ABSTRACT
We constructed a clinical decision rule to optimize the use of radiography in patients with acute knee injuries. A prospective observational study at a university hospital ED was conducted over 10 months. Patients 15 years of age and older with acute knee injuries who underwent radiography were included. Patients were excluded if they were intoxicated, had distracting injuries, previous knee surgery, or open wounds. A standardized closed question data collection instrument that recorded 12 historical and 26 physical examination criteria was used. Radiographs were interpreted by two radiologists blinded to each other's diagnosis. The main outcome parameter was the presence or absence of a fracture. We identified 18 fractures in 213 patients (8%). Patients with fractures were more likely to have severe joint line tenderness, severe localized swelling, an effusion, ecchymosis, flexion < 90 degrees, and an inability to bear weight. A clinical algorithm for the use of radiography that requires the presence of either an inability to bear weight, an effusion, or an ecchymosis was 100% sensitive for the detection of a fracture. All 76 patients without any of these criteria did not have a fracture. Limiting knee radiography to patients with these criteria would have reduced radiography by 39% without missing a fracture. In conclusion, a clinical decision rule for knee radiography that requires the presence of either an inability to bear weight, an effusion, or an ecchymosis was shown to reduce the need for radiography by 39% while still identifying all fractures. Prospective validation of this model is required.
