Reliability of and correlation between measurements of acetabular morphology

The authors sought to determine the intra- and interrater reliability of contemporary measures of acetabular morphology among a group of hip surgeons and to determine the correlations between measures of acetabular morphology. On 2 separate occasions, 3 hip surgeons independently performed blinded evaluations of anteroposterior pelvic radiographs of 40 patients. The lateral center-edge angle, medial center-edge angle, acetabular arc, extrusion index, acetabular index, Sharp's angle, posterior wall sign, crossover sign, femoro-epiphyseal acetabular roof index, acetabular wall indexes, and delta angle were assessed. A linear mixed model was used for variance estimation, and kappa and intraclass correlation coefficients were assessed for reliability. Pearson correlation coefficients were calculated for all possible pairs of radiographic measures. The acetabular index had the greatest interrater agreement (0.90; 95% confidence interval [CI], 0.84-0.93), whereas the lateral center-edge angle had the higher intrarater agreement (0.96; 95% CI, 0.92-0.98). The acetabular arc angle had the lowest interrater agreement (0.44; 95% CI, 0.29-0.57), and the Sharp's angle had the lowest intrarater agreement (0.66; 95% CI, 0.41-0.84). The posterior wall sign had an interrater agreement of 0.35 (95% CI, 0.11-0.54) and an intrarater agreement of 0.68 (95% CI, 0.49-0.86). The crossover sign had an interrater agreement of 0.66 (95% CI, 0.43-0.84) and an intrarater agreement of 0.85 (95% CI, 0.52-0.89). The acetabular index, lateral center-edge angle, and extrusion index presented with high coefficients of correlation. In addition, acetabular anteversion correlated with severity of dysplasia. Commonly used parameters such as the acetabular index, lateral center-edge angle, and extrusion index are reliable radiographic parameters to assess acetabular morphology. However, correlation between measures suggests that they may be redundant in quantifying acetabular morphology.