BACKGROUND: The optimal management of femoral fractures in adolescents is controversial. This study was performed to compare the results and complications of four methods of fixation and to determine the factors related to those complications. METHODS: We conducted a retrospective cohort study of 194 diaphyseal femoral fractures in 189 children and adolescents treated with elastic stable intramedullary nail fixation, external fixation, rigid intramedullary nail fixation, or plate fixation. After adjustment for age, weight, energy of the injury, polytrauma, fracture level and pattern, and extent of comminution, treatment outcomes were compared in terms of the length of the hospital stay, time to union, and complication rates, including loss of reduction requiring a reoperation, malunion, nonunion, refracture, infection, and the need for a reoperation other than routine hardware removal. RESULTS: The mean age of the patients was 13.2 years, and their mean weight was 49.5 kg. There was a loss of reduction of two of 105 fractures treated with elastic nail fixation and ten of thirty-three treated with external fixation (p < 0.001). At the time of final follow-up, five patients (two treated with external fixation and one in each of the other groups) had >or=2.0 cm of shortening. Eight of the 104 patients (105 fractures) treated with elastic nail fixation underwent a reoperation (two each because of loss of reduction, refracture, the need for trimming or advancement of the nail, and delayed union or nonunion). Sixteen patients treated with external fixation required a reoperation (ten because of loss of reduction, one for replacement of a pin complicated by infection, one for debridement of the site of a deep infection, three because of refracture, and one for lengthening). One patient treated with a rigid intramedullary nail required debridement at the site of a deep infection, and one underwent removal of a prominent distal interlocking screw. One fracture treated with plate fixation required refixation following refractures. A multivariate analysis with adjustment for baseline differences showed external fixation to be associated with a 12.41-times (95% confidence interval = 2.26 to 68.31) greater risk of loss of reduction and/or malunion than elastic stable intramedullary nail fixation. CONCLUSIONS: External fixation was associated with the highest rate of complications in our series of adolescents treated for a femoral fracture. Although the other three methods yielded comparable outcomes, we cannot currently recommend one method of fixation for all adolescents with a femoral fracture. The choice of fixation will remain influenced by surgeon preference based on expertise and experience, patient and fracture characteristics, and patient and family preferences.