Adjacent segment disease (ASD) was described after long-term follow-up of patients treated with cervical fusion. The term describes new-onset radiculopathy or myelopathy referable to a motion segment adjacent to previous arthrodesis and often attributed to alterations in the biomechanical environment after fusion. Evidence suggests that ASD affects between 2% and 3% of patients per year. Although prevention of ASD was one major impetus behind the development of motion-sparing surgery, the literature does not yet clearly distinguish a difference in the rate of ASD between fusion and disk replacement. Surgical techniques during index surgery may reduce the rate of ASD.
Anterior cervical decompression and fusion (ACDF) surgery for cervical spondylosis is often considered to be one of the most successful and predictable procedures that spinal surgeons perform, with high success rates reported in the literature at short-term or medium-term follow-up. Despite the relative success of this intervention, one potential source of postoperative pain and disability is symptomatic degeneration at adjacent spinal levels or adjacent segment disease (ASD). This phenomenon was first described clinically in the lumbar spine after several biomechanical cadaveric studies suggested that the greater demands placed on intervertebral disks adjacent to fused vertebrae may lead to accelerated disk degeneration. Subsequently, the same phenomenon was described in the cervical spine after ACDF. In addition to the biomechanical explanation, other potential causes of ASD include injury to musculotendinous, disk, or bony structures during index surgery and natural progression of spondylosis leading to new symptom onset.
ASD has been defined as the “development of new radiculopathy or myelopathy referable to a motion segment adjacent to the site of a previous anterior arthrodesis of the cervical spine.” This definition does not require that new symptoms are treated surgically. Symptoms treated surgically, those treated with procedures such as epidural injections, and symptoms treated nonsurgically may all qualify so long as the symptoms are related to an adjacent motion segment based on clinical and radiographic evaluation. In contrast, the similar but distinct concept of adjacent segment degeneration involves the development of new radiographic evidence of spondylosis at a motion segment adjacent to a prior fusion with or without associated symptoms. Despite the clear definition between these 2 terms, they are often used interchangeably in the literature. The reoperation rate cannot typically be used as a proxy for ASD; reoperation may address incomplete decompression at the index level, new pathology at the index level that was not present at index surgery, or new pathology at adjacent or noncontiguous levels, each of which has a distinct pathoanatomy and etiology.
Because of the relative frequency with which ACDF is performed, the consequences of ASD may affect many patients. In recent years, more than 200,000 primary ACDF procedures have been performed annually in the United States. If data from studies with the longest available follow-up after ACDF are correct and the rate of secondary surgery for ASD at 10 years is more than 15%, a rough calculation suggests that more than 30,000 patients may require surgery for ASD on an annual basis, an estimate that could underestimate the need for reoperation if the rate of ACDF continues to increase as the US population ages.
Because of the potentially large number of patients affected, ASD presents a significant challenge for spine surgeons to develop techniques that minimize its development. When ASD does occur and symptoms cannot be managed through conservative treatment options, revision operations are frequently encumbered by postsurgical scarring and previous instrumentation. This article characterizes the incidence of ASD after both fusion and motion-preserving surgery, discusses potential preventative strategies that can be used at the time of index surgery, reviews the effectiveness of various conservative therapy options, and discusses surgical considerations for treatment of ASD to select a surgical option and optimize results when conservative therapy is ineffective.
Clinical evidence of ASD
ASD After ACDF
When studying ASD, clinical research must have a sufficient duration of follow-up to capture the time period during which ASD develops. Although studies with short-term follow-up have occasionally described substantial rates of ASD, most series with short-term follow-up describe high success rates after ACDF and 1 or 2 years of surveillance is typically not sufficient to document ASD. With the passage of time, however, adjacent-level degeneration, which is often clinically silent initially, leads to associated symptoms; the need for reoperation at adjacent levels is commonly described in articles with longer follow-up. Thus, long-term studies after ACDF provide the most accurate evaluation of the incidence of ASD. Gore and Sepic described a series of 50 patients who were followed up for an average of 21 years after uninstrumented ACDF. Sixteen of the 50 patients had significant recurrent symptoms and 8 patients required surgery for degeneration at an adjacent level. The 8 patients who had recurrent symptoms but did not require surgery are not described in detail, so it is unclear whether these patients had same-level or different-level spine pathology. The rate of ASD thus lies between 16% and 32%. Hilibrand and colleagues specifically set out to characterize ASD in a series of 374 patients who underwent ACDF. This investigation clearly describes ASD treated both operatively and nonoperatively; in total, ASD developed after 14.2% of the ACDF procedures included in the study, which included 7.2% of patients who underwent secondary surgery. Hilibrand and colleagues reported an annual incidence of 2.9% and estimated the prevalence of ASD at 10 years to be approximately 25%. Ishihara and colleagues performed a study of 112 patients with an average follow-up of more than 9 years, documenting ASD in 19 patients (17%), including 7 (6.3%) who underwent a second surgery at an adjacent level. Yue and colleagues similarly described the need for reoperation due to ASD in 12/71 (17%) patients treated with ACDF at a follow-up of 7.2 years. Assuming an annualized rate of between 2% and 3%, long-term studies are necessary to accurately study ASD because most studies with short follow-up will not demonstrate stable rates of ASD.
Degeneration in the Native and Unfused Cervical Spine
The intervertebral disk begins to degenerate before any other part of the musculoskeletal system and reliably distinguishing ASD from the natural history of progressive spinal degeneration over time is difficult, especially in an individual with spondylosis severe enough to warrant the index surgery. Some insight into the question of whether ASD occurs after spinal fusion above and beyond the rate of degeneration expected without the influence of spinal fusion can be gained from the study of the natural history of patients treated either without surgery or without fusion. Although these studies could potentially provide a rough approximation of the rate of developing symptomatic cervical spondylosis unrelated to fusion, results have been variable. Gore described a series of 159 individuals without any symptoms referable to the cervical spine who were followed up longitudinally for 10 years with radiographs and clinical history. During this time period, 15% of the participants developed symptoms related to their cervical spine. In addition, Gore identified degenerative changes at C6-C7 on initial examination as a factor that increased the risk of subsequent development of symptomatic cervical spondylosis by more than 4-fold, suggesting C6-C7 should be evaluated carefully before surgery in patients set to undergo single-level ACDF at C5-C6 or C7-T1. Acikbas and colleagues performed a retrospective study of 32 patients treated with anterior cervical diskectomy without attempted fusion for myelopathy and/or radiculopathy with an average postoperative follow-up of 4.8 years. At final follow-up, 8 patients (25%) had developed symptoms attributed to adjacent-level disease. The rate of ASD was nearly identical for a cohort included in the same study treated via posterior cervical fusion (23%), suggesting that ASD was not exclusively related to the anterior approach or associated injury to anterior musculature or osteoligamentous structures. Clarke and colleagues presented a series of patients who developed ASD after posterior spinal foraminotomy. At a follow-up of 7.1 years, 15 patients (4.9% of the cohort of 303 patients) were treated for ASD and the investigators estimated a 10-year follow-up rate of 6.7%. Although the investigators suggested that this rate approached the rate of development of symptoms related to cervical spondylosis at a native motion segment, this study only identified ASD for patients who followed up at the institution where the study was performed and did not contact patients to inquire about treatment elsewhere, likely underestimating the ASD rate.
Factors Influencing ASD After ACDF
Multiple studies have attempted to identify factors that are related to the development of ASD after ACDF. Lower cervical levels have been found to have higher rates of ASD when not included in the index fusion. Hilibrand and colleagues described significantly higher rates of ASD at C5-C6 and C6-C7, the same motion segments that have the highest rates of degeneration in the native spine, and identified increased motion in levels that later degenerated. These findings linking pathologic instability and degeneration have been both supported and refuted by subsequent studies. Hilibrand and colleagues suggested using careful presurgical evaluation of these segments to consider extension of fusion when planning a single-level fusion with preexisting degeneration at C5-C6 or C6-C7 in order to prevent ASD. In a retrospective review comparing patients who had undergone ACDF with asymptomatic volunteers, Matsumoto and colleagues similarly found that patients had a higher likelihood of degenerative changes at adjacent levels 10 years after lower cervical ACDF (C5-C6 or below) compared with those who had undergone ACDF at more cranial levels. Matsumoto and colleagues found, however, that preexisting degenerative changes were not predictive of degeneration at final follow-up, a finding supported by other researchers, although it must be noted that this is only a radiographic study. There seems to be little consensus about the importance of preexisting degeneration and increased motion in the progression of radiographic degeneration, although relative strengths of the Hilibrand study are substantially longer follow-up and evaluation of clinical symptoms rather than just degenerative changes on imaging.
ASD After Total Disk Replacement
At present, there is insufficient clinical data time to thoroughly evaluate the theoretic advantages of motion preservation technology in preventing ASD. Clinical data from ACDF studies have more clearly implicated the natural history of specific motion segments rather than the effect of pathologic motion, but it remains to be seen whether studies on motion-preserving implants will draw similar conclusions. Regardless, the prevention of ASD was one of the primary rationales for the development of motion-preserving technologies. Motion-preserving surgical options, such as total disk replacement (TDR), have been studied to determine if the theoretic advantages for adjacent motion segments can be realized in clinical practice. Several short-term and medium-term follow-up studies have been published from the Investigational Device Exemption (IDE) trials associated with the US Food and Drug Administration approval process for TDR. Because these trials were designed to compare TDR with ACDF, the study cohorts provide an opportunity to compare TDR with ACDF in terms of ASD rate and factors associated with the development of ASD.
The IDE studies enrolled patients with single-level radiculopathy or myelopathy in need of single-level decompression; patients were randomized to either ACDF or TDR. Early reports from these investigations seemed to favor TDR. Two-year data from Mummaneni and colleagues comparing patients treated with ACDF with those treated using the Prestige TDR reported significantly higher rates of surgical intervention for ASD (3.4% vs 1.1%, P = .049). Only surgical intervention for ASD was reported; more rigorous methods of identifying ASD that did not require surgery were not used. The investigators attributed the decreased ASD rate to differential maintenance of physiologic motion after ACDF (allowing excessive supraphysiologic motion) and TDR (motion similar to preoperative range). More recent publications have shown less difference between ACDF and TDR in terms of ASD. In 2010, Burkus and colleagues presented a 5-year follow-up from the same Prestige cohort previously described 2 years after index surgery by Mummaneni and colleagues. At the most recent follow-up, the incidence of ASD requiring reoperation was similar between the ACDF and TDR groups, although this study had less power to detect a difference because of the smaller group size (271 total patients in this study vs 421 in the previous report). Jawahar and colleagues used a more rigorous definition of ASD in their study of 93 patients enrolled in IDE TDA trials at an average follow-up of just over 3 years and found no significant difference in the rate of ASD between ACDF (15%) and TDR (18%). Nunley and colleagues also found no difference between patients with ACDF (14%) and those with TDR (17%) who were actively treated for ASD at an average follow-up of 42 months. These last 2 studies likely contain many of the same patients based on the description of the study populations and author list but Nunley and colleagues do include an additional 77 patients and a 3-month longer follow-up. A 3-year follow-up after TDR was recently presented and showed a statistically significant decrease in ASD in the TDR group compared with the ACDF control group. The best available evidence currently does not present a clear picture to compare the rate of ASD between ACDF and cervical TDR; a longer-term follow-up is needed.
Jawahar and colleagues and Nunley and colleagues identified the presence of lumbar spondylosis as a patient-related factor associated with ASD. These studies used multiple regression methodologies for identification of factors influencing the rate of ASD, so these findings are independent of procedure; lumbar degeneration influences ASD regardless of whether a patient undergoes ACDF or TDR. This finding is consistent with contemporary evidence about factors affecting susceptibility to disk degeneration suspected to be largely dominated by genetics, which may be responsible for more than 75% of individual susceptibility to symptomatic degeneration. Nunley and colleagues also identified osteopenia as a procedure-independent factor contributing to susceptibility to ASD.
Principles of management of ASD
Preoperative Prevention Strategies
Spine surgeons evaluating patients with cervical radiculopathy or myelopathy for ACDF should pay careful attention to the history and physical examination to accurately characterize the levels involved. Updated imaging studies are similarly important to accurately identify symptom-generating motion segments, particularly for the lower cervical levels, which have greater native motion arcs, and have been shown to be more susceptible to degeneration in the native spine as described earlier. In this sense, the spinal surgeon must accurately identify all levels causing bothersome symptoms without being overly aggressive and operating on an asymptomatic degenerative disk.
The importance of including adjacent degenerative levels in the index procedure has been demonstrated by Hilibrand and colleagues who described a higher rate of ASD in patients who underwent single-level surgery. Although this finding initially seems contrary to explanations implicating lever arm length and associated pathologic adjacent-level disk stresses in the pathogenesis of ASD, this phenomenon is more likely related to undertreating patients who have multilevel disease using a single-level operation. Confirmatory evidence of this effect was presented by Komura and colleagues, who studied patients treated with long or short ACDF constructs that did or did not include the C5 to C7 motion segments, respectively. This study demonstrated an increased rate of ASD in the patients treated with longer fusions, indicating that the specific levels included in the index procedure are important in ASD; the lever arm may be of secondary importance to preexisting degeneration or susceptibility to degeneration in the development of ASD.
Intraoperative Prevention Strategies
Surgeons should attempt to limit the iatrogenic zone of soft-tissue injury during ACDF as much as possible to maintain the supporting role that musculoligamentous structures play in regulating normal intervertebral disk mobility. The longus colli is a longitudinal structure with fibers that span between multiple spinal levels and may impart substantial dynamic stability to the intervertebral disk. Dissection along the anterior vertebral border should be performed using a minimally disruptive technique and should focus on elevating, not transecting, the muscle fibers to reduce the effect on adjacent levels. When anterior plating is performed as an adjunct to ACDF, the use of a shorter plate facilitates minimal muscle dissection. The surgeon must also be particularly careful not to inadvertently injure the adjacent disk itself in any way during dissection or application of the plate. The disk has a poor capacity for regeneration and any significant injury will likely result in accelerated degeneration. Recently, this susceptibility to injury was demonstrated by an investigation in which even needle annulotomy associated with incorrect intraoperative level localization before diskectomy resulted in a 3-fold increase in degeneration.
Impingement of an anterior plate on an adjacent level may cause adjacent-level ossification disease (ALOD) as initially described by Park and colleagues. ALOD has been proposed to result from irritation and disruption of soft-tissue structures, such as the anterior longitudinal ligament or the anterior fibers of the annulus fibrosus. The severity of ALOD can range from minor involvement with focal ossification to bridging bone, which can eliminate adjacent-level motion in severe cases. Maintaining a vertical distance between the edge of the plate and the vertebral end plate of at least 5 mm has been demonstrated empirically to substantially reduce the incidence of ALOD. The avoidance of unnecessary disruption of the musculoligamentous structures along the anterior vertebral border, the reduced risk of direct injury to the disk during dissection, and the ability to maintain a safe margin between the plate edge and adjacent vertebrae all suggest that ACDF may be optimized by placement of a short plate with screws angled away from the plate, if possible, to maximize screw purchase and minimize iatrogenic soft-tissue injury.
Conservative Management Options
Because many studies focus discussion of ASD on patients treated surgically, it is often difficult to discern the success of conservative treatment. Hilibrand and colleagues followed up 46 patients for 2 or more years after the development of ASD; all patients with ASD were initially treated conservatively. Of these, 13 patients (28%) were successfully managed with conservative treatment (soft cervical collar, physiotherapy, and antiinflammatory medications) and achieved good or excellent outcomes. Six patients (13%) either refused surgery or were not offered revision and all had fair outcomes. The remaining patients (59%) failed conservative treatment and required ACDF of the adjacent level. In a smaller series, Elsawaf and colleagues were able to successfully treat 4 of the 5 patients who developed ASD after ACDF using conservative methods, although no information is provided about their conservative treatment protocol and the average follow-up averaged only 28 months. Similarly, Acikbas and colleagues found that all 19 patients who developed ASD after cervical fusion responded to an “aggressive regimen of conservative therapy” and avoided surgery. Closer investigation, however, demonstrates that 18 of these 19 patients developed ASD at 5 years or more after the index surgery. Given the average follow-up for the study was only 4.8 years, the duration of successful conservative treatment must have been short; longer follow-up may spoil these promising results after conservative treatment. More evidence is needed to determine if ASD can be reliably treated nonsurgically or if the previous history of spondylosis is associated with an accelerated clinical course for the degenerative adjacent level that frequently results in reoperation. As when treating the initial onset of radiculopathy or neck pain in the native cervical spine, conservative treatment best practice is ill-defined. A wide variety of protocols using physical therapy, limited immobilization, traction, antiinflammatory medications, and injections have been used in combination or individually with little literature to support any one treatment strategy.