The degeneration of uncovertebral joints on CT assessment and its association with postoperative heterotopic ossification in one- or two-level artificial cervical arthroplasty

Background Artificial cervical arthroplasty (ACR) preserves segmental mobility and reduces adjacent segment degenerative disease risk, yet postoperative heterotopic ossification (HO) impairs prosthetic function. Preoperative uncovertebral joint degeneration (UJD) is hypothesized to relate to HO. This study evaluated UJD on computed tomography (CT) and explored its association with HO after one- or two-level ACR, adjusting for confounders. Methods Forty patients undergoing one- or two-level ACR with a ≥1-year follow-up (2013–2019) were enrolled. UJD was graded using a modified Kellgren-Lawrence (K–L) CT system; final follow-up HO was assessed on radiographs using the McAfee classification. Generalized estimating equations (GEE) were used to address intrapatient clustering and confounders. Binary and ordinal logistic GEE, including proportional odds testing and sensitivity analyses, assessed independent UJD–HO associations. Stratified Spearman correlations were applied only for exploratory subgroup analyses. Results Overall HO incidence was 92.0%, with four segments at McAfee grade 0. After adjusting for age, sex, follow-up duration, operative level number and preoperative range of mobility, GEE showed non-significant trends linking higher-grade UJD to greater HO occurrence [adjusted odds ratio (aOR) = 12.561, 95% CI 0.885–178.247, P = 0.061] and severity (aOR = 1.493, 95% CI 0.863–2.585, P = 0.152). Partial proportional odds sensitivity analyses showed consistent results. Severe HO distribution imbalance and a scarcity of HO-negative segments led to unstable effect estimates and wide confidence intervals. Conclusions After adjusting for intrapatient clustering and confounders, preoperative UJD was not an independent predictor of HO onset or progression after ACR. Despite a positive trend, the high HO incidence caused a ceiling effect and limited statistical power, and thus, UJD's predictive value could not be confirmed. This CT-based modified K–L grading system reliably quantifies UJD and supports preoperative imaging evaluation for ACR candidates.