Survivorship Analysis of Clinical Adjacent-Segment Pathology After Single-Level Cervical Fusion.

BACKGROUND Clinical adjacent-segment pathology (CASP) is an important problem after anterior cervical surgery. The purpose of this study was to predict prevalence of CASP and determine the possible risk factors for CASP after single-level anterior cervical discectomy and fusion surgery. MATERIAL AND METHODS We retrospectively reviewed a series of patients who underwent single-level cervical discectomy and fusion surgery (ACDF). Both basic and radiographic data of patients were collected. Life-table method and Kaplan-Meier analysis were used to calculate prevalence of CASP and disease-free survival rate. Cox analysis was performed to determine the predictive factors for it. RESULTS A total of 256 patients were included in this study. The mean length of follow-up was 70.64 months. Among them, 31 patients were diagnosed as having CASP during follow-up. Nineteen of them were at the cephalad adjacent segment, and the other 12 were at the caudal segment. After ACDF procedures, 10.01% of patients developed new symptoms of CASP within 5 years, and the incidence increased to 23.89% after 10 years. The incidence rate of CASP was an average of 2.46% per year. Multivariate Cox regression analysis showed that congenital stenosis (hazard ratio [HR], 3.250; 95% confidence interval [CI], 1.538-6.867) and degeneration of adjacent segment (HR, 2.681; 95% CI, 1.259-5.709) were correlated with the incidence of CASP. CONCLUSIONS Patients with congenital stenosis and pre-existing degenerative changes of adjacent segments had a higher risk of developing CASP after single-level anterior cervical discectomy and fusion.

Background

Anterior cervical discectomy and fusion (ACDF) is a standard surgical procedure in the treatment of symptomatic cervical radiculopathy and myelopathy. Since its initial description and application by Robinson and Smith [1], ACDF procedure has shown excellent outcomes based on long-term results. However, in spite of the successful outcomes of this procedure in relieving the symptoms of cervical lesions, subsequent development of degenerative disc disease adjacent to fused discs is a matter of grave concern [2,3]. Clinical adjacent-segment pathology (CASP), which is defined as the development of new symptoms of radiculopathy or myelopathy arising from segments adjacent to the site of previous anterior cervical fusion, has been reported to affect more than one-fourth of all patients within 10 years after the index surgery [2].

The pathogenesis of CASP still remains a matter of debate [4-6], but previous studies have been performed to determine the risk factors. Several factors have been found to be associated with the development of CASP, such as age, smoking history, sagittal alignment, the number of fusion segments, and pre-existing degenerative changes of adjacent segments [7-11]. However, the majority of the studies on this topic involved relatively small numbers of patients or did not statistically control confounding variables with multivariate analyses. Although CASP is a time-dependent disease after cervical procedures, survival analysis is rarely performed to identify the potential predictive factors.

In the current study, we tried to fill some of these research gaps by analyzing patients who had only undergone single-level cervical spine procedures. As insufficient sample size and loss to follow-up is inevitable in such clinical research, we used the life-table method and Kaplan-Meier analysis to calculate the prevalence of adjacent segments. We further performed multivariate Cox analysis to determine the predictive factors for CASP.

Material and Methods

This study was approved by the Research and Ethics Committee of the Third Hospital of Hebei Medical University, and all patients gave written informed consent for their information to be stored in the database of this hospital and used for research.

Patient population

In this study, we reviewed a series of patients who underwent single-level cervical arthrodesis between January 2006 and December 2015. The inclusion criteria were adult patients with radiculopathy, myelopathy, or myeloradiculopathy from degenerative single-level cervical disc disease confirmed by correlating magnetic resonance imaging (MRI) findings, and no response to non-operative treatment for at least 3 months. Those with non-degenerative disease, such as trauma, infection, tumor, deformity, or inflammation, or undergoing any prior cervical spine operation, were excluded. Those showing herniated nucleus pulposus with or without neural compression on adjacent segments based on MRI were also excluded [2].

Operative procedure

In this study, all patients underwent anterior cervical decompression and fusion procedures. After general anesthesia, the patient was placed supine with mild cervical extension. A standard right-sided approach through a transverse incision used to expose the targeted level. After that, the compressive materials were removed, including osteophytes, herniated disc, and posterior longitudinal ligament, and then a polyetheretherketone cage filled with excised osteophytes was inserted between vertebral bodies. All procedures in this study were performed by the same surgeon. A soft cervical collar was used for 8 weeks postoperatively.

Data collection

Antero-posterior, lateral, and flexion/extension lateral plain radiographs, as well as MRI, were obtained pre-operatively. Radiographic data were also collected postoperatively at 6 weeks, at 3, 6, and 12 months, and then annually. Patients who developed additional cervical radiculopathy or myelopathy symptoms during follow-up were advised to go back to the clinic at their earliest convenience, and MRI was performed for confirmation and further evaluation.

CASP is defined as a new development of radiculopathy, myelopathy, or myeloradiculopathy symptoms, and a newly developed compressive lesion of spinal cord or nerve root in the segments adjacent to the fusion level, which is radiologically confirmed by MRI.

Basic data included age at surgery, sex, body mass index (BMI), diabetes mellitus, history of smoking and alcohol, neurological diagnosis, and fusion level. Postoperative radiographs at the 6-month follow-up were assessed to determine the congenital stenosis, curve pattern of C2–C7, Cobb angle of fused vertebrae, ROM of C2–C7, and T1 slope.

The spinal canal diameter was evaluated by measuring the Pavlov ratio. The Pavlov ratio was related to anterio-posterior diameter of the spinal canal to the anterio-posterior diameter of the vertebral body at the C5 level [12]. The presence of congenital stenosis was defined as a Pavlov ratio less than 0.80; the Cobb angle from C2 to C7 was defined as the angle formed by the inferior endplates of C2 and C7 in standing lateral radiographs. An alignment of C2–C7 Cobb angle more than 0 was defined as lordosis, and an alignment of C2–C7 Cobb angle of 0 or less was defined as kyphosis. The Cobb angle of fused vertebrae was formed by the upper endplate of cranial vertebral body and lower endplate of caudal vertebral body in the operated level. The C2–C7 range of motion (ROM) was defined as the sum of the absolute value of C2–C7 Cobb angle during flexion and extension lateral radiographs [10]. T1 slope was measured as the angle between a horizontal line and the superior endplate of T1 on standing lateral radiograph [9]. Two blinded observers independently interpreted the radiological findings twice and the mean values were used.

Pre-existing degenerative changes of adjacent segments was assessed by 2 independent authors according to preoperative radiographs and MRI based on Hilibrand definitions [2]. Patients with moderate or severe degeneration of adjacent segment were not included in this study. The existence of mild degeneration of adjacent segment was recorded as a potential risk factor.

Statistical analysis

The incidence of CASP for each year was calculated by life-table method. Kaplan-Meier analysis was used to calculate disease-free survival rate for adjacent segments. Univariate Cox proportional hazards models of association between each of the exposures of interest and the outcome were built. Hazard ratios (HRs) along with 95% confidence intervals (CIs) are reported. All variables found to be potentially associated with the outcome (P < 0.10) were entered into the Cox proportional hazards regression model to analyze their relative importance. Statistical analysis was performed with the Statistical Package for Social Sciences software (version 17.0; SPSS Inc., Chicago, IL, USA), and probability value less than 0.05 was considered statistically significant.

Results

A total of 256 patients undergoing single-level cervical spine surgery were included in this study. Among these patients, 137 were male and 119 were female. The mean age at the time of index surgery was 50.1±6.8 years. There were 167 patients diagnosed with radiculopathy, 42 had myelopathy, and 47 had myeloradiculopathy. During surgery, 234 patients were fixed with internal plates and the other 22 were not. The length of follow-up was 70.64±32.46 months. Most segments that underwent ACDF were at the C5–C6 spinal level, followed by the C6–C7, C4–C5, and C3–C4 levels in descending order of frequency.

During follow-up, 31 patients were diagnosed with CASP. Nineteen of them were at the cephalad adjacent segment, and the other 12 were at the caudal segment. Based on life-table analysis, the incidence rate of CASP was an average of 2.46% per year. The proportion of patients with CASP at different time points is listed in Table 1.

Table 1

Proportion of patients with CASP in different interval time by life-table method.

Interval time (months)	No. of patients entering interval	No. of patients with CASP	Proportion of patients with CASP (%)	Cumulative proportion surviving (%)
0–12	256	1	0.39	99.61
12–24	255	3	1.18	98.43
24–36	249	1	0.42	98.01
36–48	222	5	2.36	95.70
48–60	196	5	2.69	93.12
60–72	171	8	5.42	88.07
72–84	116	1	0.86	87.31
84–96	115	2	1.91	85.64
96–108	92	2	2.65	83.37
108–120	57	3	6.74	77.75

CASP – clinical adjacent-segment pathology.

Kaplan-Meier analysis showed an 89.99% disease-free survival rate for adjacent segments (95% CI, 85.88% to 94.10%) at 5 years and 76.11% (95% CI, 66.47% to 85.74%) at 10 years (Figure 1). In terms of the prevalence of adjacent-segment pathology, these results suggested that 10.01% (95% CI, 5.90% to 14.12%) of patients would develop new symptoms of CASP within 5 years, and 23.89% (95% CI, 14.26% to 33.52%) by 10 years.

Figure 1

Overall Kaplan-Meier survivorship curve for clinical adjacent segment pathology after single-level anterior cervical discectomy and fusion.

In the analysis of the association between patient basic data (Table 2), as well as radiographic data (Table 3) and the incidence of CASP, we found that involvement of C5–C6 level (P=0.084), congenital stenosis (P=0.067), degeneration of adjacent segment (P=0.033), and curve pattern of C2–C7 (P=0.027) were 4 potential risk factors, while age, sex, BMI, history of smoking or alcohol, diabetes mellitus, preoperative diagnosis, internal fixation, Cobb angle of fused vertebrae, ROM of C2–C7, and T1 slope were not (P>0.10). In the further multivariate Cox regression analysis, congenital stenosis (HR, 3.250; 95% CI, 1.538–6.867) and degeneration of adjacent segment (HR, 2.681; 95% CI, 1.259–5.709) were correlated with the incidence of CASP during follow-up (Table 4).

Table 2

The association between patients’ basic data and the incidence of CASP.

Characteristic	CASP (+) group (n=31)	CASP (−) group (n=225)	P value	Hazard Ratio	95% CI
Age
>60 years	7	58	0.838	0.916	0.394–2.129
≤60 years	24	167		Reference
Gender
Male	14	123	0.294	0.685	0.337–1.390
Female	17	102		Reference
BMI
>25 kg/m2	11	104	0.351	0.704	0.337–1.471
≤25 kg/m2	20	121		Reference
Smoking
Yes	13	88	0.730	1.134	0.555–2.317
No	18	137		Reference
Alcohol
Yes	10	91	0.475	0.760	0.358–1.615
No	21	134		Reference
Diabetes mellitus
Yes	5	29	0.434	1.466	0.562–3.827
No	26	196		Reference
Neurological disorder
Radiculopathy	18	149	0.179	1.770	0.769–4.073
Myelopathy	8	34	0.849	1.101	0.408–2.969
Myeloradiculopathy	5	42		Reference
C5–C6 level involved
Yes	23	131	0.084	2.032	0.908–4.544
No	8	94		Reference

CASP – clinical adjacent-segment pathology; BMI – body mass index; CI – confidence interval.

Table 3

The association between radiographic data and the incidence of CASP.

Characteristic	CASP (+) group (n=31)	CASP (−) group (n=225)	P value	Hazard ratio	95% CI
Internal fixation
Yes	29	205	0.798	1.206	0.287–5.067
No	2	20		Reference
Congenital stenosis
Yes	18	91	0.067	1.949	0.954–3.980
No	13	134		Reference
Degeneration of adjacent segment
Yes	16	73	0.033	2.152	1.063–4.356
no	15	152		Reference
Curve pattern of C2–C7
Kyphosis	5	18	0.027	2.961	1.131–7.752
Lordosis	26	207		Reference
Cobb angle of fused vertebrae
≤4°	17	99	0.218	1.561	0.769–3.169
>4°	14	126		Reference
ROM of C2–C7
>35°	18	129	0.986	1.006	0.493–2.056
≤35°	13	96		Reference
T1 slope
>20°	19	141	0.712	0.873	0.423–1.801
≤20°	12	84		Reference

CASP – clinical adjacent-segment pathology; ROM – range of motion; CI – confidence interval.

Table 4

Cox’s proportional hazards regression model for predictive factors of CASP following single-level ACDF.

	P value	Hazard Ratio	95% CI
C5–C6 level involved	0.196	2.014	0.696–5.825
Congenital stenosis	0.002	3.250	1.538–6.867
Degeneration of adjacent segment	0.011	2.681	1.259–5.709
Curve pattern of C2–C7	0.227	1.836	0.685–4.918

CASP – clinical adjacent-segment pathology; ACDF – anterior cervical discectomy and fusion; CI – confidence interval.

Discussion

After its introduction in the treatment of degenerative cervical lesions, ACDF is widely used and is reported to produce good results. CASP is a common but serious problem in patients after this procedure, and avoiding adjacent-segment pathology is an important issue among surgeons.

The pathogenesis of adjacent-segment pathology after discectomy and fusion is not completely understood. Cervical arthrodesis has been blamed for accelerating degenerative changes at adjacent spinal segments [13]. Many researchers believed that these degenerative changes might be associated with increased segmental motion and mechanical stress [4,14,15]. Some biomechanical studies showed that the loss of mobility at a given spinal level increased the range of motion and intra-disc pressure at both cephalad and caudal adjacent spinal segments [16,17]. However, other reports found that these degenerative changes are part of the natural aging process of the cervical spine, and they pointed out that there was a lack of definite evidence that arthrodesis increases the incidence of adjacent-segment degeneration [5,18].

The reported incidence of adjacent-segment pathology varied greatly with different lengths of follow-up [3,19]. Adjacent-segment pathology is not always symptomatic, and radiographic adjacent-segment pathology after fusion is not correlated with clinical outcome. Our study only focused on clinical adjacent-segment pathology of the cervical spine. In the present study, the results revealed that 10.01% of patients develop new symptoms within 5 years, and the incidence increases to 23.89% after 10 years. The incidence rate of CASP was an average of 2.46% per year. Several studies agree with our results. For example, in the study by Lee et al., 21.9% of patients would need secondary surgery on adjacent segments by 10 years postoperatively, and secondary surgery on adjacent segments occurred at a relatively constant rate of 2.3% per year [8]. Lawrence et al. reported the rate of development of new symptomatic degeneration in the cervical spine after ACDF was between 1.6% and 4.2% per year [20].

Identification of the predictive factors for CASP development would enable surgeons to identify patients at the greatest risk and adjust their monitoring and follow-up decisions. The present study showed that congenital stenosis and pre-existing degenerative changes of adjacent segments were 2 predictive factors that may be associated with postoperative CASP. Patients with congenital stenosis and pre-existing degenerative changes of adjacent segments in MRI had 3.2 times and 2.7 times greater risk, respectively, than those without.

Whether CASP is due to disease progression or a fusion-associated phenomenon cannot be determined yet. Previous studies have shown that pre-existing degeneration of disc is a risk factor for the development of CASP. Goffin et al. reported that additional radiologic degeneration at the adjacent disc levels was found in 92% of their patients, who are more likely to have clinical deterioration. The severity of this degeneration is also correlated with the length of time after surgery [21]. Park et al. calculated the rate of preoperative adjacent-segment degeneration in patients after anterior cervical surgery, and found that the rate of degeneration was significantly higher in patients with adjacent-segment disease than in those without (85% vs. 52.2%). They further concluded that adjacent-segment disease may be associated with a natural history of cervical spondylosis rather than arthrodesis [22]. Although the exact pathogenesis cannot be determined based on our study, we assumed that the pre-existing degenerative disease may act as a triggering factor for CASP.

Congenital stenosis has previously been shown to be a risk factor for degenerative disc disease [23,24], and in the present study we demonstrated it to be a predictive factor for CASP. The mechanism for this phenomenon is not very clear. Morishita et al. [25] found that a congenitally narrow canal had different effects on cervical kinematics, which means that subjects with a congenitally narrow cervical canal may be exposed to heavy mechanical loading at the cervical spine. In a recent study, Eubanks et al. reviewed a total of 635 patients retrospectively, and concluded that congenital stenosis appears to increase the incidence of radiographic adjacent-segment disease, but it does not appear to predict symptomatic adjacent-segment disease [26]. In their study, patients only had a median follow-up time of 36.2 months. We assumed that congenital narrowing is associated with a greater degree radiographic degeneration at the adjacent level initially, and it will translate into clinically significant symptoms with longer follow-up time.

Several limitations need to be considered in the present study. First, the incidence of CASP was not high; although we tried to collect all eligible patients, the sample size was relatively small, and this may weaken this study in the assessment of predictive factors. Second, were collected prospectively, but the study was retrospective in nature. We believe that a prospective, multicenter study involving a large population and long-term follow-up is needed to confirm our results.

Conclusions

In conclusion, after ACDF procedures, 10.01% of patients are predicted to develop new symptoms of CASP within 5 years, and the incidence increases to 23.89% after 10 years. The incidence of CASP is an average of 2.46% per year. Congenital stenosis and pre-existing degenerative changes of adjacent segments were 2 independent predictive factors that may be associated with postoperative CASP.