The precise orientation of the acetabular component is one of the important factors in total hip arthroplasty (THA). Computed tomography (CT) provides image data for accurate measurements of the orientation of the acetabular component. However, in many studies in the literature, the orientation of the acetabular component after THA has been expressed as a combination of inclination angle (IA) and version angle (VA) measured on radiographs. For measuring VA, an anteroposterior (AP) radiograph of the hip joint, or a cross-table lateral radiograph has been used. The accuracy of these radiological measurements was not thoroughly studied. The purpose of this study was to evaluate the accuracy of measurements on radiographs comparing to those on CT.

Materials and methods: Twenty-four hips (21 patients) after THA were recruited for this study. The same acetabular components (Trilogy, Zimmer, USA) had been used without cement in all hips. An AP radiograph of the pelvis, an AP radiograph of the operated hip and a cross-table lateral radiograph of the hip were taken. From the AP radiograph of the pelvis, the angle between an inter-teardrop line and a tangential line to the opening face of the acetabular component was measured and defined as inclination angle on radiographs. From the AP radiograph of the operated hip joint, the lengths of two axes of the ellipse of the acetabular component were measured and a version angle was calculated using the Lewinnek’s method. This version was defined as the version from the AP radiograph (VAP). From the cross-table lateral radiograph, another version angle was measured using a modified version of Woo’s method, which is the angle between a gravity line shown by a metal chain and a tangential line to the opening face of the acetabular component. This angle was defined as the version from a cross-table lateral radiograph (VCL). The CT scanning of the pelvis was performed with 64-MDCT scanner (Aquilion TSX-101A/HA, Toshiba Medical Systems. co). The plane passing through the bilateral anterior superior iliac spines and pubic tubercles was used as references for measurements of inclination angle and version angle from CT image data. Accuracy of CT measurements had been validated using a phantom model. The absolute value of the difference between the measured angle on the radiograph and that on CT was defined as an error. Mann-Whitney U-test was used for statistics, and the level of significance was set at p< 0.05.

Results: Mean of the error for inclination angle was 2.2 degrees (range 0–6, SD 1.3). Comparing to this, both of the two methods for measurements of version angle on radiographs showed large errors. Mean of the error was 6.9 degrees (range 0–18, SD 8.0) for VAP, and was 6.0 degrees (range 0–14, SD 6.1) for VCL. There were no significant differences between errors for VAP and that for VCL. These results suggest that the measurements of version angle on AP radiographs or cross-table lateral radiographs should not be used for the studies of orientation of the acetabular component.