The effect of rhBMP-2 in a novel, non-instrumented extremity nonunion model

Pre-clinical models of bony nonunion typically employ critical-length defects. However, these models may not accurately reflect clinical practice since many nonunions are diagnosed without bone loss. We developed a non-displaced rat ulna fracture model in order to examine the efficacy of recombinant human bone morphogenetic protein-2 (rhBMP-2) with an absorbable collagen sponge (ACS) for nonunion treatment. Transverse diaphyseal ulna fractures were created in 24 Sprague–Dawley rats. Eight animals (Group 1: Nonunion) received no further intervention. The remaining 16 animals were treated with 5 μg rhBMP-2/ACS at 8 weeks after the original intervention (Group 2: Nonunion + BMP) or at the time of initial injury (Group 3: Fresh fracture + BMP). In Group 1, 7 of 8 fractures demonstrated gross motion and a persistent radiographic gap (12.5% healing rate). In Groups 2 and 3, fractures healed at a rate of 75% (6 of 8 in each group) as determined by manual and radiographic evaluation. Biomechanical testing for torque load-to-failure and torsional stiffness demonstrated no significant difference between healed specimens treated with rhBMP-2. To our knowledge, this is the first description of a physiologic, non-stabilized, non-defect fracture nonunion model in a rodent. Furthermore, unlike previous nonunion models, the healing rates after treatment with rhBMP-2 are comparable to that of clinical data, suggesting that this model may provide an environment more representative of nonunions in humans.