Objectives To analyze the predictive value of inflammatory findings in hand MRI (h-MRI) and establish cutoff values of these inflammatory lesions at baseline (“acceptable inflammatory activity”) associated with absence of progression of erosions in h-RM after one year follow-up in RA remission patients
Methods Prospective study of RA patients (fulfilling ACR 1987 revised criteria) in clinical remission (defined as DAS28-ESR ≤2.6 for at least 6 months and without changes of treatment 3 months before study) and with avaliable data of h-MRI at baseline and at one year. At baseline and every 3 months clinical and laboratory evaluation was performed according to established protocol. MRI of the dominant hand were obtained at entry of study and one year after. MRI images were read by 2 experienced readers following RAMRIS recommendations. Inflammatory MRI findings (bone edema, synovitis, tenosynovitis, sum of all inflammatory lesions) and bone erosions were recorded. “Acceptable inflammatory activity in h-MRI”, was defined as the inflammatory MRI findings associated with no progression of erosions at the annual h-MRI. “No progression of erosions on h-MRI” was defined as the change in annual RAMRIS <1 point from baseline erosion score. Statistical analysis included logistic regression analysis and receiver operating characteristic (ROC). Data evaluation and statistical analysis were performed using SPSS v.18.
Results Fifty patients with RA (44 women, age 54.2±11.7 years, mean disease duration 8.2±7.1 years) and in clinical remission (mean of DAS28-ESR 1.72±0.60 and with a mean duration of remission 20.7±12.1 months) were included. Forty patients (80%) did not show progression of erosions on annual h-MRI.Univariate analysis showed a significant association between h-MRI progression of erosions at one year with disease duration, (OR 1.13, 95%CI 1.02 a 1.25, p=0.013), RAMRIS bone edema (OR 1.65, 95%CI 1.14 a 2.37, p=0.007) and total inflammation at baseline (OR 1.14, 95%CI 1.0 a 1.30, p=0.045) and a trend toward an association with baseline RAMRIS synovitis (OR 1.29, 95%CI 0.98 a 1.69, p=0.065). In the multivariate analysis, bone edema was the only significant independent predictor associated with h-MRI progression of erosions (OR 1.65 95%CI 1.14 a 2.37, p=0.007). ROC analysis identify a cut-off point of 2 for RAMRIS bone edema (AUC of 0.75, 95%CI 0.55 to 0.95, p=0.002) and 5 for RAMRIS synovitis (AUC 0.68 95%CI 0.46 to 0.89, p=0.007) as a “acceptable inflammatory activity in h-MRI” and suggest that these cutoffs were the best for to discriminate patients with or without risk of h-MRI progression at one year with a value of sensitivity >0.60 and and specificity >90%.
Conclusions In RA remission patients, baseline bone edema on h-MRI was the most important independent predictor associated with h-MRI progression of erosions at one year.The discriminative ability of inflammatory h-MRI cut-offs (2 for RAMRIS-bone edema and 5 for RAMRIS-synovitis) between patients with or without risk of h-MRI progression were acceptable. *REMAR study: Investigator Initiated Research funded by Pfizer
Disclosure of Interest None declared