2.1 Patients selection
This study is a retrospective observational cohort study. After approval from the institutional review board of Children’s Hospital of Chongqing Medical University (No.2017001). We retrospective screened patients who underwent CR due to DDH between February 2012 and November 2015 in our single tertiary medical institution. Our inclusion criteria were 1) Patients with late-presenting DDH that more than six months at diagnosis and patients who failed to the prior treatment including Pavlik harness or Ilfeld abduction orthoses; 2) DDH patients with hip subluxation and dislocation (IHDI ≥ grade II); 3) patients were received CR following the bilateral long leg hip spica cast immobilization; 4) patients and their radiographic data were followed for at least 24 months. Exclusion criteria were 1) Patients with acetabular dysplasia only or slight subluxation; 2) the hip dislocation was associated with a syndrome or other congenital hip abnormality; 3) patients with history of any open reduction procedure before initial CR; 5) patients with incomplete clinical and radiographic data at presentation.
After screened, one hundred and ten DDH patients with 138 affected hips were included in present study. There were 17 males and 93 females. There were 82 patients with unilateral DDH (82 hips) and 28 patients with bilateral DDH (56 hips). The average age at the initial treatment was 16.57 ± 4.96 months which was range from 6.40 to 33.20 months.
2.2 Closed reduction procedure
Arthrography was performed in all the affected hips of included patients through an adductor longus muscle approach using 1 cm3 of Iohexol as a contrast to evaluate hip position and assist reduction [11]. The reduction was performed by Ortolani manoeuvre gently, and CR was considered to be achieved when the centre of the femoral head had been pulled down to a position opposite the triradiate cartilage (Figure 1). Furthermore, if the adductor contracture impeding the hip reduction, the percutaneous adductor tenotomy was performed to reach a reliable safe zone [12]. Thereafter, as concentric and stable reduction was achieved, the hip was immobilized by the bilateral long leg hip spica cast at 90° to110° of flexion and 40° to 60° of abduction for 12 weeks, with a plaster change at six weeks. All patients were treated with an abduction orthoses after removal of the spica cast for a period of more than three months until concentric reduction was stable. During follow-ups, affected hips with redislocation and/or the residual acetabular dysplasia would be suggested to CR failure, and the open procedures (open reduction of the dislocated hip concomitant with innominate osteotomy and/or femoral osteotomy) would be conducted only if informed consents were obtained from these patients’ parents. All patients were followed-up every three months in the first year after removal of cast, and then followed up every six months during the second year, and every year thereafter. Anteroposterior pelvic radiographs and the frog leg lateral view were performed in all patients preoperatively and at each follow-up after removal of spica cast to assess the reduction. Nevertheless, patients were only taken an anteroposterior pelvic radiograph during Spica casting immobilization. However, for a hip with an uncertain reduction, a CT scan or MRI would be further employed for intensive evaluation. All the enrolled patients’ radiographs were reviewed individually by two researchers (Y. Z. and G. Z.) and all the classifications were determined by two authors with a consensus.
2.3 Radiographic evaluation before initial closed reduction
2.3.1 IHDI (International Hip Dysplasia Institute classification): The degree of the hip dislocation was assessed on the basis of the IHDH classification [13].
2.3.2 Presence of ossofic nucleus of femoral head: The presence of a proximal femoral ossific nucleus in each patient was reviewed and recorded based on the pelvic plain radiographs before the initial CR.
2.3.3 AI measurements: The acetabular index was measured on the AP pelvic radiographs to evaluate the acetabulum developmental situation at the time of CR [14].
2.3.4 Medial interval (MI) after CR: the medial interval was defined as the vertical distance between the medial edge of the ischium to the middle point of the proximal metaphyseal border of the femur [15].
2.3.5 Osteonecrosis of femoral head: The definition of femoral head osteonecrosis was graded according to the Bucholz-Ogden system[16]. As the Bucholz-Ogden type I and II is not currently thought to affect the functional and radiographic outcomes at skeletal maturity [17]. We therefore defined that type III and IV as the femoral head osteonecrosis in present study.
2.3.6 Severin Classification: The radiographic outcomes were assessed on the basis of the Severin radiographic classification [18]. Severin types I and II were considered to a success of CR, however, the types III, IV, V, and VI were considered to a failure of CR.
2.4 Primary outcomes
Our primary outcome was to evaluate the efficacy of CR in the treatment of DDH and to further investigate the underlying risk factors associated to the CR failure. Failure of CR was defined as follow: 1) a hip that underwent the OR procedures (open reduction of the hip with/without osteotomies) owing to the redislocation or persistent acetabular dysplasia after initial CR; 2) a hip with a grade range from III to VI according to the Severin radiographic classification at the latest follow-up. For the determination of the risk factors related to the CR failure, it is logical to adopt cases instead of hips as the independent variable because the demographic data (age, sex, etc.) was unique in each case with bilateral DDH. Therefore, cases would be defined as failure even if only single side failure occurred in the bilateral DDH.
2.5 Secondary outcomes
As osteonecrosis of femoral head after CR in the treatment of DDH was also a widely concerned issue. Therefore, we further assessed the AVN occurrence among the cases with preliminary successful CR.
2.6 Statistic analysis
All variables were analyzed by SPSS 22.0. Statistical software, and continuous data were indicated by X ± SD. Chi-square test and ANOVA analysis was used for univariate comparison and binary logistic regression analysis was used for multivariate analysis, respectively. When investigate the relevant risk factors, the ROC curve was used to determine the grouping node, and the AUC>0.5 was considered the model have predictive value. Determined the level of statistical significance with the P value set at 0.05 (P ≤ 0.05).