A new hypothesis for the bone marrow edema pathogenesis during transient osteoporosis

Transient osteoporosis is an infrequent condition of uncertain etiology with pain, limited range of motion and radiographic evidence of osteoporosis affecting one or more joints. It is self-limited, reversible and can involve only the hip (transient osteoporosis of the hip, TOH) or, less frequently, one or more joints contemporaneously or at different times (regional migratory osteoporosis, RMO). We studied four men with transient osteoporosis, including two with TOH and two with RMO. All patients underwent a standard radiographic work-up of the affected joints, arteriovenous Doppler US, computed tomography, magnetic resonance imaging (MRI) and three-phase bone scanning. In all patients, symptoms were related to bone marrow edema demonstrated at MRI and to a transitory regional arterial hyperflow observed at the early scintigraphic analysis. On the basis of our observations, we hypothesize that regional arterial hyperflow may be the cause of the bone marrow edema and therefore of the transient osteoporosis.


Introduction
Transient osteoporosis is an infrequent condition of uncertain etiology, with pain, limited range of motion and radiographic evidence of osteoporosis affecting one or more joints. It is self-limited and reversible. Transient osteoporosis can involve only the hip (transient osteoporosis of the hip, TOH) or, less frequently, one or more joints contemporaneously or at different times (regional migratory osteoporosis, RMO) [1][2][3][4][5][6].
Bone marrow edema is the main feature of this disease [7][8][9][10][11]. In an attempt to explain a cause for bone marrow edema, Takatori et al. [12] suggested that a reversible circulatory disturbance probably takes place in the bone marrow of the femoral head. Other authors [13,14] consider TOH to be an early phase of avascular necrosis (AVN) that develops into complete AVN or regresses spontaneously.
At magnetic resonance imaging (MRI) we obtained T1weighted spin echo images (TR, 500 ms; TE, 15 ms) in the three standard planes in addition to a sequence of spoiled gradient echo T2-weighted images (TR, 480 ms; TE, 18 ms; flip angle, 25°). Bone scans were obtained using a three-phase technique. Each patient was positioned supine under a large field gamma camera. An intravenous bolus of Tc99m-methylene diphosphonate (MDP, 555 MBq) was injected. During the vascular phase, we acquired 120 one-second frames of the affected joint, compared with the contralateral joint. Blood pool anteroposterior spot images were acquired after 5 minutes of the pathological joint. Whole body scan was performed after three hours to evaluate the osteoblastic activity phase; at this time, a spot image of the affected joint was also obtained. In patients with RMO, blood pool and 3-hour spot images were also obtained of the previously involved sites to evaluate the clinical course. One week after bone scanning, we performed angioscintigraphy with 555 MBq Tc99m albumin (Sorin Biomedica, Saluggia, Italy). All patients underwent arterovenous Doppler ultrasonography (US) of the lower limbs.
The patients were managed by partially discharging the joint and by active mobilization without load. When symptoms resolved, we allowed full loading.
The study was conducted according to the principles of the Declaration of Helsinki and it was approved by the Ethical Committee at Second University of Naples.

Results
The study enrolled 4 men with transient osteoporosis, including 2 with TOH and 2 with RMO ( Table 1). The typical joi-nt pain lasted from 3 to 7 months (mean, 5). In all patients the disease led to no consequences. The two patients with RMO also had involvement of other joints of the lower limbs.
Radiographic exams were unremarkable within the first 15 days of clinical onset. These, however, showed osteoporosis about one month after symptoms' onset ( Fig. 1a).  All cases presented an osteoporotic rarefaction of the joint ends (Figs. 2a, 3a, 4a). The profiles appeared hazy and difficult to define. The joint space remained constant. Demineralization increased progressively, and peaked two months after symptomatology onset. At the fifth or sixth month, remineralization was evident (Fig. 1b, 2e). After twelve months, the radiographic outline was normal.
Computed tomograms also showed reduced bone density and confirmed the standard radiographic evidence (data not shown). MRI examinations were always positive, even when carried out shortly after the onset of pain. There was a wide area at the joint ends which was difficult to define. The signal intensity on T1-weighted images was decreased with high intensity on T2-weighted images (Figs. 2c, 2d, 3c). Our opinion is that it represents bone marrow edema. It is worth noting that in the hip joint the acetabulum was not affected while the femur was altered from the head to the neck and in the trochanteric region. The results of the examination were normal at the fifth month. This corresponded to the termination of pain.
At three-phase bone scanning with Tc99m-MDP, carried out systematically at the first sign of illness, we noted an asymmetrical flow and distribution pattern of the tracer during the vascular phase with hyperflow and volume increase on the diseased side (Fig. 5). Three-hour images showed an accumulation of tracer according to increased osteoblastic activity (Fig. 5c).
Bone scanning was repeated after disappearance of pain (Fig. 6). Symmetrical flow and blood pool activity at the level of the diseased joint were observed, indicating normalized blood flow. However, 3-hour images demonstrated the persistence of an intense uptake in the same area, related to the persistence of increased osteoblastic activity (Fig. 6c).
In the two patients with RMO, images of the previously involved sites were normal in blood pool phase but showed increased activity on three-hour images (data not shown).
To better define these data, we also carried out angiography using Tc99m-albumin in all patients at the onset and at the end of the disease to confirm that the tracer's distribution at the diseased side was related to an increase in arterial blood flow. Tc99m-albumin, unlike Tc99m-MDP, does not diffuse and it does not bind to the bone tissue, therefore representing a reliable marker of the blood flow and pooling. Our data confirmed regional arterial hyperflow in the vascular phase, and an increased tracer distribution on blood pool images (Figs. 4c, 5e).
Venous and arterial Doppler US examinations of the lower limbs excluded vascular disease.
Comparison of how symptoms progressed with the results of the different diagnostic imaging exams showed that MRI and bone scanning are the most effective tests for an early diagnosis since positive results appeared at the onset of clinical symptomatology. MRI examinations rapidly turned to normal at the end of pain. The 3-hour images of the bone scans turned to normal less rapidly. Radiographs and computed tomograms gave positive results later and slowly turned to normal as the bone remineralized.

Discussion
While transient osteoporosis was initially discovered in pregnant women during the third trimester of pregnancy [5,16,17], it can affect both sexes, from late adolescence onward. A painful clinical symptomatology affects the joint, mostly the hip, with no apparent cause [4,[18][19][20]. After a gradual onset, the pain continues for several weeks and, at the same time, movement is impeded. Pain reduces or ceases at rest. Muscle hypotrophy and slight joint swelling may occur. In a few weeks, radiographs show a pattern identical to that of joint osteoporosis. Osteoporosis develops, reaches a peak in 6-12 months, and slowly regresses. In the migratory form (RMO), multiple joints are involved at the same time or, more often, at successive stages [2,21,22]. The etiology of TOH is unclear. Viral, inflammatory, metabolic and hormonal factors have been excluded. A transient ischemic insult to the bone may be responsible for transient osteoporosis [6]. Blood chemistry values are usually normal and synovial liquid is sterile. Histological examination may show areas of osteoporosis and insignificant inflammatory changes [5,17,18] or bone edema with necrotic and reparative processes [11]. Transient osteoporosis can be easily distinguished from several other diseases. During certain phases of the disease, a differential diagnosis between TOH and AVN is however possible through the use of MRI and bone scanning [5,11,23,24]. In patients with TOH, MRI shows areas of weak signal on T1-weighted images and hyperintensity on T2weighted images. These abnormalities correspond to an increase in water content in the bone marrow. The presence of bone marrow edema has led some authors to consider TOH as a transient condition which can either lead to AVN of the hip or disappear spontaneously. As a result, the definitions of "transient marrow edema syndrome" [7], "transient osteoporosis" [1,4,12] and lastly "bone-marrow edema syndrome" [11] have been proposed.
Causes of the bone marrow edema in transient osteoporosis are unclear. Some authors suggested that this is a disturbance of the venous outflow [11]. Other authors proposed a distinction between hypervascular and ischemic disorders at the onset of the edema [13].
Our study using Tc99m-MDP and Tc99m-albumin scanning confirmed regional arterial hyperflow in the vascular phase. This is concurrent with an increase in the blood pool in the early phase of the disease. We demonstrated that bone marrow edema is a consequence of arterial hyperflow, although we cannot exclude a concomitant venous outflow disorder. The spontaneous resolution of this regional arterial hyperflow, with resolution of bone marrow edema on MR images, coincides with the disappearance of pain. Successively, full weight bearing permits remineralization of the affected joint and termination of the illness.
It has to be pointed out that blood flow and volume and osteoblastic activity do not present overlapping behaviors at bone scanning. While the increased flow and volume disappear with symptoms, high uptake on the 3-hour scans, expressing high osteoblastic activity, is present also after disappearance of pain.
Since TOH is self-limited and is of brief duration, surgical core-decompression is probably not indicated. In conclusion, our data suggest that temporary arterial hyperflow may contribute to the transitory bone marrow edema observed in transient osteoporosis. Further studies are needed to elucidate the cause of arterial hyperflow in this disease.