Steroid-related osteonecrosis--an update

The use of systemic corticosteroids is common in patients with a variety of different renal diseases. Low-dose oral corticosteroids are part of many long-term treatment protocols for patients following renal transplantation or with different forms of glomerulonephritis and systemic autoimmune diseases, i.e. systemic lupus erythematosus or vasculitis. High-dose intravenous corticosteroids are typically used for short durations to treat transplant rejection episodes or for the induction of remission in highly active autoimmune diseases. Osteonecrosis (ON) is a well-known and feared complication in patients with renal diseases undergoing systemic corticosteroid therapy. It predominantly occurs in the femoral head (FH) (FH necrosis, FHN; Figure 1), but also in the proximal humerus, the knee and the ankle [1]. (a) MR scan of an FHN stage ARCO 2 in a 32-year-old male patient. The necrotic area shows a sharply demarcated signal change, but no sign of collapse of the subchondral bone. The physiologically spheric shape of the FH is still intact. (b) Bilateral FHN ARCO Stage 3 in a 39-year-old male renal transplant patient. He first received an autologous nonvascularized bone transplantation in the left hip (arrow). In the right hip, a vascularized autologous fibular graft was implanted (arrow). ON accounts for ∼10% of the >500 000 total joint replacements performed annually in the USA [2]. However, according to the author’s experience, including 17 years of extensive work in the field, a significant proportion of ON remains undiagnosed, or is first diagnosed in the late and irreversible stage of secondary arthritis (joint destruction). After renal transplantation, the incidence of ON, mainly of the FH, has been reported to be between 3 and 41% [1]. With the standard use of calcineurin inhibitors following renal transplantation, the ON incidence decreased—presumably due to reduced steroid dosage—but remained at ∼4% in renal transplant recipients treated with cyclosporine or tacrolimus [3, 4]. The average age of patients receiving total hip replacement (THR) for FHN is 38 years, with only 20% of patients being >50 years old at the time the operation was performed [5]. Bilateral involvement of the FH, reported to be 40 to 89% [6], is also typically seen at our clinic. Intravenous methylprednisolone pulse therapy to treat acute transplant rejection has been associated with FHN. Cumulative methylprednisolone, at doses of >2 g for >3 months, significantly increased the risk for FHN [7]. Patients treated with doses of prednisone <20 mg/day have a lower risk for the development of ON. This is supported by a prospective study in patients with systemic lupus erythematosus, in which only patients who have exceeded daily doses of 20 mg prednisone developed ON, while no ON was reported in patients who remained <20 mg of daily prednisone [8]. The dose dependency of steroid-induced ON is further demonstrated in renal transplant recipients. FHN was associated with the use of steroid-containing immunosuppression and occurred primarily within the first 3 months after renal transplantation [6]. More recently, a statistically significant association was described between FHN incidence and the total dose of steroids during the first 2 months after renal transplantation [9]. These authors also found a dose-response relationship and stated that FHN often occurs 3 months after renal transplantation with immunosuppressive therapy containing steroids. In a study of 2500 cases of FHN, 35% were steroid associated, 22% alcohol related and 37% were considered to be idiopathic [10]. Further aetiologies are dysbarism, also termed Caisson disease [11], sickle cell anaemia, Gaucher’s disease and inflammatory bowel disease [5]. In recent years, ON of the jaws related to bisphosphonate therapy has been reported [12]. A concise patient risk profile for the development of ON is not yet available for clinical use. ON is defined as ‘the death of all cellular elements of bone’ [11]. Ischaemia is the principal pathomechanism of ON. We could show that steroids reduce FH blood flow in pigs [13] and have a vasoconstrictive effect on lateral epiphyseal arteries of the FH, which could lead to FH ischaemia and subsequent necrosis [14]. Bone arteries show characteristics similar to other arteries; for example, blood flow to the bone is regulated by the degree of wall tension of the supplying arteries, and vasoconstriction increases perfusion pressure and reduces blood flow in bone arteries [15]. A further contributor to ischaemia may be systemic fat embolism after renal transplantation and associated steroid treatment [10]. Apart from the haemodynamic effects and embolism, local and systemic activation of intravascular coagulation based on a hypercoagulable state of plasma has been suggested as an important event leading to ischaemia [11]. It has been reported 40 years ago that increased plasma fibrinogen levels could have a systemic effect on coagulation under steroid treatment [16]. Consistent with these reports, Glueck et al. [17] found thrombophilia and hypofibrinolysis in 12 ON patients secondary to glucocorticoid treatment. Fat cell hypertrophy was found in the femur of rabbits after 5-month methylprednisolone treatment and dexamethasone treatment of a puripotent cell line derived from mouse bone marrow resulted in differentiation of these cells into fat cells in vitro [18]. The growth of fat cells in the intraosseous compartment was suggested to increase intraosseous pressure resulting in sinusoidal compression and thereby diminished perfusion of bone. Core decompression, a therapy by which a channel is drilled into the osteonecrotic lesion for decompression (see below), reduced FH intraosseous pressure after steroid treatment in pigs [19]. We demonstrated that corticosteroids reduce FH blood flow in pigs [13] and have a vasoconstrictive effect on lateral epiphyseal arteries of the FH, which could lead to FH ischaemia and subsequent necrosis [14]. A single intramuscular methylprednisolone injection of 20 mg/kg into male adult Japanese white rabbits produced characteristic histological signs of femoral and humeral ON at only 4 weeks after the injection [20]. Another model of dexamethasone-induced ON has been described recently by Relling et al. in BALB/cJ mice [21]. Dexamethasone, which was administered orally in this study, led to ON lesions in the distal femora of 40–45% of animals after 12 weeks. Typical clinical signs of ON are pain in the groin, thigh or buttock. Pain can occur during weight-bearing activities as well as at rest, or it can be induced by motion. Only few patients are asymptomatic. A new classification of ON has been developed by the Association of Research on Osseous Circulation (ARCO) (Table 1) [5]. Stage 0 is only detectable by histology. In the early stage, ARCO Stage 1, ON is reversible and can be detected by magnetic resonance imaging (MRI) and scintigraphy; X-ray examination is insufficient to detect ON in this reversible stage. Detection at this stage is crucial, as ARCO Stage 2, which is the first X-ray detectible stage, is already irreversible. The crescent sign in ARCO Stage 3 represents a separation of the subchondral plate from the underlying necrotic cancellous bone and becomes evident on MRI and radiographs (Figure 1). Stage 4 implies joint destruction and secondary arthritis of the hip. Reliable serum markers for ON have not yet been described. ARCO Classification of FHN [5] and the best-documented therapy according to stage ARCO Classification of FHN [5] and the best-documented therapy according to stage Few experimental and clinical studies exist on how to prevent ON. The best prevention, however, is avoiding steroids whenever possible and to choose the lowest possible dosage. Recently, a nitrate skin patch administered at the same time as 20 mg/kg bodyweight intramuscular methylprednisolone, significantly prevented ON of femora and humeri in a rabbit model [22]. A beneficial effect of nitrate patches in human steroid-induced ON needs to be confirmed and the potential corresponding side effects elucidated in future clinical trials. Statin treatment has been proposed for the reduction of ON risk based on data obtained from animal experiments [23, 24]. The beneficial effect of statins on the incidence of ON in humans is as yet uncertain. A retrospective MRI-based study in 284 patients found a lower ON incidence in statin-treated patients as compared to those not on statins [25]. However, statins failed to reduce ON risk in a prospective study in renal transplant patients receiving corticosteroids [26]. Taken together, effective preventive measures apart from steroid avoidance or reduction currently remain to be established. Treatment of glucocorticoid-induced ON is a challenge, as sufficiently large prospective controlled studies are lacking. The ARCO classification is a good basis for a systematic stage-related therapy (Table 1). Core decompression leads to reliable pain reduction and a good outcome when performed in Stage 1 of FHN [27]. Bisphosphonates have been shown to be beneficial in two studies [28]. However, bisphosphonates bear a risk for adynamic bone disease in renal patients and may induce ON of the jaw; consequently, their general use cannot be recommended. In Stage 2 FHN, core decompression plus autologous non-vascularized bone transplantation are performed by different techniques [5]. The authors take out a bone cylinder from the FHN via the femoral neck under fluoroscopy, the tip of which contains the centre of the necrotic area. We then debride the necrotic tissue, turn the bone cylinder and reinsert it so that the vital lateral end then lies in the necrotic zone. In Stage 2 or 3 FHN, vascularized fibular grafts have a longer FH survival, but the operative procedure is much more demanding and longer [27, 29]. An autologous vascularized fibula transplant for ON, of the right FH, is shown in Figure 1. In a long-term study of vascularized fibular grafting, after a follow-up of at least 10 years, only 10.5% of the treated hips had failed and were revised to total hip arthroplasty [29]. Intertrochanteric flexion osteotomy is another joint-preserving therapy for FHN [30] by which the osteonecrotic lesion is rotated away from the weight-bearing area. Boney non-union and pseudarthrosis are method-specific complications. The rate of failure of flexion osteotomy in our cases [30] was 10% after 5 years and 27% during the follow-up period of 10 years. However, later replacement of a hip which had undergone flexion osteotomy earlier implies technical difficulties. Therefore, we only perform flexion osteotomy in a few selected cases. In the last stage of ON, collapse of the FH results in secondary osteoarthritis of the hip. At this stage, joint replacement is indicated. The long-term results of THR are known to be worse than those in matched groups of patients who have diseases other than FHN involving the hip [31]. THR is not the preferred treatment in young patients because of shorter implant survival in FHN and the risk of repetitive surgical revisions [31]. In particular in the young patient with non-traumatic FHN, joint-preserving operative techniques should be employed if still possible. In conclusion, ON is a common complication of systemic steroid therapy. It is crucial to find ways to primarily prevent FHN, to detect it in the reversible ARCO Stage 1, and to preserve the joint as long as possible in young patients with ON. To achieve this goal, close interdisciplinary cooperation between the nephrologist and orthopaedic surgeon is important. MR is the gold standard diagnostic method to detect ON in the early stage. Core decompression is the best-documented therapy in ARCO Stage 1. In Stages 2 and 3, vascularized and non-vascularized bone grafts can delay the need for total joint replacement. In ARCO Stage 4, secondary arthritis, joint replacement is advocated. Conflict of interest statement. None declared.