Allograft versus autograft for pediatric posterior cervical and occipito-cervical fusion: a systematic review of factors affecting fusion rates.

OBJECT Fusion rates are high for children undergoing posterior cervical fusion (PCF) and occipito-cervical fusion (OCF). Autologous bone has been widely used as the graft material of choice, despite the risk of donor-site morbidity associated with harvesting the bone, possibly because very low fusion rates were reported with posterior allograft cervical fusions in children several decades ago. Higher overall fusion rates using allograft in adults, associated with improvements in internal fixation techniques and the availability of osteoinductive substances such as bone morphogenetic protein (BMP), have led to heightened enthusiasm for the use of bank bone during pediatric PCF. A systematic review was performed to study factors associated with successful bone fusion, including the type of bone graft used. METHODS The authors performed a comprehensive PubMed search of English-language articles pertaining to PCF and OCF in patients less than 18 years old. Of the 561 abstracts selected, 148 articles were reviewed, resulting in 60 articles that had sufficient detail to be included in the analysis. A meta-regression analysis was performed to determine if and how age, fusion technique, levels fused, fusion substrate, BMP use, postoperative bracing, and radiographic fusion criteria were related to the pooled prevalence estimates. A systematic review of the literature was performed according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement. RESULTS A total of 604 patients met the specific inclusion and exclusion criteria. The overall fusion rate was 93%, with a mean age of 9.3 years and mean follow-up of 38.7 months. A total of 539 patients had fusion with autograft (94% fusion rate) and 65 patients with allograft (80% fusion rate). Multivariate meta-regression analysis showed that higher fusion rates were associated with OCF compared with fusions that excluded the occiput (p < 0.001), with the use of autograft instead of allograft (p < 0.001), and with the use of CT to define fusion instead of plain radiography alone. The type of internal fixation, the use of BMP, patient age, and the duration of follow-up were not found to be associated with fusion rates in the multivariate analysis. CONCLUSIONS Fusion rates for PCF are high, with higher rates of fusion seen when autograft is used as the bone substrate and when the occiput is included in the fusion construct. Further study of the use of allograft as a viable alternative to autograft bone fusion is warranted because limited data are available regarding the use of allograft in combination with more rigid internal fixation techniques and osteoinductive substances, both of which may enhance fusion rates with allograft.