BACKGROUND: Vertical bone augmentation (VBA) remains unpredictable and challenging for most clinicians. This study aims to compare hard tissue outcomes of VBA, with and without recombinant human bone morphogenetic protein (rhBMP)-2, under space-making titanium mesh in a canine model. METHODS: Eleven male beagle dogs were used in the study. Experimental ridge defects were created to form atrophic ridges. VBA was performed via guided bone regeneration using titanium mesh and allografts. In experimental hemimandibles, rhBMP-2/absorbable collagen sponge was well mixed with allografts prior to procedures, whereas a control buffer was applied within controls. Dogs were euthanized after a 4-month healing period. Clinical and radiographic examinations were performed to assess ridge dimensional changes. In addition, specimens were used for microcomputed tomography (micro-CT) assessment and histologic analysis. RESULTS: Membrane exposure was found on five of 11 (45.5%) rhBMP-2-treated sites, whereas it was found on nine of 11 (81.8%) non-rhBMP-2-treated sites. Within 4 months of healing, rhBMP-2-treated sites showed better radiographic bone density, greater defect fill, and significantly more bone gain in ridge height (P <0.05) than controls. Experimental hemimandibles exhibited lower rates of membrane exposure and a noteworthy, ectopic bone formation above the mesh in 72% of sites. Results from micro-CT also suggested a trend of less vertical bone gain and bone mineral density in controls (P >0.05). Under light microscope, predominant lamellar patterns were found in the specimen obtained from rhBMP-2 sites. CONCLUSION: With inherent limitations of the canine model and the concern of such a demanding surgical technique, current findings suggest that the presence of rhBMP-2 in a composite graft allows an increase of vertical gain, with formation of ectopic bone over the titanium mesh in comparison with non-rhBMP-2 sites.