The institutional review board approved this retrospective study (# 015–0206). In total, 238 patients (401 hips) who showed a band pattern of the femoral heads on MRI visited our hospital from January 2010 to December 2019. Among these, SIF was diagnosed based on several published criteria [2, 23, 24] as follows: hip pain that began without any apparent history of trauma; radiographs that were normal or that showed the collapse of the femoral head, joint space narrowing and/or a linear patchy sclerotic area in the superior portion of the femoral head; a bone marrow edema pattern in the femoral head and/or neck on MRI; and a subchondral low signal-intensity band on T1 weighted MRI that was convex to the articular surface and parallel to the subchondral bone end-plate. We distinguished between SIF and ONFH via gadolinium enhanced MRI. Forty-seven hips in 44 patients (male: 10, female: 34) were diagnosed with SIF. In this study, 3 hips in 3 patients who showed rapid collapse and joint destruction, including that of the acetabular (AC), were excluded.
Among those diagnosed with SIF, patients who could be diagnosed within 3 months after hip pain were supervised to avoid weight-bearing with crutches for 6 weeks[25] and were treated on an outpatient basis every 2 weeks. Patients with late diagnosis or poor compliance could not be treated with conservative therapy as an initial treatment.
Data on patient demographics including age, sex, and body mass index (BMI), the period from onset to the first visit, history of corticosteroid intake or alcohol abuse, and medical history of osteoporosis drug intake were collected from their medical records. Data regarding whether patients could be treated conservative therapy and whether they required THA within the follow-up period were also collected. The indication of THA was the failure to relieve the pain and disability of daily life regardless of nonsteroidal anti-inflammatory medications. Alcohol abuse was defined as the consumption of more than 400 ml of alcohol per week, which is known to be a significant risk factor for osteonecrosis of the femoral head [26].
Radiographs were taken using the same technique throughout the study period; a standardized position of the beam and radiographic penetration were adopted. The radiographs of all patients were assessed using a picture archiving and communication system (PACS) on the anteroposterior (AP) radiographs. In the current study, the center-edge (CE) angle at the first visit and longitudinal joint space width (JSW) were investigated (Fig. 1A). For JSW analysis, concentric circles passing through three points set arbitrarily in the AC joint surface and the femoral head were drawn respectively (circle A and circle B in Fig. 1A). The distance between the intersection of each circle and the line, that runs through the center of the femoral head (O in Fig. 1A) and is perpendicular to the line between the bilateral teardrops (line A in Fig. 1A) was measured. The interobserver variability of the JSW between two observers (YK and TD) was 0.768.
The MRI examinations were performed by a 1.5-T system under 5-mm slice thickness. The T1- and T2-weighted spin-echo images and short tau inversion recovery (STIR) images on the coronal and axial (and/or oblique axial: paralleling the femoral neck axis) planes were available in all cases. The band lengths were measured at the slice in which the longest band was detected on T1-weighted MRI on the coronal plane as previously described [16](Fig. 1B). The interobserver variability of the band length between two observers (YK and TD) was 0.836.
Fasting blood samples were obtained to examine the biochemical markers of bone turnover related to osteoporosis, including the levels of intact type 1 procollagen-N-propeptide (P1NP) and tartrate-resistant acid phosphatase 5b (TRACP 5b). Areal BMD in the lumbar spine (LS, L2–L4) and femoral neck were assessed by dual-energy X-ray absorptiometry (DXA; Discovery A, Hologic Japan, Inc, Tokyo, Japan). Bone turnover markers and BMD were investigated at the definite diagnosis. Chi-squared or independent t-tests were used to compare the differences between the patients who required THA and those treated conservatively, and between patients who could comply with the weight-bearing limitation and those who could not. Cox regression analysis was performed to identify the risk factors for THA. Linear regression models adjusted for age, sex, anti-osteoporosis therapy, and the period from onset to diagnosis were built to determine the associations between changes in the JSW, band length, and bone metabolic markers. All statistical analyses were performed using SPSS Statistics version 23.0 (IBM Corporation, Armonk, NY) with the significance level set at 0.05.