Risk factors for upper adjacent segment degeneration after multi-level posterior lumbar spinal fusion surgery

Degenerative lumbar stenosis (DLS) is a common disease in an elderly population, characterized by back pain, radicular or claudicant leg pain, and disability [1]. DLS usually has more than one canal stenosis level and goes with spinal scoliosis or imbalance. Surgical treatment such as posterior lumbar spinal fusion is an effective way to improve patients’ clinical symptoms and correct spinal function [2]. Posterolateral fusion (PLF) and posterior lumbar interbody fusion (PLIF) have been commonly performed procedures that have been associated with substantial functional improvements for DLS. However, after multi-level posterior lumbar spinal fusion, there may be new clinical troubles, such as adjacent segment degeneration (ASD) [3]. These troubles affect patients’ prognosis and quality of life.

ASD can be classified as symptomatic and radiographic [4]. Symptomatic ASD refers to new symptoms or signs of an adjacent segment which often needs a revision surgery procedure to address adjacent level pathology. Radiographic ASD, generally defined as radiographic changes of the adjacent segment, has no clinical symptoms [5]. There are many risk factors for ASD, such as age, sex, body mass index (BMI), surgery strategy, and sagittal alignment change [6]. But there are few clinical studies about the risk factors for ASD after multi-level posterior lumbar spinal fusion surgery.

In this retrospective study, to analyze the risk factors for upper ASD, we followed up 71 patients who underwent multi-level posterior lumbar spinal fusion surgery for about 3 years.

In this study, 71 patients (34 men, 37women) met the inclusion criteria in total, with a mean age of 62.2 years. There were 22 patients in group A (11 men, 11 women) and 49 patients in group B (23 men, 26 women). The age, sex, BMI, ODI, and JOA had no statistical differences between groups A and B. As for the preoperative modified Pfirrmann grade, the number ratio of ≤ 3 versus> 3 in group A was 16:6 and in group B was 19:30; there were statistically significant differences(P = 0.011, Pearson’s chi-square test). The mean number of fixed vertebrae, decompressed levels, and interbody fusion levels had statistical differences between groups A and B (Table 1). These numbers were detailed in Tables 2, 3 and 4.

Twenty-nine patients (40.85%) presented with radiographic ASD at the last follow-up. Five cases (22.73%) came from group A, and 24 cases (48.98%) came from group B. The percentage of radiographic ASD was significantly higher in group B than in group A (P = 0.042, Pearson’s chi-square test). Clinical results had no statistically significant differences between groups A and B (Table 1). No symptomatic ASD was found during the follow-up.

As for the radiographic analysis at preoperation and last follow-up, LL, SS, PT, coronal Cobb angle, and ΔPILL had no statistically significant differences between groups A and B. But group B had a significantly higher degree of PI than group A (P = 0.042) at preoperation (Table 5) .

To analyze the risk factors for radiographic ASD, we performed the log-rank test and used the Cox proportional hazards model. The age, sex, BMI, preoperative LL, SS, PT, coronal Cobb angle, number of fixed vertebrae, and fused levels had no significant differences in the Cox proportional hazards model. But the upper disc had a preoperative modified Pfirrmann grade > 3, a high degree of PI at preoperation, and more decompressed levels which had statistically significant differences (Table 6).

There are two illustrative cases for groups A and B, respectively. Group A (Fig. 1): female, 58 years old, diagnosed with degenerative lumbar stenosis. L1–5 were fixed, L3–5 underwent laminectomy, L3–5 underwent interbody fusion, and L1–5 underwent posterolateral fusion. T12/L1 disc preoperative modified Pfirrmann grade was 4, and preoperative PI was 58°. No radiographic ASD of T12/L1 level was found at 24 months after surgery. Group B (Fig. 2): male, 60 years old, diagnosed with degenerative lumbar stenosis. L1–5 was fixed, L1–5 underwent laminectomy, L3–5 underwent interbody fusion, and L1–5 underwent posterolateral fusion. T12/L1 disc preoperative modified Pfirrmann grade was 5, and preoperative PI was 59.78°. Radiographic ASD of T12/L1 level was found because the T12/L1 disc height decreased from 7.50 to 6.0 mm at 18 months after surgery.

Degenerative lumbar stenosis (DLS) is a complex degenerative disease of the spine that often occurs in the elderly population. DLS may have more than one stenosis spinal levels or with spinal instability. Surgical treatment can restore spinal stability, relieve nerve root compression effectively, and improve symptoms and patients life quality [10]. The most ideal surgery treatment for degenerative lumbar stenosis or with scoliosis still remains controversial. But ASD which came from the posterior lumbar spinal fusion surgery always attracts the attention of spine surgeons. Many risk factors are involved in ASD, but the mechanism for its occurrence is still controversial. Moreover, there was no correlation between both the radiological development of ASD and its clinical outcome [11].

In this retrospective study, after multi-level posterior lumbar spinal fusion surgery, 29 patients (40.85%) presented with radiographic ASD, but we did not find symptomatic ASD for approximately 3 years. We observed that preoperative modified Pfirrmann grade > 3, a high degree of PI at preoperation, and more decompressed levels were the risk factors for radiographic ASD in this study.

Some researches have suggested that preoperative disc degeneration is related with radiographic ASD [9, 12]. Intrinsic disc degeneration causes the increase of disc stress at the adjacent segment [13, 14]. If this preoperative disc degenerative segment is not involved in fusion, it might aggravate the adjacent disc degeneration after fusion surgery [15]. In this study, we also find that pre-existing disc degeneration (modified Pfirrmann grade > 3) of the upper fixed segment is a risk factor for radiographic ASD(P = 0.024, 95% CI 1.149–7.392).

Sagittal alignment of the lumbar spine has an important and measurable impact on the pathogenesis of lumbar spine disorder and contributions to ASD [16]. Anatomic pelvic parameters also have an association with lumbar spinal degeneration, such as lumbar spinal stenosis and degenerative spondylolisthesis [17]. An appropriate sacral inclination is important for minimizing the incidence of ASD [18]. Di Martino et al. reported that patients with SS < 39° or PT > 21° were at higher risk for symptomatic ASD [19]. Moreover, a high degree of PI was a risk factor for developing early-onset radiographic ASD [20]. In addition, Rothenfluh et al. found that in degenerative disease of the lumbar spine, a high PI with diminished LL, especially ΔPILL ≥ 10°, seems to predispose to adjacent segment disease. If sagittal malalignment is maintained after lumbar fusion surgery, patients with such PI-LL mismatch exhibit a higher risk for undergoing revision surgery [21]. In this study, group B had a significantly higher preoperative PI than group A (49.29 ± 10.77° vs 43.64 ± 10.10°, P = 0.042), which contributes to a significant higher radiographic ASD ratio. The Cox proportional hazards model showed that a high PI is an independent risk factor for radiographic ASD ( P = 0.041, 95% CI 1.002–1.080). But there are no significant differences in SS, PT, or ΔPILL at preoperation and last follow-up. This reason might be that there is only radiographic ASD in this study but no symptomatic ASD.

The long-term RCT showed that compared with the effect of fusion with natural history, laminectomy and fusion accelerate degenerative changes at the adjacent level [22]. Patients who underwent fusion of three or four levels had a threefold increased risk of further surgery, compared with single-level fusions [23]. Biomechanical experiments found that the posterior column structure of the spine plays a key role in maintaining the stability of the spine [24]. Laminectomy of fusion segment head vertebra had been the reason for losing the balance in its flexion and axial rotation, which lead to the occurrence of ASD [25]. In this study, patients in group B, whose proximal fixed vertebral lamina and PLC were resected in surgery, had a significantly higher number of decompressed levels than group A. More decompressed levels mean that the proximal fixed vertebral posterior column structure was damaged more extensively. These also had contributions to a significantly higher radiographic ASD ratio in group B. But we did not find that the number of fixed vertebrae or interbody fusion levels have risk association with radiographic ASD. The reasons might be that (1) all patients in the two groups underwent multi-level fixation(≥ 4 fixed vertebrae) and (2) not all the proximal fixed vertebra underwent laminectomy and fusion together in group B. We just determined that more decompressed levels are a risk factor for radiographic ASD. Therefore, for patients with DLS requiring multi-level fixation and decompression, it is crucial to determine and handle the responsibility gap. In order to prevent ASD, surgeons must retain the posterior column structure such as lamina and PLC more entirely, determine the decompression range rationally, and protect the stability of the spine.

There are limitations in this study. More enrolled cases and more long-term follow-up period are needed for this study. Moreover, postoperative MR images and CT scans were not available for all patients, which might have affected our evaluation of ASD.

Resection of proximal fixed vertebral lamina and PLC might accelerate the radiographic degeneration of their upper adjacent segment after multi-level posterior lumbar spinal fusion surgery (at least 3 levels). A preoperative modified Pfirrmann grade higher than 3, a high degree of PI at preoperation, and more decompressed levels might be the risk factors for upper radiographic ASD.