Treatment of segmental bone loss remains a major challenge in orthopaedic surgery. Traditional techniques (eg, autograft) and newer techniques (eg, recombinant human bone morphogenetic protein-2 [rhBMP-2]) have well-established performance limitations and safety concerns respectively. Consequently there is an unmet need for osteoinductive bone graft substitutes that may eliminate or reduce the use of rhBMP-2. Using an established rabbit radius osteotomy defect model with positive (autogenous bone graft) and negative (empty sham) control groups, we asked: (1) whether a collagen-glycosaminoglycan scaffold alone can heal the defect, (2) whether the addition of hydroxyapatite particles to the collagen scaffold promote faster healing, and (3) whether the collagen-glycosaminoglycan and collagen-hydroxyapatite scaffolds are able to promote faster healing (by carrying a low dose rhBMP-2). A 15-mm transosseous radius defect in 4-month-old skeletally mature New Zealand White rabbits were treated with either collagen-hydroxyapatite or collagen-glycosaminoglycan scaffolds with and without rhBMP-2. Autogenous bone graft served as a positive control. Time-series radiographs at four intervals and postmortem micro-CT and histological analysis at 16 weeks were performed. Qualitative histological analysis of postmortem explants, and qualitative and volumetric 3-D analysis of standard radiographs and micro-CT scans enabled direct comparison of healing between test groups. Six weeks after implantation the collagen-glycosaminoglycan group had callus occupying greater than ½ the defect, whereas the sham (empty) control defect was still empty and the autogenous bone graft defect was completely filled with unremodeled bone. At 6 weeks, the collagen-hydroxyapatite scaffold groups showed greater defect filling with dense callus compared with the collagen-glycosaminoglycan controls. At 16 weeks, the autogenous bone graft groups showed evidence of early-stage medullary canal formation beginning at the proximal and distal defect borders. The collagen-glycosaminoglycan and collagen-glycosaminoglycan-rhBMP-2 groups had nearly complete medullary canal formation and anatomic healing at 16 weeks. However, collagen-hydroxyapatite-rhBMP-2 scaffolds showed the best levels of healing, exhibiting a dense callus which completely filled the defect. The collagen-hydroxyapatite scaffold showed comparable healing to the current gold standard of autogenous bone graft. It also performed comparably to collagen-glycosaminoglycan-rhBMP-2, a representative commercial device in current clinical use, but without the cost and safety concerns. The collagen-glycosaminoglycan scaffold may be suitable for a low load-bearing defect. The collagen-hydroxyapatite scaffold may be suitable for a load-bearing defect. The rhBMP-2 containing collagen-glycosaminoglycan and collagen-hydroxyapatite scaffolds may be suitable for established nonunion defects.