Introduction

Osteonecrosis (ON), also known as avascular necrosis of bone, is a disease with unexplained pathogenesis that disrupts joint structure and causes progressive dysfunction (1). Early recognition of ON is very important in terms of protection of joints (2). The most common symptom of ON is pain that is exacerbated by exercise. ON has many causes and corticosteroid (CS) intake is one of the most important causes (1-3). The relationship between CS dose and ON is contradictory. Although many authors claim that the cumulative CS dose is the most important factor (4-6), others have shown that ON may develop independently of dose (7,8). Vreden et al. (7) showed that even patients receiving CSs at physiological doses for adrenal insufficiency might develop ON. CS-induced ON is most commonly seen in the femoral head and it may also be seen around the knee, shoulder, elbow, hand and ankle joints. Magnetic resonance imaging (MRI) is the most sensitive and useful method for the early diagnosis of ON (1,5,9). However, there is no consensus on whether asymptomatic patients who are on CS therapy should be screened for ON.

Our aim in this study was to investigate the prevalence and imaging findings of knee ON by MRI in patients who were on CS therapy and complaining of arthralgia. We also aimed to demonstrate the diagnostic performance of plain radiography when the MRI is accepted as the gold standard for detection of knee ON and to investigate the prevalence of accompanying femoral head ON in patients who underwent hip MRI.

Methods

This study was approved by institutional Pamukkale Universtiy Ethics Committee (decision no: 60116787-020/8340). Informed consent was waived due to retrospective nature of the study. Patients with a history of knee joint surgery were excluded from the study because they could cause misinterpretation of imaging findings. One hundred and sixty-two knee MRI examinations and X-rays of 136 patients (102 females, mean age: 39.5±14.5 years, range: 16-78 years) with knee pain and history of CS intake due to various chronic diseases were retrospectively evaluated by two radiologists (with an experience of 3 and 20 years) in consensus for the presence of ON. Anteroposterior and lateral X-rays of all patients were obtained. MR images were obtained using a 1.5 Tesla superconducting magnet (GE Signa Excite HD, GE Medical Systems, Milwaukee, Wisconsin, USA) and a dedicated quad knee coil. Knee joint MRI examinations consisted of a sagittal T1-weighted (T1W) spin echo (SE), sagittal proton density-W fast spin-echo (FSE) with fat suppression (FS), axial proton density-W FSE with FS and coronal T2W FSE with FS. Other MRI parameters are shown in Table 1.

Knee joint X-rays were investigated for the presence of ON and findings such as sclerotic lesions, focal osteopenia, cortical collapses and subchondral radiolucency consistent with ON were recorded. The observers were blinded to the clinical information and MRI findings of patients. Then, at least 30 days later, knee joint MRI examinations of the patients were evaluated for the presence of ON according to predefined MRI findings (10). Localization, maximum size of the dominant lesion, multiplicity of lesions, presence of bone edema, double-line sign and subchondral extension were evaluated in these patients. Accompanying femoral head ON was also investigated in patients who underwent hip MRI.

Statistical Analysis

Statistical analysis was performed using SPSS for Windows (Version 24.0, SPSS Inc. IBM Corp, Chicago, IL). Continuous variables were expressed as mean ± standard deviation and range (minimum-maximum), and categorical variables as number and percentage. Comparison of groups with non-normal distribution was performed using Mann-Whitney U test. Independent samples t-test was used for comparisons in normally distributed data. P<0.05 was considered statistically significant.

Results

Only 13 (9.6%) knee joint X-rays showed a lesion compatible with ON. However, in a total of 162 knee joint MRI examinations, 37 knees (22.8%) (21 left and 16 right knees) of 28 patients (20.6%) (18 female, mean age: 42.8±13.4 years, range: 16-75 years) had ON. Accepting MRI as gold standard, diagnostic performance of plain radiography for diagnosis of knee ON is shown in Table 2.

MRI revealed multiple lesions in 31 knees (n=162, 19.1%) and single lesion was found in six knees (n=162, 3.7%) with ON. Dominant lesions in each knee were evaluated by MRI, and 28 of 37 lesions (75.7%) had a double-line sign, 33 of 37 knees (89.2%) had bone marrow edema and the size of dominant lesions ranged from 9 to 98 mm (mean: 37.7±19.8 mm, median: 37 mm) (Figure 1-2-3). The lesions were found to extend to the subchondral space in 24 knees (64.9%). There was no significant relationship between the number of lesions and the distribution sites (p=0.452). The distribution sites of knee ON in patients with single and multiple lesions are shown in Table 3.

The duration of CS intake was between 1 and 13 years (median: 7 years). The CS doses taken during the course of the disease varied and unfortunately these doses could not be determined. Clinical indications for CS therapy in patients with knee ON are shown in Table 4.

Only 16 out of 28 patients with knee ON had hip joint MRI (57.1%) who had hip pain. Five of 16 patients (31.3%) had femoral head ON and four of these five patients (80%) had ON of both femoral heads.

Discussion

Our results showed that approximately one third of patients suffering from arthralgia who had knee ON during CS therapy had femoral head ON. CS-induced knee ON was present in approximately one of the five patients with long-term CS intake. Double-line sign and bone marrow edema were the most common MRI findings in CS-induced knee ON. Plain radiography has low sensitivity, but high specificity for the diagnosis of knee joint ON in patients suffering from arthralgia and receiving CS therapy.

Following hip joint, the knee joint is the second most frequent joint affected by ON. Knee joint involvement will occur in 10% of patients with hip joint ON (1). Gladman et al. (2) showed that bilateral femoral head ON in 54.7% of systemic lupus erythematosus (SLE) patients treated with CS, and knee ON in 18.9% of these patients. However, our study showed that 31.3% of patients suffering from arthralgia with knee joint ON on CS therapy had hip joint ON. When the results of this study are evaluated together with the findings in the literature, our results support that the femoral head is affected more frequently than the knee in patients on CS therapy.

MRI is considered the gold standard for the diagnosis of ON and it has a very high diagnostic performance (1,5,9,11). Other imaging modalities such as plain radiography, bone scanning, positron-emission tomography and computed tomography for the diagnosis of femoral head ON have been previously studied extensively (11-14). Prompt diagnosis and early treatment of ON is critical because it is associated with a more favorable prognosis (13,15). Previous studies have shown that plain radiography has very low sensitivity for the early period of femoral head ON (10-13). Conklin et al. (16) and Deutsch et al. (17) described that plain radiography has a sensitivity of 41% and 44%, respectively, for the detection of CS-induced femoral head ON. To the best of our knowledge, there are no studies evaluating the performance of plain radiography in the diagnosis of knee joint ON. We found that the sensitivity of plain radiography for the diagnosis of knee joint ON was only 29.7% in patients receiving CS therapy with complaints of arthralgia. However, because of its high specificity (98.4%), and being cheap and easily accessible, we suggest that plain radiography should be the first choice in knee joint in patients receiving CS therapy, as well as in the hip joint.

CS-induced ON tends to be multifocal lesions at multiple joints, as illustrated by studies of patients with SLE who are on long-term CS therapy (2,18,19). Similarly, we found that five of 16 (31.3%) patients with knee joint ON had femoral head ON who were complaining of hip joint pain. In the presence of multiple joint pains in patients receiving CS therapy, clinicians should be aware of ON. We suggest that knee joint MRI should be performed to exclude the presence of ON if plain radiography is unremarkable and patient still complains of knee joint pain.

Spontaneous ON of the knee, described by Ahlbäck et al. (20), is a disorder of unknown etiology and usually affects a single location in older patients. However, CS-induced knee joint ON usually affects younger patients and presents with multiple simultaneous lesions in the femur or tibia (21). Similarly, we found multiple lesions in 31 of 37 knee joints (83.8%) with ON.

A circumscribed subchondral “band-like” hypointense lesion on T1W images and “double-line sign” on T2W images were considered pathognomonic for femoral head ON (22). The “double-line sign” consists of a hypointense outer rim (represents the reactive bone) and a hyperintense inner rim (represents granulation tissue) on T2W images. Double-line sign was found to be present in 80% of the lesions. Similarly, we found that 75.7% of dominant lesions had double-line sign on T2W images. The presence of bone marrow edema in MRI is controversial because it has two main causes: ON and bone marrow edema syndrome. The distinction between these entities is very important because treatment is different. ON is a progressive disease requiring early diagnosis and treatment, whereas bone marrow edema syndrome resolves spontaneously and surgery is not recommended. In MRI, presence of diffuse pattern in the absence of subchondral lesions seems to almost always correspond to transient bone marrow edema syndrome, which is a self-limiting disorder (23,24). In addition, the presence of bone marrow edema in patients with ON has been shown to be correlate with pain (23). In our cohort, all patients with knee joint MRI complained of pain and we found that 89.1% of knee joints with ON showed bone marrow edema. We also found that subchondral spaces were involved in 64.9% of knee joints with ON.

Study Limitations

One limitation of our study was the lack of histopathological diagnosis of the lesions. However, this did not significantly affect our results because MRI is the gold standard for the diagnosis of ON with a very high sensitivity and specificity. Second, due to the retrospective nature of our study, we could not determine the doses of CS taken by patients. However, many authors claim that the cumulative CS dose is the most important factor for the development of ON (4-6). Third, the number of patients is limited in our study. Therefore, there is a need for prospective studies with larger number of patients. Finally, all patients in our study had complaints of joint pain, thus there was a lack of pain-free patients receiving CS therapy.

Conclusion

Thirty-one point three percent of patients with knee joint ON with arthralgia who were receiving CS therapy had ON of the hip joint. Regarding its high specificity and easy accessibility, plain radiography should be the first choice in knee joint pain in patients on CS therapy. Despite a normal plain radiography, MRI should be performed in patients whose complaints are not resolved. There were multiple lesions in MRI in most of the patients who had CS-induced knee joint ON. Double-line sign and bone marrow edema were the most common MRI findings in CS-induced ON of the knee.

Ethics Committee Approval: This study was approved by institutional Pamukkale Universtiy Ethics Committee (decision no: 60116787-020/8340).

Informed Consent: Informed consent was waived due to retrospective nature of the study.

Peer-review: Externally and internally peer-reviewed.

Author Contributions: Concept - N.S.; Design - N.S.; Data Collection and/or Processing -N.S., F.U.; Analysis and/or Interpretation - F.U.; Literature Search - F.U.; Writing Manuscript - N.S., F.U.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.