Heterotopic ossification and radiographic adjacent-segment disease after cervical disc arthroplasty

J Neurosurg Spine 31:660–669, 2019

Cervical disc arthroplasty (CDA) is an accepted motion-sparing technique associated with favorable patient outcomes. However, heterotopic ossification (HO) and adjacent-segment degeneration are poorly understood adverse events that can be observed after CDA. The purpose of this study was to retrospectively examine 1) the effect of the residual exposed endplate (REE) on HO, and 2) identify risk factors predicting radiographic adjacent-segment disease (rASD) in a consecutive cohort of CDA patients.

METHODS A retrospective cohort study was performed on consecutive adult patients (≥ 18 years) who underwent 1- or 2-level CDA at the University of Calgary between 2002 and 2015 with > 1-year follow-up. REE was calculated by subtracting the anteroposterior (AP) diameter of the arthroplasty device from the native AP endplate diameter measured on lateral radiographs. HO was graded using the McAfee classification (low grade, 0–2; high grade, 3 and 4). Change in AP endplate diameter over time was measured at the index and adjacent levels to indicate progressive rASD.

RESULTS Forty-five patients (58 levels) underwent CDA during the study period. The mean age was 46 years (SD 10 years). Twenty-six patients (58%) were male. The median follow-up was 29 months (IQR 42 months). Thirty-three patients (73%) underwent 1-level CDA. High-grade HO developed at 19 levels (33%). The mean REE was 2.4 mm in the high-grade HO group and 1.6 mm in the low-grade HO group (p = 0.02). On multivariable analysis, patients with REE > 2 mm had a 4.5-times-higher odds of developing high-grade HO (p = 0.02) than patients with REE ≤ 2 mm. No significant relationship was observed between the type of artificial disc and the development of high-grade HO (p = 0.1). RASD was more likely to develop in the lower cervical spine (p = 0.001) and increased with time (p < 0.001). The presence of an artificial disc was highly protective against degenerative changes at the index level of operation (p < 0.001) but did not influence degeneration in the adjacent segments.

CONCLUSIONS In patients undergoing CDA, high-grade HO was predicted by REE. Therefore, maximizing the implant-endplate interface may help to reduce high-grade HO and preserve motion. RASD increases in an obligatory manner following CDA and is highly linked to specific levels (e.g., C6–7) rather than the presence or absence of an adjacent arthroplasty device. The presence of an artificial disc is, however, protective against further degenerative change at the index level of operation.