The availability of recombinant osteoinductive growth factors and new osteoconductive matrices offers an alternative to the use of autogenous bone (autograft) for grafting indications. This study evaluates the bone-forming activity of a mineralized collagen matrix combined with recombinant human growth and differentiation factor-5 in a rabbit posterolateral spinal fusion model. The activity of three distinct matrix-growth factor formulations is assessed by radiographic, histologic, and mechanical strength methods. Results show that the radiographic density, histologic quality, and mechanical strength of fusion at 12 weeks post-treatment rank consistently within the treatment groups. Optimal formulations are shown to perform similar to autograft in both the rate and strength of fusion. Fusion rates as high as 80% are observed within specific matrix/growth factor formulations. The average biomechanical strength of treated motion segments in the most efficacious formulation is 82% higher than that obtained with autograft, although this difference is not statistically significant. The fusion mass formed in response to matrix/ growth factor formulations is composed of normal trabecular bone with a thin outer cortical plate and modest hematopoietic bone marrow. These results demonstrate that the combination of a mineralized collagen matrix with recombinant human growth and differentiation factor-5 maximizes the inherent conductive and inductive properties of each component, respectively, to provide an effective alternative to autograft for bone grafting procedures. Anat The formation of a solid arthrodesis in spine surgery requires the orthotopic formation of bone between the transverse process, facet joint and vertebral body of adjacent motion segments. Traditional spinal fusion procedures utilize autogenous bone (autograft) as a graft to provide the osteogenic, osteoinductive, and osteoconductive components necessary for the formation of new bone at the operative site. Complications associated with the harvest of autograft, such as donor site morbidity, blood loss, and increased operative time, have prompted the search for suitable alternatives (Younger and Chapman, 1989). New bone graft materials are being designed to incorporate or mimic one or more of the bone-forming components of autograft. A common strategy is to combine a purified, recombinant form of an osteoinductive protein with an osteoconductive matrix to attract, stimulate, and support the formation of new bone by host osteogenic cells (Cook et al