Does a traction-internal rotation radiograph help to better evaluate fractures of the proximal femur?

BACKGROUND The standard radiographic series for evaluation of a suspected hip fracture in most centers includes an anteroposterior (AP) radiograph of the pelvis, as well AP and cross-table lateral views of the hip. The natural femoral neck anteversion, as well as the fracture deformity, however, may make accurate fracture classification difficult. We have noted that inexperienced physicians sometimes misclassify hip fractures based on the initial radiographic series, which may lead to errors both in surgical planning and implant choice. At our institution, we routinely obtain a physician-assisted traction-internal rotation radiograph of the affected hip in all fractures of the proximal femur. The purpose of the current study was to examine the usefulness of the traction-internal rotation radiograph for the classification of hip fractures by junior residents in our department. MATERIALS AND METHODS Forty-seven sets of complete radiographs (AP pelvis, AP hip, cross-table lateral, traction- internal rotation views) of patients who sustained a proximal femur fracture were identified. Fifteen first year orthopaedic residents (PGY2) individually reviewed the cases and classified them as one of six possible choices: 1. nondisplaced femoral neck fracture, 2. displaced femoral neck fracture, 3. stable intertrochanteric fracture, 4. unstable intertrochanteric fracture, 5. intertrochanteric fracture with subtrochanteric extension, or 6. subtrochanteric fracture. Each fracture case was classified after first reviewing the standard hip series (AP pelvis, AP hip, and cross-table lateral). A traction-internal rotation radiograph was then added to each case, and any changes in the initial classification were noted. The resident's classification was then compared with those of the senior investigators (KJK, KAE), who used all four views for classification. RESULTS Reviewing a traction-internal rotation radiograph led to a statistically significant increase in agreement between the resident and senior investigators' classification (71.9% to 77.9%, p value < or = 0.01). The residents were more accurately able to identify fracture patterns as femoral neck (from a prior 98.5% to 99.3% after reviewing a traction-internal rotation view), intertrochanteric (a prior 87.7% to 91.3%), and subtrochanteric (prior 22.9% to 28.9%) after reviewing the additional radiograph. There were a total of 57 (8.1% of all responses) changes in classification after the traction-internal rotation view, 42 of which involved a change from an incorrect to a correct classification. In 50% of the changed responses, the correct classification would have led to a change in implant or surgical procedure choice, or both. CONCLUSION The routine addition of a traction-internal rotation radiograph increased the ability to accurately classify proximal femur fractures by junior residents in our department. This has a direct impact in accurate surgical planning and implant choice.