Overview

In the late 1890s, Hadra of Galveston, Texas, stabilized a cervical fracture-dislocation in the first modern report of spine instrumentation.¹ He later used this technique in Pott disease.² For the next 100 years, cervical instrumentation remained limited to various posterior wiring techniques, of which Rogers’ technique was the most frequently employed.³ In the 1980s, wiring patterns began to include corticocancellous bone struts for added extension stiffness.⁴ In the last 2 decades, rigid, segmental fixation, including lateral mass and pedicle screws, has dominated.

The era of anterior instrumentation began after Robinson and Smith⁵ popularized the anteromedial approach to cervical disc disease in the 1950s. The simple plating systems that evolved from appendicular stabilization were fraught with loosening, backout, and other, devastating soft tissue consequences. Dedicated anterior cervical plating systems were first described in the 1970s. Bicortical screw purchase decreased screw backout and instrumentation failure rates but added the risk of canal penetration and cord injury risk. In the 1980s, unicortical locking mechanisms increased bone purchase, while preventing screw migration.⁶

Cervical instrumentation continues to evolve with new disc replacement systems, dynamic and low-profile anterior plates, cervical cages, and resorbable implants. Although promising, some of these newer technologies have been implemented without evidence of added benefit. Before recommending an implant system, American surgeons must consider its U.S. Food and Drug Administration (FDA) status (Tables 70-1 and Table 70-2). Often a device is cleared for some, but not all, of its intended indications. The reasoning for FDA decisions is often obscure and bureaucratic and does not reflect only safety or efficacy issues. The device’s package inserts should be read by surgeons, and questions should be directed to the manufacturer’s legal counsel or to the FDA (1-800-638-2041).

TABLE 70–1 U.S. Food and Drug Administration (FDA) Status of Forms of Cervical Spine Instrumentation

* FDA status refers to the most common use of the device as described in the chapter text. Virtually all of these devices have FDA-approved uses. The status of these devices is constantly evolving. A, approved; I, investigational; O, off-label.

TABLE 70–2 U.S. Food and Drug Administration (FDA) Classes

The FDA is not empowered to dictate patient care. Off-label use of cervical instrumentation may occur in two settings, each establishing different demands on the physician. The “practice of medicine” includes use of FDA-cleared, marketed devices for indications not listed on the FDA-approved labeling. When the indication or intended patient population lies outside the device’s labeling, surgeons may still legally use the device according to their best judgment, but they must be able to support the decision with reliable scientific evidence. It is prudent, although not specifically required, to discuss the FDA status and rationale of the proposed implant with the patient. No investigational device exemption or institutional review board review is needed in “practice of medicine” cases. The second off-label use setting involves experimental or investigational devices. If the FDA has not cleared the device for marketing for any indication, clinical use requires an investigational device exemption. The implant in this situation may be used only in accordance with the approved protocol’s plan of investigation. A separate, formal informed consent must be obtained from the patient. Physicians involved in the study may not share the device with other physicians.

Cervical spine implants are employed in wide-ranging indications, including trauma, tumor, deformity, infection, and degenerative disease. In each group, the goals are the same; implants are used to reduce deformity, provide stability, and share loads with host and graft tissues until healing occurs. Most of these devices aim to support the fusion process. Some implants, such as cervical laminaplasty plates, are approved as “bone graft containment devices.” Newer devices seek to improve on their predicates by emphasizing previously ancillary goals, such as minimization of adjacent segment degeneration, surgical morbidity, iatrogenic neurologic deficits, and unintended level fusion. Interest in motion-preserving devices is increasing to reduce the risk of adjacent-level degeneration. This chapter discusses the evolution, biomechanics, indications, outcomes, and complications of implants to assist spine surgeons in rational selection.

Biomechanics Introduction: Selecting a Biomechanically Correct Implant

The most common, preventable cause of instrumentation failure is related to errors in surgical judgment. Typically, modern implants are overengineered for their designated function. Direct failure of the implant is more likely because of improper selection or fatigue. The average spine cycles 3 million times per year.^7,8 If bone healing fails to occur, all implants ultimately fail, either at their anchor points in the bone or in the material itself. The novelty and technical challenge of safe implant placement should not divert the surgeon’s attention from meticulous preparation of the fusion bed and grafting technique. Occasionally, misplaced implants fail. Careful surgical exposure and intraoperative radiographic confirmation reduce misplacement (Table 70–3). More typically, failure occurs when the surgeon fails to understand fully one of the following four things:

TABLE 70–3 Important Factors in Selection of Cervical Implants

In contrast to typical fracture healing, which passes through Hunter’s stages of bone repair (inflammation, soft callus, hard callus, and remodeling), most modern cervical implants seek primary bone healing in which osteon cutting heads cross segmental gaps directly.^9,10 This approach requires near-anatomic alignment and rigid stabilization. Excessive strain or poor bone-to-bone contact stimulates fibrous tissue deposition and, ultimately, construct failure. Successful use of cervical implants requires understanding of their biomechanics.^11–13

No clear line divides a “stable” from an “unstable” spine. In serial sectioning studies, White and colleagues¹⁴ concluded that more than an 11-degree increase in sagittal angulation or more than 3.5 mm of sagittal plane translation represented instability. These values are most helpful in the acute trauma setting but are less meaningful with chronic destruction, such as infection. In many cases, cervical instrumentation is meant not to correct any innate spinal instability, but rather to reverse or prevent iatrogenic instability associated with decompression. Because each anatomic structure contributes to normal stability and kinematics, it is important during decompression to minimize surgical disruption of intact structures.¹⁵ A classic example of iatrogenic spinal destabilization is postlaminectomy kyphosis.¹⁶ The difficulty in treating this condition fostered interest in laminaplasty.

The surgeon must ask: What is unstable, which planes are affected, and how badly? When the surgeon has delineated the “personality” of the pathology, he or she must consider the implant options. That is, having decided what the implant is supposed to treat, the surgeon reviews the options with a view toward the limitations and risks of each implant. Implant risks are not universal and decrease with the experience of the surgeon. Various tools, such as fluoroscopy, navigation systems, and intraoperative monitoring, may improve safety. A surgeon who is comfortable with a wide array of implants and techniques can tailor the treatment to the patient’s needs.

In trauma, mode of failure is of paramount importance. Flexion instability is best treated with posterior stabilization. In addition, transverse atlantal ligament rupture leads to flexion instability, whereas posteriorly displaced dens fractures are unstable in extension. Posterior C1-2 wiring is mechanically more rational for transverse atlantal ligament rupture than for displaced dens fracture.

The surgeon needs to account for specific patient factors when planning cervical instrumentation. On one hand, children have excellent healing potential and may require less rigid fixation.¹⁷ On the other hand, even with excellent graft carpentry and implant placement, osteoporosis increases segmental motions and decreases construct pullout and fatigue strength.¹⁸ Insertional torque, pullout strength, and bone mineral density are highly correlated.^19,20 Good bone mineral density has a greater positive impact on pullout strength than bicortical purchase.²⁰ Adding polymethyl methacrylate (PMMA) to the screw tract significantly increases the torque and pullout strength of the screw.^21,22 Bone loss may affect pedicle screw fixation more than C1 lateral mass fixation because pedicle screws engage cortical bone rather than the cancellous bone seen in the lateral mass.²³ More rigid fixation may be required in association with other physiologic factors, such as challenged healing environments after chemotherapy and radiation therapy. Larger patients, poorly compliant patients, diabetics, and smokers may require a more aggressive, rigid implant strategy.²⁴

Early attempts at surgical fixation were complicated by infection, devascularization, inadequate metallurgy, and metal allergy. Better antisepsis, soft tissue handling, and materials evolved. Subsequently, a limited understanding of bone biology and mechanics resulted in poorly conceived implants and techniques. Over time, biomechanical studies improved implant design, although they underestimated the importance of soft tissue and muscular tension. Most biomechanical research is limited by virtue of its ex vivo nature. Cadaveric, animal, or plastic spines are tested in laboratory settings with various pure or complex loads, but these studies do not take into account the importance of muscle forces, tissue healing, or the possibility of gradual ligamentous relaxation (creep).^25,26

It is important to limit unnecessary exposure, denervation and devascularization of the paraspinal muscles. Compromised extensor musculature allows collapse into kyphosis above or below the instrumentation.²⁷ Careful muscular repair may protect the construct.²⁸ In the posterior cervical spine, a multilayered closure that includes the suboccipital triangle and ligamentum nucha improves muscular balance, decreasing eccentric implant loading.^27,29

Additional important surgical techniques to augment fixation include increasing the strength of any spinal construct with added fixation points, triangulated placement, and aiming for the dense subchondral bone of the vertebral endplate. Extending the duration of postoperative immobilization can also decrease the likelihood of implant failure.³⁰ The relative merits of unicortical versus bicortical screw purchase continue to be debated.³¹ In anterior and posterior applications, bicortical screws exhibit significantly greater holding power in terms of immediate pullout strength and fatigue resistance.^16,32–34 Even in the trauma setting, unicortical fixation maintains reduction and confers high fusion rates.³⁵ The advantages of bicortical purchase are magnified in osteoporosis, in wide decompression with potential instability, in multilevel procedures, and when fixation points are limited.^19,36,37

Relevant Anatomy for Spinal Instrumentation

For the purposes of reconstruction, the cervical spine is divided into three regions: the cervicocranium, mid-cervical spine, and cervicothoracic junction. Unique anatomic and biomechanical considerations of each region influence the instrumentation chosen.

The cervicocranium includes the skull base, atlas, and axis. The size, shape, and location of the cervicocranial joints allow more motion than the joints in the subaxial spine and render arthrodesis more challenging.^51,52 The bony elements of the cervicocranium, beginning with the occiput, are unique. The clivus ends in the basion, the anterior border of the foramen magnum. The opisthion refers to its dorsal border. From the foramen magnum, the occipital squama curves 90 degrees cranially toward the inion (or external occipital protuberance [EOP]).

Embryologically, the C1 vertebral body is absorbed into the dens. The absence of a C1 vertebral body means there is no disc between occiput and C1 or between C1 and C2. The posterior arch forms two thirds of the ring of C1. There is no C1 spinous process, but rather a posterior tubercle to which the rectus minor and suboccipital membrane attaches. The vertebral artery runs along a groove on the cranial surface of the posterior ring that becomes very shallow beyond 1.5 cm from the midline. Exposure of the ring’s superior aspect risks injury of the vertebral artery. C1 has a lateral mass on each side, but no pedicle or laminae. Although the arches are thin, the lateral masses are heavy, thick structures, each with a concave superior articulating surface.

The occiput-C1 articulation includes convex occipital condyles lateral to the foramen magnum articulating with the concave C1 lateral masses. Normal occipitocervical extension is limited to 21 degrees when the occiput abuts the C1 posterior arch.^53,54 More than 8 degrees of rotation between the occiput and C1 is pathologic. In children, the flatter occiput-C1 joints are less able to restrict motion, predisposing them to injury.⁵⁵

Atlantoaxial motion occurs through two sets of two joints. First, the slightly convex inferior facets of the axis meet the slightly convex superior facets of the atlas. These joints are oriented in the horizontal plane and have no interlocking bone to prevent subluxation. They allow 43 degrees of rotation, nearly half of normal cervical rotation.⁵⁶ The second set of atlantoaxial joints arises from the cranial projection of the odontoid projecting into the axis ring. The dens acts as the focal point of a network of ligaments providing resistance to translation, flexion, extension, and rotation. Dens resection leads to vertical and atlantoaxial instability.⁵⁷

The axis, the largest and heaviest cervical vertebra, bears a large, bifid spinous process. The bony isthmus between the facets is often called the pedicle. Technically and anatomically, this represents the C2 pars interarticularis and is a large dense structure that projects medially at 30 degrees and superiorly at 20 degrees. The short, stout, nearly horizontal C2 pedicle lies between the C1-2 facet and the vertebral body.

The lower cervicocranium (C2-3) transitions into the more homogeneous subaxial patterns.^58,59 The C3-6 vertebrae exhibit a uniform configuration but gradually increase in size distally. The vertebral bodies are roughly twice as wide as they are deep. Each contains a body; paired pedicles and articular masses; laminae; and a single, spinous process. The transverse process projects laterally from the superolateral aspect of the body and anterior surface of the articular mass and contains the foramen transversarium. The transverse process ends in anterior and posterior tubercles. At C6, the prominent anterior (carotid) tubercle can be palpated for intraoperative localization.

Morphometric data from 100 computed tomography (CT) studies revealed mean vertebral body widths of 24.6 mm in men and 23.0 mm in women.⁶⁰ The narrowest vertebral bodies measured 17 mm in men and 14 mm in women. In the midsagittal plane, the average anteroposterior diameter was in 18 mm in men (smallest 13 mm) and 16 mm in women (smallest 10 mm). A morphometric analysis of critical cervical pedicle dimensions recorded a wide range of values and only fair interobserver correlation.⁶¹ Transverse angulation was fairly constant at 40 degrees. Relative to the lateral mass axis, C3 and C4 pedicles were oriented superiorly, whereas C6 and C7 were oriented inferiorly. The dorsal entry point of the pedicle on the lateral mass, defined by transverse and sagittal offset, had similar mean values with wide ranges and variable topography. Sufficient variation exists to preclude safe pedicle instrumentation using topographic landmarks alone. Ludwig and colleagues⁶¹ recommended laminoforaminotomy or image guidance to place these screws.

Because bone size varies considerably, preoperative planning using axial and sagittal images decreases the risk of screw placement. It is important to verify that the C2 isthmus is large enough to accommodate a 3.5-mm screw. If bony element size is questionable on magnetic resonance imaging (MRI), a 2-mm cut CT scan limited to the levels of surgical interest should be obtained. It is important to specify that the CT gantry be reangled to be parallel the endplate at each disc level. Clear preoperative measurements of the bony elements allow larger screws to be employed, improving pullout and fatigue strength characteristics. Similarly, preoperative planning allows improved screw trajectory to incorporate better triangulation and subchondral bone purchase. Upper cervical anterior plating may benefit from coronal CT or MRI above C3 because anatomic variation is considerable and may make this technique inadvisable in 20% of cases.⁶²

The uncinate process projects cephalad from the inferior mid-cervical vertebral endplates. The immediately superior vertebral endplate receives the uncinate via a contiguous lateral indentation. Together, the process and indentation form the synovial uncovertebral joint (of Luschka). Biomechanically, the uncovertebral joints regulate extension and lateral bending. The posterior uncovertebral joint has a secondary role in torsional control.^63,64

Posterior midline cervical approaches first encounter the spinous process. The bifid process may be taller on one side or the other. The surgeon must pay critical attention to the midline. There is usually no sharp demarcation between the spinous process and the lamina. Surgeons performing spinous process wiring techniques must exhibit great care to avoid inadvertently entering the spinal canal. Proceeding laterally, an inferior notch is typically encountered at the junction between the lateral mass and the lamina. At the medial boundary of the lateral mass, this notch serves as an excellent landmark for lateral mass fixation and en-bloc decompression procedures. The articular masses (or pillars) are dense, heavy, rhomboid structures formed by junction of the superior and inferior articular processes. Average facet inclination is 35 degrees from vertical.

Successful and safe occipitocervical instrumentation requires a detailed understanding of spinal vascular anatomy. Placement of occipital screws risks dural sinus injury. Venous sinus injury is especially likely with screw placement within 1 cm of the EOP.⁶⁵ The internal carotid artery runs a mean 2.9 mm from the anterior C1 lateral mass and is at risk with Magerl and Harms C1 screw constructs.⁶⁶ Internal carotid artery injury could lead to life-threatening hemorrhage and stroke, although there are no clinical reports of such injury to date. In a retrospective review of 50 random contrast-enhanced CT scans of the head and neck, the artery was considered at least at moderate risk of injury in 58% of cases.⁶⁷ Medial screw angulation decreases risk.

Vertebral artery injury may occur with anterior and posterior cervical procedures. Although clinically evident injuries are rare, catastrophic consequences include fistulas, pseudoaneurysm, cerebral ischemia, and death.^68,69 Because the vertebral arteries are paired, injury to one rarely results in significant neurologic deficit. If intraoperative vertebral artery injury is suspected, it is imperative not to attempt exposure or screw placement on the contralateral side, for fear of inadvertent injury to the other vertebral artery. In anterior procedures, too lateral a dissection within the vertebral body puts the vertebral artery at risk.⁶⁹ It is important to mark the midline so that inadvertent excessive lateral dissection is avoided. Posteriorly, the vertebral arteries are vulnerable to injury during insertion of Magerl and C2 pars and pedicle screws as the drill bit traverses the C2 body. In patients with rheumatoid arthritis in particular, a high rate of ectatic and variable arterial courses has been reported and may preclude safe screw placement in 20% of cases (Fig. 70–1).^68,70,71

FIGURE 70–1 A and B, Asymmetric course of right vertebral artery at C2 (A) and C6 (B). Risks of screw placement vary significantly from right to left sides.