A prediction rule for disease outcome in patients with undifferentiated arthritis using magnetic resonance imaging of the wrists and finger joints and serologic autoantibodies

Abstract

Objective. To evaluate whether magnetic resonance imaging (MRI) of the wrists and finger joints and an analysis of serologic autoantibodies are clinically meaningful for the subsequent development of rheumatoid arthritis (RA) in patients with undifferentiated arthritis (UA). Methods. A total of 129 patients with UA, a disease status formally confirmed by a rheumatologist over a period of at least 1 year, were included. Gadolinium-diethylenetriamine-enhanced MRI of both wrists and finger joints and serologic variables were examined upon admission to our Early Arthritis Clinic at Nagasaki University. After a prospective followup of 1 year, a predictive value for the development of RA was determined for each patient. Results. The subjects were evaluated for their positive or negative status with respect to 3 objective measures at study entry: anti-cyclic citrullinated peptide (anti-CCP) antibodies and/or IgM-rheumatoid factor, MRI-proven symmetric synovitis, and MRI-proven bone edema and/or bone erosion. The patients who were positive for at least 2 of these measures progressed to RA at 1 year with a 79.7% positive predictive value (PPV), 63.0% negative predictive value, 75.9% specificity, 68.0% sensitivity, and 71.3% accuracy. Furthermore, in 22 UA patients positive for both anti-CCP antibodies and MRI-proven bone edema who were considered to have progressed to RA at 1 year, the PPV was increased to 100%. A close correlation was found between the present rule and that established in the Leiden Early Arthritis Cohort. Conclusion. MRI-proven early joint damage in conjunction with serologic autoantibodies is efficient in predicting progression from UA to RA. This method can be used to identify patients who would benefit from early treatment with disease-modifying antirheumatic drugs. INTRODUCTIONEarly undifferentiated arthritis (UA) is defined as early arthritis that does not fulfill the classification criteria for a more definitive diagnosis, according to the 1987 American College of Rheumatology (ACR; formerly the American Rheumatism Association) criteria for rheumatoid arthritis (RA) (1-3). The natural disease course of UA is variable; therefore, to minimize under-and overtreatment of patients with UA, a model was recently constructed by the Leiden Early Arthritis Cohort to estimate the likelihood of progression to RA in individual patients (2,3). Their prediction rule consists of 9 clinical variables: sex, age, localization of symptoms, morning stiffness, tender joint count, ISRCTN: 021008-1.

Bone edema on magnetic resonance imaging is an independent predictor of rheumatoid arthritis development in patients with early undifferentiated arthritis

“…However, our developed prediction model performed well and correctly predicted RA development or non-RA in 82% of our cohort, with a sensitivity of 81% and a specificity of 82%. In our cohort, the patients were less likely to develop RA over the followup than were the patients in other cohorts from comparable studies (1,6). In our study, 23.3% of the study population had UA that progressed to RA, whereas in the studies by Sugimoto et al (6) and Tamai et al (1), 63% and 58% of the patients, respectively, received a final diagnosis of RA.…”

“…The outcome of a model is dependent on the inclusion criteria of the tested cohort; i.e., the model may perform differently in cohorts with different risks of developing RA. This may explain why the model from van der Helm-van Mil et al performed rather poorly in our cohort (identified 60.2% of patients as having RA or non-RA) compared to its application in previous cohorts (1,7).…”

“…Approaches for early diagnosis, such as RA prediction models (1-3) and more specific tools, i.e., the anti-cyclic citrullinated peptide (anti-CCP) test (4,5) and magnetic resonance imaging (MRI), have been suggested (1,6). A prediction model for determining the factors that might be predictive of RA development has been created by van der Helm-van Mil et al (2), referred to as the vdHvM model; this model has been validated in several cohorts (1,7).…”

“…A prediction model for determining the factors that might be predictive of RA development has been created by van der Helm-van Mil et al (2), referred to as the vdHvM model; this model has been validated in several cohorts (1,7). The vdHvM model includes, as explanatory variables, the patient's demographic characteristics, duration of morning stiffness, measures of disease involvement in the joints, and findings from biochemical tests (i.e., tests for IgM rheumatoid factor [IgM-RF], anti-CCP, and C-reactive protein [CRP]), but does not include MRI-related variables.…”

Comparative study of the detection of joint injury in early‐stage rheumatoid arthritis by magnetic resonance imaging of the wrist and finger joints and physical examination

“…Their disease status was formally confirmed by a Japan College of Rheumatology-certified rheumatologist in our department, and a diagnosis of RA was made based on the 1987 criteria of the American College of Rheumatology for RA (4). Baseline clinical manifestations and variables included sex, age, localization of tender and/or swollen joints, morning stiffness, C-reactive protein level (measured by latex turbidimetric immunosorbent assay; Daiichi Pure Chemicals), IgM rheumatoid factor (IgM-RF) positivity (measured by latex-enhanced immunonephelometric assay, cutoff value 14 IU/ml; Dade Behring), positive status for anti-cyclic citrullinated peptide (anti-CCP) antibodies (measured by enzyme-linked immunosorbent assay, cutoff value 4.5 units/ml; DIASTAT Anti-CCP, Axis-Shield), HLA-DRB1 genotyping, and MRI of both the wrist and finger joints, as previously described (5)(6)(7)(8). All of the variables were examined on the same day, as previously reported (5)(6)(7)(8).…”