Posterior bone graft nonunion and long-term outcomes after posterior C1–2 fusion: a propensity score-matched retrospective cohort study

Posterior C1–2 fusion is a standard treatment for atlantoaxial instability. In daily practice, however, follow-up computed tomography (CT) may show no clear posterior graft union even when reduction is maintained and the implants appear stable. Whether this radiographic finding represents true surgical failure remains uncertain. We therefore examined the long-term radiographic and clinical significance of absent posterior graft union after posterior C1–2 fusion. We retrospectively reviewed adult patients who underwent posterior C1–2 fixation and fusion with the Goel–Harms technique and autologous iliac crest bone grafting between May 2012 and May 2021. Posterior graft union was assessed on 12-month postoperative CT. Patients without definite posterior graft union were all included, and comparable patients with posterior graft union were selected using 1:1 propensity score matching. Radiographic stability, implant-related complications, and clinical outcomes were then compared during long-term follow-up. Among 153 screened patients, 112 met the eligibility criteria: 92 had posterior graft union and 20 had no definite posterior graft union on 12-month CT. After matching, the final cohort included 40 patients, with 20 patients in each group. The mean follow-up duration was 67.6 ± 9.7 months. AADI and PADI improved after surgery and remained stable until the final follow-up, without significant between-group differences at any assessed time point. Mean C1–2 ROM decreased from 10.36 ± 3.66° before surgery to 1.16 ± 1.03° at 3 months and remained essentially unchanged at 1.08 ± 1.03° at the final follow-up. No patient developed segmental instability, screw loosening, or implant failure. VAS, JOA, and NDI scores improved in both groups, and the improvements were not significantly different between groups. In this propensity score-matched cohort, absence of definite posterior graft union on follow-up CT was not associated with long-term atlantoaxial instability, implant failure, or inferior clinical outcomes after posterior C1–2 fusion. When reduction is maintained and instrumentation remains stable, posterior graft nonunion on CT should be interpreted cautiously rather than regarded automatically as surgical failure.