Bone graft options for spinal fusion following resection of spinal column tumors: systematic review and meta-analysis

Objective With the advent of new adjunctive therapy, the overall survival of patients harboring spinal column tumors has improved. However, there is limited knowledge regarding the optimal bone graft options following resection of spinal column tumors, due to their relative rarity and because fusion outcomes in this cohort are affected by various factors, such as radiation therapy (RT) and chemotherapy. Furthermore, bone graft options are often limited following tumor resection because the use of local bone grafts and bone morphogenetic proteins (BMPs) are usually avoided in light of microscopic infiltration of tumors into local bone and potential carcinogenicity of BMP. The objective of this study was to review and meta-analyze the relevant clinical literature to provide further clinical insight regarding bone graft options. Methods A web-based MEDLINE search was conducted in accordance with preferred reporting items for systematic review and meta-analysis (PRISMA) guidelines, which yielded 27 articles with 383 patients. Information on baseline characteristics, tumor histology, adjunctive treatments, reconstruction methods, bone graft options, fusion rates, and time to fusion were collected. Pooled fusion rates (PFRs) and I-2 values were calculated in meta-analysis. Meta-regression analyses were also performed if each variable appeared to affect fusion outcomes. Furthermore, data on 272 individual patients were available, which were additionally reviewed and statistically analyzed. Results Overall, fusion rates varied widely from 36.0% to 100.0% due to both inter-and intrastudy heterogeneity, with a PFR of 85.7% (I-2 = 36.4). The studies in which cages were filled with morselized iliac crest autogenic bone graft (ICABG) and/or other bone graft options were used for anterior fusion showed a significantly higher PFR of 92.8, compared with the other studies (83.3%, p = 0.04). In per-patient analysis, anterior plus posterior fusion resulted in a higher fusion rate than anterior fusion only (98.8% vs 86.4%, p < 0.001). Although unmodifiable, RT (90.3% vs 98.6%, p = 0.03) and lumbosacral tumors (74.6% vs 97.9%, p < 0.001) were associated with lower fusion rates in univariate analysis. The mean time to fusion was 5.4 +/- 1.4 months (range 3-9 months), whereas 16 of 272 patients died before the confirmation of solid fusion with a mean survival of 3.1 +/- 2.1 months (range 0.5-6 months). The average time to fusion of patients who received RT and chemotherapy were significantly longer than those who did not receive these adjunctive treatments (RT: 6.1 months vs 4.3 months, p < 0.001; chemotherapy: 6.0 months vs 4.3 months, p = 0.02). Conclusions Due to inter-and intrastudy heterogeneity in patient, disease, fusion criteria, and treatment characteristics, the optimal surgical techniques and factors predictive of fusion remain unclear. Clearly, future prospective, randomized studies will be necessary to better understand the issues surrounding bone graft selection following resection of spinal column tumors.