Measuring acetabular dysplasia in plain radiographs

Several radiologic parameters have been used to qualify an acetabular coverage in studies determining whether an association exists between acetabular dysplasia and osteoarthritis of hip. However, it is not known which parameter is optimum for these epidemiologic studies. We evaluate the reliability, validity, and robustness of the radiologic parameters of acetabular coverage used in these studies. Center-edge angle (CEA), acetabular depth (AD), acetabular angle (AA), acetabular roof obliquity (ARO), and roof angle (RA) were evaluated. The components of intra- and interobserver reliability were tested. The correlations between each parameter were used to depict convergent validity. The robustness of the parameters to different projection (urogram), different definitions of the lateral acetabular margin, and a differing pelvic tilt were evaluated. The intra- and interobserver reliabilities of CEA, AD and AA ranged from 0.777 to 0.925. The CEA, AD and AA showed acceptable validity in the correlation. The AD on the urograms was 22.0% higher than those on the standing hip radiographs (P < 0.001). When the osteophyte was included in the definition of lateral acetabular margin, the CEA and AD increased significantly (P < 0.001). In simulating pelvic tilting, the AD increased significantly with the anterior pelvic tilt (P < 0.001). The ARO and RA showed poor clinical relevance. When measuring acetabular dysplasia, the AD is unsuitable for use, because it is not robust to different projection of beam and different pelvic tilts. Furthermore, one should consider that the CEA and AA are significantly influenced by different definitions of lateral acetabular margin.