Hip preserving surgery with concentrated autologous bone marrow aspirate transplantation for the treatment of asymptomatic osteonecrosis of the femoral head: retrospective review of clinical and radiological outcomes at 6 years postoperatively

A total of 222 patients (341 hips) with ONFH were treated with CABMAT between April 2003 and March 2013. Based on magnetic resonance imaging, bilateral and unilateral asymptomatic ONFH was diagnosed in 119 patients (238 of 341 treated hips; 69.8%) and 38 patients (38 of 341 treated hips, 11.1%), respectively. In this series, we retrospectively examined 31 of 38 patients with unilateral asymptomatic ONFH who had been followed-up for two years or more (two-year follow-up rate, 81.5%).

Among the 31 patients with unilateral asymptomatic ONFH included in the study (19 males and 12 females), the mean age was 40.0 years (range, 26–70 years). The study included 13 right hips and 18 left hips.

The study endpoint was set as the time point of the most recent follow-up or the time when the patients required additional surgery (total hip arthroplasty).

Osteonecrosis of the hip was induced by corticosteroids and alcohol in 28 and 3 hips, respectively. Corticosteroids were used in 13 of the 28 hips to manage systemic lupus erythematosus, whereas the remaining 15 patients received corticosteroids for various reasons (rheumatoid arthritis, 1; nephrotic syndrome, 1; facial nerve palsy, 1; dermatomyositis, 1; collagen disease, 1; idiopathic thrombocytopenic purpura, 2; ulcerative colitis, 1; IgA nephropathy, 2; iritis, 1; brain infarction, 1; myasthenia gravis, 1; interstitial pneumonia, 1; and primary biliary liver cirrhosis, 1).

The diagnosis, classification, and staging of ONFH were based on the 2001 Japanese Orthopaedic Association (JOA) classification, which refers to findings on anteroposterior and lateral plain radiographs or magnetic resonance imaging scans [10]. Based on these criteria, ONFH was classified into four types by the location of necrotic lesions on T1-weighted or X-ray images. Type A lesions were defined as occupying the medial one third or less of the weight-bearing portion. Type B lesions were defined as occupying the medial two thirds or less of the weight-bearing portion. Type C1 lesions were defined as occupying more than the medial two thirds of the weight-bearing portion without extending laterally to the acetabular edge. Finally, type C2 lesions were defined as occupying more than the medial two thirds of the weight-bearing portion and extending laterally to the acetabular edge. Staging was based on the anteroposterior and lateral views of the femoral head on radiographs. Stage 1 was defined in hips with no specific osteonecrosis findings on radiograph but with specific findings on magnetic resonance imaging, bone scintigram, or histological examination. Stage 2 was defined in hips with demarcating sclerosis without collapse of the femoral head. Stage 3 was defined in hips with collapse of the femoral head, including crescent sign, but without joint space narrowing. Stage 3 was further divided into substages 3A and 3B, according to the extent of femoral head collapse (<3 mm and ≥3 mm, respectively). Finally, stage 4 was defined in hips with osteoarthritic changes. Preoperatively, 5, 6, 10, and 10 of 31 hips analyzed had osteonecrosis of type A, B, C1, and C2, respectively. Because all hips were asymptomatic, osteonecrosis of stage 3 or 4 was not observed. Specifically, of the 31 hips included in the study, 11 and 20 hips had ONFH stage 1 and 2, respectively.

Bone marrow aspiration, concentration, and transplantation were performed based on the method developed and previously performed by Sakai et al. and Yoshioka et al. [7, 8]. Specifically, the bone marrow was aspirated from both anterior iliac crests by using a bone marrow harvest needle, and subsequently transferred from the bone marrow collection kit into a collection bag. The blood fraction containing the BMA was processed by a two-step centrifugation method (KUBOTA 9800, Kubota, Japan) at room temperature. After inverse centrifugation of the blood bag at 1200 g for 10 min, the erythrocytes were transferred into the satellite bag until the interface between the plasma and erythrocyte layer was 15 mm from the bottom of the bag. Subsequently, using high-speed centrifugation at 3870 g for 7 min, plasma was transferred slowly into a satellite bag until approximately 4 cm of plasma were left on top of the buffy coat layer. Before transplantation, we typically created three multi-directional holes—central, anteromedial, and posterolateral—by percutaneous drilling with a Kirschner wire (diameter, 2.4 mm) to perforate the interface between the necrotic lesion and healthy bone. Next, core decompression was performed via the percutaneous technique with a 4.8-mm diameter trephine (Iso Medical Systems, Tokyo, Japan), which was inserted into the center of the necrotic site through the greater trochanter. Its position in the femoral head and necrotic site was monitored with biplane fluoroscopy, and transplantation was performed. After the operation, weight bearing was limited for 6 weeks, while non-weight bearing exercise was allowed.

Immediately before the procedure and at each follow-up, patients were evaluated in terms of the JOA hip score [11], which is an objective index of hip joint function quantified using items in three categories including (i) pain, (ii) hip range of motion, (iii) walking, and (iv) ADL performance. Clinical outcomes were defined in terms of the change in JOA hip score between the preoperative evaluation and the most recent follow-up evaluation. The JOA hip score is used as part of the preoperative and postoperative clinical assessment of hip function [12]. The 100-point scale includes subcategories for pain (40 points), range of motion (20 points), walking ability (gait; 20 points), and ADL performance (20 points), with higher scores indicating better hip function. The rate of conversion to total hip arthroplasty was also evaluated.

The anteroposterior and lateral radiographs of the affected hip were obtained at each clinical evaluation. Radiographic progression of femoral head collapse (from pre-surgery to the most recent follow-up) was evaluated in consideration of ONFH classification and staging. The extent of femoral head collapse was measured by using a template overlay of circles with increasing diameter (1-mm increments) [13] (Fig. 1). All radiographs were independently evaluated by two observers who were specialists in orthopedic surgery and authorized by the JOA.

The JOA scores before CABMAT and at the most recent follow-up evaluation after CABMAT were compared using the paired Student’s t test. Radiological outcomes of CABMAT-treated hips were compared against those of untreated hips (natural course of ONFH) and against those of hips treated with core decompression, and such comparisons used χ² tests, Fisher’s tests, and log rank test (for survival curves). Data were presented as mean ± standard deviation. A P-value of <0.05 was considered to indicate statistical significance. All statistical analyses were performed with IBM SPSS Statistics version 19.0 (IBM Corporation, Armonk, USA).

Overall, the mean follow-up period was 5.8 years (range, 2.0–6.9 years). On stratification, according to type of osteonecrosis, the mean and range of follow-up duration were 4.1 (2.0–6.5), 3.0 (2.0–3.5), 4.2 (2.0–6.4), and 5.8 (3.7–6.9) years for ONFH of type A, B, C1, and C2, respectively.

The mean and range of the JOA hip score was 88 (72–100) and 92 (45–100) points before surgery and at the most recent follow-up evaluation, respectively. One patient who underwent total hip arthroplasty of the contralateral side was excluded from this evaluation. While the mean JOA score improved following CABMAT, the improvement was not statistically significant (P ≥ 0.05, Student’s t test). The pain score was 40 (40–40) and 36 (20–40) points before surgery and at the most recent follow-up, respectively. Although the pain score decreased significantly after surgery (P < 0.05), only 6% (2/31) of patients reported pain during walking, whereas the other patients reported feeling no pain or slight discomfort in the affected hip.

The JOA walking and ADL scores, respectively, were 15 (10–20) and 19 (10–20) points before surgery, but 16 (10–20) and 18 (6–20) points after surgery, indicating significant improvement (P < 0.05, Student’s t test). Finally, the JOA range of motion score was 20 (17–20) and 19 (9–20) points before and after surgery, respectively, which did not indicate a significant change following surgery (Fig. 2).

A total of 9.6% (3/31) of patients underwent total hip arthroplasty at an average of 33 months after surgery.

Among hips with ONFH of types C1 and C2, 6 of 10 hips (60%) and 5 of 10 hips (50%), respectively, exhibited progression of femoral head collapse. The extent of collapse is listed in Table 1.

By comparing the outcomes of CABMAT treatment for asymptomatic, pre-collapsed ONFH with those associated with the natural course of ONFH, we could verify that CABMAT changed the natural course of the disease. Compared to the natural course and the outcomes of core decompression reported in the past, the outcomes of ONFH treated with CABMAT were better, as it is discussed in detail in the following paragraphs. Though CABMAT was not 100% successful, we found clear evidence that such a procedure can alter the natural course of the disease in some patients with ONFH.

We further discuss our findings in the context of current knowledge in the field of ONFH management. Nevertheless, since it is not possible to accurately match our study sample to those from previous investigations in terms of extent and location of ONFH, the comparisons should be considered qualitative.

This study has several limitations. First, because this was a retrospective study and because ONFH is a rare disease with large variations in clinical status (stage, type, etiology), defining a control group was difficult. The lack of a control group represents a limitation of this study. Furthermore, we used a small sample size (31 patients). To validate the efficacy of CABMAT, prospective studies are warranted, preferably in the form of multicenter, randomized clinical trials comparing CABMAT outcomes with the natural course of ONFH and with the outcomes of core decompression. Additional study limitations include the lack of adequate randomization and matching based on ONFH stage, ONFH type, sex, and age, which should be considered in future investigations.