The Effect of Preoperative Cervical Spondylosis on Heterotopic Ossification After Cervical Disc Replacement

Abstract

Study Design

Retrospective cohort study.

Objectives

This study aimed to explore the effect of preoperative cervical spondylosis on the heterotopic ossification (HO) formation in different locations after cervical disc replacement (CDR).

Methods

The degree of preoperative cervical spondylosis was evaluated radiologically, including the intervertebral disc, uncovertebral joints, facet joints and ligaments. The effects of cervical spondylosis on the HO formation after CDR were analyzed according to the location of HO. Multivariate logistic regression was performed to identify the independent factors.

Results

149 patients with a total of 196 arthroplasty segments were involved in this study. HO, anterior HO (AHO), and posterior HO (PHO) developed in 59.69%, 22.96%, and 41.84% levels, respectively. The significant factors in univariate analysis for PHO after CDR included the disc height loss, anterior osteophytes, preoperative uncovertebral joint osteophytes and facet joint degeneration. The incidence of adjacent segment degeneration (ASD) was significantly higher in the PHO group compared to that without PHO at the last follow-up (P = .003). The disc height loss in high-grade HO was significantly more than that in low-grade group (P = .039). Multivariate analysis identified disc height loss was the only independent factor for PHO (P = .009). No significant degenerative factors related to the formation of AHO were found.

Conclusions

Preoperative cervical spondylosis predominantly affected the HO formation in the posterior disc space after CDR. The disc height loss was an independent risk factor for PHO formation. Rigorous criteria for the extent of preoperative disc height loss should be used when selecting appropriate candidates for CDR.

Introduction

Cervical disc replacement (CDR) is a relatively new technique in cervical spine surgery as an alternative treatment to anterior cervical discectomy and fusion (ACDF) for cervical degenerative disc disease at 1 or more levels. The initial design concept of CDR is to preserve motion of the index level and diminish the adjacent segments’ stress transfer, which theoretically reduce degeneration at the adjacent levels compared to ACDF.^1-3 Unfortunately, heterotopic ossification (HO) raised as an adverse outcome after CDR that may decrease the motion of index level and even develop to complete intervertebral fusion, which is against the design philosophy of disc replacement with motion-preserving. ⁴ An extensive body of literatures have emerged to investigate the incidence, predisposing factors and clinical implications since the phenomenon was noted. However, the underlying mechanism of HO has not been clearly elucidated until present.

Degenerative cervical spondylosis refers to age-related wear and tear that affect elements of the cervical spine over time, including the intervertebral discs, facet joints, and other connective-tissue structures. ⁵ Preoperative degenerative spondylosis is considered 1 of the contributors to HO formation after CDR. Wu et al. ⁶ reported that the rate of HO formation was significantly higher in patients with spondylosis compared to that with soft-disc herniations. In recent years, several qualitative or quantitative studies confirmed the degree of degeneration in the surgical level was positively correlated with postoperative HO,^7-10 whilst some studies found preoperative degenerative changes exert no significant influence on the occurrence of HO.^11,12 However, few studies investigated the effect of all the components of cervical spondylosis on HO. When evaluating the degenerative degree of cervical spondylosis, most investigations tended to involve the intervertebral discs and ligament elements, the effects of degeneration in uncovertebral joint and facet joint have been rarely studied. Chung et al. ¹³ identified preoperative uncovertebral hypertrophy as a significant risk factor for the occurrence of motion-affecting HO through a survivorship analysis. Cao et al. ¹⁴ found there was significant positive correlation between preoperative uncovertebral joints degeneration and HO. Due to the interdependent process among disc degeneration, uncovertebral joint degeneration and facet joint degeneration, the independent factors in preoperative spondylosis causing HO formation was still unclear. Furthermore, previous studies suggested that the mechanism of HO in anterior and posterior disc space may be different.^15,16. The impacts of preoperative spondylosis on the occurrence of HO in different location has been poorly studied.

In the current study, preoperative spondylosis of intervertebral disc, uncovertebral joint, facet joint and ligament elements were comprehensively evaluated in patients who underwent CDR with a minimum follow-up of 24 months. This study aimed to investigate the influences of preoperative spondylosis at surgical segment on postoperative HO, and to explore whether the influence of degeneration on HO formation in anterior and posterior disc space is different.

Methods

Patient Population

This was a retrospective study and the study protocol was approved by the Ethics Committee of West China Hospital of Sichuan University (Project License Number 20190946). All patients provided written informed consent. Patients who underwent single-level and two-level cervical disc replacement using the Prestige-LP (Medtronic Sofamor Danek USA, Inc., Memphis, Tennessee, USA) from 2009 to 2018 and had a minimum of 2 years of follow-up data were included in this study. The inclusion criteria were as follows: 1) 18-65 years old; 2) single- or 2-level cervical disc disease causing symptomatic radiculopathy or myelopathy between C3 and C7; 3) failure of conservative treatment for at least 12 weeks. The exclusion criteria consisted of: 1) instability, irreducible kyphosis, or severe degeneration at the surgical segment; 2) ossification of the posterior longitudinal ligament; 3) ankylosing spondylitis, rheumatoid arthritis, or other autoimmune or metabolic bone disease; 4) osteoporosis (T-score ≤ -2.5); 5) prior cervical spine surgery.

Surgical Techniques

All operations were performed by the same senior spine surgeon. The patient was placed with the neck in a neutral position after general anesthesia, and a right-side Smith-Robinson approach was performed at the index level. After exposure, the anterior longitudinal ligament, intervertebral disc, posterior longitudinal ligament and osteophytes were removed to achieve a complete discectomy and decompression. A high-speed burr was used to carefully prepare the endplate and a trial implant was used to determine the size of prosthesis. Then, a rail cutter guide and bit were used to drill the fixation channels in the endplate and a correctly sized Prestige-LP disc was inserted into the prepared intervertebral space. After verification of the proper placement of implant through anterior-posterior and lateral fluoroscopy, copious irrigation with normal saline was used to wash away the bone dust during endplate preparation in every case. The same procedure was performed at other level in patients with double segmental lesions. A drainage tube was inserted after meticulous hemostasis before closing the incision. Nonsteroidal anti-inflammatory drugs were not routinely used in this cohort.

Radiological Evaluation

All the radiological data were collected preoperatively, 1 week and at last follow-up postoperatively. The degeneration degree of the index level was evaluated preoperatively in the neutral lateral radiographs using a quantitative scoring system proposed by Walraevens et al., ¹⁷ which consists of height loss (score 0-4), anterior osteophytes (score 0-3) and endplate sclerosis (score 0-2), with higher scores representing more severe degeneration. Height loss is defined as the middle disc height with respect to a normal middle disc height at an adjacent level. Anterior osteophytes are assessed through its length relative to the anteroposterior diameter of the middle of the corresponding vertebral body. The endplate sclerosis is distinguished among no apparent sclerosis, just detectable and definite sclerosis according to the degree of detectability.^10,17 The uncovertebral joint was assessed preoperatively using the coronal reconstruction computed tomography (CT) to identify osteophytes at the uncovertebral joint area. Facet joint degeneration was evaluated using a computed tomographic grading system reported by Park et al. ¹¹ (grade 1-4), with higher grade denoting more advanced degree of degeneration. HO was graded based on the McAfee classification criteria and classified into low-grade (grade 1-2) and high-grade HO (grade 3-4). ⁷ According to the location of HO on lateral neutral radiographs, index segments were divided into anterior heterotopic ossification (AHO) and non-AHO groups, posterior heterotopic ossification (PHO) and non-PHO groups, respectively ¹⁵ (Figure 1). Adjacent segment degeneration (ASD) was defined by the presence of at least 1 of the following conditions^7,18: 1) new anterior or enlarging osteophyte formation, 2) narrowing of the disc space ≥ 10% compared to preoperative radiographs, 3) ossification of the anterior longitudinal ligament (OALL). The ligamentum nuchae ossification progression was defined as new or enlarging ossification of ligamentum nuchae at last follow-up compared with pre-operation. Shell angle was defined as the angle formed between the superior and inferior endplate of the prosthesis, and segmental range of motion (ROM) was measured on the flexion-extension radiographs. The above assessments were performed by 2 surgeons independently, and the discrepancies were settled by consulting a senior surgeon to achieve consensus.

Figure 1.

Low-grade heterotopic ossification (HO): anterior HO (AHO) (A) and posterior HO (PHO) (B). High-grade HO: both AHO and PHO (C), and PHO (D). Abbreviation: HO, heterotopic ossification; AHO, anterior heterotopic ossification, PHO, posterior heterotopic ossification.

Clinical Outcomes

Clinical outcomes were assessed using the Japanese Orthopaedic Association (JOA) score, Neck Disability Index (NDI) functional score, and visual analog scale (VAS) score. The JOA score was used to evaluated the neurological status of patients with myelopathy, the NDI score was used to assess the function of the neck, and the VAS score was used to assess the neck and arm pain of the patients. The data were collected preoperatively and at the last follow-up postoperatively.

Statistical Analysis

Statistical analyses were performed using SPSS Version 25.0 software (IBM Corp., Armonk, New York, USA). The results were presented as mean ± standard deviation for continuous variables and number of cases and percentage for categorical variables. A paired t-test was used to compare pre- and post-operative parameters. The independent t-test or the Mann-Whitney U test was used to compare continuous variables depending on the normality of data. The Chi-square or Fisher’s exact test was used for categorical variables. A binary logistic regression analysis was used to identify risk factors for HO. Variables with P < .10 in the univariate analysis were included in the multivariate logistic regression analysis. Due to the lack of eligible variable in AHO group, variables with P < .15 in the univariate analysis were in included in the multivariate analysis. A 2-tailed P-values < .05 was considered statistically significant.

Results

Demographic Data

This retrospective study included 149 patients (64 male, 85 female) with a minimum 24 months follow-up data. The mean age of the patients was 43.32 ± 8.24 years. The median follow-up time was 41.0 months (25% quartile: 35.0, 75% quartile: 56.0 months; range 24-131 months). A total of 196 segments in 149 patients were evaluated, of whom 102 patients underwent single-level and 47 patients underwent 2-level CDR. The most common operated level was C5/6 (n = 116), followed by C4/5 (n = 42), C6/7 (n = 26) and C3/4 (n = 12). Overall, HO was diagnosed in 88 levels (constituent ratio: 44.90%) with low-grade HO and 29 levels (constituent ratio: 14.80%) with high-grade HO were identified. AHO occurred in 43 patients and in 45 levels (22.96%) and PHO was observed in 68 patients and in 82 levels (41.84%). The clinical outcomes including the JOA, NDI, VAS neck and VAS arm scores showed significant improvement at last follow-up compared to pre-operation (Table 1). No patients required a reoperation at the surgical level or adjacent levels during the course of this study.

Table 1.

Demographic Data.

Variable	Value
Number of patients (segments)	149 (196)
Age (years)	43.32 ± 8.24
Sex
Male	64
Female	85
Follow-up (months)	48.95 ± 23.14
CDR level
1	102
2	47
Level distribution
C3/4	12
C4/5	42
C5/6	116
C6/7	26
JOA score
Preoperative	12.17 ± .95
Last follow-up	15.98 ± .66*
NDI score
Preoperative	22.43 ± 3.26
Last follow-up	5.66 ± .60*
VAS neck score
Preoperative	5.87 ± .50
Last follow-up	1.58 ± .51*
VAS arm score
Preoperative	5.68 ± .79
Last follow-up	.97 ± .53*

Abbreviation: CDR, cervical disc replacement; JOA, Japanese Orthopaedic Association; NDI, neck disability index; VAS, visual analog scale.

^*Significant difference compared with the preoperative values.

Univariate Analysis of HO According to the Location and the Influence on clinical Outcomes

All the index level were divided into AHO and non-AHO group, PHO and non-PHO group, respectively, according to the location of HO occurrence. The factors of age, sex, follow-up time, level distribution and the assessment of preoperative spondylosis were involved in the univariate analysis. For the analysis of AHO, no significant factors were found. The degree of overall disc degeneration, disc height loss, anterior osteophytes were significantly higher in the PHO group than in the non-PHO group (P<.001 for overall disc degeneration score; P=.003 for mean disc height loss score; P=.006 for mean anterior osteophyte score) (Figure 2). Similarly, the presence of preoperative uncovertebral joint osteophytes and facet joint degeneration were also significantly frequent in the PHO group compared to non-PHO group (P = .045 for uncovertebral joint osteophyte; P = .035 for facet joint degeneration). There was no significant difference in the shell angle or ROM postoperatively neither comparing the AHO and non-AHO group nor the PHO and non-PHO group. At the last follow-up, the ROM at index level was significantly lower in the PHO group than the non-PHO group (6.92 ± 4.43 vs 8.81 ± 4.82, P = .003) while that in AHO group was lower compared to non-AHO group, but the difference was not significant (6.95 ± 5.18 vs 8.35 ± 4.58, P = .053). In addition, the incidence of ASD was significantly higher in the group with PHO compared to that without PHO (32.9% vs 14.9%, P = .003) at the last follow-up (Figure 3) (Table 2). There were no significant differences in JOA, NDI, VAS scores between patients with and without AHO, and patients with and without PHO at the last follow-up. Although larger ROM of C2-C7 was noted in patients without AHO than those with AHO and patients without PHO than those with PHO, the differences were not statistically significant at the last follow-up. The incidence of ASD was significantly higher in patients with PHO compared with those without PHO (41.2% vs 18.5%, P = .002) (Table 3).

Figure 2.

A 46-year-old woman who underwent cervical disc replacement using Prestige-LP at the C5/6 level. The preoperative lateral radiograph showed that a decreased intervertebral disc height, anterior osteophyte and ossification of anterior longitudinal ligament (OALL) at C5/6 level (A). The preoperative coronal reconstruction computed tomography (CT) images indicated the uncovertebral joint osteophyte (B). The preoperative magnetic resonance image (MRI) showed cervical spondylosis at the C5/6 level (C). One week postoperatively, lateral radiograph showed cervical disc replacement at C5/6 level (D). At 50 months follow-up, neutral X-ray (E), magnified view of C5/6 region (F) and dynamic X-rays indicated low-grade posterior heterotopic ossification at the C5/6 level and the range of motion of the prosthesis was 13.19° (G and H).

Figure 3.

A 52-year-old man who underwent cervical disc replacement using Prestige-LP at the C5/6 level. The preoperative lateral radiograph showed that a decreased intervertebral disc height, anterior osteophyte and ossification of anterior longitudinal ligament at C5/6 level (A). The preoperative coronal reconstruction computed tomograph images indicated the uncovertebral joint osteophyte (B). The preoperative MRI showed cervical spondylosis at the C5/6 level (C). One week postoperatively, lateral radiograph showed cervical disc replacement at C5/6 level (D). At 40 months follow-up, neutral X-ray (E), magnified view of C5/6 region (F) and dynamic X-rays (G and H) indicated high-grade posterior heterotopic ossification at the C5/6 level with limited prosthesis mobility and newly formed anterior osteophytes at caudal adjacent segment. The range of motion of the prosthesis was 3.07°.

Table 2.

Univariate Analysis of Heterotopic Ossification of Each Segment Depending on the Location at Last Follow-up.

	AHO			PHO
	(−) n=151	(+) n=45	P	(−) n=114	(+) n=82	P
Age	43.93 ± 8.34	42.16 ± 7.68	.203	42.76 ± 8.20	44.59 ± 8.14	.236
Sex			.301			.908
Male	64	23		51	36
Female	87	22		63	46
Follow-up (months)	47.84 ± 21.62	54.82 ± 29.69	.498	47.48 ± 20.53	52.17 ± 27.64	.882
Level distribution			.384			.309
C3/4	9	3		8	4
C4/5	35	7		20	22
C5/6	90	26		68	48
C6/7	17	9		18	8
Overall degree of disc degeneration	1.53	1.67	.723	1.25	2.00	<.001*
Mean height loss score	.54	.49	.718	.39	.71	.003 *
Mean anterior osteophyte score	.79	.96	.256	.68	1.04	.006 *
Mean endplate sclerosis score	.20	.22	.687	.17	.26	.196
Pre-op posterior osteophyte			.339			.278
Yes	76 (50.3%)	19 (42.2%)		59 (51.8%)	36 (43.9%)
No	75 (49.7%)	26 (57.8%)		55 (48.2%)	46 (56.1%)
Pre-op uncovertebral joint osteophyte			.575			.045 *
Yes	81 (53.6%)	22 (48.9%)		53 (46.5%)	50 (61.0%)
No	70 (46.4%)	23 (51.1%)		61 (53.5%)	32 (39.0%)
Facet joint degeneration			.720			.035 *
1	117 (77.5%)	36 (80.0%)		95 (83.3%)	58 (70.7%)
2	34 (22.5%)	9 (20.0%)		19 (16.7%)	24 (29.3%)
OALL	26 (17.2%)	10 (22.2%)	.447	21 (18.4%)	15 (18.3%)	.982
Pre-op ligamentum nuchae ossification	28 (18.5%)	4 (8.9%)	.124	14 (12.3%)	18 (22.0%)	.071
Progression of ligamentum nuchae	13 (8.6%)	4 (8.9%)	1.000	8 (7.0%)	9 (11.0%)	.331
Post-op
Shell angle (°)	4.94 ± 4.79	3.70 ± 5.17	.137	4.86 ± 5.04	4.38 ± 4.69	.739
ROM at index level (°)	6.77 ± 3.70	7.55 ± 3.59	.144	6.90 ± 3.74	7.02 ± 3.61	.757
Last follow-up
Shell angle (°)	1.64 ± 4.59	1.92 ± 5.67	.739	2.13 ± 4.86	1.12 ± 4.80	.151
ROM at index level (°)	8.32 ± 4.58	6.95 ± 5.16	.053	8.81 ± 4.82	6.88 ± 4.42	.003*
Changes of ROM	1.55 ± 5.25	−.60 ± 6.44*	.024	1.91 ± 5.80	−.14 ± 5.11	.011*
ASD	32 (21.2%)	12 (26.7%)	.440	17 (14.9%)	27 (32.9%)	.003*

Abbreviation: AHO, anterior heterotopic ossification; PHO, posterior heterotopic ossification; OALL, ossification of anterior longitudinal ligament; post-op, values at 1 week after surgery; ROM, range of motion; ASD, adjacent segment degeneration.

Bold values indicate these values would be included in multivariable logistic regression analysis.

^*Significant difference between groups.

Table 3.

Clinical Outcomes of Patients with or Without Anterior Heterotopic Ossification and Patients with or Without Posterior Heterotopic Ossification at Last sollow-up.

	AHO			PHO
	(−) n=106	(+) n=43	P	(−) n=81	(+) n=68	P
JOA score	15.95 ± .64	16.08 ± .76	.555	16.04 ± .69	15.92 ± .64	.530
NDI score	5.69 ± .61	5.60 ± .58	.537	5.65 ± .60	5.68 ± .61	.672
VAS neck score	1.60 ± .51	1.53 ± .50	.474	1.54 ± .50	1.63 ± .52	.314
VAS arm score	.97 ± .53	.95 ± .53	.847	.93 ± .59	1.01 ± .44	.289
ROM of C2-C7	50.79 ± 12.99	46.59 ± 13.13	.077	51.29 ± 12.55	47.53 ± 13.59	.081
ASD	27 (25.5%)	16 (37.2%)	.152	15 (18.5%)	28 (41.2%)	.002*

Abbreviation: AHO, anterior heterotopic ossification; PHO, posterior heterotopic ossification; JOA: Japanese Orthopaedic Association; NDI: neck disability index; VAS: visual analog scale; ROM, range of motion; ASD, adjacent segment degeneration.

^*Significant difference between groups.

The univariate analysis was also performed between low-grade HO and high-grade HO. The disc height loss score of high-grade HO group was significantly higher than that of the low-grade HO group (P = .008) (Table 4). Regarding the clinical outcomes, no significant differences were noted for JOA, NDI, VAS scores, and ASD at the last follow-up between patients with and without high-grade HO. The ROM of C2-C7 was significantly lower in the high-grade HO groups than group without high-grade HO (41.41 ± 13.17° vs 51.38 ± 12.47°, P<.001) (Table 5).

Table 4.

Effects of Preoperative Spondylosis on the Degree of Heterotopic Ossification.

	Low-grade HO (n = 88)	High-grade HO (n = 29)	P
Follow-up (months)	48.47 ± 23.90	60.66 ± 33.71	.129
Level distribution			.086
C3/4 and C4/5	30	5
C5/6 and C6/7	58	24
Overall degree of disc degeneration	1.80	2.17	.164
Mean height loss score	.52	.97	.008*
Mean anterior osteophyte score	.99	1.07	.751
Mean endplate sclerosis score	.28	.14	.227
Pre-op posterior osteophyte			.710
Yes	49 (55.7%)	15 (51.7%)
No	39 (44.3%)	14 (48.3%)
Pre-op uncovertebral joint osteophyte			.619
Yes	38 (43.2%)	11 (37.9%)
No	50 (56.8%)	18 (62.1%)
Facet joint degeneration			.523
1	66 (75.0%)	20 (69.0%)
2	22 (25.0%)	9 (31.0%)
OALL	19 (21.6%)	4 (13.8%)	.359
Ossification of ligamentum nuchae	15 (17.0%)	7 (24.1%)	.397
Progression of ligamentum nuchae	8 (9.1%)	3 (10.3%)	.545
ASD	24 (27.3%)	11 (37.9%)	.277

Abbreviation: AHO, anterior heterotopic ossification; PHO, posterior heterotopic ossification; OALL, ossification of anterior longitudinal ligament; ASD, adjacent segment disease.

^*Significant difference between 2 groups.

Table 5.

Clinical Outcomes of Patients With or Without High-Grade Heterotopic Ossification at Last Follow-up.

	High-grade HO (−) (n = 122)	High-grade HO (+) (n = 27)	P
JOA score	16.00 ± .65	15.90 ± .74	.665
NDI score	5.69 ± .58	5.56 ± .70	.361
VAS neck score	1.57 ± .50	1.63 ± .56	.705
VAS arm score	.95 ± .54	1.04 ± .44	.430
ROM of C2-C7	51.38 ± 12.47	41.41 ± 13.17	<.001*
ASD	32 (26.2%)	11 (40.7%)	.132

Abbreviation: HO: heterotopic ossification; JOA: Japanese Orthopaedic Association; NDI: neck disability index; VAS: visual analog scale; ROM, range of motion; ASD, adjacent segment degeneration.

^*Significant difference between groups.

Multivariate Logistic Regression Analysis of HO According to the Location

Based on the results of univariate analysis, factors with P < .10 were involved in the multivariate logistic regression analysis of PHO group, including disc height loss, anterior osteophytes, preoperative osteophytes in the uncovertebral joint, facet joint degeneration and ligamentum nuchae ossification. Due to the lack of qualified factors in AHO group, we expanded the criteria to involve factors with P < .15 for multivariate analysis, including preoperative ligamentum nuchae ossification, postoperative shell angle and ROM at index level. No significant factors were identified in the analysis of AHO group (Table 6). For PHO group, height loss was identified as independent risk factors (P=.009, OR=1.820) (Table 7). The optimal cutoff score for disc height loss was .5, and the most suitable cutoff for the degree of disc height loss denoted as percentage ¹⁷ to predict PHO was 18.52%, with disc height loss ≥18.52% indicating a significant higher risk of PHO (P<.001, odds ratio [OR] 3.23, 95% CI 1.76-5.91). Besides, the incidence of PHO in segments with different disc height loss score were further investigated. As shown in Table 8, the incidence of PHO was significantly higher in segments with larger height loss score (P = .005).

Table 6.

Multivariate Logistic Regression Analysis of Risk Factors for Anterior Heterotopic Ossification.

Risk factors	p	OR	95% CI
Pre-op ligamentum nuchae ossification	.147	.437	.143-1.338
Post-op shell angle	.225	.956	.890-1.028
Post-op ROM at index level	.206	1.059	.969-1.158

Abbreviation: Post-op, values at 1 week after surgery; ROM, range of motion.

Table 7.

Multivariate Logistic Regression Analysis of Risk Factors for Posterior Heterotopic Ossification.

Risk factors	p	OR	95% CI
Mean height loss score	.009*	1.820	1.160-2.856
Mean anterior osteophyte score	.052	1.454	.997-2.120
Pre-op uncovertebral joint osteophyte	.257	1.438	.767-2.696
Facet joint degeneration	.107	1.814	.879-3.743
Pre-op ligamentum nuchae ossification	.195	1.720	.757-3.909

^*Statistical significance.

Table 8.

The incidence of PHO in segments with different disc height loss score before surgery.

	0	1	2	P
PHO				.005*
No	75 (65.8%)	33 (54.1%)	6 (28.6%)
Yes	39 (34.2%)	28 (45.9%)	15 (71.4%)

Abbreviation: PHO, posterior heterotopic ossification.

^*Statistically significant.

Discussion

The theoretical advantages of CDR are based on the premise that motion preservation and load-sharing at the index level would mitigate the risk of ASD. ¹⁹ However, the occurrence of HO can be an intractable issue after CDR. Although previous studies have shown that HO did not worsen the patient-reported clinical outcomes, the progressive character of HO may lead to motion restriction at the treated level, which is contrary to the fundamental goal of CDR.^2,7,20 Thus, identifying the real risk factors for HO and selecting appropriate candidates are of critical importance in the improvement of CDR. In the present study, we found that only HO in the posterior disc space is affected by preoperative degeneration. In the multiple elements of preoperative spondylosis, disc height loss may be the only independent risk factor for PHO formation.

A large number of studies has been conducted to explore the cause and risk factors for HO with different focus since it was first reported by Parkinson and Sekhon, ⁴ among which preoperative cervical spondylosis accounted for an important aspect. Cervical spondylosis is a generalized aging process encompassing a sequence of degenerative changes of cervical spine. The three-joint complex, including the 2 posterior joints and the intervertebral disc, and the bilateral uncovertebral joints are involved in the pathogenesis. ²¹ Although insignificant correlations were reported in a few studies, most studies emphasized the significant role of preoperative degeneration in the formation of HO following CDR. However, few studies investigated the effect of all the components of cervical spondylosis on HO. Chung et al ¹³ retrospectively reviewed 19 patients who underwent single-level CDR with Bryan. They found preoperative uncovertebral hypertrophy was a significant risk factor for high-grade HO through survivorship analysis. Wu et al. ⁶ evaluated forty patients who underwent single-level cervical arthroplasty using Bryan discs and divided them into a soft-disc herniation group and a spondylosis group. A significant higher rate of HO formation in spondylosis group was found compared with that with a soft-disc herniation. They defined the spondylosis in this study as disc disease with osteophytes, calcified disc, or marginal spurs. Jin et al. ¹⁵ reported HO in posterosuperior disc space was significantly related with the presence of preoperative ossification and ossification in the ligamentum nuchae. However, in a study of 170 patients, Yi et al. ¹¹ evaluated the preoperative existing degenerative change including osteophytes, ossification of the ligaments and disc degeneration through detailed grades to investigate HO. Their results suggested only male gender and prosthesis type were the predisposing factors for HO and the preoperative degeneration of cervical spine exerted no significant influence on the occurrence of HO. In recent years, Tian et al. ⁹ revealed the only predisposing factor for high-grade HO after surgery was pre-existing paravertebral ossification before operation through a retrospective review of 71 patients treated with Bryan discs. Zhou et al. ⁷ found patients with more severe preoperative spondylosis had significant higher rates of postoperative HO using the Kellgren-Lawrence scale. Li et al. ⁸ also concluded the occurrence of HO after CDR is correlated to the severity of preoperative degeneration. In addition, a significant positive correlation between preoperative uncovertebral joints degeneration and postoperative HO was found in the study of Cao et al. ¹⁴ However, Guo et al. ¹² analyzed 57 levels of cervical arthroplasty and no significant association was found between the HO formation and preexisting cervical degeneration. The discrepancy in results is probably due to the differences in the sample size, prosthesis type, population group, evaluating approach and components of preoperative cervical spondylosis.

In the analysis of 196 segments replaced by Prestige-LP discs, this study comprehensively assessed the preoperative spondylosis including the components of three-joint complex, uncovertebral joints and ligaments. The results of univariate analysis suggested preoperative height loss, anterior osteophytes, uncovertebral joint osteophytes and facet joint degeneration was significant degenerative factors for the occurrence of PHO after CDR. Further multivariate logistic regression analysis, however, identified the disc height loss before surgery was the only independent factor for PHO formation. This is probably due to the interdependently degenerative process among components in each cervical segment. The relationships between disc height loss and osteophyte formation in the vertebrae margins, uncovertebral joint osteophytes and facet joint degeneration have not been clearly defined. Walraevens et al. ¹⁷ indicated that height loss is the strongest predictor for disc degeneration. Previous study also demonstrated the importance of preoperative disc height loss and postoperative degree of distraction on the development of HO in CDR. ²² With aging, the degenerative cascade causes gradual desiccation of the intervertebral disc resulting in the loss of disc height, which transfer the biomechanical loads from the nucleus pulposus to the surrounding annulus fibers. ²³ The decreased disc and facet joint stability lead to an altered biomechanical environment and abnormal motion. Traction stress related to instability may cause the development of anterior and posterior marginal endplate and uncinate spurs. ²⁴ It was also stated that traction at the site of osseous attachment of annulus fibers to the vertebral surface may result in the reactive development of osteophytes.^21,25 Sohn et al ²⁶ found a significant correlation between the osteophytes formation and disc height loss in a cadaveric study. Besides, a numerical simulation study revealed that osteophyte formation is an adaptive bone remodeling process in response to the adjusted mechanical environment, which were mainly caused by the intervertebral disc degeneration. ²⁷

Spondylosis of uncovertebral joint is characterized by osteophytes on the uncinate process ²⁸ ; hence, the degeneration of uncovertebral joint was evaluated according to the presence of osteophyte in this study. This factor showed statistical significance in univariate analysis yet was insignificant in logistic regression analysis, which is probably because of its degenerative process resulted from the intervertebral disc degeneration. Traction stress due to instability may lead to the formation of uncinate osteophytes. ²⁴ Besides, Cao et al. ¹⁴ also found a significant positive correlation between preoperative intervertebral space degeneration and uncovertebral joint degeneration. With progressive diminution of disc height related to degeneration of the cervical intervertebral joint, the bone protuberances about the uncovertebral articulations approach each other and are pressed firmly together, which might lead to the occurrence of osteophytes. ²¹ The impact of facet joint degeneration presented analogous pattern with uncovertebral degeneration in univariate and multivariate analysis. The three-joint complex in each cervical motion segment are highly functional interdependent, such that abnormalities of the disc tend to eventually affect the facet joints. ²⁹ The diminution in disc height could cause pathologic loading of the facet joints, as less weight is supported by the disc when it becomes functionally incompetent.^28,30 Nonetheless, the correlation between the degeneration of intervertebral disc and facet joints remains unclear. ⁵ In addition, our study suggested the disc height loss in high-grade HO was significantly higher in comparison with low-grade HO. Kim et al. ³¹ also found that the postoperative distraction height was significantly higher in the high-grade HO group, which usually accompany with more preoperative disc height loss. Therefore, the formation of high-grade HO is likely to restabilize the altered biomechanical environment at surgical segments.

Surgical procedures and postoperative biomechanical changes due to disc height loss may also contribute to the development of HO. Patients with more preoperative disc height loss may produce more bone dusts during endplate preparation and decompression procedures, which is considered an important predisposing factor for HO formation.^32-34 Besides, segments with more disc height loss usually experience higher degree of distraction, resulting in greater changes in the postoperative biomechanical environments such as intervertebral height and segmental mobility.^22,31 Ganbat et al ¹⁶ found that the addition of compressive force predominantly increased HO formation at the posterior parts of vertebrae, while HO was mainly formed in the anterior parts of vertebrae for anteriorly directly shear force through a three-dimensional finite element analysis. Likewise, prosthesis implantation at heigh loss segments could lead to greater pressure at the bone-implant interface compared to those with normal preoperative disc height, ³⁵ which probably increase the risk for posterior HO. In the present study, no significant correlation was found between preoperative degeneration and the occurrence of AHO neither in the univariate analysis nor in the multivariate regression. Previous studies suggested that HO in the anterior disc space may be more related to the increased shear force rather than the degenerative factors.^15,16,36 Thus, the mechanism and predisposing factors of AHO and its avoidance needed further investigation.

The prevention of ASD is 1 of the main goals of CDR. In the current study, we found that the rate of ASD was significantly higher in patients with PHO compared to those without PHO at the last follow-up. Interestingly, the correlation was not significant when further comparing the low-grade HO and high-grade HO affecting motion, which is consonant with previous report. ⁷ Zeng et al ³⁷ found the progression rate of paravertebral ossification at the adjacent level was not significantly different from those of non-surgical and non-adjacent levels. These results may reveal that ASD tend to be the natural degeneration of cervical spine during follow-up rather than the complication caused by the changes of biomechanical environment. ³⁸

However, there were several limitations to the study. First, the retrospective nature and a single institution study presented inherent weaknesses and limited generalizability of the results. Second, we use lateral X-ray rather than CT to detect HO and the ossification of uncovertebral joint area was not considered in our study. Previous studies have suggested a high incidence of HO in the uncovertebral region,^9,37 which can be uneasy to detectable using plain films. However, radiography has a lower radiation dose and is more acceptable for patients when they do not have any clinical symptoms. Third, the follow-up time of our study was relatively short while the incidence and grade of HO may be more accurate and stable in a longer follow-up. Fourth, the study was limited to the Prestige-LP discs; thus, the results of HO formation may not be suitable in other type of prostheses with different design. Fifth, we did not identify significant factors for anterior HO neither in the univariate analysis nor in the logistic regression. Further studies concentrating on the biomechanical impacts are needed to be conducted.

Conclusion

In conclusion, this study revealed that preoperative cervical spondylosis predominantly affected the HO formation in the posterior disc space after CDR. The degree of preoperative disc height loss, anterior osteophytes, uncovertebral joint osteophytes and facet joints degeneration were more severe in patients with posterior HO after CDR. However, in the multiple elements of preoperative cervical spondylosis, only the disc height loss was an independent risk factor for posterior HO formation. Rigorous criteria for the degree of preoperative disc height loss should be used when selecting appropriate candidates for CDR.

List of abbreviations

CDR

cervical disc replacement

ACDF

anterior cervical discectomy and fusion

HO

heterotopic ossification

CT

computed tomography

AHO

anterior heterotopic ossification

PHO

posterior heterotopic ossification

ASD

adjacent segment degeneration

OALL

ossification of the anterior longitudinal ligament

ROM

range of motion

JOA

Japanese Orthopaedic Association

NDI

neck disability index

VAS

visual analog scale; SVA: sagittal vertical axis

ORCID iDs

Yi-Wei Shen https://orcid.org/0000-0003-1308-4584

Hao Liu https://orcid.org/0000-0002-2992-6232