ABSTRACT
BACKGROUND: Ultrasound has been utilized in various settings for evaluation and treatment of skeletal injuries. Bone has different tissue acoustic impedance than soft tissue allowing visualization of the cortical disruption found in fractures. OBJECTIVE: To determine emergency physicians' accuracy in diagnosing cranial and long bone fractures using ultrasound. METHODS: This multi-center prospective double-blinded study used high-frequency linear ultrasound to detect induced fractures among eight test locations from eight cadaver models. After a standard orientation, blinded emergency physicians interpreted real-time sonographic images of test locations. RESULTS: Proximal tibia combined sensitivity (SE)/specificity (SP) was 87.3/69.8% with a combined positive predictive value (PPV)/negative predictive value (NPV) of 84.6/74.3%. Distal radius combined SE/SP was 93.7/93.5% with a combined PPV/NPV of 93.4/90.8%. Frontal combined SE/SP was 84.1/88.9% with a PPV/NPV of 84.9/88.3%. Temporal-parietal combined SE/SP was 95.2/87.9% with a PPV/NPV of 94.8/88.2%. Time to decision varied from less than 10 seconds to 357 seconds. Mean time to decision was 43 to 63 seconds depending on fracture site. CONCLUSION: Ultrasound by trained emergency medicine physicians can reliably identify fractures in the radius, tibia, frontal, and temporal bones in a very short amount of time, allowing for triage, treatment, and resource management.
