The osteoconductive capacity of fibrillar collagen-biphasic calcium phosphate composition was compared to autogenous bone in a canine spinal fusion model. All animals underwent a standard intervertebral body fusion (L2-L4) with rigid internal fixation and received either autogenous bone alone or a mixture of the ceramic and autogenous bone (3:1) as the graft material. Animals were followed for 12 months and the quality of fusion in each animal assessed by biomechanical testing and histological analysis. The fused L2-L4 segment of each dog was embedded in bone cement and mounted in a specially designed mechanical tester for testing in flexion, extension, and side bending. Overall, the mean rigidity of the fusion mass was not significantly different between the two groups [10.5 +/- 4.1 (SD) for autogenous bone vs. 11.3 +/- 1.7 for the ceramic plus autogenous bone, p greater than 0.05]. Similar findings were obtained for mean bending moment, compressive load, angular deformation, and energy absorbed for the two groups. Histological analysis was performed on transverse nondecalcified specimens. Quantitation of bone ingrowth using back-scattered electron imaging disclosed no significant differences in the amount of new bone formed at the graft site between autogenous bone and the ceramic plus autogenous bone recipients (23.4 +/- 10% vs. 25.8 +/- 8.8%) when correction for the autogenous bone volumes was performed. Light microscopic analysis of toluidine blue-stained transverse sections demonstrated new bone growth around and through the ceramic bone graft material. These results suggest that use of a collagen-biphasic calcium phosphate ceramic and autogenous bone mixture (3:1) provides a suitable osteoconductive alternative to the use of autogenous bone and results in the formation of a mechanically competent fusion mass not significantly different from that obtained with autogenous bone alone.