The fusion of lumbar vertebrae has been associated with changes in the biomechanical forces applied to adjacent segments.27 The area of adjacent segment biomechanical forces and motion has been studied since the 1980s.28,29 In 1984, Lee and Langrana28 put 16 cadaveric models under combined compression and bending loads to observe motion. They found increased stress on adjacent segments and increased loading on facet joints within unfused segments. Further evidence of facet joint strain postfusion was demonstrated by Ha et al30 in an experiment with in vitro canine spines. They found increased segmental mobility and changed contact patterns within the joint. Using a finite model, Chen et al31 also found increased forces in adjacent disks postfusion and noted larger stress increases at the upper adjacent disk rather than the lower disk.
Increased motion at adjacent segments has been postulated as a reason for ASD. Axelsson et al32 studied 6 patients undergoing radiographic analysis of segment motion before and after L4–L5 fusion surgery. Increased mobility occurred in adjacent segments for 2 of the 6 patients. In an in vivo model, Hayes et al33 specifically examined mobility and found increased translational motion in adjacent segments when L3–L4 was fused, and this motion correlated with lower back pain. Esses et al34 studied changes in motion with different fusion techniques in cadaveric models. Posterolateral fusions were found to have less motion in adjacent segments as compared with anterior fusions.
In both cadaveric and finite models, increased intradiskal pressure has been found within adjacent segments.35–37 Weinhoffer et al36 studied intradiskal pressure in adjacent segments during flexion in cadaveric models. Intradiskal pressure increased during flexion, and the more segments that were fused, the greater the intradiskal pressure. Similar results were found by Cunningham et al,38 who studied 11 cadaveric models and found that intradiskal pressure increased by as much as 45%.
Using an in vivo mouse model and a finite element model, Lotz et al39 studied both the changes in stress in adjacent segments and the biochemical changes within adjacent segments. The finite element model predicted an increase in hydrostatic stress in the middle regions of the annulus by nearly ten-fold. Numerous harmful responses occurred at the histological and cellular levels, including disorganization of the annulus fibrosus, an increase in apoptosis with associated loss of cellularity, and damaging changes in gene expression. In an animal model by Phillips et al,40 results included loss of chondrocytes and notochordal cells within the nucleus pulposus.
Clinical and Radiographic Diagnosis
Separating the progression of osteoarthritis in patients postfusion from lumbar ASD is difficult. In a study by Kumar et al,41 patients were followed postfusion with the Short Form 36 and Oswestry Disability Index, functional testing, and radiographs. Their results were compared with those from age- and sex-matched controls. Radiographic changes above the level of fusion were worse in those with fusion compared with those not fused. There was no statistically significant difference in clinical outcomes. Wai et al42 studied MRI results over 20 years for patients undergoing lumbar fusion surgery and found that the prevalence of degenerative changes in patients undergoing surgery was similar to age-matched controls. Hambly et al43 reported that radiographic changes occur in the transition zone cephalad to lumbosacral fusion, regardless of whether surgery occurred.
As with ASD in the cervical spine, the correlation between clinical, MRI, and radiographic signs of ASD is controversial. Interpretation of radiographic evidence is also complicated by the fact that asymptomatic patients have been found to have substantial abnormalities on MRI.44 Plain radiographs have been found to be of little help in the diagnosis of ASD. Frymoyer et al45 examined plain radiographs from 96 patients who had undergone fusion surgery. Results showed no correlation between radiographic findings and clinical symptoms. The progressive changes seen on radiographs of adjacent segments are illustrated in Figure 3.46 In the current authors’ experience treating ASD, severe degenerative changes cause symptoms of spinal stenosis. Figure 4 shows an area of neural foraminal narrowing at the L2–L3 level.
Radiographs showing progressive changes after lumbar spinal fusion preoperatively (A) and 2 years (B), 5 years (C), and 9 years (D) postoperatively. In this patient, definite signs of adjacent segment disease are present by 13 years after L4–L5 spinal fusion (E). However, it is unclear whether these radiographic changes signify clinical significance. (Reprinted with permission from Cheh G, Bridwell KH, Lenke LG, et al. Adjacent segment disease following lumbar/thora-columbar fusion with pedicle screw instrumentation. Spine [Phila Pa 1976]. 2007; 32:2253–2257.)
Radiograph of a patient with a L3–L5 posterior spinal fusion with instrumentation who presented to clinic after months of severe back pain and radiating pain down bilateral legs (A). Computed tomography myelogram showing spinal stenosis at L2–L3 with bilateral facet hypertrophy (B). This patient eventually had her spinal instrumentation removed and underwent fusion of her L2–L3 segment with instrumentation.
Many studies have examined the prevalence of clinical symptoms associated with ASD. In a long-term retrospective study by Ghiselli et al,47 215 patients were followed postfusion for an average of 6.7 years. The rate of reoperation at adjacent segments was 16.5% in the first 5 years and 36.1% at 10 years. Similar results were reported by Gillet48 in a retrospective study that found a reoperation rate of 20% over a range of 2 to 15 years of follow-up. Penta et al49 studied MRIs for 10 years after lumbar interbody fusion, and of the 81 patients studied, 32% had radiographically diagnosed ASD. In a large retrospective cohort study by Sears et al50 of 1000 consecutive posterior lumbar interbody fusion procedures with a mean follow-up of 63 months, the 10-year prevalence of further surgery for ASD was 22.2%. The annual incidence of surgery for ASD was 2.2%.
The rate of ASD after lumbar fusion varies widely between studies.5 In a review by Park et al5 of 22 studies, the reported range of ASD was 5.2% to 100% based on either radiographic or clinical diagnosis. The authors attributed this wide variation to the range of patient populations, differing clinical or radiographic definitions of ASD, and the retrospective nature of many of the studies.
Several studies have examined the risk factors that put a portion of patients at risk for ASD. Aota et al22 reported that the factor most influential in postfusion instability was age. The study showed that 11 of 30 patients older than 55 years developed ASD postfusion, whereas only 3 of 25 patients younger than 55 years developed ASD postfusion. Similarly, a retrospective study by Etebar and Cahill23 found that ASD was higher in postmenopausal women. The link between menopause and worsening osteoarthritis is not entirely clear, but active research is ongoing into the mechanism by which a lack of estrogen influences osteoarthritis development and progression.24 In a retrospective review of 49 patients, Rahm and Hall25 found that increasing age, as well as inter-body fusion, was associated with a higher risk of ASD.
Investigators have studied whether the length of fusion in fusion surgery factors into the development of ASD. Penta et al49 examined MRIs and radiographs from 52 patients who had normal preoperative findings in adjacent levels.49 In this group of patients, the length of fusion did not factor into whether ASD occurred. Wimmer et al51 studied the effect of polysegmental fusions on anteroposterior translation in 120 patients with painful spondylolisthesis who were treated with combined anterior and posterior fusions. The study group was divided into a monosegmental fusion group (n=46) and a polysegmental fusion group (n=74). Polysegmental fusion was associated with a greater degree of anteroposterior translation.
Fusion involving the L4–L5 segment has been cited as a possible cause of disk degeneration at L5-S1. This was specifically studied by Ghiselli et al52 in a group of 32 patients over 7.3 years. Their investigation found a general progression of degenerative changes in the L5-S1 segments. However, the authors concluded that there was no need for routine fusion of L5-S1 in patients with isolated L4–L5 symptoms.
Another significant factor in determining ASD postfusion is the anatomical alignment of spinal segments preoperatively. Investigators have noted increased ASD in patients with a preoperative L1 and S1 axis distance greater than 35 mm.53 In a retrospective study, Kumar et al41 reported that patients with a normal C7 plumb line and sacral inclination had the lowest rate of ASD. In a study by Lazennec et al54 of 81 patients who had undergone fusion, there was a statistically significant relationship between postoperative pain and sacral tilt. Oda et al55 used a sheep model to determine the biomechanical changes in adjacent segments in kyphotic spines. Results showed that kyphotic posterolateral fusion significantly altered supra-adjacent segments by inducing more stiffness in the posterior ligamentous complex and increasing lamina strain under flexion-extension loading. There were significant degenerative changes in the supra-adjacent segments in these sheep. Lehmann et al56 found a correlation between preoperative segmental instability in the segment above fusion and lumbar spinal stenosis. The greater the translation instability, the more likely the adjacent segment would become unstable post-fusion. Biomechanical studies have also shown that once degeneration occurs at one segment, the risk of degeneration at other adjacent segments increases.57 Schlegel et al58 reported that segments that were 2 segments removed from fusion were just as likely to show degeneration as adjacent segments.
The alignment of fusion in the lumbar spine was investigated in a biomechanical cadaveric study by Akamaru et al.59 They reported that fusion in either hyper- or hypolordotic alignment of L4–L5 resulted in different loads on adjacent spinal segments. Hypolordotic alignment of L4–L5 resulted in the greatest amount of flexion-extension motion at L3–L4, whereas hyperlordotic alignment of L4–L5 resulted in the greatest amount of flexion-extension motion at L5-S1. In a study of patients undergoing 360° fusion with healthy adjacent segments preoperatively, it was found that maintaining the lordotic angle at approximately 20° postoperatively was associated with the prevention of ASD.60
Several studies have investigated ASD in the lumbar spine following different treatment decisions. In a prospective study of 111 patients with spondylolisthesis who were randomized to exercise, uninstrumented posterolateral fusion, or instrumented posterolateral fusion, accelerated adjacent segment degeneration was found in the patients that had fusion and laminectomy.61 Abdu et al62 studied 380 surgical candidates with degenerative spondylolisthesis who underwent 1 of 3 surgical interventions: posterior in situ fusion, posterolateral instrumented fusion with pedicle screws, or posterolateral instrumented fusion with pedicle screws plus interbody fusion (360°). At 3- and 4-year follow-up, there was no statistical difference in outcome criteria (ie, Short Form 36 bodily pain, physical function scales, and the modified Oswestry Disability Index) between the different operative techniques. Furthermore, in a randomized, controlled trial performed by Videbaek et al,63 patients who underwent anterior lumbar interbody fusion combined with posterolateral lumbar fusion or posterolateral lumbar fusion alone showed no increased risk of ASD. Therefore, it does not appear that a specific treatment technique accelerates the onset of increases the likelihood of developing ASD compared with any other.
Edwards et al64 investigated long adult fusions of the thoracolumbar spine to L5 and the resultant degenerative changes to S1. Thirty-four patients were followed for a mean of 5.6 years. Results showed that 61% of patients had resultant degenerative disk disease. Risk factors for degenerative changes were a positive sagittal balance, younger age at operation, and any signs of radiographic degeneration at L5-S1.
Added instrumentation during spinal fusion has been studied as a possible risk factor for ASD. Shono et al65 investigated whether fusion surgeries augmented with instrumentation such as a compression hook and a transpedicular screw fixation system increased motion in adjacent segments. The study was conducted with a calf lumbosacral spine model. As spinal instrumentation increased, higher segmental displacement occurred at the upper residual intact motion segment. Interestingly, different results were reported by Wiltse et al66 with the use of pedicle screws. They studied pedicle screws and their relationship to transition zone changes above or below the fused segment. No increase in ASD incidence was found with pedicle screw placement.
Ahn et al60 retrospectively studied 3188 patients who underwent thoracolumbar spinal fusion to find the rate of ASD, as well as risk factors. Results showed a failure rate of 6% at 10 years at adjacent segments. Risk factors noted were multiple level fusions, old age, degenerative disease prior to fusion, and male sex. A study by Lee et al21 investigated overall risk factors for revision surgery due to ASD. Of 1069 patients undergoing spinal fusion surgery, 2.62% required revision surgery. Postoperative facet degeneration was associated with revision surgery, and the incidence of ASD in proximal segments was found to be higher than that in more distal segments. Preexisting degeneration was found to be a significant risk factor for ASD requiring surgery.
The intraoperative decision of distracted disk height of the fused segment caused by cage or bone insertion has been studied as a risk factor for ASD. Kaito et al67 studied 84 patients with L4 spondylolisthesis who were treated with posterior lumbar interbody fusion. Patients who developed ASD were found to have higher L4–L5 disk space distraction.
Nassr et al68 investigated whether incorrect needle localization during anterior cervical diskectomy and fusion caused ASD. Of 87 patients undergoing anterior cervical diskectomy and fusion, 15 had incorrect needle localization. The group with incorrect needle localization was associated with a 3 times higher rate of ASD. The authors concluded that either needle trauma or unnecessary surgical dissection may have caused ASD during the study. A recent systematic review for risk factors for ASD after lumbar fusion showed that age older than 60 years was associated with an increased risk of ASD after spinal fusion.19 Similarly, if the patient had preexisting facet degeneration, degenerative disk disease, multilevel fusion, laminectomy performed adjacent to a segment, a construct stopped at L5, or excessive disk height distraction, then the risk of ASD may be higher.