Comparison of 1-Level to 2-Level Cervical Disc Arthroplasty

Study Design. Systematic review with network meta-analysis (NMA). Objective. To determine if there is a difference in adjacent segment disease (ASD) and all-cause reoperations between 1-level and contiguous 2-level cervical disc arthroplasty (CDA1, CDA2). Summary of Background Data. There are conflicting results from previous studies that there is a difference in reoperations between CDA1 and CDA2 possibly reflecting undertreatment of adjacent cervical degenerative disc disease with its natural history in CDA1. Methods. We conducted a random-effects NMA with only FDA Investigational Device Exemption (IDE) studies and international randomized controlled trials (RCTs) with more than 5 years of follow-up. Direct comparisons included CDA1 versus 1-level anterior cervical discectomy with fusion (ACDF1), CDA2 versus 2-level ACDF (ACDF2), and ACDF2 versus ACDF1; together, these formed a connected network in which ACDF served as the anchoring comparator, enabling anchored indirect comparisons between CDA1 and CDA2. Binary outcomes (ASD and all-cause reoperations) were analyzed using odds ratios within a frequentist framework in Stata. Results. NMA for ASD reoperations had 13 studies directly comparing CDA1 vs ACDF1 (n=4,135), 2 studies comparing 2-level CDA2 vs ACDF2 (n=789), and 1 study comparing ACDF2 vs ACDF1 (n=207). Results demonstrated no significant difference in ASD reoperation (OR 0.98, 95% CI 0.38–2.54) for CDA1 versus CDA2. For all-cause reoperations there were 14 studies directly comparing CDA1 vs ACDF1 (n=4,246), 2 studies comparing CDA2 vs ACDF2 (n=837), and 1 study comparing ACDF2 vs ACDF1 (n=223). CDA1 versus CDA2, demonstrated no significant difference in all-cause reoperation (OR 0.98, 95% CI 0.40–2.39). Conclusions. In our comprehensive NMA of selected FDA IDE studies and international RCTs with long-term follow-up, we found no significant difference between 1-level CDA and 2-level CDA with respect to ASD and all-cause reoperations. Future research is needed to see if our results are replicated in real world data.