This is a retrospective consecutive case series of 138 Gustillo-Anderson type IIIB and IIIC segmental tibial fractures treated at Walter Reed Army Medical Center in soldiers injured in Iraq between March 2003 and March 2005. Five patients with a head injury and four who were lost to follow-up were excluded. The patients were treated definitively with either a ringed external fixator or a reamed intramedullary nail, evaluated in terms of supplementary bone grafting with either autogenous bone (group 1, 67 patients) or recombinant human bone morphogenetic protein-2 at 1.50 mg/ml applied to an absorbable collagen sponge (group 2, 62 patients). The mechanism of injury, defect size and classification, associated injuries, presence of infection, preliminary treatment/fixation, number of procedures before definitive management, time to and details of definitive management, subsequent infection, re-operation, smoking history and other complications were noted. Radiographs were assessed for union, delayed union or nonunion by an independent investigator. All the patients were male. Their mean age was 26.6 years (20 to 42) and the mean follow-up was for 15.6 months (12 to 32). Group 2 had a slightly higher profile of concomitant injuries and a slightly worse fracture classification, but these were not significant. The rate of union was 76% (51 of 67) for group 1 and 92% for group 2 (57 of 62; p = 0.015). There was also a higher rate of subsequent infection in group 1 (14.9%) compared with group 2 (3.2%; p = 0.001) and a higher rate of re-operation (28%) in group 1 (p = 0.003). There were no observed hypersensitivity reactions to the recombinant human bone morphogenetic protein-2 implant.