Commercially available recombinant human bone morphogenetic protein 2 (rhBMP2) has demonstrated efficacy in bone regeneration, but not without significant side effects. The authors used rhBMP2 encapsulated in poly(lactic-co-glycolic acid) (PLGA) microspheres placed in a rabbit cranial defect model to test whether low-dose, sustained delivery can effectively induce bone regeneration. The rhBMP2 was encapsulated in 15% PLGA using a double-emulsion, solvent extraction/evaporation technique, and its release kinetics and bioactivity were tested. Two critical-size defects (10 mm) were created in the calvaria of New Zealand white rabbits (5 to 7 months of age, male and female) and filled with a collagen scaffold containing either (1) no implant, (2) collagen scaffold only, (3) PLGA-rhBMP2 (0.1 μg per implant), or (4) free rhBMP2 (0.1 μg per implant). After 6 weeks, the rabbits were killed and defects were analyzed by micro-computed tomography, histology, and finite element analysis. The rhBMP2 delivered by means of bioactive PLGA microspheres resulted in higher volumes and surface area coverage of new bone than an equal dose of free rhBMP2 by micro-computed tomography (p=0.025 and p=0.025). Finite element analysis indicated that the mechanical competence using the regional elastic modulus did not differ with rhBMP2 exposure (p=0.70). PLGA-rhBMP2 did not demonstrate heterotopic ossification, craniosynostosis, or seroma formation. Sustained delivery by means of PLGA microspheres can significantly reduce the rhBMP2 dose required for de novo bone formation. Optimization of the delivery system may be a key to reducing the risk for recently reported rhBMP2-related adverse effects.