Paget's disease of bone is a pathologic condition marked by increased bone resorption and formation 1,2 . The exact etiology is not fully understood; however, a slow paramyxovirus infection occurring in patients with a genetic predisposition is suspected 3 . The proximal part of the femur and the acetabulum are involved in as many as 80% of patients with Paget's disease 4 . Also, osteoarthritis of the hip joint is frequently associated with Paget's disease with medial or concentric wear patterns 5 . Four major studies on the efficacy of total hip arthroplasty for the treatment of Paget's disease of the hip have been reported 6-10 . Complications associated with the procedure have included varus placement of the femoral component, protrusio acetabuli, increased blood loss, difficulty with reaming due to sclerotic bone, heterotopic ossification, and an increased risk of mechanical failure. Recently, extremely rapid osteolysis in the proximal part of the femur was reported following total hip arthroplasty in a patient with extensive Paget's disease of the pelvis and femur 11 .
We report the case of a patient in whom monostotic Paget's disease of the pelvis was transferred to the middle part of the uninvolved, ipsilateral femur following total hip arthroplasty.
A sixty-six-year-old man was referred to our department in 1993 because of intractable pain in the left part of the groin related to osteoarthritis of the hip. The patient had untreated monostotic Paget's disease involving the left hemipelvis. Plain radiographs demonstrated no evidence of the disease in any other location, including, notably, the ipsilateral femur ( Fig. 1 ). The patient underwent a total hip arthroplasty through a transtrochanteric approach with use of a Charnley-Marcel Kerboull prosthesis (CMK; Benoist Girard, Howmedica, Herouville Saint Clair, France). Both components were inserted with use of CMW 1 bone cement (DePuy, Devon, Exeter, United Kingdom). The femoral medullary canal was blocked with use of a bone plug made of cancellous bone that had been retrieved from the acetabulum and the proximal part of the femur, as is performed routinely at our institution ( Fig. 2 ). Bone was retrieved during the preparation of the acetabulum with a curet as two primary anchor holes were made to increase rotational stability of the socket. Macroscopically, osteoarthritis of the femoral head was noted. Histological analysis of specimens from the capsule and the synovial membrane showed degenerative fibrosis with no inflammatory cells. Postoperative treatment included anticoagulation therapy, systemic antibiotics, and nonsteroidal anti-inflammatory drugs (ketoprofen, 100 mg per day for five days) to prevent heterotopic ossification. Passive-motion exercises for the involved joint were begun immediately postoperatively. The patient was free to walk with two supports and to bear partial weight on the involved limb after three days. The postoperative course was uneventful, with union of the greater trochanter occurring six weeks after hip replacement, allowing total weight-bearing. The hip was graded as excellent according to the Merle d'Aubignç?¼12 hip score, meaning that there was no pain, no limp, and at least 130°ree; of flexion. At one year after the hip replacement, an asymptomatic lytic lesion was observed distal to the tip of the stem of the femoral prosthesis ( Fig. 3-A ). The functional outcome continued to be satisfactory despite the extension of the osteolytic lesion into the proximal and distal parts of the femur ( Fig. 3-B ). The stem had not migrated, and wear of the acetabular component was undetectable on plain radiographs of the pelvis. Laboratory tests, including determination of the C-reactive protein level and the erythrocyte sedimentation rate, showed no sign of infection. The level of serum alkaline phosphatase was twice the normal level. A bone scan performed three years postoperatively with technetium 99-labeled methylene diphosphonate showed intense tracer uptake in the left hemipelvis and in the medial aspect of the diaphysis of the left femur around the tip of the stem ( Fig. 4 ). No other sites of involvement were apparent.
To avoid a fatigue fracture of the femur distal to the stem, the total hip arthroplasty was revised through a transtrochanteric approach in 1996. The acetabular component was left in place, as it showed no sign of macroscopic wear and was not loose. The femoral stem, which was judged to be stable on intraoperative examination, was removed. With the patient's informed consent (which had been obtained prior to the procedure), multiple specimens were taken from the distal part of the femur, including the bone plug. The biopsy specimens were fixed in 10% formalin and were embedded in paraffin. Sections, 5 µm thick, were obtained and surface-stained with hematoxylin and eosin for standard and polarized light microscopy. A long-stemmed femoral component was used to bypass the osteolytic region of the middle of the diaphysis.
There was no evidence of deep infection, and all of the cultures remained sterile.
Histological analysis of the specimens revealed findings typical of Paget's disease of bone 13 , with a mosaic pattern associated with multinucleated osteoclasts ( Figs. 5-A and 5-B ). No birefringent particles suggesting polyethylene wear debris were observed under polarized light microscopy. The patient underwent treatment for Paget's disease, including administration of calcitonin (100 IU, three times per week) for two months. However, the level of alkaline phosphatase remained at 1.78 times the normal level. Although no further extension of the osteolytic lesion was noted, there was no radiographic bone-remodeling indicating new-bone formation. Bisphosphonate treatment was therefore started, with administration of etidronate at a dose of 400 mg per day for five months. The alkaline phosphatase level decreased progressively and remained normal thereafter. The lytic lesion of the femur was progressively replaced with new bone ( Figs. 6-A and 6-B ). The hip continued to be graded as excellent five years after the femoral revision, at which time Paget's disease had not appeared in any other location.
The clinical signs and symptoms of Paget's disease of bone were first described in 1877 14 , but despite extensive research the etiology of the disease remains unknown. A number of hypotheses have been advocated, including a genetic autosomal dominant component, involvement of human leukocyte antigen (HLA-DQw-1) 15 , and ethnic and geographic factors 16,17 . Moreover, a number of studies have supported a viral etiology 3,18 . Indeed, it has been proposed that a localized disorder of bone-remodeling occurs as a result of a viral infection of osteoclasts in pagetic bone. Inclusions similar to nucleocapsids have been described in the nuclei and cytoplasm of osteoclasts at pagetic sites but not in nonpagetic osteoclasts from the same patients 19,20 . The virus-like particles are similar to those of the paramyxovirus family 3 . However, debate has continued as to whether the alleged virus is a respiratory syncytial virus, measles virus, or canine distemper virus 3,21-23 . A current unifying hypothesis has suggested that the functionally hyperactive osteoclasts in pagetic bone are a product of a virus-mediated increase in cell fusion between osteoclasts and osteoclast progenitor cells that have migrated to pagetic sites. Many investigators have claimed that a common viral infection, in a genetically susceptible host, predisposes to an osteoclastic lesion that is expressed in adulthood as the abnormality that produces Paget's disease 1,13,24-26 .
In the case of our patient, it is highly probable that the occurrence of the pagetic femoral lesion in a previously uninvolved bone was related to the use of autogenous pagetic bone graft from the pelvis. This hypothesis is supported by the fact that in most instances, the sites that are affected with Paget's disease when the diagnosis is made are the only ones that show radiographic changes over time 13,26,27 . Although progression of the disease within a given bone may occur, the sudden appearance of new sites of involvement years after the initial diagnosis is uncommon. Moreover, extensive analysis of the dynamic radiographic patterns of Paget's disease of bone 28 has shown that, in long bones, Paget's disease usually begins at one end of the involved bone and subsequently extends toward the other end. Therefore, although the possibility of a suddenly appearing new pagetic site in the mid-shaft of the femur cannot be eliminated, this hypothesis is not likely.
To the best of our knowledge, transfer of Paget's disease through autograft bone has never been described in the literature as a complication of total hip arthroplasty. O'Driscoll and Hastings 29 reported a case of extension of monostotic Paget's disease from the femur to the uninvolved tibia after arthrodesis of the knee. Cardaci 30 described a case of translocation of Paget's disease to the tibia from an autogenous bone graft retrieved from the iliac crest. Cameron et al. 31 recently reported a case of strut allograft invasion by Paget's disease of the host femur.
To the best of our knowledge, transfer of Paget's disease has not been reported in association with the use of deep-frozen or irradiated allograft bone, although pathological lesions other than osteoarthritis, including Paget's disease, can be present in osteoarthritic femoral heads that are considered suitable for bone-bank donation 32 .
The case of our patient indicates that orthopaedic surgeons should be aware of the risk of transfer of Paget's disease associated with the use of pagetic autograft bone. Moreover, early postoperative osteolysis in a patient with Paget's disease who undergoes total hip arthroplasty could result from transfer of Paget's disease rather than from wear debris.