Retroversion of the hemipelvis rather than hypoplastic posterior wall decreases acetabular anteversion in hips affected by Perthes disease

The acetabular retroversion has a moderate incidence of 31–60% in all patients of the Perthes disease. It might be caused by posterior wall dysplasia based on recent animal researches. However, some studies support that hemipelvic retroversion is the main factor for the acetabular retroversion. The primary pathological factor of increasing retroversion angle is still controversial anatomically. This study aimed to identify whether there is acetabular retroversion in children with Perthes disease,and to find a method to distinguish version types. Forty children with unilateral Perthes disease who were admitted to our hospital from January 1, 2012 to December 31, 2018 were enrolled, and 40 controls were matched based on sex and age. The acetabular anteversion angle (AAA), internal wall anteversion angle (IWAA), anterior wall height of the acetabulum (A), acetabular posterior wall height (P), and acetabular width (W) were assessed on computed tomography (CT) at the level of the femoral head center. The acetabular wall difference index (AWDI; AWDI = P-A)/W*100) was calculated. The mean AAA was significantly lower in Perthes disease hips (10.59 (8.05–12.46)) than in contralateral hips (12.04 (9.02–13.33)) (p = 0.002) but did not differ from control hips (9.68 ± 3.76) (p = 0.465). The mean IWAA was significantly lower in Perthes hips (9.16 ± 3.89) than in contralateral hips (11.31 ± 4.04) (p = 0.000) but did not differ from control hips (9.43 ± 3.82) (p = 0.753). The mean AWDI did not differ between Perthes hips (0.41 ± 4.94) and contralateral hips (− 1.12 (− 4.50, 2.17)) (p = 0.06) or control hips (− 0.49 ± 5.46) (p = 0.437). The mean W was significantly higher in Perthes hips (44.61 ± 5.06) than in contralateral hips (43.36 ± 4.38) (p = 0.000) but did not differ from control hips (45.02 ± 5.01) (p = 0.719). The mean A and P did not differ between Perthes hips and contralateral hips or control hips. Correlation analysis of all hip joints revealed a significant correlation between AAAs and IWAAs (r = 0.772; r = 0.643; r = 0.608; and r = 0.540). Linear regression analysis revealed that AAAs increased with IWAAs. Multiple linear regression showed that IWAAs and AWDIs have good predictive value for AAAs in both Perthes and control hips (R(2) = 0.842, R(2) = 0.869). In patients with unilateral Perthes disease, the affected acetabulum is more retroverted than that on the contralateral side, which may be caused by hemipelvic retroversion. The measurements in this study could distinguish the form of acetabular retroversion. IWAAs and AWDIs can be used as new observations in future studies of acetabular version.