advancements, high pseudarthrosis rates—up to 41% with complications as high as 82%—still persisted (21,22,23). Biomechanically, this instrumentation method failed in the lumbosacral junction due to lack of torsional and axial stability. It offered little resistance to flexion moments at L5-S1 disc space. Also, numerous neurologic complications were noted to sacral roots due to passage of Luque wires into the spinal canal (21).

In the 1980s, Cotrel et al. introduced the concept of fixation with hooks and screws at every level of fusion (CD instrumentation) (24,25). Distal sacral fixation was achieved with S1 pedicle screws or iliosacral screws. This allowed for the concept of spinal derotation. Biomechanically, the CD instrumentation allowed for greater rigidity, axial compression, and torsional stability. Yet despite this, it was still poor in controlling flexion at the lumbosacral junction. This lead to high distal pseudarthrosis rates at the lumbosacral junction and S1 screw loosening and pull-out. Devlin et al. reported pseudarthrosis rates of 33%, with 70% of the complications directly related to the instrumentation (26).

The last major advancement with regard to improving instrumentation at the lumbosacral junction was introduced with the Galveston technique (22,27,28). This new instrumentation technique improved upon the distal fixation difficulties of all prior systems. In this system, the longitudinal rods are contoured and placed between the two bony tables of the iliac bone, starting from the posterior superior iliac spine and aiming for the anterior inferior iliac spine. This intrapelvic placement of the hardware allows for the biomechanical advantages of controlling fixation at the lumbosacral junction due to the fixed angle of the rod. This allows for a larger lever arm distally, and a larger area of contact between the rod and the ilium. Numerous studies have noted low lumbosacral pseudarthrosis rates with this method of distal instrumentation (29,30). This has paved the way for more modern techniques of distal lumbosacral fixation with iliac rods, screws, and bolts placed in a Galveston fashion.

To better understand the biomechanics of lumbosacral fixation, McCord et al. has developed the concept of the lumbosacral pivot point (31). This point is anatomically defined as the posterior border of the vertebral body of L5 at the middle of the osteoligamentous column at the L5-S1 disc space. They concluded from their ex vivo biomechanical testing that posterior implants that are ventral to this point provide a more effective moment arm to resist flexion forces and improve the ultimate fixation strength than do implants that are dorsal to it. Furthermore, O’Brien has divided the sacropelvic region into three zones (32). Zone I includes the S1 vertebral body and the cephalad margins of the sacral ala. Zone II extends from the inferior margins of the sacral ala and extends into S2 and to the tip of the coccyx. Zone III encompasses the ilium bilaterally. In his biomechanical testing, fixation strength was noted to significantly improve from Zone I to Zone III.

Fixation and fusion to the sacrum can be divided into two categories: long and short fusions. Long fusions to the sacrum have a higher likelihood of failing with pseudarthrosis and hardware failure unless the fusion includes the sacropelvic region (6,30,33,34,35,36,37). Long fusions are typically those that extend from L2 or more cephalad levels to the sacrum. Other indications for sacropelvic fixation include cases of scoliosis with an oblique takeoff at the L5-S1 level, rigid structural curves, paralytic curves, neuromuscular curves, and congenital deformities (6,36,38). Less common reasons for sacropelvic fixation are flat-back syndrome requiring corrective osteotomy, decompression below a long fusion ending at L5, osteopenic patients with multisegment
fusion, high-grade spondylolisthesis, pelvic obliquity (such as in poliomyelitis, cerebral palsy), fractures and dislocations of the lumbar spine, postsacrectomy tumor reconstruction, and sacral fractures with or without neurologic injury (6,38). Numerous options exist for distal sacral and/or sacropelvic fixation. These may include casting and bracing, sacral sublaminar devices (wires, cables, hooks), sacral pedicle screws, sacral alar screws, iliosacral screws, S-rods, Jackson intrasacral rods, Kostuik sacral bars, the Galveston technique with iliac rod, posts, screws, or bolts (4,6,18,38,39). However, biomechanically, iliac fixation with intrailiac screws placed in a Galveston fashion offers the most stable construct. The surgical approach of extending the fusion to the sacrum can include posterior, anterior, posterior lumbar interbody, transforaminal lumbar interbody, or a combined anteroposterior approach.

Pseudarthrosis continues to be the most common and devastating complication of lumbosacral fusion (40,41). As the number of fusion procedures has increased, so has the incidence of pseudarthrosis in this region. Typically, patients present with ongoing pain after a period of pain relief from their index procedure. Failure of the instrumentation is one of the first clues to the presence of pseudarthrosis (42,43,44). Lack of implant failure by 3 to 5 years has been reported as a good evidence of stable arthrodesis (45). Implant failure typically presents with pull-out of screws or breakage of the hardware. There are many potential intraoperative factors that can predispose to pseudarthrosis in this area (Table 15.1). Additionally, many patient factors as outlined in Table 15.2 can further increase the risk of pseudarthrosis.

Pseudarthrosis workup is very controversial. No single test is 100% sensitive or specific. Radiographically, a pattern of continuous trabeculae that traverses the grafted region of the adjacent vertebral bodies with a lack of motion has been defined as the radiographic criterion for fusion (46). Dawson et al. have recommended the use of anteroposterior (AP) and oblique radiographs (47). AP and lateral radiographs correlated with the presence of pseudarthrosis in only 48% of cases, whereas when oblique radiographs were added, the sensitivity increases to 82% due to better visualization of the facet joints (47). Oblique radiographs can, however, be problematic to interpret in cases of scoliosis, kyphosis, or when instrumentation blocks appropriate imaging (47). While these types of imaging assess the structural integrity, they do not assess the functional status of the fusion (48,49,50). Flexion and extension films have been recommended to better and more functionally evaluate a solid arthrodesis. This, however, is very controversial. Flexion and extension views may be insensitive when there is subtle motion, muscle guarding, or both. Brodsky et al. have noted a greater correlation with AP, lateral, and oblique radiographs versus bending films (51). Similarly, Lauerman et al. have not found bending radiographs useful (52).