The discovery of bone morphogenetic proteins (BMPs) by Marshall Urist (7) in 1965 has led to a diverse area of research dedicated to the identification and characterization of osteoinductive growth factors. Members of the TGF-β superfamily, BMPs have been proposed for a number of applications in orthopaedic surgery (8). Although at least 20 different BMPs have been reported (9), much of the clinical study has focused on recombinant human BMP-2 (rhBMP-2) (INFUSE; Medtronic, Sofamor-Danek, Memphis, TN) and rhBMP-7 (osteogenic protein-1 or OP-1).

rhBMP-2 and BMP-7 have been evaluated in numerous preclinical models, and successful healing in long bone defects has been reported (8,10, 11 and 12). Similar findings have been demonstrated in spinal arthrodesis models in animals (13, 14, 15 and 16). Food and Drug Administration (FDA) approval has been granted for the use of rhBMP-2 to enhance anterior lumbar spinal fusion (17) and rhOP-1 to supplement posterior spine fusions. In a prospective, randomized controlled pilot study, Baskin et al. (18) compared cancellous iliac crest autograft to rhBMP-2 (0.6 mg per level) reconstituted within a type I collagen sponge in a structural fibular allograft (CORNERSTONE-SR Allograft Ring; Regeneration Technologies, Alachua, FL) in 33 patients after a one- or two-level anterior cervical discectomy and fusion (ACDF) with 2-year follow-up. Statistically significant improvement for both groups at all postoperative intervals were noted in Neck Disability Index (NDI), composite neck pain, and composite arm pain scores. Although both groups significantly improved overall, NDI and arm pain scores demonstrated statistically greater improvement in the rhBMP-2 group at the 24-month time point (p< 0.055). Donor-site pain in the control group was statistically significantly greater up until 6 weeks postoperatively (p < 0.007), but not at the 24-month follow-up. All patients in both groups met radiographic criteria for fusion at 6, 12, and 24 months.

Similarly, Buttermann (19) reported results from a nonrandomized prospective study of rhBMP-2 and allograft (0.9 mg per level) compared with tricortical iliac crest autograft in one-, two-, and three-level ACDF. Notably, in the investigational group, a portion of the BMP-soaked type I collagen sponge was placed outside the allograft ring in the disk space along with autograft obtained from anterior osteophytes. Furthermore, the iliac crest donor site in control patients was reconstructed with fibular allograft. Over the 2- to 3-year period of follow-up, both groups showed statistically significant improvement in all primary outcome measures (p < 0.01), and at no point was there a significant difference between groups.

Although no postoperative complications were reported by Baskin et al. (18), the rhBMP-2 group in Buttermann’s study experienced a significantly higher rate of neck swelling, dysphagia, and ICU readmissions for postoperative monitoring. Based on this experience, the author concluded that postoperative complications related to anterior cervical edema were concerning enough to warrant further investigation (19).

Although the quantity of published data is limited, preliminary evidence suggests that fusion rates with the
use of rhBMP-2 in the ventral cervical spine are at least equivalent, if not superior, to iliac crest autograft. Ongoing randomized controlled clinical trials studying the use of rhBMP-2 in multilevel ventral cervical fusions, where a greater difference in pseudarthrosis rate may be seen, will further develop guidelines for the potential role for growth factor therapy in anterior cervical spine procedures.