Although periacetabular osteotomies are widely used for the treatment of symptomatic acetabular dysplasia, the surgical outcomes after long term follow-up are still limited. Thus, we assessed hip survival and patient-reported outcomes (PROMs) at 20 years after the transposition osteotomy of acetabulum (TOA).

Among 260 hips in 238 patients treated with TOA, 172 hips in 160 patients were evaluated at average 20.8 years, excluding patients who died or lost to follow-up. Kaplan-Meier analysis was used to assess survivorship with an end-point of THA. PROMs were evaluated using the VAS satisfaction, VAS pain, Oxford hip score (OHS), and Forgotten joint score (FJS). The thresholds of favorable outcomes of FJS and OHS were obtained using the receiver-operating characteristic curve with VAS satisfaction ≥ 50 and VAS pain < 20 as anchors.

Thirty-three hips (19.2%) underwent THA at average 13.3 years after TOA. Kaplan-Meier analysis revealed hip survival rate at 20 years was 79.7%. Multivariate analysis showed the preoperative Tönnis grade significantly influenced hip survival. Survival rates with Tönnis grade 0, grade 1, and grade 2 were 93.3%, 86.7%, and 54.8% at 20 years, respectively. More than 60% of the patients showed favorable PROMs (VAS satisfaction ≥ 50, VAS pain < 20, OHS ≥ 42, FJS ≥ 51). Advanced Tönnis grade at the latest follow-up and higher BMI were significantly associated with unfavorable OHS, but not with other PROMs.

This study demonstrated the durability of TOA for hips with Tönnis grade 0–1 and favorable satisfaction in majority of the patients at 20 years after surgery. Current presence of advanced osteoarthritis is associated with the lower hip function (OHS), but not necessarily associated with subjective pain and satisfaction. Higher BMI also showed a negative impact on postoperative function.

In order to investigate the mechanisms of collapse in osteonecrosis of the femoral head, we examined which part of the femoral head was the key point of a collapse and whether a collapsed region was associated with the size of the necrotic lesion. Using 30 consecutive surgically removed femoral heads we retrospectively analysed whole serial cut sections, specimen photographs, specimen radiographs and histological sections.

In all of the femoral heads, collapse consistently involved a fracture at the lateral boundary of the necrotic lesion. Histologically, the fractures occurred at the junction between the thickened trabeculae of the reparative zone and the necrotic bone trabeculae. When the medial boundary of the necrotic lesion was located lateral to the fovea of the femoral head, 18 of 19 femoral heads collapsed in the subchondral region. By contrast, when the medial boundary was located medial to the fovea, collapse in the subchondral region was observed in four of 11 femoral heads (p = 0.0011). We found that collapse began at the lateral boundary of the necrotic lesion and that the size of the necrotic lesion seemed to contribute to its distribution.