Heterotopic ossification in cervical disc arthroplasty: Is it clinically relevant?

STUDY
DESIGN: Retrospective cohort study.
OBJECTIVE: To analyze the presence and clinical relevance of heterotopic ossification (HO) at 3 years mean follow-up.
METHODS: Thirty patients suffering from cervical radiculopathy and/or myelopathy treated with anterior disc replacement (ADR) were studied. HO was classified using the McAfee grading system. Range of motion was measured from flexion and extension x-rays. Short-form 36 and neck disability index (NDI) assessed functional outcome.
RESULTS: Forty-five prostheses were implanted in 30 patients with cervical radiculopathy and/or myelopathy, mean age 40.9 years. Nineteen patients received 1 level and 11 patients received multilevel disc replacement. The incidence rate of HO was 42.2% (19 levels). Segmental range of motion was ≥3° in 93.8% of patients with HO. There was no significant difference in functional scores between those who did and those who did not develop HO. Males tended to develop HO more frequently than females, though this was not statistically significant. The indication for surgery (soft disc hernia or spondylosis) was not associated with the formation of HO.
CONCLUSIONS: Functional improvement is maintained despite the presence of HO following cervical disc arthroplasty. Indications for arthroplasty should not be halted by the risk of HO. [Table: see text] The definiton of the different classes of evidence is available on page 83.

Study Rationale and Context

Maintenance of motion following arthroplasty can be hindered by the development of heterotopic ossification (HO), whose causes remain uncertain.1,2,3,4,5 Whether the formation of HO following cervical disc arthroplasty is of clinical or functional importance is also unknown. The aims of this study are to investigate the rate of HO at a longer follow-up (mean 3 years), to determine if the presence of HO negatively affects functional outcome and to identify factors that may precipitate the formation of HO following cervical disc arthroplasty.

What is the incidence of HO following cervical arthroplasty?

Does the development of HO negatively influence the clinical outcome in patients receiving arthroplasty?

Are there factors that influence the formation of HO following cervical arthroplasty?

Methods

Retrospective cohort study.

Patients suffering from radiculopathy, myelopathy or myeloradiculopathy secondary to soft disc hernia and/or mild degree spondylosis operated between October 2004 and December 2006.

Patients with prior surgery at the instrumented level were excluded.

(Fig. 1)

Fig. 1

Patient population and intervention

Thirty-nine patients suffering from cervical radiculopathy and/or myelopathy treated with anterior disc replacement (ADR) were selected. Nine patients were excluded from this analysis because of insufficient outcome data at different time intervals (n = 7) or were not available for follow-up (n = 2).

Evidence of HO using the McAfee grading system6 (Table 1).

Table 1

McAfee grading of heterotopic ossification (HO)6

Grade 0	Absence of HO
Grade 1	Presence of HO in front of vertebral body but not in the anatomic disc space
Grade 2	Presence of HO in the disc space, possibly affecting the prosthesis's function
Grade 3	Bridging HO with prosthesis's motion still preserved
Grade 4	Complete fusion of the segment with absence of motion in flexion/extension

Prostheses range of motion (ROM). ROM was calculated by using two lines parallel to the Prodisc-C keels, or to the endplates of the Prestige LP, and measuring the intervening angle in flexion and extension on lateral radiographs. Lordosis was assigned a negative value and kyphosis a positive one; the difference between the two values gave the final ROM. ROM was measured three times per level and a mean value was obtained.

CT scan on patients with grade two or higher HO to assess localization of HO.

Short-form 36 (SF-36) and neck disability index (NDI). A comparison of scores was made between those who developed HO and those who did not using a two-sided Student t-test.

All images were independently reviewed by both a radiologist and a neurosurgeon not directly involved in the surgical procedures.

Forty-five disc prostheses (seven Prestige-LP and 38 Prodisc-C) were implanted in 30 patients (18 males) ranging in age from 28–63 years (mean age 40.9). The underlying disease included radiculopathy (n = 13), myelopathy (n = 4) or myeloradiculopathy (n = 13), secondary to soft-disc hernia (n = 17), spondylosis (n = 8) or a combination of disc hernia and spondylosis (n = 5). Nineteen patients received one-level disc replacement, seven patients received two-level and four patients received three level disc replacement. The following distribution of prostheses/level was recorded: C3–4:5; C4–5:4; C5–6:23; C6–7:13 (Table 2).

Table 2

Patient characteristics

Characteristic	N = 30
Age, years, mean (range)	40.9 (28–63)
Male, n (%)	18 (60)
Soft disc hernia, n (%)	17 (57)
Spondylosis, n (%)	8 (27)
Soft disc hernia & spondylosis, n (%)	5 (17)
Radiculopathy, n (%)	13 (43)
Myelopathy, n (%)	4 (13)
Myeloradiculopathy, n (%)	13 (43)
No. of levels treated1 level2 levels3 levels	4519 (patients) 7 (patients) 4 (patients)
Levels implantedC3–4C4–5C5–6C6–7	542313

The incidence rate of HO using the number of prostheses as the denominator was 42.2% (19 levels in 16 patients, Fig. 6). The risk of a patient developing HO after receiving one or more prostheses was 53.3% (16/30).

Fig. 6

Number of ossified levels and stratification for McAfee ossification grades (HO: heterotopic ossification)

Ten patients showed progressive ossification over time: in six patients de novo HO was demonstrated at 23–46 months (mean 33.3) postoperatively (Figs. 2, 3).

Fig. 2

Double-level arthroplasty with Prodisc-C at C5–6 and C6–7. Absence of HO at 1 month after surgery (left), presence of grade two HO 1 year postoperatively (middle) and of grade three HO (right) 3 years postoperatively, respectively.

Fig. 3

Double-level arthroplasty with Prodisc-C at C5–6 and C6–7. Note the progression of HO at the C4–5 disc space at 1 month (left), 12 (middle) and 48 (right) months, respectively, postoperatively (arrow).

Among the 16 patients with HO, 15 (93.8%) demonstrated a range of motion ≥3° (Fig. 4).

Fig. 4

Flexion and extension x-rays (same patient as in Fig 3) showing mobile disc prostheses at 48-months follow-up despite the presence of heterotopic ossification.

Comparison of pre- and postoperative SF-36 and NDI scores revealed functional improvement from baseline to follow-up at 3, 6, 12, 18, 24, and 36 months. There was no significant difference in functional scores between those who did and those who did not develop HO, P > .05 (Fig. 7).

Fig. 7

Differences of pre- and postoperative functional outcome in the two groups

Risks and relative risks by factor related to patient characteristics associated with HO formation were analyzed: males tended to develop HO more than females (twice as likely), though this did not reach statistical significance (Table 3).

Table 3

The risk (%) and unadjusted relative risk (RR) of HO by patient characteristics

	n/N (%)	RR	95% CI	P-value
Sex
Female	4/12 (33.3)	1.0
Male	12/18 (66.7)	2.0	0.8, 4.7	.13
Indication
Soft disc hernia	8/17 (47.1)	1.0
Spondylosis	4/8 (50.0)	1.1	0.4, 2.5	1.0
Soft disc hernia & spondylosis	3/5 (60.0)	1.3	0.5, 3.1	1.0
Neurological involvement
Myelopathy	2/4 (50.0)	1.0
Radiculopathy	7/13 (53.8)	1.1	0.4, 3.2	1.0
Myeloradiculopathy	7/13 (53.8)	1.1	0.4, 3.2	1.0

No patients required revision surgery for adjacent segment disease or persisting/recurrent symptoms.

Our rate of grade two and three HO (42.2%) is consistent with that previously reported.7 It is higher than in other series8,9 but lower than Sola's et al.10

HO is prevalent on the anterolateral surface of vertebral bodies (Fig. 5).

Fig. 5a–b

Same case as in Fig 4. Axial CT scan showing the presence of laterally sited HO (arrows) at C4–5 level (a) and at C6–7 level (b)

HO does not influence the clinical and functional outcome.

Strengths: This study reports 3 years mean follow-up data, to date the longest available. For all included patients complete follow-up data are available. Furthermore, we have shown that a delayed progression, or de novo onset, of HO is possible.

Limitations: This study was not conducted prospectively and is based on a relatively small albeit consecutive sample size. Lack of potential prognostic factors available for analysis and a 77% follow-up rate are further limitations.

A longer follow-up is necessary to assess whether arthroplasty can reach the goal of maintaining motion and if development of high-grade HO at the target level infers an increased risk of adjacent segment degeneration.

Clinical and functional improvement following cervical disc arthroplasty is maintained despite the presence of HO.

The indication for surgery (soft-disc herniation or “hard-disc” spondylosis) is not associated with the formation of HO.

Indications for arthroplasty are not affected by the risk of HO.

Preventative measures such as prophylactic nonsteroidal medications, local application of bone wax or intentional early range of motion measures were not routinely utilized in our study. While these are interesting factors, actual prospective studies will be needed to compare their effect. Our experiences may serve as a valuable baseline comparison.

Methods evaluation and class of evidence (CoE)
Methodological principle:
Study design:
Prospective cohort
Retrospective cohort	•
Case-control
Case series
Methods
Patients at similar point in course of treatment	•
Follow-up ≥85%
Similarity of treatment protocols for patient groups	•
Patients followed for long enough for outcomes to occur	•
Control for extraneous risk factors*
Evidence class:	III
*Authors must provide a description of robust baseline characteristics, and control for those that are potential prognostic factors.