AGGRESSIVE GRANULOMATOUS LESIONS AFTER HIP ARTHROPLASTY KAJ TALLROTH, ANTTI ESKOLA, SEPPO SANTAVIRTA, YRJ T. KONTTINEN, T. SAM LINDHOLM From the Orthopaedic Hospitalofthe Invalid Foundation, Helsinki We reviewed 19 patients who presented with aggressive granulomatosis around the femoral stem after hip replacement. All had experienced stress pain and had required revision arthroplasty on average 8.8 years after the primary operation. Fifteen patients were men and four were women; none had rheumatoid arthritis. One patient had an uncemented Moore hemiprosthesis; the others all had cemented total hip replacements. When first detected, the granulomatous lesions were multifocal in 13 patients. The first granuloma was in the region of the lesser trochanter in 10, and near the tip of the stem in only two. Speed of growth varied but on average there was doubling of the area on anteroposterior films in 2.2 years (range 6 months to 4.6 years). Aggressive granulomatous lesions in replaced hips are a distinct condition, different from simple loosening or infection; the lesions may grow rapidly, so revision surgery is indicated soon after diagnosis. Recently a number of patients have been reported with localised aggressive resorption of bone around the cemented femoral component of a total hip replacement. In the first report, Harris et a! (1976) reported four such patients with tumour-like bone resorption after hip replacement, and described this condition as a benign, non-inflammatory, adverse tissue reaction. The condition usually leads to loosening of the prosthesis, and other authors have reported radiographic and histological investigations (Remus et a! 1985; Scott, Riley and Dorfman 1985; Jasty et a! 1986; Griffiths, Burke and Bonfiglio 1987). Aggressive granulomatous lesions after hip arthroplasty appear to be a separate entity which can be differentiated from simple implant loosening. Our study K. Tallroth, MD, Head of Radiology Department A. Eskola, MD, Research Fellow S. Santavirta, MD, Assistant Professor T. S. Lindholm, MD, Senior Orthopaedic Surgeon The Orthopaedic Hospital of the Invalid Foundation, Tenholantie 10, SF-00280 Helsinki, Finland. Y. T. Konttinen, MD, Senior Investigator The Fourth Department of Medicine, University Central Hospital, Unioninkatu 38, SF-00170 Helsinki, Finland Correspondence should be sent to Dr S. Santavirta at W. Aaltosentie 9, SF-00570 Helsinki, Finland. 1989 British Editorial Society of Bone and Joint Surgery 0301-620X/89/4 1 58 $2.00 J Bone Joint Surg [Br] 1 989 :7 1 -B :71-5. aimed to determine and analyse the natural history and radiographic outcome. MATERIALS AND METHODS We report 19 patients who required revision of a hip replacement for aggressive granulomatosis between 1982 and 1986 at the Orthopaedic Hospital of the Invalid Foundation, Helsinki. During that period 417 patients required revision arthroplasty for all causes. We defined aggressive granulomatosis as the radiographic appearance of large, focal and usually ovoid lytic areas around the prosthesis in the definite absence of infection. The lytic areas do not correspond in outline to the general shape of the cement around the prosthesis and sometimes appear to grow rapidly. Ofthe patients, 15 were men and four women. Their mean age at the time of primary arthroplasty was 56.2 years (range 36 to 76). Eighteen of the patients had the granulomatosis after primary hip surgery and one after a revision arthroplasty. In 15, the hip arthroplasty had been for primary osteoarthritis, in two for congenital acetabular dysplasia, in one for old tuberculosis and in one after a traumatic fracture of the femoral neck. This latter patient had an uncemented Moore hemiprosthesis, the other 18 had had cemented total hip replacements of various types. The cement around the femoral component had been inserted digitally, without an intramedullary VOL. 71-B, No. 4. AUGUST 1989 571
572 K. TALLROTH, A. ESKOLA, S. SANTAVIRTA, Y. T. KONTTINEN, T. S. LINDHOLM Fig. 2 Fig. 1 Figure 1 - Diagram of the zones around the stem of a prosthesis. In 10 patients, granulomas were first detected in zone 7 around the lesser trochanter and in two around the tip of the stem. Figure 2 - Radiograph of a 70-year-old man two years after total hip replacement for primary osteoarthritis. There are ovoid granulomas in all zones, replacing 42% of the planimetric area of bone around the cemented prosthesis. This is an example of fast granulomatous growth. Fig. 3 Fig. 4 Fig. S Figure 3 - A 70-year-woman six weeks after a cemented total hip replacement for primary osteoarthritis. Figure 4 - Two years later, the hip was painful under stress and radiographs showed granulomatous lesions at lesser trochanter level and medial to the mid part of the stem. Figure 5 - At three years postoperatively, the granulomas have increased in size and there is a further focus lateral to the mid stem. THE JOURNAL OF BONE AND JOINT SURGERY
AGGRESSIVE GRANULOMATOUS LESIONS AFTER HIP ARTHROPLASTY 573 Fig. 6 Fig.7 Fig. 8 Figure 6 - A 53-year-old man three months postoperatively. Figure 7 - Six years after operation, granulomas are seen at lesser trochanter level and around the distal part of the stem. Figure 8 - At 10 years, these granulomas had become very large. The hip was painful, but at revision, the prosthesis was stable in its new position. record the size of granulomas on anteroposterior films, related to the different zones around the femoral stem (Fig. 1). In five cases the granuloma was examined histopathologically. RESULTS Fig. 9 Figure 9 - Radiograph 12 years after insertion of an uncemented Moore prosthesis. At revision ovoid granulomas were found. plug. The first postoperative radiograph in each case had shown the femoral stem in a technically correct, neutral position. Our radiographic evaluation included radioplanimetry (Planix 5.6, Tamaya Technics Inc, Japan) to The first clinical sign of a granuloma was usually stress pain, which led to radiographic examination and detection of the lesions around the stem of the prosthesis (Figs 2 to 9). In 1 3 patients, the granulomatous lesions were multifocal at first diagnosis, 12 had lesions around the upper stem (zones 1, 2, 6 and 7 in Figure 1); 10 of these appeared to have started in the region of the lesser trochanter (zone 7). There were granulomas around the lower part of the stem in 1 1 and around the tip in two. On the first radiographs which showed the granulomas, the planimetric area of the lesions around the upper stem (zones 1, 2, 6 and 7) averaged 5.4 cm2 (range 1.7 to 18.6), and around the lower stem 4.8 cm2 (range 1.1 to 16.7). For the whole proximal femur the area averaged 9. 1 cm2 (range 2.8 to 22.5 cm2). Revision arthroplasty of these hips was performed on average 8.8 years (range 9 to 14) after the primary hip operation and on average 14 months (range 4 to 30) after aggressive granulomatosis had first been diagnosed. During the period between diagnosis and revision, one patient had a spontaneous femoral fracture around the mid part of the femoral stem. This patient was a 6 1-yearold woman who had had a Lubinus-type prosthesis inserted for primary osteoarthritis 1 3 years earlier. VOL. 71-B, No. 4, AUGUST 1989
574 K. TALLROTH, A. ESKOLA, S. SANTAVIRTA, Y. T. KONTTINEN, 1. S. LINDHOLM In some cases, the granulomas grew rapidly in size; in the 14 patients with adequate radiographs radioplanimetry showed doubling of the area of the lesions in an average of 2.2 years (range 6 months to 4.6 years) (Fig. 10). Eight ofthe 10 patients who had first presented with granulomas in the region of the lesser trochanter only had massive multifocal granulomatosis by the time of revision. Nine patients showed measurable sinking of the stem (average 9 mm, range 3 to 23 mm) and five patients had some varus tilting of the prosthesis. One patient also had the radiographic appearance of granulomas around the acetabular component. At the revision operations, 14 stems were found to be loose ; the others were judged to be firm. All histopathological specimens showed collagen deposition, with a large number of histiocytes and giant cells containing microscopic cement particles. None of the patients showed elevation of ESR or CRP levels, and all intra-operative bacterial cultures were negative. cause of this histiocytic response in certain patients with technically well-implanted and even firmly fixed prostheses is still unclear. Wear of cement caused by micromotion, and patient-specific hypersensitivity are hypothetical trigger mechanisms. The synovium-like biomembrane around the cement is known to have the capacity to produce prostaglandin E2 and col!agenase; these may mediate the resorption of bone (Goldring et a! 1983). 15 Planimetry 2 cm 12 9 DISCUSSION 6 Fewer than SO cases of aggressive granulomatosis around cemented hip prosthesis have been reported (Harris et a! 1976; Bell et a! 1983; Remus et a! 1985; Scott et a! 1985; Jasty et al 1986; Griffiths et a! 1987). These accounts confirm that in some patients aggressive granulomatous lesions may lead to rapid bone lysis around apparently stable cemented arthroplasties in the absence of sepsis or malignant disease. On the basis of this review of the literature, these lesions appear to be rare complications of hip replacement. We found, however, that 4.6% of our revision arthrop!asties of the hip had radiographic evidence of these lesions. One of our main findings is that aggressive granulomatosis should be recognised as an established entity, which differs in character from localised bone resorption caused by infection or mechanical failure (Carlsson, Genii and Linder 1983). The cause ofaggressive granulomatosis is debatable. Five of our patients still had firm fixation of the femora! stem at revision, arguing against mechanical causes. Fragmentation of the cement mantle and failure of implant fixation have been blamed (Harris et al 1976), but our series does not confirm this, especially as one of our patients had a cementless Moore prosthesis. Charn!ey (1979) considered that acrylic cement was relatively inert, causing little tissue response as long as there is sound fixation. It seems possible, however, that some patients may develop a foreign body reaction to acrylic cement, and possibly to a metallic implant. The reported histopathologica! appearance ofthe granulomas has been uniform : histiocytes ofvarying size, numerous giant cells, and local areas of eosinophilic necrotic debris including microscopic particles of cement (Harris et a! 1976; Bell et a! 1983; Remus et a! 1985; Griffiths et al 1987). The 3 0 16 32 43 64 80 Time Fig. 10 (months) Figure 10 - Diagram showing the speed of growth of granulomas measured by planimetry of anteroposterior radiographs in 14 patients. Regression lines (produced by the BMDP 6D-program) indicate the general trends. The heavy line shows the total area ; the dotted line represents proximal granulomas (zones 1, 2, 6 and 7), and the thin line the distal ones (zones 3, 4, and 5). Most previous studies are based on few cases (Harris et a! 1976; Remus et a! 1985 ; Jasty et a! 1986; Griffiths et a! 1987) and show an equal incidence in men and women. We found a clear predominance of men, 1 5 of 19 patients. In addition, though about half of our primary replacements are in patients with rheumatoid arthritis, we have, to date, seen no evidence ofaggressive granulomatosis in these patients. The speed of the growth of granulomas varies: previous authors have described revision arthroplasties from about 2 to 6 years after the primary operation. In our series the speed of growth of the lesions was unpredictable, some doubling their size in a few months, while others grew very slowly. All our patients had pain as the first clinical sign but in some at that time the granulomas were already fairly large. Lesions may be multifocal, and the region of the lesser trochanter appeared to be a common site. Since the lesions sometimes grow rapidly and there may be risk of spontaneous fracture, patients should be THE JOURNAL OF BONE AND JOINT SURGERY
AGGRESSIVE GRAN ULOMATOUS LESIONS AFTER HIP ARTHROPLASTY 575 considered for revision arthrop!asty soon after the diagnosis has been made. At present we treat patients with aggressive granulomatous lesions by revision arthroplasty, using uncemented titanium prostheses and bone grafting. Financial support for this study was received from the Finska Lflkaresallskapet, the Tampere Tuberculosis Foundation and the Paulo Foundation. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. REFERENCES Bell RS, Ha eri GB, Goodman SB, Fornasier VL. Case report 246: osteolysis of the ilium associated with a loose acetabular cup following total hip arthroplasty, secondary to foreign body reaction to polyethylene and methyl methacrylate. Skeletal Radiol 1983; 10 :20 1-4. Carlsson As, Gentz C-F, Linder L. Localized bone resorption in the femur in mechanical failure of cemented total hip arthroplasties. Acta Orthop Scand 1983;54:396-402. Charnley J. Low friction arthrop/asty of the hip : theory and practice. Berlin etc : Springer-Verlag, 1979. Goldring SR, Schiller AL, Roelke M, Rourke CM, O Neill DA, Harris WH. The synovial-like membrane at the bone-cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg [Am] 1983 ;65-A :575-84. Griffiths HJ, Burke J, Bonfiglio TA. Granulomatous pseudotumors in totaljoint replacement. Skeletal Radio/ 1987 ;16 :146-52. Harris WH, Schiller AL, Scholler JM, Freiberg RA, Scott R. Extensive localized bone resorption in the femur following total hip replacement. J Bone Joint Surg [Am] 1976:58-A :612-8. Jasty MJ, Floyd WE III, Schiller AL, Goldring SR, Harris WH. Localized osteolysis in stable, non-septic total hip replacement. J Bone Joint Surg [Am] 1986;68-A :912-9. Reinus WR, Gilula LA, Kyriakos M, Kuhiman RE. Histiocytic reaction to hip arthroplasty. Radiology 1985;l55:31S-8. Scott WW Jr, Riley LH Jr, Dorfman HD. Focal lytic lesions associated with femoral stem loosening in total hip prosthesis. AJR 1985 ; 144:977-82. VOL. 71-B, No. 4, AUGUST 1989