Radiographic detection of hip and pelvic fractures in the emergency department.

OBJECTIVE Our purpose was to evaluate the detection of hip and pelvic fractures with radiography in the emergency department. MATERIALS AND METHODS All MR images of the lower extremity or pelvis ordered from July 2005 through June 2008 by the emergency department after the patient had undergone radiography were retrospectively reviewed in consensus by two musculoskeletal radiologists. All radiographs and MR images were reviewed in regard to patient age and sex, MRI protocol, site of fracture, fracture identification, other possible pain generators, and unexpected findings. Receiver operating characteristics analysis was performed on the findings. RESULTS A total of 92 patients and 97 examinations were included. Our patient sample (77 women, 15 men; average age, 70.8 years; range, 19-94 years) had an elderly female bias. Sixty-five of the patients had a history of trauma. Thirteen patients (14%) with normal radiographic findings were found to have 23 fractures at MRI (six hip and 17 pelvic fractures). In 11 patients (12%) MRI showed no fracture after radiographic findings had suggested the presence of a fracture. In another 15 patients who had abnormal findings on radiographs, MRI depicted 12 additional pelvic fractures not identified on radiographs. In 43 of the 59 patients (73%) without MRI evidence of a fracture, the MRI findings suggested the presence of a potential pain generator, including muscle edema and tears, trochanteric bursitis, and hamstring tendinopathy. Receiver operating characteristics analysis for the detection of hip fractures with radiography showed an area under the curve of 0.74. CONCLUSION Our study showed poor sensitivity of radiography in the evaluation of hip and pelvic pain in the emergency department.