Necrotic Volume, Bone Marrow Edema and the Development of Symptoms in Patients with Osteonecrosis of the Femoral Head

Introduction: In recent studies using MR imaging, poor results of worsening pain in patients with osteonecrosis of the femoral head were reported in 23% to 62% [1–3]. The necrotic volume was reported to be one of the most significant factors for worsening of hip pain [4]. Bone marrow edema at the femoral head and neck was reported to be observed in patients with osteonecrosis, having a strong association with hip pain [1–3]. The purpose of this study was to prospectively investigate the risk factors on MR images to predict the outcome, particularly assessing necrotic volume and bone marrow edema. Materials and Methods: Eighty-three asymptomatic or minimally symptomatic hips in 61 consecutive patients were followed prospectively. Inclusion criteria included osteonecrosis of the femoral head identified by typical MR findings, with no radiographic evidence of progression of collapse, and equal to or greater than 85 points of Harris hip score. Symptomatic hips were defined as those complaining of mild or more severe pain with the hip score below 70 points. Every 3 months, the patients underwent clinical and radiographic examination, and MR images were obtained at 6to 12-month intervals. The mean patient age when the MR images initially revealed necrotic lesions was 35 years (range, 15–74 years). The necrotic volume was measured on the first abnormal MR images. To calculate the necrotic lesions of the entire femoral head on each image, an image-analysis software was used. Initially, the circumference of the femoral head and the necrotic area as seen on sequential T1-weighted coronal images were outlined using a non-permanent fine tip marker. The program measures the area of the outlined structure on each particular slice. The necrotic volume was calculated by the integration of each sequential coronal image. Bone marrow edema was defined as an ill-defined area of low signal intensity on T1weighted images with corresponding high signal intensity on T2-weighted images that involved the femoral head and neck beyond the necrotic zone extending to intertrochanteric regions. The radiographic stage was determined according to the classification of Steinberg et al [5]. Followup was continued until worsening of hip pain with less than 70 points of Harris hip score or surgery. Asymptomatic hips were followed for at least 24 months. The mean clinical and radiographic follow-up period after the initial diagnosis was 60 months (range, 3–168 months). Results: Thirty-six (43%) of the 83 hips were symptomatic at the latest follow-up. With a hip score below 70 points or the need for surgery as the end point, the cumulative rates of survival were 72.3% (with 95% confidence interval, 67.4–77.2%) at 1 year, 61.4% (56.1–66.7%) at 2 years, 57.3% (51.8–62.8%) at 5 years, and 54.7% (48.9–60.5%) at 10 years. Correlation of the initial and final radiographic stage was summarized in Table 1. The average necrotic volume was 28.5% (range, 3.0–70.2%): 39.5 ± 13.3% (mean and standard deviation) in the 36 symptomatic hips and 19.9 ± 16.5% in the 47 asymptomatic hips. In 35 hips with large necrotic volume of equal to or greater than 30%, 25 (71%) hips were symptomatic at the final follow-up, which were significantly higher compared to 48 hips with small necrotic volume of less than 30%; only 11 (23%) of these hips were symptomatic (p < 0.0001). Bone marrow edema was present in 28 (34%) hips during the follow-up period. Twenty-one of these 28 hips showed bone marrow edema on diagnostic MR images in which osteonecrosis was initially identified. Seven hips did not show bone marrow edema on those images. The average period between initial diagnosis of osteonecrosis and development of bone marrow edema in these seven hips was 16 months (range, 6–24 months). Bone marrow edema was typically identified around the time of onset of hip pain. Bone marrow edema was seen at stage I in three (11%) hips, stage II in five (18%) hips, stage III in 18 (64%) hips and stage IV in 2 (7%) hips. Twenty-seven (96%) of the 28 hips were symptomatic (Table 2). Bone marrow edema was significantly correlated with worsening of hip pain (p < 0.0001). The necrotic volume of hips with bone marrow edema was significantly larger than that without bone marrow edema (p < 0.0001).Bone marrow edema was found to be the most significant risk factor for worsening of pain (p < 0.0001). Discussion: With the use of computer software, we measured the necrotic volume on sequential coronal MR images, and found that the necrotic volume was one of the possible risk factors to predict the outcome. The necrotic volume might be important factors, however, some hips with large necrotic volume remained asymptomatic without progressive collapse and some hips with moderate necrotic volume became symptomatic with progressive collapse. Bone marrow edema was reported to have a strong association with hip pain [1–3], however, the mechanism of the edema remains unknown. The final radiographic stage of the 28 hips that showed bone marrow edema in this study was significantly advanced compared to those without bone marrow edema. Our results suggest that bone marrow edema might represent an inflammatory change in the reactive process during the progression of collapse of the femoral head and might be an unfavorable sign for prognosis of the disease. In conclusion, the large necrotic volume of equal to or greater than 30% may be the useful indicator to prognosticate future worsening of hip pain. Bone marrow edema was strongly correlated with necrotic volume and was the most significant risk factor for worsening of hip pain. References: 1. Kubo et al. Magn Reson Imaging. 15, 1017, 1997 2. Koo et al. Radiology, 213, 715, 1999. 3 Radke et al. Skeletal Radiol. 33, 519, 2004. 4. Steinberg et al. Clin Orthop. 367, 262, 1999. 5. Steinberg et al. J Bone Joint Surg Br. 77, 34, 1995. Table 1. Correlation of Initial and Final Radiographic Stage