Definitions
Adjacent segment disease (ASD) has been broadly and inconsistently defined in the literature. Other terms that have been used to describe ASD include junctional disease, junctional stenosis, or a transitional lesion. ASD has been applied to a plethora of radiographic findings suggestive of degeneration in a motion segment immediately adjacent to a spinal fusion construct with or without symptoms. Although the many symptoms included in studies range from minor symptoms to those necessitating surgical intervention, definitions generally do not include axial pain, numbness, or muscle spasms from the index operation. For the purposes of this chapter, we will define ASD as radiographic findings of adjacent segment degeneration with accompanying and correlative symptoms. We will define adjacent segment degeneration (ASDeg) as radiographic evidence of pathology in the adjacent segment without clinical sequelae.
Several imaging findings are consistently associated with ASDeg. On plain radiographs, one can observe disc height collapse, end-plate sclerosis, osteophyte formation, or spondylolisthesis. On computed tomography (CT) scanning and magnetic resonance imaging (MRI), the most common finding is disc degeneration, although facet hypertrophy and central or foraminal stenosis are also seen. There are two systems in the literature to describe ASD in the cervical spine, but there is no such accepted scale for ASD of the lumbar spine. The radiographic cervical ASD grading system proposed by Hilibrand and colleagues ranges from 1 to 4, with the higher numbers reflecting the increasing severity of adjacent disc and end-plate changes, as well as neural element compression on MRI, CT, and plain radiographs. The radiographic system proposed by Park and colleagues ranges from 0 to 3, with the higher numbers reflecting increasing degrees of ossification across the adjacent disc space. ASD in the lumbar spine has sometimes been described using scales designed for degenerative disc disease, such as the Pfirrmann or UCLA disc degeneration grades, although this is far from uniform. Unfortunately, our current conceptualization of ASD includes a heterogeneous group of patients, which makes it difficult to thoroughly, yet succinctly, encapsulate their radiographic patterns.
Biomechanical Pathogenesis of Adjacent Segment Disease
ASD can be viewed as an accelerated version of age-related lumbar spondylosis, which is briefly reviewed here. Most of the axial loading in the spine is transmitted through the disc space, which is comprised of the annulus fibrosus and nucleus pulposus. Through the aging process, the hydrated proteoglycan matrix of the nucleus desiccates, leading to the loss of disc height and increasing the load transmitted to the facet joints. Under supraphysiological loads, the facets degenerate, resulting in abnormal motion that can accelerate disc degeneration and spondylolisthesis. Disc herniation, reactive osteophyte formation, and buckling and hypertrophy of the ligamentum flavum can lead to neural element compression and clinical symptoms.
Lumbar arthrodesis alters physiological biomechanics and is thought to lead to ASD by transferring motion to adjacent mobile segments. A canine in vivo study by Dekutoski and colleagues found increased facet loading and motion at levels proximal to a fusion construct. Bastian and colleagues fused human cadaveric spines from T12 to L2 and found that the range of motion for the adjacent levels increased following fixation. Other human cadaveric studies not only corroborated these findings of increased motion but also found evidence of increased intradiscal pressure in adjacent segments. Although there may be contributions from the normal progression of degenerative disease, these biomechanical studies support the iatrogenic hypothesis in the pathogenesis of ASD.
Minimally invasive techniques probably represent the most studied attempt at preventing ASD, the prevailing theory being that reduced soft tissue disruption leads to less iatrogenic instability. Older studies, however, contend that the scar tissue formed from open surgery leads to less stability through the increased stiffness from scar tissue formation. There have been several attempts to study how traditional open surgery versus minimally invasive spine surgery affects the risk of developing ASD. One study by Yee et al. reported a trend toward a decreased risk of ASD in minimally invasive transforaminal lumbar interbody fusion (TLIF) compared with an open approach, although this trend was not statistically significant. Other studies, however, have failed to demonstrate any difference in ASD related to the surgical technique. A prospective study by Ekman and colleagues randomized 111 patients with isthmic spondylolisthesis to exercise or posterolateral fusion, and although there was an increased rate of radiographic ASDeg, there was no difference in the clinical outcome or reoperation rate, with a mean follow-up of 12.6 years.
Incidence
Describing the natural history of ASD is confounded by expected age-related degeneration of the spine. Data are conflicting and definitions are heterogeneous across studies. In a study of asymptomatic individuals, Boden and colleagues found degenerative changes in 57% of adults over 60 years of age. Rates of ASD range from 2% to 100% depending on the series. The exact timeline for ASD versus ASDeg has not been well established in the literature. In general, ASDeg that is symptomatic and that appears faster than the natural history of ASDeg is considered ASD. Of course, the length of time that is considered faster than expected is ambiguous and subjective. Cheh et al. found that 43% of their patients had ASDeg, but only 24% had clinical symptoms. Conversely, 6.3% had clinical symptoms without a radiographic correlation. A review of the literature by Park and colleagues found that the rate of ASDeg was significantly higher than the rate of ASD, suggesting that not all radiographic progression is clinically symptomatic or meaningful. When narrowing the focus to symptomatic patients, a review by Radcliff et al. and the retrospective study by Lee et al. both found an approximately 2% to 3% risk per year of developing ASD. Again, a similar result was reported in the retrospective series of Ghiselli and colleagues, who demonstrated an ASD risk of 16% at 5 years and 36% at 10 years. The time course for ASD has been studied both retrospectively and prospectively but, again, data are conflicting. The average duration from index operation to reoperation for ASD ranges from 6 months to 5 years.
Risk Factors
Risk factors for developing ASD have been inconsistently demonstrated in the literature. Age appears to be frequently associated with increased rates of ASD; one may rationalize that the aged spine may be less able to accommodate the altered biomechanical loads imposed by a fusion. ASD may also correlate with the surgical approach. Most studies have demonstrated a lower rate of ASD in anterior approaches versus posterior approaches ; this is an intuitive result, as anterior approaches avoid disruption of the posterior tension band at the adjacent levels. Unfortunately, in further tests of this hypothesis, minimally invasive posterior fusion techniques do not appear to convincingly decrease the rate of ASD. Furthermore, although loss of motion is considered a primary driver of ASD, motion preservation techniques such as lumbar disc arthroplasty or dynamic stabilization have not consistently been shown to reduce the incidence of ASD.
Existing deformity may also play a role, as sagittal imbalance and pelvic incidence to lumbar lordosis mismatch of greater than 10 degrees have been shown to increase loading of the posterior column in biomechanical studies and have been associated with higher rates of ASD in retrospective series. Indeed, two consistent risk factors identified in a review of the literature by Radcliff and colleagues were decompression adjacent to a fusion construct and ending the fusion at the apex of a deformity. Less consistent associations have been found for obesity, bone density, length of fusion, gender, and smoking status.
Preexisting adjacent segment spondylosis also has been examined; however, the findings have been heterogeneous and inconclusive. Although some studies have suggested disc degeneration at the level adjacent to the index operation as a risk factor, others have found that facet arthropathy predicts the development of ASD. It may therefore seem intuitive to prophylactically include these degenerated levels into the index fusion. However, Throckmorton and colleagues retrospectively studied patients who underwent fusion adjacent to a degenerated level versus adjacent to a normal level and did not find any clinical difference with a minimum of 2 years’ follow-up. These data suggest that surgeons should not routinely include levels for surgery that have radiographic evidence of degeneration but are not symptomatic, given the additional morbidity.
Association With Patient-Reported Outcomes
The association of ASD and ASDeg with validated metrics of pain and function has not been thoroughly quantified in the literature; this is likely owing to the heterogeneous definitions of ASD, ranging from including all patients with either radiographic or clinical evidence of degeneration to narrowly including only those patients requiring revision. Unsurprisingly, results have been equivocal or conflicting. In a highly cited study, Throckmorton and colleagues retrospectively studied 25 patients at least 2 years after a posterior lumbar fusion and paradoxically found higher Short Form 36 (SF-36) scores in patients with a degenerated adjacent segment on an MRI compared with patients with a normal-appearing adjacent level. Although they concluded that degenerative disc disease adjacent to a lumbar fusion “may not pose a significant clinical problem,” the study was limited in its small sample size, short minimum follow-up relative to the natural history of ASD, omission of a comparison between symptomatic and asymptomatic patients, and omission of preoperative SF-36 scores for patients. Cheh and colleagues retrospectively studied 188 patients who underwent lumbar or thoracolumbar pedicle screw instrumentation with a minimum of 5 years follow-up; they found that patients with ASDeg had worse Oswestry Disability Index (ODI) scores than those without ASDeg and that patients with ASD had worse scores than those without ASD. They noted that the preoperative ODI values were similar for the ASDeg and ASD groups. The paucity and quality of data do not allow firm conclusions to be drawn, especially regarding the relationship between ASDeg and patient-reported outcomes. By its very definition, ASDeg is asymptomatic, yet the fact that some data show worse function in these patients compared with those without ASDeg suggests either misclassification or that ASDeg and ASD lie on a spectrum of disease. We await future studies that examine these discrepancies and refine these diagnoses.
Prevention of Adjacent Segment Disease
Devices to prevent ASD have been developed on two premises: (1) that multisegmented constructs create a rigid, supraphysiological moment arm that stresses the rostral segment, and (2) that posterior decompression destabilizes rostral levels, leading to excessive motion and accelerated degeneration of facets and intervertebral discs. Destabilizing maneuvers include disruption of the posterior tension band (i.e., the supraspinous/interspinous ligaments and ligamenta flava) during decompression and disruption of the superior facet during pedicle screw placement. The two main categories of devices are designed to either stabilize the rostral adjacent segment or dynamically stabilize the segments that would have traditionally been fused.
Interspinous devices are typically polyetheretherketone (PEEK) or silicon blocks that are placed and secured between spinous processes. Commonly used devices include the Coflex (Paradigm Spine, New York, NY), Wallis (Zimmer, Warsaw, IN), and Device for Intervertebral Assisted Motion (DIAM; Medtronic, Memphis, TN). These devices were initially designed to be used as standalone methods of minimally invasive indirect decompression for lumbar stenosis; utilization for this purpose has declined dramatically in recent years secondary to higher reoperation rates, higher costs, and the inability to produce superior patient-reported outcomes compared with standard laminectomy. However, biomechanical and clinical evidence is growing for the use of a hybrid, or “topping-off” technique, in which an interspinous device is placed at the segment immediately proximal to a fusion construct. The goal of this technique is to create a gradual transition from fused to mobile segments and to offload the intervertebral disc and facet complexes. Results of retrospective and small prospective studies, as well as a recent meta-analysis, suggest decreased rates of both radiographic and symptomatic ASD using the topping-off technique in comparison with rigid fixation-only constructs. However, larger randomized studies are warranted to confirm these findings before the method can become more widely adopted.
Dynamic stabilization is a nonfusion technique that aims to maintain physiological load transfers to the adjacent segment by reducing the rigidity of the moment arm. No currently available system has been able to convincingly reduce the development of ASD. The most commonly implanted system worldwide is Dynesys (Zimmer, Warsaw, IN), the first iteration of which was first implanted in 1994 by Schwarzenbach and colleagues. Based on the ultimately unsuccessful Graf ligamentoplasty system, the Dynesys system comprises titanium pedicle screws linked by flexible polyethylene terephthalate cords to prevent excessive flexion; these cords are threaded within polycarbonate-urethane spacers to prevent excessive extension. Although biomechanical studies have demonstrated the ability of the Dynesys system to preserve range of motion, recent long-term data have offered conflicting evidence regarding its ability to reduce ASD ; one study even found an extremely high 29% 5-year risk of ASD in patients undergoing Dynesys implantation. Furthermore, several studies have raised concerns about the increased rate of surgical site infection in patients who underwent Dynesys implantation, which is possibly secondary to increased bacterial adherence to the braided cords and the increased length of operative time compared with traditional instrumented fusion with smooth titanium or cobalt-chromium rods. With yet unclear indications and revision rates comparable with or slightly higher than traditional fusion, it appears unlikely that Dynesys or any other dynamic stabilization system will soon be incorporated into standard practice.
Clinical Evaluation and Diagnosis
Although the most common presentation of ASD is central stenosis, a variety of symptoms and signs are associated with lumbar ASD, including back pain, deformity, neurogenic claudication, and lower extremity pain, numbness, or paresthesias. Whether a patient presents with myelopathy, claudication, or radiculopathy is of course predicated on the level and structural etiology of the lesion (e.g., herniated disc, hypertrophied facets, and/or thickened ligamentum flavum). Back pain is often a symptom of generalized spondylosis, however, pain of a mechanical nature or pain associated by progressive kyphosis may indicate adjacent segment instability; this may be further evaluated with imaging incorporating physiological axial loading, such as upright lumbar radiographs with flexion and extension. A patient presenting with progressive deformity warrants upright radiographs of the entire spine to evaluate global coronal and sagittal balance. Patients presenting with symptoms of neural element compression, such as conus medullaris syndrome, neurogenic claudication, or lower extremity radiculopathy warrant an MRI to refine the localization and severity of stenosis suspected on history and neurological examination.
Nonoperative Management
In patients with ASD who present with mechanical back pain, neurogenic claudication, or lower extremity radiculopathy in the absence of deficit, a trial of conservative measures is generally warranted. Common recommended modalities include physical therapy, antiinflammatory medications, and facet or epidural glucocorticoid injections. No strict guidelines exist on what constitutes a reasonable trial of conservative management, but studies commonly report a 3- to 9-month period without relief of symptoms as sufficient to warrant offering operative management. To date, there are no studies comparing the efficacy and durability of conservative versus operative management of ASD. This paucity of data is unsurprising, as revision surgery with its inherent morbidity is generally undertaken as a last resort or when conservative management is unreasonable (e.g., in the setting of a progressive neurological deficit).
Operative Management
Several surgical options are available for patients who fail conservative management of pain and in particular for those who develop neurological deficit, progressive deformity, or mechanical instability. Decompression-only surgery is generally not recommended because of high rates of reoperation for restenosis and symptom recurrence. The current standard approach is a posterior exploration of fusion, laminectomy, and/or foraminotomy for decompression of neural elements, and extension of instrumentation with pedicle screws to the newly degenerated segment. Pedicle screw fixation may also be supplemented by TLIF or posterior lumbar interbody fusion (PLIF), although the latter has been associated with higher rates of recurrent ASD than posterolateral fusion only—ostensibly because of excessive segmental rigidity. Patient-reported outcomes of posterior approach revision surgery from case series are favorable, with satisfactory or better results in up to 85% of patients and with up to 92% of patients resuming gainful employment. Unfortunately, these demonstrated benefits of posterior revision operations can be offset by higher rates of complications, blood loss, and healthcare costs. For example, over half of all incidental durotomies occur in patients undergoing revision surgery, with historical data showing a 5% to 8% rate of durotomy. Furthermore, Smorgick et al. found that posterior revisions are associated with 16% greater blood loss than index operations. Exposure of and determining compatibility with existing instrumentation only adds to the difficulties of revision.
One method to potentially reduce the large exposures associated with posterior revision and the risks of instrumentation incompatibility (especially in the cases of legacy instrumentation) is the use of cortical screws. Biomechanical studies of cortical screw trajectories in the lumbar spine have shown similar pullout strength and construct stiffness as pedicle screws. Although there is a relative paucity of long-term clinical studies of cortical screw constructs, early studies and a recent meta-analysis have shown similar patient-reported outcomes and radiographic fusion rates as pedicle screw constructs. Benefits of a cortical screw-based construct appear to include lower blood loss, shorter hospitalizations, and shorter incision lengths. When extending a fusion with cortical screws, one could choose to expose only the most rostral screw of the existing pedicle screw-based construct, place a cortical screw at that level and link them with an offset cross connector. This hybrid construct method has been used as a pedicle screw rescue technique and as an osteosynthetic technique for flexion distraction injuries, but it is not yet proven in revision surgeries for ASD.
Owing to the morbidity of posterior revision surgery, the less invasive lateral lumbar interbody fusion (LLIF) has been proposed as an alternative method of decompression and stabilization ( Fig. 22.1 ). This approach avoids disruption of the posterior ligamentous complex, dissection through postoperative scar, and the perils of inadvertent durotomies given previous laminectomy defects. Biomechanical studies suggest that although extension of posterior instrumentation provides the most stability in all planes of motion, LLIF with a lateral plate provides sufficient stability to allow bony union to occur. Early retrospective case studies of LLIF using standalone cages for ASD suggested high rates of symptom relief related to neural element compression, as well as high rates of bony fusion. Louie et al. conducted a single-institution retrospective cohort study of 47 patients suffering from symptomatic ASD after lumbar fusion surgery who underwent standalone LLIF or posterolateral fusion. They found that improvement in patient-reported outcomes (visual analog scale and ODI) and radiographic fusion rates were similar, and that patients undergoing LLIF benefited from shorter operative times, less intraoperative blood loss, and shorter hospitalizations. In other small retrospective studies, similar clinical and radiographic outcomes have been observed when comparing LLIF with TLIF and PLIF for revision of rostral ASD. Whereas these LLIF studies are promising, selection bias has likely played a major role in generating positive results; the pathology in patients who undergo LLIF must be amenable to indirect decompression and must not require large osteotomies for correction of an ASD-related deformity. We foresee a growing role for LLIF in the treatment of ASD, but the posterior revision will likely remain the workhorse approach and the one chosen for more severe cases of ASD, despite its relative morbidity.
