Dynamic Stabilization of the Lumbar Spine: Indications and Techniques

Fusion remains the standard method of controlling pain from lumbar spinal instability and is commonly performed in conjunction with decompression procedures. However, unintended consequences of otherwise successful lumbar fusion procedures may occur, including transfer of forces to adjacent segments, possibly resulting in expedited lumbar spondylosis, stenosis, and further instability at adjacent segments and potentially requiring reoperation. Less invasive surgical fusion techniques may reduce associated surgical trauma but ultimately do not alter the basic biomechanical changes associated with fusion.

There have been significant efforts to develop spinal systems that control a painful “dysfunctional” spinal unit without causing fusion, and in fact, a recent entire issue of Neurosurgical Focus was devoted to this topic. Some examples of these devices include facet replacements, interspinous process devices, and pedicle screw–based motion-controlling systems, and the more commonly used classes of devices will be discussed later in this chapter. Most available systems provide the surgeon with the ability to perform a fusion, use a nonfusion construct, or use a combination hybrid construct, although nonfusion pedicle screw–based constructs represent off-label use of the devices. Hybrid approaches allow the surgeon to “top off” a fusion level with a nonfusion construct, with the thought that the nonfusion device may help mitigate the stresses on the adjacent level and reduce adjacent-level facet and disc disease.

Regulatory Status

Currently, several nonfusion interspinous devices, including the Medtronic X-Stop (Minneapolis, MN), Paradigm Spine coflex (New York, NY), and Vertiflex Superion InterSpinous Spacer (San Clemente, CA), have been approved through the US Food and Drug Administration (FDA) premarket approval (PMA) pathway following investigational device exemption (IDE) trials. However, the FDA has not approved any pedicle screw–based system as a stand-alone device for nonfusion applications.

Currently, there are multiple pedicle screw–based motion-controlling devices that have been cleared by the FDA through the 510(k) process, but only when intended as an adjunct to fusion. The FDA recently provided further guidance and clarification regarding their regulatory stance on dynamic stabilization devices in their final order on December 30, 2016, defining these devices as semirigid systems (SRS). In this final order, they define rigid pedicle screw systems as “providing immediate rigid fixation to the spinal column as an adjunct to spinal fusion procedures,” while semirigid systems are defined as “systems that contain one or more non-uniform and/or non-metallic longitudinal elements (e.g., polymer cords, moveable screw heads, springs) that allow more motion or flexibility (e.g., bending, rotation, translation) compared to rigid systems and do not provide immediate rigid fixation to the spinal column as an adjunct to spinal fusion procedures.” Both rigid systems and SRS, which are both intended as adjuncts to spinal fusion, were downgraded to class II devices. However, clinical data will still be required as a special control for future clearance of SRS with any indication. Furthermore, for all currently legally marketed SRS with a prior 510(k) clearance, the manufacturers are required to submit clinical data as special controls in a 510(k) amendment by June 28, 2018, in order to continue legally marketing their devices.

However, the final order did not address truly dynamic systems, which are intended for nonfusion use, and these devices have remained class III devices that will still require approval through the PMA pathway.

Biomechanical Rationale

Both the spine surgeon and the implant design engineer benefit substantially from a clear understanding of the native function of the spine and the contribution of each of the spinal elements. The motion of the spine can be studied in the most basic form at a single level that begins with the functional spinal unit (FSU). The FSU comprises two vertebral bodies with their articulations, including the intervertebral disc, as well as the two posterior facet joints. The intervertebral disc forms an integral part of the FSU and has a high potential of becoming problematic, especially with age. Anatomically, the disc consists of fibrous layers arranged in alternating lamellar structures, in conjunction with a gelatinous inner core referred to as the nucleus pulposus. The other important articulations within the FSU are the facet joints, which are also susceptible to disease. Facets, in the normal condition, play a role in controlling the motion of the FSU. This three-joint complex within each FSU controls the kinematic response to load. The primary modes of loading taken into consideration are those associated with axial compression, flexion–extension bending, lateral bending, and axial torsion, as shown in Fig. 152.1 .

FIGURE 152.1, Demonstration of biomechanical testing parameters.

Over time, as degeneration occurs, the disc becomes fibrotic and less able to dissipate and distribute loads. Consequently, nonphysiologic loads are then distributed to the vertebral endplates and the annulus of the disc, which may lead to morphologic endplate changes and annular fissuring. With the onset of the degenerative cascade, both the intervertebral disc and the facets become compromised. Currently, it is unclear whether the onset of anterior column degeneration leads to posterior degeneration, and vice versa. Regardless, the degeneration within the FSU may lead to the inability to withstand even physiologic loads, and eventually, depending on the severity, instability may develop. Both clinically and biomechanically, instability can be defined as the inability of the FSU to control physiologic displacement. With instability, the neurological structures are prone to impingement and injury. Instability of the intervertebral discs shifts greater-than-normal motion and load to the facet joints. With time, these structures undergo hypertrophy, which combined with ligamentum flavum buckling from disc height loss can lead to substantial narrowing of the central neural canal, as well as of the lateral recesses and neural foramina.

Fusion has been the gold standard in the surgical treatment of many spinal pathologies with instability. Fixation instrumentation including screws, rods, plates, and intervertebral cages have served as adjuncts to fusion and positively contributed to radiographic outcomes (i.e., radiologic fusion assessment). However, the techniques have not always improved patient clinical outcomes—hence the advent of posterior, dynamic stabilization constructs that provide immediate stability but not necessarily rigidity that would allow transmission and propagation of nonphysiologic loads to adjacent segments.

Recent attention has been focused on preventing adjacent-level degeneration, as well as obtaining a positive functional outcome in the treatment of degenerative spinal pathology. To this end, numerous implants have been developed with the aim of effectively controlling physiologic loading and ensuring the appropriate kinematic motion at the index level. Posterior pedicle-based dynamic stabilization—as well as other motion preservation devices, such as total disc replacement, nucleus augmentation, total facet arthroplasties, and combinations thereof—are at various stages of evaluation and consideration, including as potential alternatives to fusion.

There has been a proliferation of pedicle screw–based dynamic or semirigid designs over the past 20 years. These systems all have some degree of similarity in that they employ pedicle screws connected by a rod or other linear element, which allows some form and degree of motion. As such, these devices appear to limit segmental motion but provide varying degrees of “stability” to the spine (biomechanically equivalent to increased stiffness in one or more of the six generally assessed biomechanical ranges of motion). There is some belief that standard rigid constructs may lead to too much stress shielding within the construct with insufficient loading to promote fusion, which is part of the design rationale for SRS used as an adjunct to fusion. For the purposes of this chapter, the recent FDA terminology defined above will be used; devices that are intended as an adjunct to fusion will be referred to as SRS, while devices intended for nonfusion will be referred to as dynamic stabilization devices.

A major design concern with any truly dynamic stabilization device remains that of robustness. Essentially, these devices are cantilever in design, as they transfer loads from the vertebral bodies through the pedicle screws and connecting units. Although this may “offload” the anterior and middle columns of the spine to a certain extent, the devices are theoretically required to bear this load for the life of the patient. Interestingly, traditional pedicle screw systems used as an adjunct to fusion are FDA cleared as “temporary” fixation devices, as they must only bear a load until fusion of the FSU occurs, which effectively shields the device from further stresses. They are tested and expected to perform for 2 years, in the absence of pseudarthrosis, after which they may be removed. Complicating the testing procedures for dynamic stabilization devices, the number of cycles that the lumbar spine undergoes in the normal patient is unknown and likely widely variable. Estimates vary from 250,000 to 2 million cycles per year. A dynamic stabilization device may need to withstand millions of cycles without breakage to last for the life of the patient. The device would need to do this without producing harmful wear debris or the patient suffering implant failure. This remains a difficult design paradigm to fulfill.

For instance, in a recent animal study, Luo et al. implanted Cosmic (hinged) pedicle screws (Arthur N. Ulrich Company, Germany) bilaterally in the L2–5 pedicles of goats. A unilateral rod or bilateral rods were placed, and the unilateral screws not connected by a rod were used as controls. Following 3 months of in vivo loading, the animals were sacrificed and the lumbar spines were explanted. While the screw pullout force for the unilateral and bilateral rod groups were similar, both these groups had an almost 70% reduction in screw pullout force compared to the control screws without a rod, and the formation of fibrous connective tissue at the bone–screw interface was observed. However, Karakoyun et al. found no difference in flexion, lateral bending, and axial rotation when comparing a Cosmic construct with a polyethylethylketone (PEEK) rod construct, though they remained less than the intact spine and were not dynamically loaded prior to biomechanical testing. Clearly, screw and construct stability of a dynamically loaded system is an area that needs to be investigated further.

Another design parameter that a potential manufacturer needs to address is that of the amount of stiffness required in an individual patient, which is often variable among different patients. An objective measure of assessment was described by Brown et al., who utilized an intraoperative spinal stiffness gauge that measured force displacement (Spinal Stiffness Gauge, Mekanika, Boca Raton, FL). Their findings of 655 motion segments demonstrated wide variability in degrees of stiffness. Generally, the L5–S1 segment was stiffer than L4–5, which was stiffer than L3–4. Male patients had stiffer spinal segments than did female patients. Older patients had stiffer spinal segments than did younger patients. Finally, stiffness was found to decrease by 20% following decompression. These basic findings raise the issue of patient-specific requirements and the potential need to customize implants based on intraoperative findings. Additionally, it remains unknown as to the required stiffness in an individual situation that will lead to pain relief and enough stabilization so as to prevent recurrence of symptoms and clinical (and/or potential radiographical) instability.

Along these lines, there remains variation among different SRS and dynamic systems, and at different numbers of instrumented levels. Perez-Orribo et al. compared 1-level and 2-level constructs with rigid and SRS with a mobile polyaxial screw-rod interface and found that the 1-level SRS construct maintained normal range of motion while the rigid system decreased range of motion by 51%. However, in 2-level constructs, SRS had significantly less range of motion than the intact level in all movements except axial rotation, but still maintained significantly more motion than the rigid system. Obid et al. compared three hybrid constructs and found no significant differences in range of motion from L3–5 in cadaveric spine explants in flexion and extension, though the specimens were not loaded dynamically prior to testing. Similarly, in an anterior lumbar interbody fusion (ALIF) cadaveric spine model, no differences in graft loading were detected between the SRS and rigid titanium rod construct, though again, the testing was performed immediately after implantation without dynamic loading. However, in a finite element study comparing the coflex interspinous device and Dynesys device at the L4–5 segment in flexion, extension, lateral bending, and axial rotation, both devices decreased range of motion by at most 21% in any direction, and the coflex device actually increased range of motion by 19% in flexion.

System Designs

Complex biomechanical and clinical needs have been addressed by designers with varying degrees of success. As previously stated, all pedicle screw–based SRS are similar in that they anchor an implant that allows for some degree of mobility to pedicle screws. Perhaps the most basic design type was the Graf ligament. This system limited movement in flexion, and had good early clinical results, but ultimately it never received larger clinical use.

More complex designs have included multiarticulated metal or metal-polymer systems that allow movements in different planes. These systems may include ball-and-socket components and spring assemblies and were designed to meet many of the aforementioned biomechanical goals, but the systems were quite complex in their mechanical design. They also have not been widely applied in clinical settings.

The systems that are intermediate or simpler in their design complexity include but are not limited to the N-Hance spine system (Synthes, West Chester, PA), Isobar system (Alphatec Spine, Carlsbad, CA), Dynesys Spinal System (Zimmer Spine, Warsaw, IN), and CD Horizon Legacy PEEK rods attached to titanium screws (Medtronic, Minneapolis, MN) ( Figs. 152.2 to 152.6 ).