The etiology of ASD has been the subject of considerable research and controversy in recent years. The pivotal question is whether or not fused motion segments biomechanically cause ASD or if this disease is simply part of the natural history of cervical spondylosis. Many biomechanical and clinical studies have attempted to answer this question, and the truth likely lies in the middle, meaning that ASD is due to a combination of altered biomechanics and natural progression of arthritis.

Biomechanical investigations have attempted to address the claim that cervical fusions alter the biomechanical environment at adjacent segments, leading to accelerated disk degeneration due to increased loading and excessive motion. While many studies substantiate this claim, there is not conclusive evidence that this is true. Cadaver studies evaluating simulated one- (1) and two-level anterior cervical plating (2) found that fusion increases intradiskal pressures and motion in adjacent levels, and it appears that this increased pressure and motion is amplified with fusion of additional levels. On the other hand, other cadaver studies have found that one- to three-level anterior cervical fusions did not significantly increase motion at adjacent segments, suggesting that degeneration may be due to natural history of the disks (3,4). Part of the challenge in testing this hypothesis is that for technical reasons, these cadaver studies evaluated spines without the occipitoatlantoaxial articulations, which accounts for 50% of sagittal and rotational motion of the cervical spine. Biomechanical studies of simulated cervical fusion, which do not include the upper cervical spine, likely overestimate the amount of motion and stress transferred to adjacent segments because a portion of that motion is normally accommodated by the upper cervical spine (occiput-C2). Additionally, there are inherent weaknesses in comparing cadaver specimens, which vary widely in terms of ligamentous and joint laxity. An in vivo biomechanical study prospectively evaluated patients before and after one- to four-level anterior cervical discectomy and fusion (ACDF) with fluoroscopy, finding that in the first 1 to 2 years after surgery there was no difference in motion at levels adjacent to cervical fusions (5). Overall, these biomechanical studies suggest that anterior cervical fusion may inherently increase the risk of ASD but do not offer conclusive evidence.

The first extensive clinical description of cervical ASD was that of Hilibrand et al. (6), who explained it 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.” Theirs was a retrospective review of 374 patients with 409 anterior cervical fusion procedures for myelopathy or radiculopathy. The incidence of symptomatic ASD was 2.9% per year, and survivorship analysis predicted that within 10 years of the index procedure, 25.6% would have new disease at an adjacent segment. Patients with multilevel fusions were at lower risk of developing new disease at an adjacent
level than patients who had single-level fusions. Risk factors for new disease included preexisting spondylosis and single-level fusions involving C5 or C6. More than two-thirds of the 58 patients with ASD required revision surgery.

Ishihara et al. (7) evaluated 112 patients for symptomatic ASD at minimum 2-year follow-up after ACDF. Nineteen percent of these patients developed symptomatic ASD, and survivorship analysis showed 16% incidence at 10 years and 33% at 17 years. Of the 19% with new disease, 37% required revision surgery. The incidence was higher in patients with preexisting indentation of the dura at the adjacent level, although cervical alignment and the number of levels fused did not influence the incidence of ASD. Similarly, Yue et al. (8) reported on 71 patients who were followed for 7 years after ACDF. The incidence of adjacent segment degeneration, as a radiographic finding, was 73%, although only 17% underwent revision surgery for symptoms of myelopathy and/or radiculopathy. Additionally, there was no significant correlation between number of levels fused and development of adjacent segment degeneration.

One of the risk factors for developing adjacent segment degeneration may be the proximity of the plate to the disk space, as reported by Park et al. (9). This was a retrospective review of 118 patients who had ACDF with a plate and were evaluated for the extent of ossification adjacent to the anterior plate as a sign of adjacent segment degeneration. At a mean of 25.7 months, adjacent level ossification occurred in 59% of the cephalad disks and 29% of the caudal disks, and the rate of ossification was higher in patients with a plate-to-disk distance less than 5 mm at both the cephalad level (67% compared with 24%) and the caudal level (45% compared with 5%). Positioning the plate less than 5 mm away from the adjacent disk space may speed adjacent level ossification and degeneration.

One group has suggested that incorrect level needle localization during ACDF may contribute to ASD. Nassr et al. (10) retrospectively reviewed 87 two-level ACDF procedures with 15 patients having needle placement in a nonoperative level. The groups were no different in terms of age, sex, or length of follow-up, but the incorrectly marked group demonstrated an increased risk of development of ASD by an odds ratio of 3.2.

Although no formal classification exists that accounts for both radiographic and clinical symptomatology, Hilibrand et al. (6) suggested a radiographic grading system for adjacent level degenerative changes. Grade I has normal radiographic findings, while Grade II has x-rays showing narrowing of the disk spaces without posterior osteophytes and signal changes in the disks on MRI. Grade III has x-rays with less than 50% of normal disk height and posterior osteophytes, while MRI or CT myelogram shows herniated nucleus pulposus without neural compression. Finally, Grade IV has the same x-rays as Grade III with MRI findings of spinal cord compression with or without nerve root compression and CT myelogram evidence of nerve root cutoff with or without spinal cord compression.

Many patients who have undergone ACDF have radiographic evidence of spondylosis at adjacent levels without clinical symptoms of myelopathy or radiculopathy. Goffin et al. (11) followed 180 patients who had undergone ACDF for an average of 100.6 months, determining that 92% had evidence of adjacent level disk degeneration on a lateral cervical spine x-ray (measured by disk height and osteophytes). While “time since surgery” was positively correlated with degeneration, there was no significant correlation between degeneration and clinical symptoms, as measured by the Odom score and Nurick scale. A study by Dohler et al. (12) found that 14 of 21 (67%) patients had translation at the adjacent level 27 months after ACDF but there was no significant correlation with pain. Furthermore, Boden et al. (13) showed that in 63 subjects without any prior history of cervical symptoms, MRI revealed disk degeneration in 25% of people under 40 and 60% of those older than 40. Based upon such a high prevalence of radiographic degeneration among those having undergone prior surgery and among the general population, it is clear that asymptomatic patients should be managed conservatively.

Clinical findings of myelopathy or radiculopathy are necessary in addition to radiographic evidence of degeneration to satisfy the diagnosis of ASD. Patients who have failed a trial of conservative management and who have clinical symptoms limiting their daily activities and patients experiencing functional decline due to progressive myelopathy may benefit from adjacent level surgery. Hilibrand et al. (6) reported that over two-thirds of their 58 patients with ASD underwent revision surgery. Seven of nineteen (37%) patients with ASD in Ishihara’s study underwent revision surgery (7).