This work evaluated gelatin microparticles and biodegradable composite scaffolds for the controlled release of bone morphogenetic protein-2 (BMP-2) in vitro and in vivo. Gelatin crosslinking (10 and 40mM glutaraldehyde), BMP-2 dose (6 and 60ng BMP-2 per mg dry microparticles), buffer type (phosphate buffered saline (PBS) and collagenase-containing PBS), and gelatin type (acidic and basic) were investigated for their effects on BMP-2 release. Release profiles were also observed using poly(lactic-co-glycolic acid) (PLGA) microparticles with varying molecular weights (8300 and 57,500). In vitro and in vivo studies were conducted using radiolabeled BMP-2; the chloramine-T method was preferred over Bolton-Hunter reagent for radioiodination with this system. BMP-2 release from PLGA microparticles resulted in a moderate burst release followed by minimal cumulative release, while BMP-2 release from gelatin microparticles exhibited minimal burst release followed by linear release kinetics in vitro. Growth factor dose had a small effect on its normalized release kinetics probably because of an equilibrium between gelatin-bound and unbound BMP-2. Differences in release from acidic and basic gelatin microparticles may result from the different pretreatment conditions used for gelatin synthesis. The in vitro release kinetics for both gelatin microparticles alone and within composite scaffolds were dependent largely on the extent of gelatin crosslinking; varying buffer type served to confirm that controlled release relies on enzymatic degradation of the gelatin for controlled release. Finally, in vivo studies with composite scaffolds exhibited minimal burst and linear release up to 28 days. In summary, dose effects on BMP-2 release were found to be minimal while varying gelatin type and release medium can alter release kinetics. These results demonstrate that a systematic control of BMP-2 delivery from gelatin microparticles can be achieved by altering the extent of basic gelatin crosslinking.