Overview

As the terminal segment of the spine, the sacrum is subjected to substantial forces as it transmits axial loads from the lumbar spine to the adjacent ilia via the sacroiliac joints. This relationship leads to several pathologies that are specific to the sacrum, in addition to the extra demands applied as a consequence of surgical treatments. Fixation into the sacrum remains somewhat more challenging than other regions of the caudal spine. The remnant vertebral bodies of the sacrum are relatively small, and the cancellous bone quality is often poor. This leaves comparatively little bone for fixation, and generally only the S1 and S2 segments are amenable to screw placement. Additionally, the sacral instrumentation may be the most distal fixation for long constructs that span the entire thoracolumbar spine, which places great strain upon these two or four screws. These factors have led to the routine use of supplemental iliac bolts to prevent both failure of sacral fixation and sacral insufficiency fractures. In spite of these limitations for fixating to the sacrum, numerous options include pedicle screws, alar screws, posterior sacral plates, and screws combined with intrasacral rods.

Historical Background

Although internal fixation for lumbosacral fusion was described as early as the 1940s, effective sacral fixation involves relatively new technology. In 1948 King described the use of bilateral facet screws for lumbosacral fusions, which allowed early mobilization. These screws were not widely adopted in part because there was a high incidence of early failure and pseudarthrosis.

The first commonly used form of spinal internal fixation was the Harrington rod, used for fusions to the sacrum on many occasions; but fusing to the sacrum proved difficult. The bone of the S1 lamina is thin, and often it would not support a Harrington hook, plus there were concerns of S2 nerve root irritation from the tip of the laminar hooks. This limited fixation led to the development of the Harrington sacral bar, effectively the first iliac bolt. These Harrington bars were transiliac rods that bridged across the posterior superior iliac spine and gave a strong foundation for Harrington constructs to the sacrum. Such bar constructs had much better purchase in bone but also had many limitations: they were prone to migration over time and required a separate skin incision, which often resulted in wound complications. Furthermore, they could freely rotate in the ilium and did little to resist flexion and extension movements. Additionally, long fusions to the sacrum that used this distraction instrumentation system had a pseudarthrosis rate that approached 40%, and it flattened the patient’s lumbar lordosis, leading to the now well-known flat back syndrome. Because of these limitations, most surgeons strongly avoided long fusions to the sacrum, and short lumbosacral fusions were often not instrumented. The Harrington system was later modified to allow some rod contouring and angling of the distal hook to allow for lumbar lordosis. This helped, but loss of lumbar lordosis was still the norm, and flat back syndrome remained a common occurrence.

Anterior fixation also proved more challenging within the lumbar spine. Dwyer instrumentation was developed and implemented in the late1960s, followed by Zielke instrumentation in the 1970s, principally for lumbar deformity correction and to aid arthrodesis for long fusions. These anterior instrumentation systems were effective and rapidly grew in popularity for lumbar deformity, although early series showed loss of lumbar lordosis as a common occurrence. However, the shape of the sacrum and the position of the iliac crest precluded extension of these systems past the lumbosacral junction. Some authors attempted hybrid implant systems, using Zielke or Dwyer instrumentation in the lumbar spine, with anterior lumbar interbody fusion (ALIF) at L5-S1, and applying an anterior staple across the lumbosacral junction. These early staples were conceptually similar to modern anterior lumbosacral plating systems, but they lacked threads or tines and were easily displaced and often migrated. They also were not rigid, therefore they did not provide much stability. Pseudarthrosis was common.

The 1970s saw the first widespread interest in the treatment of spinal stenosis, first described by Verbiest in 1951. Many spine surgeons appreciated the need for concomitant arthrodesis in addition to multilevel decompressions for stenosis. With few fixation options available, surgeons experimented with both Harrington rods and Knodt rods to stabilize decompressions and fusions. These implants were not well suited to the task. First, both were distraction systems that typically flattened the normal lumbar lordosis. Second, both systems relied on hooks and supplemental use of sublaminar wires, but with loss of the posterior elements for the decompression, there was no place to apply these hooks or wires; this meant surgeons had to instrument an unaffected vertebra above the decompression. Distal fixation often consisted of supralaminar hooks applied to S1 or hooks tamped into the ala. Supplemental wires could be placed around the S1 or S2 laminae via small drill holes or through the dorsal sacral foramina. Again, implant dislodgement, breakage, and pseudarthrosis were very common.

Effective sacral fixation was not available until the advent of pedicle screws. Roy-Camille was among the first to experiment with the concept of screw-and-plate constructs, and Harrington began using L5 lag screws fixed with wire to a Harrington distraction rod for treating severe spondylolisthesis. It would take another two decades before such devices saw widespread use and were popularized in the United States by the work of Arthur Steffee, mostly during the 1980s. Pedicle screws offered the first effective means of capturing quality bone in the sacral bodies via posterior instrumentation; these delivered much more robust fixation than prior hook or wire constructs, and they allowed for multiaxial control for reduction of deformity and maintenance of normal sagittal contours.

The S1 screw is still the only pedicle screw that is amenable to routine bicortical purchase, which greatly enhances pullout strength. However, in spite of the benefits of bicortical purchase, S1 screws are often a weak link in a lumbosacral construct. Much of this is due to sacral anatomy, because the cancellous bone is not dense, and the cortex is often rather thin. The S1 pedicles are usually quite large relative to the lumbar pedicles, and overall screw purchase is often less than lumbar pedicle screws. These weaknesses have led to several additions to enhance S1 screws. These include sacral plates (Chopin plate, Colorado system), S2 alar screws, Jackson intrasacral rods, and routine use of anterior column support and supplemental iliac screws for long fusions to the sacrum (more than three levels).

Sacral Anatomy

The sacrum is a physiologically and anatomically unique component of the axial skeleton. It functions to transmit forces from the mobile spine to the adjacent ilia to convert compressive forces from the lumbar spine to shear forces at the sacroiliac joints. Functionally the sacrum serves as a bipod with forces entering at the sacral promontory, bifurcating, and exiting through the alae and sacroiliac joints. This unique function necessitates unique anatomy. As a result, the sacral promontory and alae must withstand significant stress, and this drives the bony architecture of the sacrum on both a macroanatomic and microanatomic level.

The surface anatomy of the sacrum is unique, and it demonstrates the remnants of key vertebral features from the cephalad spine. It is formed by the fusion of five sacral vertebrae that retain many homologous features of the mobile spine but also exhibit features unique to sacral vertebrae ( Fig. 53-1 ).

Key bony landmarks of the dorsal surface of the sacrum.

The bodies of the sacral vertebrae are highly tapered and rapidly decrease in cross-sectional area, from cranial to caudal. Although the average anteroposterior (AP) length of the S1 body is 50 mm for men and 47 mm for women, at S2 this figure decreases dramatically, to 31 mm for men and 28 mm for women. Sacral vertebrae are also larger in the coronal than in the sagittal plane and have relatively small bodies.

The anterior costal and posterior transverse processes of S1–S3 are very large and broad, fusing to form the paired alae, which contain some of the denser bone found within the sacrum. Because of the fusion of the anterior and transverse processes, each sacral level must contain two foramina to transmit both the ventral and dorsal primary spinal rami. The ventral foramina are much larger, because they convey the substantial root contributions of the sciatic nerve. The comparatively small dorsal roots are accompanied by branches of the sinuvertebral artery and can be a source of brisk bleeding during dissection.

In spite of the substantial forces transmitted through the sacrum, much of the sacral cancellous bone is of low density and poor quality. This is consistent with the bipodlike function of the sacrum, because the densest bone is found within the sacral promontory and paired alae, but the remainder of the sacrum has poor quality bone, because it is functionally shielded from stress.

Anterior to the sacrum lie many important visceral structures that include the paired common and internal iliac arteries and veins, paired L5 nerve roots, sigmoid colon, middle sacral artery, and sympathetic plexus. All of these structures may be subject to risk during approaches to and surgical instrumentation of the lumbosacral junction. The safest region of the anterior sacrum is the paracentral portion of the S1 body; the vascular structures bifurcate above the S1 body and trace laterally, traversing far from the midline at the level of S1. The L5 nerve roots also course laterally and lie at the junction of the S1 vertebral body and ala. The only midline structure is the middle sacral artery, which lies directly on the midline of the anterior sacrum. With this in mind, the safest location for implants that exit the anterior aspect of the sacrum is just lateral to the midline within the S1 body. Implants that exit the cortex within the ala place the neurovascular structures at much greater risk. S1 screws are the safest of the bicortical implants, because their tips should lie within the medial aspect of the S1 body ( Fig. 53-2 ).

Neurovascular structures located on the anterior aspect of the sacrum.

Sacral Biomechanics

Several important biomechanical principals drive the choice of sacral fixation, especially for long fusions. One of the most important is the concept of the lumbosacral pivot point, described by McCord and colleagues. Centered within the posterior longitudinal ligament at the level of the L5–S1 disk space, it is a powerful predictor of the efficacy of lumbosacral fixation. As an implant extends further anteriorly from the lumbosacral pivot point, it will provide more construct stiffness and will support a higher load to failure. Implants that remain posterior to the lumbosacral pivot point impart little additional stability; those that extend far anterior, such as iliac bolts, impart substantial biomechanical benefits.

Due to the normal anatomic relation of the sacrum to the lumbar spine in an ambulatory adult, a typical S2 pedicle screw will not extend anterior to the lumbosacral pivot point, whereas an S2 alar screw does extend anterior and is biomechanically preferred. Among the common implants and techniques available for sacral fixation, only S1 pedicle screws, S2 alar screws, and iliac bolts will reach or extend anterior to the lumbosacral pivot point.

Triangulation of implants dramatically increases the pullout strength of a construct. Ruland and colleagues found that medially directed and crosslinked S1 pedicle screws outperformed either hooks or stand-alone pedicle screws. This reinforces to the biomechanical benefits of the medial S1 pedicle screw over alar screws or sacral hook constructs.

S1 Pedicle Screws

Pedicle screws placed into the S1 vertebra remain the mainstay of direct sacral fixation in modern spine surgery. The technique for placement is familiar to the majority of spine surgeons, and the quality of fixation is quite good if placed properly; it harmoniously integrates with posterior instrumentation of the lumbar spine, provides direct sacral fixation, and is relatively safe.

Several important details are required for placing S1 screws properly for both maximum fixation and safety. Placement of S1 screws is much like placement of lumbar pedicle screws, but a few anatomic factors can make placing an ideal S1 pedicle screw more difficult than screws at more cranial levels. The wide interpedicular distance at S1 and the need to medialize the tip of the screw to capture the dense bone of the promontory both necessitate a very medial projection of the S1 screw. However, the iliac crest can block the ideal path for the screw, especially in men with large iliac crests ( Fig. 53-3 and 53-4 ).

Ideal S1 pedicle screw placement.

Ideal S1 pedicle screw placement.