LETTERS Antibodies against mutated citrullinated vimentin fail to predict anti-TNFa treatment response in rheumatoid arthritis C Dejaco 1 , C Duftner 1 , W Klotz 2 , M Schirmer 1 , M Herold 2 Department of Internal Medicine, 1 General Hospital of the Elisabethinen, Klagenfurt and 2 Clinical Division of General Internal Medicine, Medical University Innsbruck, Austria Antibodies to citrullinated peptides (ACPA), such as anti-mutated citrullinated vimentin (anti-MCV) and anti-cyclic citrullinated peptide antibodies (anti- CCP), may reflect disease activity in rheumatoid arthritis (RA) (1). Although these ACPA tests showed comparable diagnostic value (2), anti-MCV correlates better with disease activity than anti-CCP and is superior in the prediction of radiographic destruction (3, 4). So far, only anti-CCP levels have been analysed in RA patients under treatment with tumour necrosis factor (TNF)a blocking agents (5). This study was performed to test whether anti- MCV and anti-CCP2 antibodies are differentially influenced by TNFa blocking agents, and to evaluate a possible role of these assays in predicting clinical response to TNFa blocking agents. The results were then compared with immunoglobulin (Ig)G-, IgA-, and IgM-rheumatoid factor (RF). We included 42 RA patients [mean age 48.5 (¡10.9) years; 81.0% female; median disease duration 6.5 (range 0.5–34.0) years] treated with a TNFa blocking agent [infliximab (n511), etanercept (n57), adalimumab (n524)] in addition to disease-modifying anti-rheumatic drugs (DMARDs) and/or prednisone. Anti-CCP2 (Phadia GmbH, Freiburg, Germany), anti-MCV, IgG- RF, IgA-RF, and IgM-RF (ORGENTEC Diagnostica GmbH, Mainz, Germany) were measured using com- mercially available enzyme-linked immunosorbent assay (ELISA) kits (2). Statistical analysis was performed using SPSS, version 11.0 (SPSS Inc., Chicago, IL, USA). The non-parametric Mann–Whitney U-test, Wilcoxon rank tests, receiver operating character- istic (ROC) curve analysis and logistic regression (maximum likelihood ratio method) were performed as appropriate. This study was approved by the local ethics committee. After 6 months of TNFa blocking treatment, median Disease Activity Score (DAS)-28 decreased from 5.7 (range 1.4–7.7) to 3.4 (1.5–7.7) (pv0.001). Thirty-one (73.8%) of the 42 RA patients were classified as EULAR (European League Against Rheumatism) responders [pooled moderate (n53) and good responders (n528)] (6). Elevated levels of anti-MCV and anti-CCP2 were found in 76.2% and 81.0% of patients, respectively (difference not sig- nificant). After TNFa blocking therapy, anti-MCV and anti-CCP2 levels did not change in either responders or non-responders (Table 1). To compare anti-MCV and anti-CCP2 tests as possible predictors of a EULAR-response ROC curve analysis, single predictor and covariant-adjusted logistic regression models (with patients’ age, sex, disease duration and baseline DAS-28 score as possible negative confounders) were performed: neither anti-MCV nor anti-CCP2 tests were associated with anti-TNFa treatment response (data not shown). IgG-, IgA-, and IgM-RF were present in 78.6%, 73.8%, and 78.6% of TNFa blocker naı ¨ve RA patients, respectively. Reductions in IgG-, IgA-, and IgM-RF concentrations were observed in EULAR responders but not in non- responders (Table 1). The underlying regulatory mechanisms of RA-associated autoantibodies with different effects of TNFa blocking agents on ACPA and RF are still not well understood (7). It has been speculated that RF production could be at least partially dependent on inflammation whereas ACPA production might be more ‘constitutive’, even though in loco produc- tion of antibodies directed to citrullinated antigens has been reported in rheumatoid synovitis (8). Blockage of TNFa may also favour autoimmune phenomena due to an increase of interleukin (IL)-10 and failure of cytotoxic T-lymphocyte activity, normally inhibiting autoreactive B cells. Indeed, induction of antinuclear antibodies occurred in anti-TNFa-treated RA patients despite reduced RF and decreased or stable anti-CCP levels (7, 9). TNFa blocking treatment thus seems to interfere selectively with the regulation of autoantibodies rather than inducing pan-suppression. Our study was limited by the retrospective design, the short follow-up period, the pooled analysis of patients with different anti-TNFa agents, conven- tional DMARDs and/or prednisone, as well as the analysis of patients with long-standing disease. Previous studies have shown greater reductions of 66 Scand J Rheumatol 2009;38:66–76 www.scandjrheumatol.dk # 2009 Taylor & Francis on license from Scandinavian Rheumatology Research Foundation DOI: 10.1080/03009740802220067