Osteonecrosis of the knee: differences among idiopathic and secondary types

Transcription

1 Rheumatology 000;39: Osteonecrosis of the knee: differences among idiopathic and secondary types J. Narváez, J. A. Narváez, J. Rodriguez-Moreno and D. Roig-Escofet Department of Rheumatology, Hospital Príncipes de España, Ciudad Sanitaria y Universitaria de Bellvitge and Department of Computed Tomography and Magnetic Resonance Imaging, Institut de Diagnostic per la Imatge, Hospital Duran i Reynals, Ciudad Sanitaria y Universitaria de Bellvitge, Barcelona, Spain Abstract Objective. To describe the clinical and imaging features of patients with osteonecrosis of the knee, emphasizing the differences among idiopathic and secondary types. Methods. A retrospective chart review of 37 consecutive patients (4 knees) with osteonecrosis of the knee confirmed by bone scintigraphy and/or magnetic resonance imaging (MRI), and a comparison of idiopathic and secondary types of osteonecrosis. Results. Twenty-four patients had idiopathic osteonecrosis, and in 3 patients one or more predisposing factors were identified (secondary osteonecrosis). Idiopathic osteonecrosis of the knee was typically a disease of the elderly, characterized by severe knee pain of sudden onset, unilateral involvement, and restriction of the lesions generally to one femoral condyle or tibial plateau, with predilection for the medial compartment of the joint. Secondary osteonecrosis generally occurred in younger patients and frequently had an insidious onset with mild or vague pain, the lateral compartment of the knee was often involved, and the lesions were generally larger than lesions arising spontaneously; in the great majority of cases they involved the femoral condyles and/or tibial plateaus. Bilateral distribution and multifocal involvement was also seen in these forms. Magnetic resonance imaging was helpful in confirming the suspected diagnosis when conventional radiographs were normal or equivocal, and demonstrated different patterns of abnormalities in idiopathic and secondary types. Conclusion. There are significant differences between idiopathic and secondary osteonecrosis, especially in regard to clinical presentation and the location, extent and MRI appearance of the lesions. These differences are probably due to a difference in the pathogenetic mechanism. KEY WORDS: Osteonecrosis of the knee, Idiopathic osteonecrosis, Secondary osteonecrosis, Magnetic resonance imaging. forms of this condition: the idiopathic, primary or spontaneous ON (ION) of the knee, which appears in patients without risk factors for osteonecrosis, and secondary ON (SON), caused by a number of well- recognized predisposing factors, often acting alone but at times acting in concert. The first of these forms has been characterized as a distinct entity since its first description by Ahlbäck et al. in 968 [ 9]. The second form is poorly documented; few reports have been devoted to SON of the knee [0 3], in contrast with the great number of published papers that have focused on SON of the hip. To describe the clinical and imaging features of patients with ION and SON of the knee, we present a clinical review of a series of 37 patients (4 knees), emphasizing the differences among the two types of osteonecrosis. Osteonecrosis (ON) is a relatively common condition that usually results from circulatory impairment of an area of bone, leading to the ischaemic death of the cellular constituents of the bone and marrow. There are a few specific areas of the skeleton for which this condition seems to have a predilection. These include the femoral and humeral heads, the femoral condyles, the proximal tibia and some of the small bones of the foot and ankle and of the wrist. The area most commonly affected is the femoral head, followed by the humeral head and the knee. In the knee, it may involve either the distal femur or the proximal tibia, and one or both condyles or plateaus. There seem to be two distinct Submitted 7 May 999; revised version accepted 3 March 000. Correspondence to: F. J. Narváez Garcia, C/Llobregat, no a, Hospitalet de Llobregat 08904, Barcelona, Spain British Society for Rheumatology

2 Idiopathic and secondary osteonecrosis of the knee 983 Patients and methods for the presence of focal deformities of the cartilage and subchondral cortical bone, tearing or degeneration of Our study was a retrospective chart analysis of 37 the menisci, and joint effusion. patients ( 4 knees) with ON of the knee, confirmed by Twenty-five patients had a minimum follow-up of bone scintigraphy and/or magnetic resonance imaging 6 months after the initial diagnosis was made; the (MRI). The patients comprised 5 women and men remaining were lost to follow-up. In these patients, and the mean age at diagnosis was 60 ± 3 yr (range the evolution of ON was monitored with periodical 9 85 yr). According to the presence of risk factors for conventional radiographs and clinical examinations. osteonecrosis, the patients were grouped into two Follow-up MRI examination was performed in only categories: ION (4 cases) and SON ( 3 cases). Four four patients. patients with SON had bilateral involvement. The medical charts of all patients were reviewed Statistical analysis comprehensively to obtain clinical, laboratory, imaging Continuous data were described as mean ± S.D. and and disease evolution data. Symptoms were usually categorical variables as percentages. Patients were classidescribed carefully and adequately in the medical fied in idiopathic and secondary groups and their clinical records. All patients had conventional radiographs of manifestations were compared. Comparisons were made the knees taken at the initial visit and at follow-ups. using Student s t-test for independent continuous vari- The necrotic lesion was staged according to the classiables or the Mann Whitney U-test when the assumption fication of Aglietti et al. [6]: stage = normal; stage of normality was not realized. To analyse categorical = flattening of the affected weight-bearing part of the data, we performed the x test with Yates correction or condyle; stage 3 = a subchondral lesion comprising an Fisher s exact test when the expected values were less area of radiolucency of variable size and depth surthan 5. Statistical significance was defined as P rounded proximally and distally by some sclerosis; stage We also calculated the wr coefficient as a measure of 4 = the radiolucent area is surrounded by a sclerotic the magnitude of the association and the relative risk halo and the subchondral bone has collapsed and is ( RR) between groups for some variables, together with visible as a calcified plate; stage 5 = secondary degenerthe 95% confidence interval (CI). ative changes with osteophyte formation and subchondral sclerosis on both femur and tibia, and some erosion. The degree of osteoarthritis was graded radiologically Results as follows: none = no visible changes; minimal = spurring of the tibial spine, mild joint space narrowing, Results are summarized in Table. some flattening of the femoral condyles, and small marginal osteophytes; moderate = appreciable joint Aetiology space narrowing, subchondral sclerosis and larger osteostudied; ION was observed in 4 (64.9%) of the 37 patients phytes; severe = joint space obliteration with deformity in 3 (35.%) of them one or more aetiological or even destruction of the articular surface, and occatransplantation factors were identified: alcoholism (two cases), renal sionally lateral subluxation. Radiographs were reviewed (two cases were receiving treatment with to establish the degree of ON and osteoarthritis in a cyclosporin A and 5 mg prednisone per day), connective blinded manner by the first two authors (J.N. and tissue disorders (two cases of systemic lupus eryth- J.A.N.); in cases of disagreement a consensus opinion ematosus and two cases of rheumatoid arthritis, all of was obtained. them treated with low doses of glucocorticosteroids), Confirmatory [99mTc]methylene diphosphonate bone thalassaemia (one case), leukaemia (one case treated scintigraphy was performed in 9 patients. MRI examplantation), with chemotherapy and allogeneic bone marrow trans- ination was carried out in 34 patients, four of whom and systemic steroid administration (one had bilateral involvement, and a total of 38 knees were case of myasthenia gravis, one case of giant cell arteritis studied. MRI images were reviewed in a blinded manner and one case of eosinophilic fasciitis). by one radiologist (J.A.N.), who did not know if the patient was classified as having ION or SON (he Location and extent of the lesions described only the appearance of the lesion without In 33 ( 89.%) patients the lesions were unilateral (the classifying the patients as having ION or SON). The right knee was affected in 3 cases and the left knee in diagnosis of osteonecrosis by the use of MRI was based 0); the remaining four patients (0.8%), all of them mainly on two patterns: (i) a focal subcortical loss of with secondary forms, had bilateral disease. signal intensity on T - and T -weighted images (Fig. ), The most frequent location was the medial femoral and (ii) a focal lesion of variable signal intensity sur- condyle, which was affected, alone or in combination rounded by a demarcation rim. This demarcation rim with involvement of the lateral condyle and/or the appears as a low signal margin on T -weighted images, proximal tibial, in 36/4 (87.8%) knees (Table ). With and a double halo ( high-signal inner border inside a regard to the aetiology and extent ( Table 3), all the low-signal peripheral rim) on T -weighted images ION were unilateral, with lesions restricted to one ( Fig. ). These two patterns may be associated with condyle or tibial plateau in 3/4 (95.8%) cases. The bone marrow oedema. MRI images were also assessed medial side of the joint was most commonly involved,

3 984 J. Narváez et al. (A) (B) FIG.. (A) Sagittal T -weighted MRI showing a low-signal subchondral lesion (arrows) of linear morphology. This lesion is associated with ill-defined bone marrow oedema. (B) On T -weighted sequences the subchondral lesion also shows a low signal (arrowheads) and the ill-defined bone marrow oedema shows a high signal. Joint effusion (arrows) is also well demonstrated, with a homogeneous high signal. with predilection for the weight-bearing surface of the femoral condyle. Involvement of the lateral femoral femoral condyle (0/4 cases). condyle occurred in up to 75% of knees, alone or in SON was unilateral in nine (69.%) patients and combination with changes in the medial femoral condyle bilateral in four (30.8%). Generally, the SON lesions and/or tibial plateaus. The four patients with bilateral were of greater size than the idiopathic ones, involving disease had diffuse involvement of both femoral condyles at least one femoral condyle and tibial plateau or both and tibial plateaus in each knee; in all of them, the femoral condyles (with or without concomitant involve- symptoms began simultaneously. ment of tibial plateaus) in 5/7 (88.%) knees. Only in Four patients (0.8%), all of them with secondary /7 (.8%) knees were the lesions restricted to one forms, had polytopic ON at the time of diagnosis, with

4 Idiopathic and secondary osteonecrosis of the knee 985 TABLE. Clinical features of 37 patients with ON of the knee Patient Age (yr) Sex Aetiology Uni-/bilateral LFC MFC LTP MTP Polytopic 77 F Idiopathic Unilateral + 69 F Idiopathic Unilateral M Alcoholism Unilateral F Corticosteroids Bilateral Humeral head 5 44 F Renal transplant Bilateral Femoral heads 6 67 F Idiopathic Unilateral F Idiopathic Unilateral F SLE Bilateral Femoral heads 9 60 M Idiopathic Unilateral F RA Unilateral M Alcoholism Unilateral M Idiopathic Unilateral F Idiopathic Unilateral F Renal transplant Unilateral M Idiopathic Unilateral M Corticosteroids Unilateral M Idiopathic Unilateral M Idiopathic Unilateral F Thalassaemia Unilateral F Idiopathic Unilateral + 7 F Idiopathic Unilateral + 9 M Leukaemia Bilateral Hips + shoulders 3 48 F SLE Unilateral F Idiopathic Unilateral F Idiopathic Unilateral F RA Unilateral F Idiopathic Unilateral F Idiopathic Unilateral F Idiopathic Unilateral M Corticosteroids Unilateral F Idiopathic Unilateral M Idiopathic Unilateral F Idiopathic Unilateral F Idiopathic Unilateral M Idiopathic Unilateral F Idiopathic Unilateral F Idiopathic Unilateral LFC, lateral femoral condyle; MFC, medial femoral condyle; LTP, lateral tibial plateau; MTP, medial tibial plateau; SLE, systemic lupus erythematosus; RA, rheumatoid arthritis. simultaneous involvement of the right humeral head in onset of well-localized severe pain occurred in only 5/3 patient 4, both femoral heads in patients 5 and 8, and (38.5%) patients; in the remaining eight patients (6.5%) both hips and both shoulders in patient. the onset of symptoms was gradual, with mild and Clinical features diffuse pain (Table 4). These differences may explain the longer delay until diagnosis observed in the group with The disease was more frequent in women (5 women and men). The mean age at onset in the group with ION was 66 ± 8 yr, whereas in the group with SON it TABLE. Extent of the ON lesions was 49 ± 4 yr. The interval between the onset of symp- No. of patients: No. of knees: toms and diagnosis was 05 ± 0 days (range (%) days). Unilateral 33 Pain was the dominant symptom in all cases. It MFC 0 0 (48.8) appeared suddenly in 7/37 (73%) patients ( remem- LFC (4.9) bered the exact moment it occurred) and gradually in MFC+LFC (4.9) MTP 3 3 (7.3) 0/37 ( 7%). None of the patients had a history of MFC+MTP 4 4 (9.7) minor knee trauma. In the ION group, the onset of MFC+LFC+MTP+LTP (4.9) symptoms was acute in /4 (9.6%) patients and Bilateral 4 gradual in /4 (8.4%). Pain was usually severe MFC+LFC+MTP+LTP 4 8 (9.5) and localized, occurring predominantly on active motion The table shows the number (percentage) of patients or knees or weight-bearing and causing significant functional affected. impairment in all cases; pain at rest was seen in 58% of MFC, medial femoral condyle; LFC, lateral femoral condyle; MTP, patients. In the SON group, the typical history of sudden medial tibial plateau; LTP, lateral tibial plateau.

5 986 J. Narváez et al. TABLE 3. Aetiology and extent of the lesions (46.3%), minimal changes were seen in (9.3%), n moderate changes in six (4.6%) and severe changes in four (9.8%). Radiological evidence of osteoarthritis was ION, n = 4 more frequent in the ION group. In these patients, the Restricted to one femoral condyle or tibial plateau initial radiographs revealed arthrosis in 9/4 (79.%) MFC 0 knees: /4 (45.8%) presented minimal changes, 6/4 MTP 3 Simultaneous involvement of both femoral condyles (5%) moderate changes and /4 (8.3%) severe changes. SON, n = 3/7 knees In SON, the initial radiographs revealed osteoarthritis Unilateral: 9 in only 3/7 (7.6%) knees: in one knee (5.8%) there Restricted to LFC were minimal changes and in two (.7%) there were Simultaneous involvement of both femoral condyles Involvement of MFC and MTP 4 severe changes. Diffuse involvement of both femoral condyles At the follow-up examination, the Aglietti stage had and plateaus not changed in 6 of the 0 knees, but had increased in Bilateral: 4 four despite the conservative treatment. None of the Diffuse involvement of both femoral condyles 8 and plateaus patients with initial normal radiographs developed osteonecrotic changes. MFC, medial femoral condyle; LFC, lateral femoral condyle; MTP, Scintigraphy. Confirmatory bone scintigraphy was medial tibial plateau. carried out in 9 patients. It was very useful in conn = number of patients. firming the suspected diagnosis in patients without radiological evidence of ON at the time of presentation, and SON ( ± 3 versus 96 ± 6 days). In the cases of in detecting multiple ON in patients with polytopic steroid-related ON, the interval between steroid intake involvement. We did not find differences in the bone and the appearance of symptoms ranged from 9 months scan pattern between ION and SON. to yr. MRI. The mean interval from the onset of symptoms Physical examination showed tenderness over the to MRI examination was 5.6 ± 4 months. In patients involved area in 00% of patients, small or moderate with ION, the most frequent MRI pattern was a subcorteffusion in 80.5% of knees (the synovial fluid was ical focal loss of signal, which was observed in 95.6% examined in 8 patients and was non-inflammatory in of knees on T -weighted images and in 69.5% of knees all of them) and flexion contracture or mild restricted on T -weighted images (Fig. ). The demarcation rim motion in 0%. (double-halo sign) was not seen in any patient. Imaging modalities Additional frequent findings in ION patients were homo- Radiography. The mean duration from the onset of lateral meniscal lesion (degeneration and/or tearing), symptoms to the first radiographic examination was which was observed in all knees but one (95.6%), and 5 ± 7 weeks. Initially, plain roentgenograms showed focal deformity of the subchondral plate, which was evidence of ON in only 0 (48.8%) of the 4 knees, and seen in 65% of cases. On the contrary, in patients with were normal in the remaining (5.%). Aglietti s SON the demarcation rim was observed in 73.3% of stage II was observed in nine knees (%), stage III in knees ( Fig. ). Homolateral meniscal lesion was found six (4.6%), stage IV in two (4.9%) and stage V in three in 33% of cases, and focal deformity of the subchondral (7.3%). The initial radiographs were normal in 9 knees bone plate was seen in 0% of knees. The extent of bone TABLE 4. Differences between ION and SON ION SON No. of patients, 4 No. of patients, 3 No. of knees, 4 No. of knees, 7 Age (yr) 66 ± 8 49± 5 P < 0.00 Diagnostic delay (days) 96 ± 6 ± 3 n.s. (P > 0.50) Onset Sudden 9.6% 38.5% (/7) P < 0.005, wr = 0.5 Gradual 8.4% 6.5% (8/3) RR = 7.38a (95% CI.83, 9.77) Lesion of medial meniscus 9.6% 3.5% P < 0.00, wr = 0.70 RR = 3.89 (95% CI.63, 9.4) Extentc Limited 95.8%.8% (/7) P < 0.00, wr = 0.85 Extensive 4.% 88.% (5/7) RR =.8b (95% CI 4.0, 45.4) Bilateral involvement 0% 30.7% (4/3) P < 0.05, wr = 0.39 Polytopic involvement 0% 30.7% (4/3) arisk of gradual onset. brisk of extensive lesions. cthe lesion was categorized as limited in cases restricted to one femoral condyle or tibial plateau; if the involvement was greater the lesion was considered extensive.

6 Idiopathic and secondary osteonecrosis of the knee 987 necrosis measured on MRI scans at presentation was patients with well-recognized risk factors for ischaemic significantly greater in SON than in ION. Evidence of necrosis of bone (SON of the knee); however, its clinical an osteochondral defect at presentation was infrequent and imaging features are not yet fully understood, in both groups (6.6% of cases with ION and 4.3% of mainly because of the limited number of reports devoted cases with SON), and was observed only in cases with to this condition [0 3]. longer diagnostic delay. Joint effusion was seen in 80.5% Despite the comparatively small number of patients of knees. included in our study, which may not have been large Outcome enough to detect minor differences, we found significant differences between the idiopathic and secondary forms Twelve patients were lost to follow-up. In the remaining of osteonecrosis, especially with regard to its clinical 5 patients (four with bilateral involvement), the average presentation and the location and extent of the lesions follow-up period was 8 ± 3 months (range 6 60 ( Table 4) and their MRI appearance. These results months) after the initial examination. Initially, surgical indicate that two distinct forms of the disease can be treatment (arthroplasty) was performed in four patients identified. with Aglietti s stage IV or V; an operation was recom- ION occurs in older individuals, usually after the mended for other knee but it was refused by the patient. sixth decade of life [, 4 8]. Women are affected three In addition, one case with steroid-related ON was treated times more frequently than men [, 4, 6, 8]. The pain is with core decompression. The remaining 8 patients ( usually characterized by sudden onset (frequently the knees) were managed conservatively ( limitation of patient could remember the exact moment or the activity weight-bearing, and analgesics). Progression of the during which the symptoms started) and is initially symptoms/disease led to a knee arthroplasty in six of severe, causing significant functional impairment; pain these knees within yr. In the other patients is frequently present at rest [, 3 8]. Unilateral involvesymptoms resolved or improved. ment predominates over bilateral involvement. The con- At the time of last follow-up, 6/4 (5%) knees in the dition is generally restricted to one femoral condyle or ION group and 4/7 (3.5%) in the SON group required tibial plateau. The medial side of the joint is the side arthroplasty, although two of the patients with SON that is most commonly involved, and the lesion occurs also had severe articular damage at presentation due to more often in the femoral condyle than in the tibial erosive rheumatoid arthritis. plateau [4, 6 8]. Less frequently, the lateral condyle is affected [6 8]. In SON, the age and sex distribution is Discussion dependent upon the primary disease, although it generally affects a younger age group. The onset of symptoms Whereas ION of the knee is an accepted entity, it has is gradual in more than 60% of patients, and often mild become clear that a similar condition can develop in or vague pain occurs only with active motion or (A) (B) FIG.. (A) Sagittal T -weighted MRI demonstrating a large subchondral lesion (arrowheads) isointense with normal fatty marrow and surrounded by a serpiginous low-signal band. (B) On T -weighted sequences this lesion (arrows) is demarcated by a band of high signal intensity inside a hypointense margin the double halo sign.

7 988 J. Narváez et al. weight-bearing. Generally, the lesions of SON are larger osteoporosis is frequent, and its characteristic predilection than those of ION; fewer than 5% of cases show for the area of maximum weight-bearing of the changes of a limited extent. Both knees are affected in medial femoral condyle or tibial plateau, strongly suggest 30 80% of patients [6, 0,, 3] and involvement of a traumatic lesion as the most likely underlying the lateral femoral condyle occurs in up to 60% of cause of ION. In this respect, more authors [3, 7, ] patients [6, 0, ], alone or in combination with are of the opinion that the initial insult may be stress simultaneous involvement of the medial femoral condyle microfractures occurring within the weak subchondral [0, ]. Proximal tibial involvement is less common. bone plate, caused by the repeated impact of the articular This is contrast to the almost invariably unilateral surfaces. Additional support for this assertion comes distribution of ION and its predilection for the weightbearing from the similarity in MRI appearance between ION segment of the medial femoral condyle. The and stress fractures, which appear as band-like areas simultaneous presence of ON in other locations ( hips of low signal intensity in the intramedullary space on and shoulders) is not rare [0, ]. T - and T -weighted images [4, 5]. Our results were similar to those reported in other Moreover, meniscal tears have been reported in associstudies. However, contrary to the earlier series of ION, ation with this condition, and a prominent role of in which there were appreciable radiographic osteonecrotic meniscal injury in the pathogenesis of ION has also changes in 76 00% of patients at the time of been proposed [6]. It is well established that the injured diagnosis [4, 6, 7, 4], we found a higher proportion of meniscus has a diminished load-bearing capacity. The necrotic lesions in the early stages, with normal initial repeated impact of the articular surface against a fragmented radiographs in more than half of our patients. This meniscus during everyday activity could result difference could be explained by the fact that scinti- in microfractures and subsequent vascular insufficiency graphy and MRI allowed us to make an early diagnosis, of the subchondral bone. The appearance of ION about before the development of radiographic abnormalities. the knee after meniscal surgery also supports this pathogenetic We also observed different patterns of MRI abnormalities relationship [7]. The high rate of medial menis- in ION and SON. It is remarkable that the cus lesion (95.6%) observed in our patients with ION is demarcation rim, which is considered specific and pathognomonic in agreement with this hypothesis. However, this cannot for avascular ON [5], was observed in most be the only factor involved in its pathogenesis, because of the patients with SON but in none of those with some patients with ION had no evidence of a torn ION. The demarcation rim or reactive interface separ- meniscus. In this sense, it has been suggested that the ates necrotic bone and marrow from adjacent viable presence of osteoarthritis and/or osteoporosis may also tissue, and shows a low signal intensity on T -weighted be a significant contributing factor in the development of subchondral plate microfractures [3, 9, ]. images and a double-halo appearance on T -weighted images. The double halo consists of a high-signal inner It is commonly recognized that osteoporosis and border inside a low-signal rim surrounding the lesion. osteoarthritis (as a result of violent impact between Histological studies have demonstrated that the low- osseous surfaces not protected by articular cartilage) signal outer rim corresponds to sclerotic bone, whereas can weaken the load-bearing capacity of bone such that the high-intensity inner border probably consists of fracture can result from normal activity. The exact hypervascular granulation tissue resulting from a hyperaemic temporal and causal relationship between ON and response adjacent to thickened trabeculae [6]. osteoarthritis remains controversial, because is difficult In contrast, the most common MRI pattern observed to establish whether ON is simply a predisposing factor in ION was a focal subcortical area of low signal for osteoarthritis or whether the presence of moderate intensity on both T - and T -weighted images. To date, to severe osteoarthritis predisposes to spontaneous there have been few reports on the use of MRI in ON osteonecrosis of bone [7]. In our series, radiological of the knee [ 7 9]. Of interest, similar observations evidence of osteoarthritis at the time of diagnosis was have been reported increasingly in ION of the knee more frequent in the group with ION ( 79%) than in the [8, 0, ]. Among the few reports of MRI in SON, group with SON (7%), although this difference could Chancelier et al. [9] described MRI signal alterations be explained by the older age of the ION patients. In similar to those we saw in a series of patients receiving initial radiographs, only 33% of patients with ION long-term corticosteroid therapy. The cause of these were found to have moderate to severe osteoarthritis apparent differences among idiopathic and secondary in the compartment where osteonecrosis developed types is not clear, although they may be ascribable to a subsequently. difference in the underlying pathogenesis. By contrast, the pathophysiology of SON of the knee The aetiopathogenesis of ION of the knee remains seems to be comparable with that which occurs in the unclear. Two main theories have been proposed: primary femoral head [7]. In SON of the knee, several aetiolo- vascular insufficiency leading to infarction of bone gical factors have been identified, all of them causing, [6, ], and a minor, unrecognized traumatic insult by different mechanisms (embolization, thrombosis, producing microfractures in the subchondral bone and vessel wall injury, marrow encroachment or venous subsequent osteonecrosis [3, 7, 9,, ]. The usual occlusion) [ 5], an interruption of the blood supply. sudden onset of symptoms that characterizes ION, its Interruption of the vascular supply to the subchondral occurrence in middle-aged and elderly patients, in whom bone then leads to death of bone and marrow cells over

8 Idiopathic and secondary osteonecrosis of the knee 989 a larger volume than is observed in ION [7]. This similar. Mont MA, Myers TH, Krackow KA, Hungerford DS. pathogenesis can explain the similar MRI appearance Total knee arthroplasty for corticosteroid associated avas- of the two conditions, with evidence of the demarcation cular necrosis of the knee. Clin Orthop 997;338:4 30. rim in up to 80% of cases of avascular necrosis of the 3. Mont MA, Tomek IM, Hungerford DS. Core decompres- sion for avascular necrosis of the distal femur. Clin Orthop hip [8] and in 73% of our cases of SON of the knee; 997;334:4 30. this value agrees with the results of other authors [9]. 4. Ahuja SC, Bullough PG. Osteonecrosis of the knee: A In summary, although the number of patients included clinicopathological study in twenty-eight patients. J Bone is not large, this study does indicate a sharp contrast Joint Surg 978A;60:9 7. between two different types of ON of the knee, which 5. Froberg PK, Braunstein EM, Buckwalter KA. presumably have different underlying pathogenetic Osteonecrosis, transient osteoporosis, and transient bone mechanisms. The more extensive lesions in patients with marrow edema. Radiol Clin North Am 996; 34: the secondary form of disease and the facts that they 6. Mitchell DG, Joseph PM, Fallon M et al. Chemical shift occur in the lateral compartment and that other joints MR imaging of the femoral head: an in vitro study of may be affected argue for a generalized metabolic normal hips and hips with avascular necrosis. Am J Roentgenol 987;48: aetiology as opposed to some more local mechanical 7. Pollack MS, Dalinka MK, Kressel HY, Lotke PA, Spritzer cause in the idiopathic type, which may or may not be CE. Magnetic resonance imaging in the evaluation of related to pre-existing degenerative changes and/or suspected osteonecrosis of the knee. Skeletal Radiol meniscal injury. 987;6: Björkengren AG, Al-Rowaih A, Lindstrand A, References Wingstrand H, Thorngren KG, Petterson H. Spontaneous osteonecrosis of the knee: value of MR imaging in determining. Ahlbäck S, Bauer GCH, Bohne WH. Spontaneous prognosis. Am J Roentgenol 990:54:33 6. osteonecrosis of the knee. Arthritis Rheum 968;: 9. Chancelier MD, Hélenon O, Page B, Rousselin B, Legendre C, Moreau JF. Ostéonécroses aseptiques des. Muheim G, Bohne WH. Prognosis in spontaneous genoux induites par les corticoïdes. Aspects IRM. J Radiol osteonecrosis of the knee. Investigation by radionuclide 99;73:9 0. scintimetry and radiography. J Bone Joint Surg 970; 0. Lecouvet FE, Van de Berg BC, Maldague BE et al. Early 5B:605. irreversible osteonecrosis versus transient lesions of the 3. Lotcke PA, Ecker ML, Alavi A. Painful knees in older femoral condyle: prognostic value of subchondral bone patients: radionuclide diagnosis of possible osteonecrosis and marrow changes on MR imaging. Am J Roentgenol with spontaneous resolution. J Bone Joint Surg 977;59A: 998;70: Blum GM, Tirman PFJ, Cruess JV. Osseous and cartila- 4. Rozing PM, Insall J, Bohne WH. Spontaneous ginous trauma: In: Mink JH et al., eds. Magnetic resonance osteonecrosis of the knee. J Bone Joint Surg 980:6A: 7. imaging of the knee. nd edn. New York: Raven Press, 5. Houpt JB, Alpert B, Lotem M et al. Spontaneous 993: osteonecrosis of the medial tibial plateau. J Rheumatol. Resnick D, Niwayama G. Osteonecrosis: diagnostic tech- 98;9:8 90. niques, specific situations, and complications. In: Resnick 6. Aglietti P, Insall JN, Buzzi R, Deschamps G. Idiopathic D, Niwayama G, eds. Diagnosis of bone and joint disosteonecrosis of the knee: aetiology, prognosis and treat- orders. nd edn. Philadelphia: W.B. Saunders, 988; ment. J Bone Joint Surg 983;65B: Houpt JB, Pritzker KPH, Alpert B, Greyson ND, Gross 3. Fazzlari NL. Trabecular microfracture. Calcif Tissue Int AE. Natural history of spontaneous osteonecrosis of the 993;53(Suppl. ):S43 6. knee (SONK ): a review. Semin Arthritis Rheum 983; 4. Lee JK, Yao L. Stress fractures: MR imaging. Radiology 3: ;69: Al-Rowaih A, Lindstrand A, Björkengren AG, 5. Anderson MW, Greenspan A. Stress fractures. Radiology Wingstrand H, Thorngren KG. Osteonecrosis of the knee. 996;99:. Diagnosis and outcome in 40 patients. Acta Orthop Scand 6. Norman A, Baker ND. Spontaneous osteonecrosis of the 99;6:9 3. knee and medial meniscal tears. Radiology 978;9: 9. Ecker ML, Lotcke PA. Osteonecrosis of the medial part of the tibial plateau. J Bone Joint Surg 995; 77A: Brahme SK, Fox JM, Ferkel RD, Friedman MJ, Flannigan 0. Sasaki T, Yagi T, Monji J, Yasuda K, Masuda T. Steroidscopic BD, Resnick DL. Osteonecrosis of the knee after arthroinduced osteonecrosis of the femoral condyle: A clinical surgery: diagnosis with MR imaging. Radiology study of eighteen knees in ten patients. J Jpn Orthop 99;78:85 3. Assoc 986;60: Mitchell DG, Rao VM, Dalinka MK et al. Femoral head. Kelman GJ, Williams GW, Colwell CW, Walker RH. avascular necrosis: correlation of MR imaging, radio- Steroid-related osteonecrosis of the knee. Two case reports graphic staging, radionuclide imaging, and clinical findand a literature review. Clin Orthop 990;57:7 6. ings. Radiology 987;6:709 5.