Sacrum

The sacrum is formed from five fused vertebrae in which the specially adapted and large transverse processes merge into thick lateral masses, the alae. The sacral spinal canal has four pairs of dorsal and ventral foramina. The subdural and subarachnoid spaces terminate as the thecal sac tapers at the caudal margin of S2. The filum terminale internum is an extension of the pia arachnoid of the conus medullaris, extending from the tip of the conus to the end of the subdural space. At the termination of the subdural space, the thecal sac tapers to invest the filum terminale internum and form the filum terminale externum. The filum terminale externum extends to the end of the sacral canal and attaches to the rostral portion of the coccyx.

Structures Adjacent to the Sacrum

For the safe placement and attachment of instrumentation constructs in the lumbosacralpelvic region, a thorough knowledge of the anatomic relationships of the neural, vascular, and visceral structures in the region is important. The common iliac arteries begin at the aortic bifurcation (L4 level) and pass along the lateral surface of the L5 vertebral body. They then bifurcate at the level of the LSJ, giving rise to the internal and external iliac arteries. The iliac arteries lie ventral and lateral to the iliac veins and therefore do not actually make contact with the spine. The internal and external iliac arteries pass ventral to the sacral alae. The internal iliac arteries pass close to the bony surface of the ala, whereas the external iliacs are separated from the bony surface by the psoas muscles. The lumbosacral trunk is formed by the ventral branches of the L4 and L5 nerve roots. It is joined by the sacral nerves located on the ventral surface of the alae between the iliac veins and the sacroiliac joint (SIJ). The sigmoid colon is also found in approximation to the ventral surface of the sacrum. It loses its mesentery and becomes far less mobile as it reaches the ventral aspect of the S3 vertebral body and becomes the rectum.

Sacroiliac Joint

The SIJ is formed by the interdigitating surfaces of the sacral alae and the iliac bones. It is predominantly a fibrocartilaginous amphiarthrodial (no synovial capsule) joint. There is a small diarthrodial (synovial capsule present) portion located at the ventral aspect of the SIJ. The interdigitation and matching contours of the iliac and sacral alar surfaces create an interlocking mechanism to help stabilize the joint. The wedgelike shape of the sacrum helps stabilize the SIJ and serves to transfer loads from the spine to the pelvis (Fig. 152-2).

FIGURE 152-2 The keystone configuration of the pelvis allows the transfer of weight from the spine to the pelvis and ultimately to the lower extremities.

The SIJ is essentially an immobile joint that functions as a shock absorber for the spine. In studies on fresh cadavers, there was minimal motion in pediatric specimens, and none in adults.⁵ Another cadaveric study has demonstrated that in adults older than 50 years of age, autofusion of the joint is observed in 75% of specimens.⁶

The major biomechanical function of the pelvis is that of transferring loads from the SIJ to the hip joints. The stable transfer of these loads is dependent on the ligaments connecting the lumbar vertebrae and the sacrum to the pelvis. The ligamentous structures spanning the SIJ include the interosseous, dorsal, and ventral sacroiliac ligaments (Fig. 152-3). The interosseous, sacroiliac, and dorsal sacroiliac complex provides the major stabilization for the SIJ.

FIGURE 152-3 Dorsal (A) and ventral (B) views of the major ligamentous attachments of the sacroiliac joint.

The iliolumbar ligament passes from the transverse process of the L5 vertebra to the iliac crests. A less substantial part of the ligament may span to the transverse process of L4 as well. The position of this ligament allows a wide range of motion in flexion and extension across the LSJ, but it severely restricts lateral bending and axial rotation.

The force vector of axial load from the spine is located ventral to the SIJ. This causes a ventral rotational tendency of the sacrum at the level of the SIJ. The center point of this rotational vector is located near the center of the S2 vertebral body (Fig. 152-4). The sacrospinous and sacrotuberous ligaments pass from the lower sacrum to the ischial bones. The position of these ligaments creates a long moment arm through which they are able to resist sacral rotation and are thereby able to maintain the lordotic lumbosacral posture despite the gravitational sagittal plane vector.

FIGURE 152-4 The center point of the rotational vector is located near the center of the S2 vertebral body. The diagram depicts the rotational tendency of the sacrum about this point. It is through the long lever arms created by the sacrospinous and sacrotuberous ligaments that this rotational tendency is counteracted and the sacrum is stabilized.

Muscular Interactions

The musculature of the lumbosacralpelvic region acts on the spine in a complex multidirectional fashion. The muscles and the weight of the upper body act in many instances via long moment arms and may place substantial forces on the spine. An example of such action is the force exerted on the spine by the rectus abdominis musculature. These muscles act by a moment arm extending from the pubic symphysis to the sternum, producing a spinal vector toward kyphosis. A pendulous abdomen does the same in providing a constant exaggerated spinal load in the upright position and to some extent during sitting.

In the resection of sacral tumors, stability of the sacropelvic region can be jeopardized because portions of the sacrum and possibly the SIJ are removed. Resection of the caudal portion of the sacrum up to the S1-2 interspace and removal of up to one third of the SIJ can be performed with only a 30% loss of weight-bearing capacity. The lower half of the S1 body and up to one half of the SIJ can be resected with a 50% loss of weight-bearing stability.⁷ Preservation of 50% of weight-bearing capacity is adequate for early ambulation in the postoperative period, and further stabilization is not likely to be necessary. In general, in cases of tumor or other destructive lesions of the sacral (e.g., infection), the bilateral alae should be evaluated. If one ala or more than 50% of bilateral alae are destroyed, the patient will require lumbopelvic fixation.

Fixation with Sublaminar Devices

The addition of sublaminar devices such as hooks, wires, or cables at the sacrum is a simple means of enhancing fixation. It should be borne in mind that the sacral laminae are often quite thin and may not provide substantial strength to hold such devices. All of these devices, including hooks, are tension-band fixators. Dorsally placed tension-band constructs provide limitation of flexion but do not provide substantial torsional stability or resistance to extension. With fusions at the LSJ, it has been shown that fixation with sublaminar devices alone does not improve the fusion rate when this rate is compared with that of noninstrumented fusions.²⁶ Therefore, sublaminar devices should be used at the LSJ only in conjunction with other devices.

Although the bony thickness at the sacral lamina is meager, the bone stock is usually quite substantial at the nearby dorsal foramina. Thus, the dorsal foramina are excellent sites to place hooks for the purpose of enhancing construct strength. When added to a sacral pedicle screw construct, this technique has been shown to significantly increase rigidity.²⁷

Fixation with Screws

A number of simple techniques use bone screws to enhance sacral fixation. Screws provide rigid three-dimensional stabilization and can be used for short-segment fixation at the LSJ. Shorter constructs provide a major advantage because immobilization of long segments of the spine increases the load, not only at the immediately adjacent segment but also at more distal segments.²⁸

Directing screws medially (pedicular) or laterally (alar) and then rigidly attaching them to the rod or plate greatly enhances pull-out resistance. Studies are in conflict regarding which of these two methods is better, but both offer a substantial biomechanical advantage over a straight sagittal-plane trajectory.^25,29 Oblique trajectories dramatically increase the cross-sectional area of bone that resists screw dislodgement via the cantilevered forces applied by the spine in flexion. If the longitudinal members are then linked with cross-members, a triangulation effect is created. This enhances both the rigidity of the construct and its pull-out resistance.^25,30,31 Torsional stability is particularly enhanced by cross-fixation of the rods.^30,32,33

Perhaps the easiest method for enhancing sacral fixation with screws is the placement of an additional pair of laterally directed bone screws into the sacral alae below the S1 level. This method provides a biomechanical advantage over a single pair of pedicle screws.^13,17 These screws are easily added to a construct without the need for preplanning or special devices, and fluoroscopy is not required. For this method, a type of screw whose attachment site to the rod is somewhat narrow is optimal. Screws with a broad attachment site may be difficult to place sufficiently close together for the optimal trajectory of the screws to be attained (Fig. 152-8).

FIGURE 152-8 If the site of attachment of the second screw is narrow, sufficient latitude is present to allow proper trajectory (A). If the attachment is too broad, it impedes proper placement of the screw (B).

The placement of screws into the S2 and even the S3 levels has been advocated to enhance the security of sacral fixation. Biomechanical testing has demonstrated that these screw placement sites do not add sufficiently to the security of fixation to warrant their use.^17,25 The thickness of the sacrum in the sagittal plane diminishes at the lower sacral levels. Therefore, inadvertent penetration of the ventral surface of the sacrum with such screws is more likely, and injury to the anatomic structures immediately adjacent to the ventral cortical surface may occur. However, such S2 alar screws can be used as more substantial anchor points if they are used as S2 sacroiliac screws, in which the screw transverses the SIJ and functions as an iliac screw in terms of longer length and iliac purchase.

The pedicular transvertebral screw fixation technique has been described for treatment of lumbosacral spondylolisthesis.³⁴ This technique involves placement of a long screw through the pedicle of S1, passing through the rostral end plate of S1 and into the caudal end plate of the L5 vertebral body. When combined with screws in the L4 pedicles, this provides a unique biomechanical construct. It appears to be a simple method for stabilizing high-grade spondylolisthesis. The simplicity of this technique is attractive, and the clinical results appear to be encouraging (Fig. 152-9).

FIGURE 152-9 With the pedicular transvertebral screw fixation technique, the sacral screw passes through the pedicle of S1 and into the inferior portion of the end plate of the L5 vertebral body. The upper end of the construct consists of L4 pedicle screws.

Along with placement of screws spanning from vertebra to sacrum, fibular bone struts can be applied in a similar manner. Placement of fibular grafts to act as dowels can be performed from a dorsal or ventral approach.^35–39

The procedure has been reported in use with autografts, allografts, and even with vascularized autografts.^36,37,40,41 Transvertebral fibular dowels should always be augmented with screw fixation.

Along with screw placement, sacral fixation can be supplemented by insertion of the rods directly into the lateral sacral bony masses. Jackson has described a method whereby screws are directed through the S1 pedicle, and cortical purchase is obtained in the vertebral end plate of S1 as well. The rods are then passed into the lateral sacrum toward the SIJ and affixed rigidly to the screws.

Devices for Sacral Fixation with Multiple Screws

A number of devices especially designed for sacral and sacropelvic fixation have been marketed. The two basic types are those that cross the SIJ and those that affix directly to the sacrum only. The sacral fixation devices are generally platelike structures that are secured to the sacrum with screws. Asher and Strippgen described one early example of these.⁴² It is called the Isola plate because of its butterfly shape (and is named after a species of Rocky Mountain butterfly). This design contained holes for the passage of multiple sacral screws. The position of the holes dictated placement of the screws, and anthropometric sacral measurements taken in male and female cadavers served as the basis for spacing of the holes.

The Tacoma plate (Medtronic Sofamor Danek, Memphis, TN) is designed to direct the additional screws to the sacral alae rather than to the pedicles. It is available in a two-screw and a three-screw model. Such devices are easily attached without special training beyond that required for pedicle screw placement.

Fixation across the Sacroiliac Joint

The most commonly used method for sacropelvic attachment is the Galveston technique or a variation of it. The procedure was originally described by Allen and Ferguson for the treatment of scoliosis.^43,44 It is accomplished by inserting angled rods into the iliac bones and passing them into the hard cortical bone above the sciatic notch (Fig. 152-10). This technique is useful for providing extremely rigid fixation in difficult cases, such as reconstruction after tumor resection or patients with lumbosacral agenesis.⁴⁵

FIGURE 152-10 The Galveston technique for sacropelvic attachment uses angled rods that are placed into the iliac bones along the transverse bar of the ilium. The rods are passed into the cortical bone above the sciatic notch.

In conditions of large preload, such as the correction of pelvic obliquity in neuromuscular diseases, loss of correction can occur because of the slippage of one rod relative to the other. Several manufacturers have developed cross-fixation devices that attach to both rods and prevent such slippage and markedly enhance the rigidity of the constructs. Although the attachment mechanism, ease of use, and biomechanics of these devices differ substantially, they all enhance construct rigidity and lessen the risk of failure in multilevel constructs.

Placing Galveston rods into the ilium can be cumbersome and technically difficult. Therefore, specialized screws for iliac insertion have been developed (Isola iliac screws, DePuy AcroMed, Raynham, MA). Iliac screws are placed by using fluoroscopy, and the rods are then either directly attached or joined to the screw by a connector fitting. These devices markedly enhance the rigidity of lumbosacral instrumentation constructs. The increases in rigidity are similar for both the iliac screws and the Galveston rods.¹⁷ Multiple studies have confirmed that iliac fixation provides the most effective means to supplement sacral screws.^46,47

Bicortical Iliac Screw Fixation

A simple method of sacropelvic fixation is the placement of long, variable-angle bone screws obliquely through the iliac bones (Fig. 152-11). When combined with additional sites of fixation at the sacrum, this technique creates a tripod effect for load sharing. This method has not been studied for its biomechanical characteristics, but early clinical results appear satisfactory.⁴⁸

FIGURE 152-11 The tripod geometry of variable angle screws that are passed obliquely through the iliac bones as well as the sacrum provides an increased load-sharing capability of the implant (A). Adding a rigid cross-member increases the rigidity and stability of the construct (B).

Bicortical iliac fixation is a simple technique to perform and does not require preplanning or specialized devices. It is important to pay strict attention to the screw trajectory and to ascertain that the screw tip will pass ventral to the lumbosacral pivot point. Also, a small section of the dorsal sacroiliac ligamentous complex must be removed to allow the screw to be placed in a low-profile configuration.

Revision Surgery

One of the major complications of lumbosacropelvic instrumentation is loosening or failure of hardware. In some cases, however, such loosening or failure may not prevent solid bony fusion. Therefore, it is a matter of judgment about when these constructs should be revised. Patients with loosened hardware should be followed closely with serial radiographic studies. Reparation is recommended if evidence of either progressive unacceptable deformity or symptomatic pseudarthrosis is observed. Because late healing may occur, it is advisable to delay revision surgery until it is certain that an acceptable result does not occur without revision.

Revision surgery of instrumentation constructs should address the reason for the failure. If there is a fractured implant, more fixation points should be included to improve load sharing. If screw pull-out is observed, the use of larger-diameter or longer screws provides greater thread purchase. If the pull-out appears to be largely due to poor bone quality, pressurized injection of polymethylmethacrylate may improve screw pull-out resistance.²⁵

In general, surgery for failed spinal instrumentation involves the implantation of more hardware. For failed sacral fixation, creation of a construct that spans the SIJ is advisable. Further bone grafting should always be performed by using autologous bone, if possible. Harvest of iliac crest bone may not be possible because many of these patients have had a previous harvest at the iliac sites. One limited alternative is the use of rib grafts. If iliac harvest is performed in patients with iliac fixation sites, the harvest sites should be placed as far ventrally as possible.

In cases of Galveston rod failure, additional fixation points can usually be secured. These rods tend to loosen by pull-out because they do not have purchase into the deep portion of the iliac bones. Further fixation should be attempted by placement of screws into the sacral pedicles or alae and attachment to the rod. Also, rigid cross-members should be added if these are not already present in the construct.

An alternative surgical approach may also be required for revision surgery. An interbody fusion is sometimes a reasonable alternative to augment a dorsolateral fusion. Ventral fusion at the lumbosacral junction may be accomplished by using interbody grafts after ventral discectomy or vertebrectomy for tumor, infection, or grade III or IV spondylolisthesis. Allograft bone is the usual substrate for grafting in cases of interbody fusion, but newer interbody cage devices that allow the use of autograft will soon be available. Because no rigid ventral stabilization device can be easily applied to the sacrum, either ventrally or laterally, such interbody grafts or cages should be supplemented with dorsal hardware, unless the decompression consisted of only simple lumbosacral discectomy in the absence of pathology in the posterior elements.

Screws may be placed through the sacral promontory retrograde into the S1 pedicles, but the bulk of the attached hardware usually produces risks of iliac vessel injury that outweigh any potential benefits. Because of this anatomic stumbling block, interbody cages should be viewed as supplemental devices rather than as primary stabilization hardware in any case of complex lumbosacral reconstruction. Many different types of interbody devices have been developed for placement via either open or endoscopic techniques (BAK cage, Ray cage, Brantigan device, Novus device, Harms cage). Some of these have received clearance from the Food and Drug Administration, and others are seeking clearance.

At the LSJ, it is far easier to place interbody grafts or devices via a true ventral (transperitoneal or extraperitoneal) approach or via a dorsal lumbar interbody approach than via a lateral (retroperitoneal) approach. The crossing iliac veins ventral to L5, the bony iliac wing that blocks the view, and the bulky psoas muscle at this level make the lateral approach to the lumbosacral junction uncomfortable at best. The ventral approach usually sacrifices the sympathetic plexus over the sacral promontory, sometimes producing ejaculatory dysfunction in males, and it may occasionally require sacrifice of one or both internal iliac vessels. The dorsal interbody approach usually involves the sacrifice of some or all of both facet complexes at L5-S1 and has a relatively high risk of nerve root injury. Regardless of which approach is chosen for interbody fusion at the lumbosacral junction, a good working knowledge of the pertinent anatomy and potential complications is critical.

Postoperatively, patients who have undergone revision surgery should be placed in thoracolumbosacral orthoses. The addition of a hip extension device to these orthoses has traditionally been recommended. However, it has been shown that spinal motion is not reduced by the addition of this cumbersome and uncomfortable attachment.⁵⁰

In the management of sacral tumors, involvement of the S1 segment is often the only impediment to what might otherwise be a potential surgical cure. Particularly in chordomas, the most common of sacral tumors, complete surgical resection with margins is usually considered curative because the tumor is usually indolent, metastasizing beyond its site of origin only very late in its course. Many such tumors are detected only when they are quite large, although they are still localized. The ability to perform a total sacrectomy for cure is impaired by the difficulty of reestablishing spinopelvic stability in the absence of the sacrum.

In a few cases, such stability has been accomplished by attaching ventral allograft bone struts to transpelvic plates at the lower end and to the remaining lumbar vertebrae above.⁵¹ Pelvic rim stabilization with transverse plates and stabilization with oblique allograft struts from the dorsal iliac wings to the telescoped lumbar vertebrae have also been attempted with some success.⁵² Graft positioning oblique to the primary ventral rotational force vector in the upright posture is the main clinical and theoretical impediment to the success of these techniques.

Research continues regarding new techniques of lumbopelvic reconstruction after total sacrectomy. In cadavers, whole allograft sacral transplants may be plated between the iliac wings ventrally to reestablish anterior column support and pelvic ring integrity. The transplants are supplemented with iliolumbar rods (Galveston technique) dorsally and generous volumes of autograft bone. Much work remains, however, before this technique can be considered for clinical use. A number of manufacturers are also offering advanced plating and connector devices for enhancement of fixation in these very difficult cases.