PURPOSE: The aim of the present study was to test whether or not a synthetic matrix consisting of a polyethylene glycol (PEG) hydrogel containing recombinant human bone morphogenetic protein-2 (rhBMP-2) combined with grafting materials enhances bone regeneration compared with grafting alone or empty control sites. MATERIAL AND METHODS: In each of 10 rabbits, four titanium cylinders were screwed in perforated slits made in the external cortical bones of the calvaria. The following four treatment modalities were randomly allocated: (1) empty control, (2) a combination of a PEG matrix and hydroxyapatite/tricalciumphosphate (HA/TCP) granules and a combination of a PEG matrix containing either 10 microg/ml (3) or 30 microg/ml (4) of BMP-2 and HA/TCP granules. After 8 weeks, the animals were sacrificed and ground sections were obtained for histological analysis. For statistical analysis repeated measures ANOVA and subsequent pairwise Student's t-test were applied (P<0.01). RESULTS: Histomorphometric analysis showed an average area fraction of newly formed bone of 13.96+/-5.98% for the empty control, 15.16+/-7.95% for the PEG and HA/TCP group, 26.32+/-8.56% for the group containing 10 mug rhBMP-2/ml, and 30.15+/-7.63% for the group containing 30 microg rhBMP-2/ml. Statistical analysis revealed significantly more newly formed bone in the two rhBMP-2 groups compared with the PEG and HA/TCP group and with the empty control. Regarding the surface fraction of the HA/TCP graft particles covered with newly formed bone the addition of rhBMP-2 revealed a more than two-fold increase compared with cylinders containing HA/TCP granules without rhBMP-2. This difference reached statistical significance. CONCLUSIONS: It is concluded that rhBMP-2 significantly enhances bone regeneration in rabbits when delivered by a synthetic matrix containing HA/TCP. This synthetic PEG matrix containing HA/TCP granules apparently fulfills a number of criteria required for an ideal carrier system for rhBMP-2.