Abstract Recombinant human BMP-2 (rhBMP-2) was immobilized non-covalently and covalently as a monolayer on plasma vapour deposited (PVD) porous commercially pure titanium surfaces in amounts of 5–8 μg cm −2, providing a ca. 10-fold increase vs. previously reported values . Dissociation of the immobilized [ 125I]rhBMP-2 from the surface occurred in a two-phase exponential decay: a first rapid phase (ca. 15% of immobilized BMP-2) with a half-life of 1–2 days and a second slow sustained release phase (ca. 85% of immobilized BMP-2) with a half-life of 40–60 days. Dissociation rate constants of sustained release of k −1 = 1.3–1.9 × 10 −7 s −1 were determined, allowing an estimation of the binding constants ( K A) for the adsorbed rhBMP-2 monolayer, to be around 10 12 M −1. The rhBMP-2-coated surfaces showed a high level of biological activity, as demonstrated by in vitro epifluorescence tests for alkaline phosphatase with MC3T3-E1 cells and in vivo experiments. In vivo osteoinductivity of rhBMP-2-coated implants was investigated in a gap-healing model in the trabecular bone of the distal femur condylus of sheep. Healing occurred without inflammation or capsule formation. The calculated concentration of released rhBMP-2 in the 1 mm gap ranged from 20 to 98 nM – well above the half-maximal response concentration ( K 0.5) for inducing alkaline phosphatase in MC3T3-E1 cells. After 4, 9 and 12 weeks the bone density (BD) and bone-to-implant contact (BIC) of the explanted implants were assessed histomorphometrically. Implants with immobilized rhBMP-2 displayed a significant (2- to 4-fold) increase in BD and BIC values vs. negative controls after 4–9 weeks. Integration of implants by trabecular bone was achieved after 4 weeks, indicating a mean “gap-filling rate” of ∼250 μm week −1. Integration of implants by cortical bone was observed after 9 weeks. Control implants without rhBMP-2 were not osseointegrated. This study demonstrates the feasibility of enhancing peri-implant osseointegration and gap bridging by immobilized rhBMP-2 on implant surfaces which may serve as a model for future clinical applications.