Although decision rules for radiographs of pediatric knee injuries have been suggested from retrospective studies, prospective evaluations of such rules have been limited. We sought to prospectively assess the performance of a rule in children presenting with acute knee injuries. Eligible participants were children aged 3-18 years with an acute knee injury. The settings for the study were a tertiary pediatric emergency department (ED), a community hospital ED, and a pediatric urgent care center. All of the participants received standard knee radiographs. Before radiography, each patient was assessed by a pediatrician or pediatric emergency physician for presence of the following: 1) inability to bear weight, 2) inability to flex the knee to 90( degrees ), 3) presence of bony tenderness. The radiographs were interpreted by a radiologist blinded to the study; those with findings reported as consistent with acute fracture were considered positive. A total of 146 patients were enrolled (65% male, mean age 11.6 years). Of these, 15 (10.3%) had a fracture on their radiograph, 6 of which were related to trampoline use. Seventy-seven (53%) were negative for criterion 1 (i.e., able to bear weight immediately after the accident and in the ED), none (0%) of whom had fractures. The negative predictive value of this criterion was 1.0 (95% CI 0.94-1.0). The positive predictive value was 0.22 (95% CI 0.13-0.34). The sensitivity was 1.0 (95% CI 0.82-1.0). The specificity was 0.59 (95% CI 0.50-0.67). Three patients negative for criterion 3 were found to have fractures. The proximal tibia was the most common fracture site (47%). In conclusion, assessment of the ability to bear weight would have decreased the use of radiography by 53% without missing any fractures in our study population. No additional value to the rule was found by adding assessment of the ability to flex the knee or bony tenderness.