OBJECTIVE The objective of this study was to evaluate bone grafts consisting of rhBMP-2 on an absorbable collagen sponge with a ceramic composite bulking agent, rhBMP-2, directly on a ceramic-collagen sponge carrier or iliac crest bone graft (ICBG) in combination with local bone graft to effect fusion in a multisegmental instrumented ovine lumbar intertransverse process fusion model. METHODS Thirty-six sheep had a single treatment at 3 spinal levels in both the right and left intertransverse process spaces. Group 1 sheep were treated with 7.5 cm3 of autograft consisting of ICBG plus local bone for each intertransverse process space. For Groups 2-4, 4 cm3 of local bone was placed within the intertransverse process space followed by 4.5-5 cm3 of the rhBMP-2 graft material. Group 2 animals received 1.5 mg/cm3 rhBMP-2 on an absorbable collagen sponge with a commercial bone void filler consisting of Type I lyophilized collagen with a biphasic hydroxyapatite/β-tricalcium phosphate ceramic with local bone. Group 3 animals received 0.75 mg/m cm3 of rhBMP-2 on a collagen ceramic sponge carrier with local bone. Group 4 animals received 1.35 mg/cm3 of rhBMP-2 on the same collagen ceramic sponge carrier with local bone. Sheep were euthanized 6 months postoperatively. Manual palpation, biomechanical testing, CT, radiography, and undecalcified histology were performed to assess the presence of fusion associated with the treatments. RESULTS All animals in Groups 2-4 that received grafts containing rhBMP-2 achieved radiographic and CT fusion at all 3 levels. In Group 1 (bone autograft alone), only 19% of the levels demonstrated radiographic fusion, 14% resulted in possible radiographic fusion, and 67% of the levels demonstrated radiographic nonfusion. Biomechanical testing showed that Groups 2-4 demonstrated similar stiffness of the L2-5 segment in all 6 loading directions, with each of the 3 groups having significantly greater stiffness than the autograft-only group. In Group 1, only 2 of 18 levels were rated as achieving bilateral histological fusion, with an additional 3 levels showing a unilateral fusion. The majority of the treated levels (13/18) in Group 1 were scored as histological nonfusions. There were no histological nonfusions in Groups 2 through 4. All 18 levels in Group 2 were rated as bilateral histological fusions. A majority (34/36) of the levels in Group 3 were rated as bilateral histological fusions, with 2 levels showing a unilateral fusion. A majority (35/36) of the levels in Group 4 were rated as bilateral histological fusions, with 1 level showing a unilateral fusion. CONCLUSIONS In the ovine multilevel instrumented intertransverse process fusion model, rhBMP-2 was able to consistently achieve CT, radiographic, biomechanical, and histological fusion. Compared with ICBG, the gold standard for bone grafting, rhBMP-2 was statistically superior at achieving radiographic and histological fusion.
Keywords: ACS = absorbable collagen sponge; CCS = collagen ceramic sponge; HA = hydroxyapatite; ICBG = iliac crest bone graft; OLR = ordinal logistic regression; PA = posteroanterior; TCP-HA = tricalcium phosphate; autograft; lumbar fusion; multilevel spine fusion; rhBMP-2; rhBMP-2 = recombinant human bone morphogenetic protein–2.