Construct length rather than interbody spacer material is associated with reoperation after anterior cervical...

We compared reoperation after primary anterior cervical discectomy and fusion (ACDF) performed with structural allograft versus synthetic interbody cage, and we characterized the indications for revision in each group. Comparative data on interbody spacer selection derive largely from series with approximately one year of observation in which radiographic nonunion, rather than reoperation, is the endpoint. Whether spacer material alters the need for revision over a longer horizon remains unresolved.

We identified all consecutive patients undergoing primary ACDF for degenerative disease at a single academic center between January 2016 and July 2021, requiring a minimum of 24 months of follow-up. Pseudarthrosis and adjacent segment disease were defined as symptomatic failure requiring reoperation and were abstracted from operative reports by two independent reviewers. The primary analysis was a multivariable Cox proportional hazards model adjusting for age, Charlson Comorbidity Index (CCI), and number of fused levels. Propensity score matching was retained as a sensitivity analysis.

Of 588 procedures screened, 248 patients met criteria (211 allograft, 37 cage). Mean age was 55.1 ± 10.2 years and 39.9% were male, with a median follow-up of 44.5 months (IQR 33.3-55.9). Revision was performed in 32 of 211 allograft patients (15.2%) and 4 of 37 cage patients (10.8%, P = 0.618). Estimated freedom from revision at two years was 90.9% and 91.9% (log-rank P = 0.619). Cage was not associated with revision after adjustment (hazard ratio 0.90, 95% CI 0.30-2.65). Three-level fusion was associated with revision relative to single-level fusion (hazard ratio 2.96, 95% CI 1.13-7.76, P = 0.028). Adjacent segment disease was the leading indication in both groups. Anterior plate fixation was used in 99.5% of allograft constructs but in only 33.3% of cage constructs with documented plate status.

Interbody spacer material was not associated with reoperation in this cohort, whereas construct length was. With 36 events and a confidence interval extending to a 2.2-fold higher hazard, these data represent a null result under limited power rather than evidence of equivalence. Construct length warrants explicit discussion during preoperative counseling. Spacer material and construct configuration travelled together in this cohort, which has implications for how the existing comparative literature should be read.