Bone morphogenetic proteins (BMPs) are important mediators of fracture repair. They include a group of bioactive proteins that are synthesized locally at a fracture site and work mostly by upregulating stem-cell differentiation, thereby stimulating bone formation. As such, BMPs are of major interest to orthopaedic surgeons as osteoinductive agents in patients with a difficult fracture in whom bone-healing is potentially compromised. Several BMPs, including recombinant human BMP-2 (rhBMP-2), have been characterized and synthesized via recombinant technology: bone formation is induced in a more or less dose-dependent fashion and may be given in quantities that vastly exceed physiologic amounts. rhBMP-2 has U.S. Food and Drug Administration approval for use in anterior lumbar spine fusions (2002) and for open tibial fractures (2004). A small number of studies have shown efficacy for achieving bone-healing in these areas, although recently the methodology of some of these studies has been called into question.

Lyon et al. put together a fascinating study. Functionally, it is a well-designed randomized controlled trial of “injectable” BMP in closed tibial shaft fractures in patients treated with intramedullary nailing: the authors compared the radiographic and clinical healing in patients treated with injection at the fracture site, using two quantities of BMP-2 in calcium phosphate matrix, with that in controls (nailing only). The importance of this study, however, is that the ultimate goal of the study seems to be to design an injectable “cocktail” that will improve fracture-healing, by speeding up the healing process and minimizing the rate of nonunion. This goal has been the vision of surgeons and scientists much longer than the half century since Urist’s discovery of BMP.

Unfortunately for all of us, the findings of the study indicate that the “cocktail” (as prepared) was unlikely to make a difference: the trial was shut down prematurely when a planned preliminary review demonstrated no treatment effect. There are a number of reasons these results may have been found. rhBMP-2 might not significantly affect healing in the optimal circumstances of closed fractures treated with the standard of care. Perhaps more likely though, the recipe needs to be refined to improve the cocktail’s effect. There are many variables in play here. For example, it is known that the efficacy of rhBMP-2 is carrier-dependent, and the authors suggest that the carrier may have allowed dispersal of the BMP away from the fracture site (seen radiographically). Additionally, changes in the preparation of the BMP product were made between this and prior clinical trials that may have altered clinical effects, although details of those changes are proprietary and thus unavailable.