The effects of a large collagen-hydroxyapatite (COL-HA) substitute with or without isolated bone marrow nuclear cells (BMNC) on spine fusion were evaluated using a two-level posterolateral spinal fusion model in sheep. Six sheep each were included in both the COL-HA and autograft groups. In each animal, spine fusions were performed at two levels by randomization with or without BMNC. After an observation period of 4.5 months, the fusion blocks (length = 50 mm) were harvested and evaluated by micro-CT with respect to fusion rate and microarchitecture, and histology was performed to qualitatively assess new bone morphology. BMNC supplementation did not lead to significant differences in fusion rates or microarchitectural properties in the both groups (p = 0.56). Porosity decreased significantly in the COL-HA group (p = 0.002), suggesting a relatively fast degradation of COL-HA. Qualitative histology assessment revealed quick reabsorption of COL-HA replaced by newly formed bone, and this new bone morphology was similar to autograft. Interestingly, COL-HA generated significant amounts of new bone in vivo which were similar to autograft, and the presence of BMNC seemed to enhance this process. Apart from porosity, the comparable fusion rate and similar microarchitectural parameters suggested equal performance of COL-HA and autograft with respect to spine fusion and bone quality. In conclusion, the COL-HA with BMNC demonstrated comparable fusion rates as the autograft. COL-HA was capable of forming new bone and was revealed to be similar to the autograft with respect to microarchitectural properties apart from porosity of the fusion blocks.