External fixation is the current standard treatment for skeletal stabilization of open tibial fractures, but intramedullary fixation techniques have become increasingly popular. The aim of this study was to compare, in an animal model, the susceptibility to infection of contaminated fractures stabilized with external fixation with that of contaminated fractures fixed with intramedullary locking nails with or without reaming. A unilateral osteotomy of the tibia was performed in 15 goats under general anesthesia. Each osteotomy was stabilized with either (a) a unilateral biplanar external fixator, (b) an 8 mm diameter intramedullary rod inserted without reaming of the medullary cavity, or (c) a 10 mm diameter rod inserted after reaming. A standardized inoculum of Staphylococcus aureus, 10(3) colony forming units per milliliter, was placed at each osteotomy site on a piece of absorbable gelatin sponge, to simulate contamination of an open fracture. Antibiotics were not administered. The animals were allowed full activity after the procedure. Fourteen days postoperatively, the animals were killed, radiographs of the tibiae were taken, and the tibiae were harvested in a sterile manner. Multiple specimens for quantitative microbiological analysis were taken from the fracture site and from sites 3 cm distal and 6 cm proximal to the fracture. Additional specimens of bone were taken for histological study. Clinical, radiographic, and microbiological analysis demonstrated that, in this animal model, there were significantly fewer and less severe infections in fractures fixed with external fixation than in those fixed with an intramedullary nail with or without reaming. There was marked cortical necrosis in tibiae that had been fixed with nailing and reaming.