MRI Finds Hip or Pelvic Fractures After Initial Negative Plain X-rays

Abstract & Commentary

Commentary by Jacob Ufberg, MD, Assistant Professor of Emergency Medicine, Residency Program Director, Department of Emergency Medicine, Temple University School of Medicine, Philadelphia, PA. Dr. Ufberg is on the Editorial Board of Emergency Medicine Alert.

Source: Dominguez S, et al. Prevalence of traumatic hip and pelvic fractures in patients with suspected hip fracture and negative initial standard radiographs — a study of emergency department patients. Acad Emerg Med 2005;12:366-370.

The purpose of this study was to evaluate the incidence of hip fractures presenting to the emergency department (ED) with negative initial radiographs. This was a retrospective cohort study performed in an academic community ED. The cohort included all adult patients presenting to the ED during a one-year period with hip pain who had received standard plain radiographs. Eligible patients were collected by a query of an electronic radiology log and medical records.

All plain radiographs and magnetic resonance imaging (MRI) studies were ordered solely at the discretion of the treating emergency physician. Initial plain films were read in real-time by a board-certified radiologist; all MRI studies were read by a fellowship-trained musculoskeletal radiologist, and re-read by a similarly trained radiologist to assess interobserver agreement. A fracture was defined as any fracture seen on plain film or MRI, or subsequent diagnosis of hip fracture as determined by follow-up phone call or review of subsequent medical records.

A total of 895 patients had hip films performed during the study period, with 764 (85%) completing follow-up. The mean age of the 764 patients included was 66.7 years, and 62% were female. Of the patients enrolled, 95% had a history of blunt trauma, most of whom (85%) were injured in a fall. Two hundred-nineteen patients (29%) had fractures identified on plain films; the most common fractures involved the femoral neck (36%) and intertrochanteric region (29%).

Of the 545 patients with negative plain films, 62 (11%) had MRI studies ordered by the treating physician. MRI identified 29 fractures in 24 patients (10% of all fractures identified in the study) with negative initial plain films. Overall, 4.4% of all patients with negative initial plain films had a fracture identified by MRI. Of the patients with negative plain films and positive MRI study, 92% were 65 years of age or older. Characteristics of the fractures identified by MRI were as follows: 35% involved the pubic rami, 28% involved the sacrum, 14% involved the femoral neck, 7% involved the intertrochanteric region, 3% involved the femoral head, 3% involved the acetabulum, and 10% involved other areas. The interobserver agreement for the presence of fracture on MRI was very good.

Among patients with initial negative plain films who did not receive MRI in the ED, no patient was determined to have a fracture on follow-up review of records or one-month post-visit phone call. The authors suggest that there is considerable potential for the development of a clinical decision rule to identify patients presenting to the ED with hip pain and negative plain films who remain at risk for fracture.


This study noted that approximately 10% of fractures in the cohort were not recognized by initial plain radiography, and that 4.4% of patients with hip pain and negative initial plain films had fractures identified by MRI. At first glance, these numbers suggest that MRI (or perhaps computerized tomography?) should be used more liberally by emergency physicians in an attempt to reduce the number of missed fractures.

However, it is important to note several aspects of this study. It appears that by using only physician discretion, all fractures were identified on the initial visit.

No patients were diagnosed with hip fracture subsequent to the initial visit, and the fracture rate among the patients who had MRI studies was 39%. It appears that the clinical judgment of emergency physicians (at least the ones in this particular ED) is pretty good, and may not need the help of a clinical decision rule.

Additionally, we should note that the majority of fractures identified by MRI were types of fractures that are managed conservatively. However, was every fracture truly identified? I find it hard to believe that while 10% of fractures were missed by initial radiograph in this study, there were truly no fractures among the nearly 500 patients with negative initial radiographs who did not undergo MRI. Most likely, some number of patients went home with clinically insignificant or conservative management-type fractures and recovered uneventfully.

What we really need is a clinical decision rule that targets the use of advanced imaging toward identifying fractures that are truly significant: ones that require operative management or carry a high risk of complication. My guess is that the medical/legal risks of not diagnosing ANY fracture will keep us from reaching this goal.