Osseointegration [direct bone-implant contact (BIC)] is a primary goal following installation of endosseous dental implants. Such bone contact provides stability for the dental implant over time. The objective of this study was to evaluate bone formation and BIC at long-term, functionally loaded, endosseous dental implants placed into bone induced by recombinant human bone morphogenetic protein-2 (rhBMP-2) in an absorbable collagen sponge (ACS) carrier. Mandibular, saddle-type, alveolar ridge defects (approximately 15 x 10 x 10 mm), two per jaw quadrant, were surgically induced in each of six young adult American fox hounds. The defects were immediately implanted with rhBMP-2/ACS. Two defects per animal additionally received a nonresorbable expanded polytetrafluoroethylene (ePTFE) membrane or a bioresorbable polyglycolide fiber membrane. Healing was allowed to progress for 3 months, when the ePTFE membrane was removed, and machined, threaded, titanium dental implants were installed into the rhBMP-2/ACS induced bone and into the adjacent resident bone. At 4 months of osseointegration, the implants were exposed to receive abutments and prosthetic treatment (two- or three-unit bridges). Some implants were removed for histologic analysis. The remainder of implants were exposed to functional loading for 12 months at which time the animals were killed for histometric analysis. One animal died prematurely due to kidney failure unrelated to the experimental protocol and was not included in the analysis. The 12-month block sections from a second animal were lost in the histological processing. Four sites receiving rhBMP-2/ACS and ePTFE or resorbable membranes experienced wound failure and membrane exposure, and subsequently exhibited limited bone formation. Defects without wound failure filled to contour with the adjacent alveolar bone. The newly formed bone exhibited features of the resident bone with a re-established cortex; however, it commonly included radiolucent areas that resolved over time. Dental implants block biopsied at 4 months exhibited limited, if any, crestal resorption, whereas those exposed to functional loading for 12 months exhibited some crestal resorption. Implants biopsied at 4 months exhibited a mean (+/- SD) BIC of 40.6 +/- 8.2% in rhBMP-2/ACS induced bone vs. 52.7 +/- 11.4% in resident bone. Dental implants exposed to 12 months of functional loading exhibited a mean BIC of 51.7 +/- 7.1% in rhBMP-2/ACS induced bone vs. 74.7 +/- 7.0% in resident bone. There were no significant differences between dental implants placed into rhBMP-2/ACS induced bone and resident bone for any parameter at any observation interval. In conclusion, rhBMP-2/ACS-induced bone allows installation, osseointegration, and long-term functional loading of machined, threaded, titanium dental implants in dogs.