International Journal of Rheumatic Diseases 2014; 17: 872–877
ORIGINAL ARTICLE
IgA rheumatoid factor as a serological predictor of poor response to tumour necrosis factor a inhibitors in rheumatoid arthritis Rajalingham SAKTHISWARY,1 Syahrul S. SHAHARIR,1 Mohd S. MOHD SAID,1 Abdul W. ASRUL2 and Nor S. SHAHRIL3 1
Department of Medicine, Universiti Kebangsaan Malaysia Medical Centre, 2Department of Microbiology and Immunology, Universiti Kebangsaan Malaysia Medical Centre, and 3Department of Medicine, Putrajaya Hospital, Kuala Lumpur, Malaysia
Abstract Aim: The main objective of this study is to elucidate the role of immunoglobulin A (IgA) rheumatoid factor (RF) in predicting the clinical response to tumour necrosis factor a inhibitors (TNFi) among patients with rheumatoid arthritis (RA). Method: We recruited all patients with RA who were ever on TNFi for a minimum duration of 3 months at our centre. Based on the European League Against Rheumatism response criteria, subjects were further divided into responders and non-responders. Age-matched RA patients who were on conventional disease-modifying antirheumatic drugs and in remission were enrolled as controls. Subjects were tested for quantitative values of IgA, IgM, IgG RF and anti-citrulinated cyclic peptides (CCP). Further, all subjects were assessed for the disease activity score that includes 28 joints (DAS28) and Stanford Health Assessment Questionnaire (HAQ) 8-item Disability Index (HAQ-DI). Results: A total of 31 subjects with RA who had received TNFi and 15 controls were enrolled in this study. There was a trend for the non-responders (n = 10) to have higher levels of all isotypes of RF and anti-CCP. However, only the IgA RF and anti-CCP levels were significantly higher in the non-responder group compared to the responders and controls (P = 0.001, P = 0.034, respectively). On multivariate analysis, only the IgA RF remained significant (OR 0.989; 95% CI 0.980–0.999; P = 0.026). Conclusion: IgA RF is potentially a novel predictor of response to TNFi in RA patients. Testing for pretreatment IgA RF levels could be a reasonable consideration before commencement of TNFi. Key words: IgA rheumatoid factor, rheumatoid arthritis, rheumatoid factor isotypes, tumour necrosis factor a inhibitors.
INTRODUCTION Tumour necrosis factor a inhibitors (TNFi) are a step forward in the therapeutic armamentarium of rheumaCorrespondence: Associate Professor Dr Rajalingham Sakthiswary, MRCP(UK), Department of Medicine, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Jalan Yaacob Latif, Bandar Tun Razak 56000, Cheras, Kuala Lumpur, Malaysia. Email: [email protected]
toid arthritis (RA). Despite the promises of clinical trials and pharmaceutical companies, clinicians are often faced with poor response to TNFi. Up to one-third of patients may not respond optimally to this costly form of treatment.1 A question which is yet to be answered is how to predict patients’ responses to treatment before and during the early phase of treatment with TNFi. To date, no reliable indices have been identified as predictive factors in this regard.
© 2014 Asia Pacific League of Associations for Rheumatology and Wiley Publishing Asia Pty Ltd
IgA rheumatoid factor and response to TNFi
A high titer of rheumatoid factor (RF) is associated with a more aggressive course of the disease with progressive joint damage and extra-articular involvement. This has especially been observed in patients with raised immunoglobulin A (IgA) RF.2 Yet the sequential measurement of IgA RF is not performed routinely in clinical practice. In 2007, Bobbio-Pallavicini reported that high pretreatment levels of IgA RF were associated with poor clinical responses to TNFi.3 Unfortunately, there were no subsequent verification studies to confirm the above finding. A few existing studies have demonstrated a fall in RF levels following control of the disease with both the traditional disease-modifying antirheumatic drugs (DMARDs) and TNFi.4–7 The remaining studies have reported contradictory results.8,9 The overall paucity of data does not allow for convincing conclusions to be made on the definite relationship between the levels of IgA RF and clinical response to TNFi. Studies have consistently pointed toward genetic heterogeneity in RA across different ethnic groups.10,11 There is hardly any data on autoantibody pattern in Asian RA patients treated with TNFi. So far, we have only been armed with lessons from the West in this area of research. Therefore, this study is probably the first in this region to deal with this aspect of RA. Identification of predictors of clinical response to TNFi may potentially pave the way for earlier recognition of nonresponders. This would guide clinicians to make prompt decisions on switching of biologics, not only to arrest disease progression but also to save on costs. The main objective of this study is therefore to elucidate the role of IgA RF in predicting the clinical response to TNFi.
from baseline for those with low to moderate disease activity (DAS28 ≤ 5.1) pre-TNFi, or improvement of more than 1.2 in subjects with high disease activity (DAS28 > 5.1) pre-TNFi. As for subjects who responded to the second TNFi (after failing to respond to one previous TNFi), they were classified as responders. Subjects were tested for quantitative values of IgA, IgM, IgG RF and anti-citrulinated cyclic peptides (antiCCP). Further, all subjects were assessed for DAS28 and Stanford Health Assessment Questionnaire (HAQ) 8-item Disability Index (HAQ-DI).14 The subjects were categorized as high disease activity (DAS28 > 5.1), moderate disease activity (3.2 < DAS28 ≤ 5.1), low disease activity (2.6 < DAS28 ≤ 3.2) and remission (DAS28 ≤ 2.6).13 For the non-responders who had already discontinued the TNFi, the last DAS28 and HAQ-DI scores while on TNFi were taken into account. The DAS28 scores pre-TNFi were gathered from the medical records. Fifteen age-matched controls who were RA patients on conventional DMARDs and in remission were enrolled and tested for IgA, IgM, IgG RF and anti-CCP.
METHODOLOGY
Anti-CCP
Study design and subjects
Anti-CCP were tested using a commercially available ELISA kit (Axis-Shield, Dundee, UK). The upper normal limit (5 U/mL) was determined in accordance with the manufacturer’s recommendations.
We recruited all patients with RA who were ever on TNFi for a minimum duration of 3 months at our centre, Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur. The inclusion criteria of this study included: (i) patients who fulfill the American College of Rheumatology (ACR) 2010 criteria for RA;12 (ii) patients ever on TNFi due to intolerance or inadequate response to conventional DMARDs; and (iii) patients with seropositive disease. Based on the European League Against Rheumatism (EULAR) response criteria, subjects were further divided into responders and non-responders.13 Responders were patients who achieved improvement in Disease Activity Index of 28 joints (DAS28) of more than 0.6
International Journal of Rheumatic Diseases 2014; 17: 872–877
Rheumatoid factors The different RF isotypes (IgM, IgA and IgG) were tested using an indirect solid-phase enzyme-linked immunosobent assay (ELISA) involving the binding of Fc fragments of highly purified human IgG to microwells (Dade Behring, Marburg, Germany). The quantitative analysis for IgM, IgG and IgA RF were calibrated as per the kit manual. The procedure was carried out in triplicate. According to the manufacturer’s recommendations, RF concentrations of above 15 IU/mL were considered positive.
Statistical analysis The data analysis was performed using SPSS (Statistical Package for Social Sciences) version 20 (SPSS Inc., Chicago IL, USA). Continuous variables were described as mean standard deviation (SD), or median (range), while the categorical variables as counts and percentages. The analysis of variance (ANOVA) test was used to make a comparison across the three groups, that is, responders, non-responders and controls. Variables which were significant on univariate analyses were
873
R. Sakthiswary et al.
further analyzed using multivariate analysis while controlling for confounders. The association between two continuous variables were determined using bivariate correlation and linear regression analyses. A P-value of < 0.05 was considered significant.
RESULTS Sociodemographic and clinical characteristics of RA patients on TNFi A total of 31 subjects with RA who have received TNFi and 15 controls were enrolled in this study. There were 21 responders and 10 non-responders to TNFi. Table 1 summarizes the characteristics of the RA patients treated with TNFi. The mean age of these patients was 56.90 8.95 years. It took more than a decade on the average (10.45 6.01 years) from time of diagnosis to treatment with TNFi. The most widely prescribed TNFi in our centre was adalimumab (54.8%). A significant proportion of the subjects had moderate to high disease activity (38.7%).
Table 1 Sociodemographic and clinical characteristics of rheumatoid arthritis patients who have been on tumor necrosis factor a inhibitors (TNFi) Parameters Age (years) Gender, n (%) Female Male Ethnicity, n (%) Malays Chinese Indians Disease duration pre-TNFi (years) Duration on TNFi (months) TNF inhibitors, n (%) Infliximab Adalimumab Etanercept Golimumab CRP (mg/dL) ESR (IU/mL) DAS28 Remission–low disease activity (n[%]) Moderate–high disease activity (n[%]) HAQ DI
n = 31 56.90 8.95 30 (96.80) 1 (3.20) 17 (54.80) 5 (16.10) 9 (29.00) 10.45 6.01 15.74 11.58 2 (6.45) 17 (54.80) 10 (32.36) 2 (6.45) 1.08 1.34 65.29 28.03 19 (61.30) 12 (38.70) 0.11 0.22
Data presented as either counts (percentages) or mean SD. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; HAQ-DI; Health Assessment Questionnaire Disability Index; DAS28, 28 jointbased Disease Activity Score.
874
RF isotypes across the study groups Table 2 illustrates the clinical and serological parameters in the three study groups. There was a trend for the non-responders to have higher levels of all isotypes of RF. This pattern was most evident for IgA RF levels, which were significantly higher in the non-responder group compared to the responders and controls (P = 0.001). The significant association between the IgA RF levels and the non-responders was confirmed on multivariate analysis (OR 0.989; 95% CI 0.980–0.999; P = 0.026) with adjustment for confounders, that is, HAQ-DI, duration on biologics and anti-CCP levels. The IgA RF levels demonstrated a significant linear relationship with DAS28 levels while on TNFi (P = 0.005, standardized beta coefficient = 0.490). However, changes in DAS28 (pre-TNFi DAS28–on TNFi DAS28) showed no linear association with the IgA RF levels (P = 0.071, standardized beta coefficient = 0.334) (Fig. 1).
IgA RF versus anti-CCP Although the anti-CCP levels were much higher among the non-responders (P = 0.034) on univariate analysis, this variable was insignificant on multivariate analysis (OR = 0.500; 95% CI 0.979–1.004; P = 0.199). Unlike IgA RF, anti-CCP levels showed no significant association with DAS28 while on TNFi (P = 0.093, standardized beta coefficient = 0.307). Similar results were obtained for anti-CCP and changes in DAS28 with TNFi (P = 0.147, standardized beta coefficient = 0.272).
DISCUSSION The findings of this study depict the clinical importance of IgA RF in predicting response to TNFi among RA patients. Higher levels of IgA RF have a significant association with poor response to this form of biologics. Our results concur with the findings of Bobbio-Pallavicini et al. who reported that the non-responder group had significantly (P = 0.003) higher levels of IgA RF (130.4 U/mL [13.8–276.7 U/mL]) compared to the responder group (24.8 U/mL [10.2–90.8 U/mL]). Further, in keeping with our findings, the above-named study found no significant relationship between the lack of response and the other RF isotypes (IgG RF and IgM RF),although all RF isotypes had a significant drop among the responders after completing 1 year of TNFi therapy.3 However, Lequerre et al. had conflicting results with comparable positive rates (P = 0.35) of IgA RF among their responders and non-responders. It is
International Journal of Rheumatic Diseases 2014; 17: 872–877
IgA rheumatoid factor and response to TNFi
Table 2 Clinical and serological parameters of the responders, non-responders and the controls Parameter
Responders (n = 21)
Age (years) Disease duration pre-TNFi (years) Duration on TNFi (months) DAS28 pre-TNFi DAS28 on TNFi Changes in DAS28 (pre-TNFi–on TNFi) IgA RF (U/mL) IgG RF (U/mL) IgM RF (U/mL) Anti-CCP (U/mL) HAQ-DI
56.90 9.53 19.14 5.17 2.68 2.74 42.33 88.33 84.72 182.54 0.45
9.86 5.55 12.17 1.39 0.81 1.1.3 77.07 92.14 118.53 143.98 0.40
Non-responders (n = 10) 56.89 12.20 8.60 5.41 5.37 0.19 152.64 99.31 98.93 278.48 0.80
6.86 6.75 5.87 0.86 0.69 0.97 100.34 94.64 111.86 46.91 0.47
Controls (n = 15)
P-value
58.64 7.88 6.10 4.10 NA NA NA NA 29.71 72.21 76.35 48.91 88.39 114.51 138.97 145.86 0.11 0.22
0.856 0.061 0.564 < 0.05 < 0.05 0.001 0.756 0.950 0.034 0.002
Data presented as either counts (percentages) or mean SD. TNFi, tumor necrosis factor a inhibitors; IgA RF, immunoglobulin A rheumatoid factor; HAQ-DI, Health Assessment Questionnaire Disability Index; DAS28, 28 joint-based Disease Activity Score; CCP, citrulinated cyclic peptide; RF, rheumatoid factor. Significant P-values are in bold.
2
(a)
R Linear = 0.240
y = 2.97 + 7.39E–3*x
Ig A rheumatoid factor
(b)
2
R Linear = 0.112
y = 2.31 + (–6.35)E–3*x
Figure 1 Relationship between immunoglboulin A rheumatoid factor (IgA RF) levels and (a) Disease Activity Score of 28 joints (DAS28) while on tumor necrosis factor a inhibitors (TNFi) and (b) changes in DAS28 (pre-TNFi–post-TNFi).
International Journal of Rheumatic Diseases 2014; 17: 872–877
noteworthy that unlike Bobbio-Pallavicini et al.3 and our study which compared the average values (mean or median) of IgA RF, Lequerre et al.15 analyzed the frequency of IgA RF positivity. This could partially explain the discrepancy in their findings in this regard. The non-responders were on a shorter duration of TNFi (8.60 5.87 months) compared to the responders (19.14 12.17 months) given that the TNFis were discontinued in the former group of patients usually after 6 months owing to the lack of or suboptimal clinical response. This practice of ours was in accordance with the EULAR 2013 recommendations.16 Although the non-responders had the highest disease duration (12.20 6.75 years), the difference in this parameter among the three groups failed to reach statistical significance. Although IgA RF was first identified in the sera of RA patients as early as 1963,17 the clinical significance of this biomarker is somewhat under-recognized and therefore has not received the attention it deserves in terms of research. Several studies have indicated that IgA RF is more specific for RA compared to the classic IgM RF.18,19 Moreover, it is predictive of a more aggressive course of the disease with early radiographic erosions.20,21 The precise mechanism by which circulating IgA RF influences the disease process remains unclear. The association observed between IgA RF levels and poor response to TNFi does not necessarily imply causation. It is pertinent to mention here that five out of ten of the non-responders showed marked improvement with tocilizumab, an interleukin 6 receptor antagonist.
875
R. Sakthiswary et al.
Data on the response to non-TNFi with high IgA RF is currently not available. The non-responders tended to have higher levels of IgG RF and IgM RF. Lack of statistical significance of these variables could reflect a type II error. Apart from IgA RF, anti-CCP and HAQ-DI were associated with poor response at univariate analysis. Across the studies, there is remarkable inconsistency with regard to the predictive value of anti-CCP for the response to TNFi. While Bobbio-Pallavicini et al.,3 Lequerre et al.15 and De Rycke et al.6 found no association between anti-CCP levels and clinical response to TNFi, Alessandri et al.4 and Atzeni et al.7 demonstrated a decrease in anti-CCP levels in parallel with clinical improvement with TNFi therapy. Diversity of the study populations and methodological variations may partially explain the discordance in the results of the aforementioned studies. With respect to functional disability and response to TNFi, in agreement with our findings, Kristensen et al.22 stated that lower HAQ-DI scores were related to better response to TNFi. This study is not without limitations. The levels of autoantibodies tend to fluctuate in the blood circulation. Hence, the average of serial readings is preferred over a single measurement. Moreover, there are other possible unmeasured confounders that might affect the production of the RF isotypes, such as cigarette smoking. Comprehensive data on the smoking habits of the subjects were not collected. We did not include subjects with seronegative disease (subjects who tested negative for classic IgM RF) and therefore our results may not be generalizable to this subset of RA patients. In conclusion, IgA RF is potentially a novel predictor of response to TNFi in RA patients. Testing for pretreatment IgA RF levels could be a reasonable consideration before commencement of TNFi. Clinicians may consider non-TNFi biologic therapy such as tocilizumab, abatacept and rituximab, as second line or even first line in patients with high levels of the above or static levels, despite TNFi. However, this recommendation should be on the basis of strong supportive evidence by more prospective clinical studies.
ACKNOWLEDGEMENT This study was funded by the APLAR Research Grant Award 2013. RS who is the first and corresponding author is the recipient of the award.
876
REFERENCES 1 Hyrich KL, Watson KD, Silman AJ, Symmons DP (2006) Predictors of response to anti-TNF-alpha therapy among patients with rheumatoid arthritis: results from the British Society for Rheumatology Biologics Register. Rheumatology (Oxford) 45 (12), 1558–65. 2 Teitsson I, Valdimarsson H (1984) Use of monoclonal antibodies and F(ab’)2 enzyme conjugates in ELISA for IgM, IgA and IgG rheumatoid factors. J Immunol Methods 71, 149–61. 3 Bobbio-Pallavicini F, Caporali R, Alpini C et al. (2007) High IgA rheumatoid factor levels are associated with poor clinical response to tumour necrosis factor alpha inhibitors in rheumatoid arthritis. Ann Rheum Dis 66, 302–7. 4 Alessandri C, Bombardieri M, Papa N et al. (2004) Decrease of anti-cyclic citrullinated peptide antibodies and rheumatoid factor following anti-TNFalpha therapy (infliximab) in rheumatoid arthritis is associated with clinical improvement. Ann Rheum Dis 63 (10), 1218–21. 5 Mikuls TR, O’Dell JR, Stoner JA et al. (2004) Association of rheumatoid arthritis treatment response and disease duration with declines in serum levels of IgM rheumatoid factor and anti-cyclic citrullinated peptide antibody. Arthritis Rheum 50 (12), 3776–82. 6 De Rycke L, Verhelst X, Kruithof E et al. (2005) Rheumatoid factor, but not anti-cyclic citrullinated peptide antibodies, is modulated by infliximab treatment in rheumatoid arthritis. Ann Rheum Dis 64 (2), 299–302. 7 Atzeni F, Sarzi-Puttini P, Dell’Acqua D et al. (2006) Adalimumab clinical efficacy is associated with rheumatoid factor and anti-cyclic citrullinated peptide antibody titer reduction: a one-year prospective study. Arthritis Res Ther 8 (1), R3. 8 Yazdani-Biuki B, Stadlmaier E, Mulabecirovic A et al. (2005) Blockade of tumour necrosis factor alpha significantly alters the serum level of IgG- and IgA-rheumatoid factor in patients with rheumatoid arthritis. Ann Rheum Dis 64 (8), 1224–6. 9 Bobbio-Pallavicini F, Alpini C, Caporali R, Avalle S, Bugatti S, Montecucco C (2004) Autoantibody profile in rheumatoid arthritis during long-term infliximab treatment. Arthritis Res Ther 6, R264–72. 10 Taneja V, Mehra NK, Kailash S, Anand C, Malaviya AN (1992) Protective & risk DR phenotypes in Asian Indian patients with rheumatoid arthritis. Indian J Med Res 96, 16–23. 11 Barton A, Bowes J, Eyre S et al. (2004) A functional haplotype of the PADI4 gene associated with rheumatoid arthritis in a Japanese population is not associated in a United Kingdom population. Arthritis Rheum 50 (4), 1117–21. 12 Aletaha D, Neogi T, Silman AJ et al. (2010) 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 69 (9), 1580–8.
International Journal of Rheumatic Diseases 2014; 17: 872–877
IgA rheumatoid factor and response to TNFi
13 Fransen J, van Riel PL (2005) The Disease Activity Score and the EULAR response criteria. Clin Exp Rheumatol 23 (Suppl. 39), S93–9. 14 Thompson PW, Pegley FS (1191) A comparison of disability measured by the Stanford Health Assessment Questionnaire disability scales (HAQ) in male and female rheumatoid outpatients. Br J Rheumatol 30(4), 298–300. 15 Lequerre T, Jouen F, Brazier M et al. (2007) Autoantibodies, metalloproteinases and bone markers in rheumatoid arthritis patients are unable to predict their responses to infliximab. Rheumatology (Oxford) 46 (3), 446–53. 16 Smolen JS, Landewe R, Breedveld FC et al. (2014) EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2013 update. Ann Rheum Dis 73 (3), 492–509. 17 Teitsson I (1988) IgA rheumatoid factor as predictor of disease activity. Scand J Rheumatol Suppl 75, 233–7. 18 Bas S, Perneger TV, Kunzle E, Vischer TL (2002) Comparative study of different enzyme immunoassays for measure-
International Journal of Rheumatic Diseases 2014; 17: 872–877
19
20
21
22
ment of IgM and IgA rheumatoid factors. Ann Rheum Dis 61, 505–10. Swedler W, Wallman J, Froelich CJ, Teodorescu M (1997) Routine measurement of IgM, IgG, and IgA rheumatoid factors: high sensitivity, specificity, and predictive value for rheumatoid arthritis. J Rheumatol 24 (6), 1037–44. Bas S, Genevay S, Meyer O, Gabay C (2003) Anti-cyclic citrullinated peptide antibodies, IgM and IgA rheumatoid factors in the diagnosis and prognosis of rheumatoid arthritis. Rheumatology (Oxford) 42 (5), 677–80. Berglin E, Johansson T, Sundin U et al. (2006) Radiological outcome in rheumatoid arthritis is predicted by presence of antibodies against cyclic citrullinated peptide before and at disease onset, and by IgA-RF at disease onset. Ann Rheum Dis 65 (4), 453–8. Kristensen LE, Kapetanovic MC, Gulfe A, Soderlin M, Saxne T, Geborek P (2008) Predictors of response to anti-TNF therapy according to ACR and EULAR criteria in patients with established RA: results from the South Swedish Arthritis Treatment Group Register. Rheumatology (Oxford) 47 (4), 495–9.
877
Copyright of International Journal of Rheumatic Diseases is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
ORIGINAL ARTICLE
IgA rheumatoid factor as a serological predictor of poor response to tumour necrosis factor a inhibitors in rheumatoid arthritis Rajalingham SAKTHISWARY,1 Syahrul S. SHAHARIR,1 Mohd S. MOHD SAID,1 Abdul W. ASRUL2 and Nor S. SHAHRIL3 1
Department of Medicine, Universiti Kebangsaan Malaysia Medical Centre, 2Department of Microbiology and Immunology, Universiti Kebangsaan Malaysia Medical Centre, and 3Department of Medicine, Putrajaya Hospital, Kuala Lumpur, Malaysia
Abstract Aim: The main objective of this study is to elucidate the role of immunoglobulin A (IgA) rheumatoid factor (RF) in predicting the clinical response to tumour necrosis factor a inhibitors (TNFi) among patients with rheumatoid arthritis (RA). Method: We recruited all patients with RA who were ever on TNFi for a minimum duration of 3 months at our centre. Based on the European League Against Rheumatism response criteria, subjects were further divided into responders and non-responders. Age-matched RA patients who were on conventional disease-modifying antirheumatic drugs and in remission were enrolled as controls. Subjects were tested for quantitative values of IgA, IgM, IgG RF and anti-citrulinated cyclic peptides (CCP). Further, all subjects were assessed for the disease activity score that includes 28 joints (DAS28) and Stanford Health Assessment Questionnaire (HAQ) 8-item Disability Index (HAQ-DI). Results: A total of 31 subjects with RA who had received TNFi and 15 controls were enrolled in this study. There was a trend for the non-responders (n = 10) to have higher levels of all isotypes of RF and anti-CCP. However, only the IgA RF and anti-CCP levels were significantly higher in the non-responder group compared to the responders and controls (P = 0.001, P = 0.034, respectively). On multivariate analysis, only the IgA RF remained significant (OR 0.989; 95% CI 0.980–0.999; P = 0.026). Conclusion: IgA RF is potentially a novel predictor of response to TNFi in RA patients. Testing for pretreatment IgA RF levels could be a reasonable consideration before commencement of TNFi. Key words: IgA rheumatoid factor, rheumatoid arthritis, rheumatoid factor isotypes, tumour necrosis factor a inhibitors.
INTRODUCTION Tumour necrosis factor a inhibitors (TNFi) are a step forward in the therapeutic armamentarium of rheumaCorrespondence: Associate Professor Dr Rajalingham Sakthiswary, MRCP(UK), Department of Medicine, Faculty of Medicine, Universiti Kebangsaan Malaysia Medical Centre, Jalan Yaacob Latif, Bandar Tun Razak 56000, Cheras, Kuala Lumpur, Malaysia. Email: [email protected]
toid arthritis (RA). Despite the promises of clinical trials and pharmaceutical companies, clinicians are often faced with poor response to TNFi. Up to one-third of patients may not respond optimally to this costly form of treatment.1 A question which is yet to be answered is how to predict patients’ responses to treatment before and during the early phase of treatment with TNFi. To date, no reliable indices have been identified as predictive factors in this regard.
© 2014 Asia Pacific League of Associations for Rheumatology and Wiley Publishing Asia Pty Ltd
IgA rheumatoid factor and response to TNFi
A high titer of rheumatoid factor (RF) is associated with a more aggressive course of the disease with progressive joint damage and extra-articular involvement. This has especially been observed in patients with raised immunoglobulin A (IgA) RF.2 Yet the sequential measurement of IgA RF is not performed routinely in clinical practice. In 2007, Bobbio-Pallavicini reported that high pretreatment levels of IgA RF were associated with poor clinical responses to TNFi.3 Unfortunately, there were no subsequent verification studies to confirm the above finding. A few existing studies have demonstrated a fall in RF levels following control of the disease with both the traditional disease-modifying antirheumatic drugs (DMARDs) and TNFi.4–7 The remaining studies have reported contradictory results.8,9 The overall paucity of data does not allow for convincing conclusions to be made on the definite relationship between the levels of IgA RF and clinical response to TNFi. Studies have consistently pointed toward genetic heterogeneity in RA across different ethnic groups.10,11 There is hardly any data on autoantibody pattern in Asian RA patients treated with TNFi. So far, we have only been armed with lessons from the West in this area of research. Therefore, this study is probably the first in this region to deal with this aspect of RA. Identification of predictors of clinical response to TNFi may potentially pave the way for earlier recognition of nonresponders. This would guide clinicians to make prompt decisions on switching of biologics, not only to arrest disease progression but also to save on costs. The main objective of this study is therefore to elucidate the role of IgA RF in predicting the clinical response to TNFi.
from baseline for those with low to moderate disease activity (DAS28 ≤ 5.1) pre-TNFi, or improvement of more than 1.2 in subjects with high disease activity (DAS28 > 5.1) pre-TNFi. As for subjects who responded to the second TNFi (after failing to respond to one previous TNFi), they were classified as responders. Subjects were tested for quantitative values of IgA, IgM, IgG RF and anti-citrulinated cyclic peptides (antiCCP). Further, all subjects were assessed for DAS28 and Stanford Health Assessment Questionnaire (HAQ) 8-item Disability Index (HAQ-DI).14 The subjects were categorized as high disease activity (DAS28 > 5.1), moderate disease activity (3.2 < DAS28 ≤ 5.1), low disease activity (2.6 < DAS28 ≤ 3.2) and remission (DAS28 ≤ 2.6).13 For the non-responders who had already discontinued the TNFi, the last DAS28 and HAQ-DI scores while on TNFi were taken into account. The DAS28 scores pre-TNFi were gathered from the medical records. Fifteen age-matched controls who were RA patients on conventional DMARDs and in remission were enrolled and tested for IgA, IgM, IgG RF and anti-CCP.
METHODOLOGY
Anti-CCP
Study design and subjects
Anti-CCP were tested using a commercially available ELISA kit (Axis-Shield, Dundee, UK). The upper normal limit (5 U/mL) was determined in accordance with the manufacturer’s recommendations.
We recruited all patients with RA who were ever on TNFi for a minimum duration of 3 months at our centre, Universiti Kebangsaan Malaysia Medical Centre, Kuala Lumpur. The inclusion criteria of this study included: (i) patients who fulfill the American College of Rheumatology (ACR) 2010 criteria for RA;12 (ii) patients ever on TNFi due to intolerance or inadequate response to conventional DMARDs; and (iii) patients with seropositive disease. Based on the European League Against Rheumatism (EULAR) response criteria, subjects were further divided into responders and non-responders.13 Responders were patients who achieved improvement in Disease Activity Index of 28 joints (DAS28) of more than 0.6
International Journal of Rheumatic Diseases 2014; 17: 872–877
Rheumatoid factors The different RF isotypes (IgM, IgA and IgG) were tested using an indirect solid-phase enzyme-linked immunosobent assay (ELISA) involving the binding of Fc fragments of highly purified human IgG to microwells (Dade Behring, Marburg, Germany). The quantitative analysis for IgM, IgG and IgA RF were calibrated as per the kit manual. The procedure was carried out in triplicate. According to the manufacturer’s recommendations, RF concentrations of above 15 IU/mL were considered positive.
Statistical analysis The data analysis was performed using SPSS (Statistical Package for Social Sciences) version 20 (SPSS Inc., Chicago IL, USA). Continuous variables were described as mean standard deviation (SD), or median (range), while the categorical variables as counts and percentages. The analysis of variance (ANOVA) test was used to make a comparison across the three groups, that is, responders, non-responders and controls. Variables which were significant on univariate analyses were
873
R. Sakthiswary et al.
further analyzed using multivariate analysis while controlling for confounders. The association between two continuous variables were determined using bivariate correlation and linear regression analyses. A P-value of < 0.05 was considered significant.
RESULTS Sociodemographic and clinical characteristics of RA patients on TNFi A total of 31 subjects with RA who have received TNFi and 15 controls were enrolled in this study. There were 21 responders and 10 non-responders to TNFi. Table 1 summarizes the characteristics of the RA patients treated with TNFi. The mean age of these patients was 56.90 8.95 years. It took more than a decade on the average (10.45 6.01 years) from time of diagnosis to treatment with TNFi. The most widely prescribed TNFi in our centre was adalimumab (54.8%). A significant proportion of the subjects had moderate to high disease activity (38.7%).
Table 1 Sociodemographic and clinical characteristics of rheumatoid arthritis patients who have been on tumor necrosis factor a inhibitors (TNFi) Parameters Age (years) Gender, n (%) Female Male Ethnicity, n (%) Malays Chinese Indians Disease duration pre-TNFi (years) Duration on TNFi (months) TNF inhibitors, n (%) Infliximab Adalimumab Etanercept Golimumab CRP (mg/dL) ESR (IU/mL) DAS28 Remission–low disease activity (n[%]) Moderate–high disease activity (n[%]) HAQ DI
n = 31 56.90 8.95 30 (96.80) 1 (3.20) 17 (54.80) 5 (16.10) 9 (29.00) 10.45 6.01 15.74 11.58 2 (6.45) 17 (54.80) 10 (32.36) 2 (6.45) 1.08 1.34 65.29 28.03 19 (61.30) 12 (38.70) 0.11 0.22
Data presented as either counts (percentages) or mean SD. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; HAQ-DI; Health Assessment Questionnaire Disability Index; DAS28, 28 jointbased Disease Activity Score.
874
RF isotypes across the study groups Table 2 illustrates the clinical and serological parameters in the three study groups. There was a trend for the non-responders to have higher levels of all isotypes of RF. This pattern was most evident for IgA RF levels, which were significantly higher in the non-responder group compared to the responders and controls (P = 0.001). The significant association between the IgA RF levels and the non-responders was confirmed on multivariate analysis (OR 0.989; 95% CI 0.980–0.999; P = 0.026) with adjustment for confounders, that is, HAQ-DI, duration on biologics and anti-CCP levels. The IgA RF levels demonstrated a significant linear relationship with DAS28 levels while on TNFi (P = 0.005, standardized beta coefficient = 0.490). However, changes in DAS28 (pre-TNFi DAS28–on TNFi DAS28) showed no linear association with the IgA RF levels (P = 0.071, standardized beta coefficient = 0.334) (Fig. 1).
IgA RF versus anti-CCP Although the anti-CCP levels were much higher among the non-responders (P = 0.034) on univariate analysis, this variable was insignificant on multivariate analysis (OR = 0.500; 95% CI 0.979–1.004; P = 0.199). Unlike IgA RF, anti-CCP levels showed no significant association with DAS28 while on TNFi (P = 0.093, standardized beta coefficient = 0.307). Similar results were obtained for anti-CCP and changes in DAS28 with TNFi (P = 0.147, standardized beta coefficient = 0.272).
DISCUSSION The findings of this study depict the clinical importance of IgA RF in predicting response to TNFi among RA patients. Higher levels of IgA RF have a significant association with poor response to this form of biologics. Our results concur with the findings of Bobbio-Pallavicini et al. who reported that the non-responder group had significantly (P = 0.003) higher levels of IgA RF (130.4 U/mL [13.8–276.7 U/mL]) compared to the responder group (24.8 U/mL [10.2–90.8 U/mL]). Further, in keeping with our findings, the above-named study found no significant relationship between the lack of response and the other RF isotypes (IgG RF and IgM RF),although all RF isotypes had a significant drop among the responders after completing 1 year of TNFi therapy.3 However, Lequerre et al. had conflicting results with comparable positive rates (P = 0.35) of IgA RF among their responders and non-responders. It is
International Journal of Rheumatic Diseases 2014; 17: 872–877
IgA rheumatoid factor and response to TNFi
Table 2 Clinical and serological parameters of the responders, non-responders and the controls Parameter
Responders (n = 21)
Age (years) Disease duration pre-TNFi (years) Duration on TNFi (months) DAS28 pre-TNFi DAS28 on TNFi Changes in DAS28 (pre-TNFi–on TNFi) IgA RF (U/mL) IgG RF (U/mL) IgM RF (U/mL) Anti-CCP (U/mL) HAQ-DI
56.90 9.53 19.14 5.17 2.68 2.74 42.33 88.33 84.72 182.54 0.45
9.86 5.55 12.17 1.39 0.81 1.1.3 77.07 92.14 118.53 143.98 0.40
Non-responders (n = 10) 56.89 12.20 8.60 5.41 5.37 0.19 152.64 99.31 98.93 278.48 0.80
6.86 6.75 5.87 0.86 0.69 0.97 100.34 94.64 111.86 46.91 0.47
Controls (n = 15)
P-value
58.64 7.88 6.10 4.10 NA NA NA NA 29.71 72.21 76.35 48.91 88.39 114.51 138.97 145.86 0.11 0.22
0.856 0.061 0.564 < 0.05 < 0.05 0.001 0.756 0.950 0.034 0.002
Data presented as either counts (percentages) or mean SD. TNFi, tumor necrosis factor a inhibitors; IgA RF, immunoglobulin A rheumatoid factor; HAQ-DI, Health Assessment Questionnaire Disability Index; DAS28, 28 joint-based Disease Activity Score; CCP, citrulinated cyclic peptide; RF, rheumatoid factor. Significant P-values are in bold.
2
(a)
R Linear = 0.240
y = 2.97 + 7.39E–3*x
Ig A rheumatoid factor
(b)
2
R Linear = 0.112
y = 2.31 + (–6.35)E–3*x
Figure 1 Relationship between immunoglboulin A rheumatoid factor (IgA RF) levels and (a) Disease Activity Score of 28 joints (DAS28) while on tumor necrosis factor a inhibitors (TNFi) and (b) changes in DAS28 (pre-TNFi–post-TNFi).
International Journal of Rheumatic Diseases 2014; 17: 872–877
noteworthy that unlike Bobbio-Pallavicini et al.3 and our study which compared the average values (mean or median) of IgA RF, Lequerre et al.15 analyzed the frequency of IgA RF positivity. This could partially explain the discrepancy in their findings in this regard. The non-responders were on a shorter duration of TNFi (8.60 5.87 months) compared to the responders (19.14 12.17 months) given that the TNFis were discontinued in the former group of patients usually after 6 months owing to the lack of or suboptimal clinical response. This practice of ours was in accordance with the EULAR 2013 recommendations.16 Although the non-responders had the highest disease duration (12.20 6.75 years), the difference in this parameter among the three groups failed to reach statistical significance. Although IgA RF was first identified in the sera of RA patients as early as 1963,17 the clinical significance of this biomarker is somewhat under-recognized and therefore has not received the attention it deserves in terms of research. Several studies have indicated that IgA RF is more specific for RA compared to the classic IgM RF.18,19 Moreover, it is predictive of a more aggressive course of the disease with early radiographic erosions.20,21 The precise mechanism by which circulating IgA RF influences the disease process remains unclear. The association observed between IgA RF levels and poor response to TNFi does not necessarily imply causation. It is pertinent to mention here that five out of ten of the non-responders showed marked improvement with tocilizumab, an interleukin 6 receptor antagonist.
875
R. Sakthiswary et al.
Data on the response to non-TNFi with high IgA RF is currently not available. The non-responders tended to have higher levels of IgG RF and IgM RF. Lack of statistical significance of these variables could reflect a type II error. Apart from IgA RF, anti-CCP and HAQ-DI were associated with poor response at univariate analysis. Across the studies, there is remarkable inconsistency with regard to the predictive value of anti-CCP for the response to TNFi. While Bobbio-Pallavicini et al.,3 Lequerre et al.15 and De Rycke et al.6 found no association between anti-CCP levels and clinical response to TNFi, Alessandri et al.4 and Atzeni et al.7 demonstrated a decrease in anti-CCP levels in parallel with clinical improvement with TNFi therapy. Diversity of the study populations and methodological variations may partially explain the discordance in the results of the aforementioned studies. With respect to functional disability and response to TNFi, in agreement with our findings, Kristensen et al.22 stated that lower HAQ-DI scores were related to better response to TNFi. This study is not without limitations. The levels of autoantibodies tend to fluctuate in the blood circulation. Hence, the average of serial readings is preferred over a single measurement. Moreover, there are other possible unmeasured confounders that might affect the production of the RF isotypes, such as cigarette smoking. Comprehensive data on the smoking habits of the subjects were not collected. We did not include subjects with seronegative disease (subjects who tested negative for classic IgM RF) and therefore our results may not be generalizable to this subset of RA patients. In conclusion, IgA RF is potentially a novel predictor of response to TNFi in RA patients. Testing for pretreatment IgA RF levels could be a reasonable consideration before commencement of TNFi. Clinicians may consider non-TNFi biologic therapy such as tocilizumab, abatacept and rituximab, as second line or even first line in patients with high levels of the above or static levels, despite TNFi. However, this recommendation should be on the basis of strong supportive evidence by more prospective clinical studies.
ACKNOWLEDGEMENT This study was funded by the APLAR Research Grant Award 2013. RS who is the first and corresponding author is the recipient of the award.
876
REFERENCES 1 Hyrich KL, Watson KD, Silman AJ, Symmons DP (2006) Predictors of response to anti-TNF-alpha therapy among patients with rheumatoid arthritis: results from the British Society for Rheumatology Biologics Register. Rheumatology (Oxford) 45 (12), 1558–65. 2 Teitsson I, Valdimarsson H (1984) Use of monoclonal antibodies and F(ab’)2 enzyme conjugates in ELISA for IgM, IgA and IgG rheumatoid factors. J Immunol Methods 71, 149–61. 3 Bobbio-Pallavicini F, Caporali R, Alpini C et al. (2007) High IgA rheumatoid factor levels are associated with poor clinical response to tumour necrosis factor alpha inhibitors in rheumatoid arthritis. Ann Rheum Dis 66, 302–7. 4 Alessandri C, Bombardieri M, Papa N et al. (2004) Decrease of anti-cyclic citrullinated peptide antibodies and rheumatoid factor following anti-TNFalpha therapy (infliximab) in rheumatoid arthritis is associated with clinical improvement. Ann Rheum Dis 63 (10), 1218–21. 5 Mikuls TR, O’Dell JR, Stoner JA et al. (2004) Association of rheumatoid arthritis treatment response and disease duration with declines in serum levels of IgM rheumatoid factor and anti-cyclic citrullinated peptide antibody. Arthritis Rheum 50 (12), 3776–82. 6 De Rycke L, Verhelst X, Kruithof E et al. (2005) Rheumatoid factor, but not anti-cyclic citrullinated peptide antibodies, is modulated by infliximab treatment in rheumatoid arthritis. Ann Rheum Dis 64 (2), 299–302. 7 Atzeni F, Sarzi-Puttini P, Dell’Acqua D et al. (2006) Adalimumab clinical efficacy is associated with rheumatoid factor and anti-cyclic citrullinated peptide antibody titer reduction: a one-year prospective study. Arthritis Res Ther 8 (1), R3. 8 Yazdani-Biuki B, Stadlmaier E, Mulabecirovic A et al. (2005) Blockade of tumour necrosis factor alpha significantly alters the serum level of IgG- and IgA-rheumatoid factor in patients with rheumatoid arthritis. Ann Rheum Dis 64 (8), 1224–6. 9 Bobbio-Pallavicini F, Alpini C, Caporali R, Avalle S, Bugatti S, Montecucco C (2004) Autoantibody profile in rheumatoid arthritis during long-term infliximab treatment. Arthritis Res Ther 6, R264–72. 10 Taneja V, Mehra NK, Kailash S, Anand C, Malaviya AN (1992) Protective & risk DR phenotypes in Asian Indian patients with rheumatoid arthritis. Indian J Med Res 96, 16–23. 11 Barton A, Bowes J, Eyre S et al. (2004) A functional haplotype of the PADI4 gene associated with rheumatoid arthritis in a Japanese population is not associated in a United Kingdom population. Arthritis Rheum 50 (4), 1117–21. 12 Aletaha D, Neogi T, Silman AJ et al. (2010) 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 69 (9), 1580–8.
International Journal of Rheumatic Diseases 2014; 17: 872–877
IgA rheumatoid factor and response to TNFi
13 Fransen J, van Riel PL (2005) The Disease Activity Score and the EULAR response criteria. Clin Exp Rheumatol 23 (Suppl. 39), S93–9. 14 Thompson PW, Pegley FS (1191) A comparison of disability measured by the Stanford Health Assessment Questionnaire disability scales (HAQ) in male and female rheumatoid outpatients. Br J Rheumatol 30(4), 298–300. 15 Lequerre T, Jouen F, Brazier M et al. (2007) Autoantibodies, metalloproteinases and bone markers in rheumatoid arthritis patients are unable to predict their responses to infliximab. Rheumatology (Oxford) 46 (3), 446–53. 16 Smolen JS, Landewe R, Breedveld FC et al. (2014) EULAR recommendations for the management of rheumatoid arthritis with synthetic and biological disease-modifying antirheumatic drugs: 2013 update. Ann Rheum Dis 73 (3), 492–509. 17 Teitsson I (1988) IgA rheumatoid factor as predictor of disease activity. Scand J Rheumatol Suppl 75, 233–7. 18 Bas S, Perneger TV, Kunzle E, Vischer TL (2002) Comparative study of different enzyme immunoassays for measure-
International Journal of Rheumatic Diseases 2014; 17: 872–877
19
20
21
22
ment of IgM and IgA rheumatoid factors. Ann Rheum Dis 61, 505–10. Swedler W, Wallman J, Froelich CJ, Teodorescu M (1997) Routine measurement of IgM, IgG, and IgA rheumatoid factors: high sensitivity, specificity, and predictive value for rheumatoid arthritis. J Rheumatol 24 (6), 1037–44. Bas S, Genevay S, Meyer O, Gabay C (2003) Anti-cyclic citrullinated peptide antibodies, IgM and IgA rheumatoid factors in the diagnosis and prognosis of rheumatoid arthritis. Rheumatology (Oxford) 42 (5), 677–80. Berglin E, Johansson T, Sundin U et al. (2006) Radiological outcome in rheumatoid arthritis is predicted by presence of antibodies against cyclic citrullinated peptide before and at disease onset, and by IgA-RF at disease onset. Ann Rheum Dis 65 (4), 453–8. Kristensen LE, Kapetanovic MC, Gulfe A, Soderlin M, Saxne T, Geborek P (2008) Predictors of response to anti-TNF therapy according to ACR and EULAR criteria in patients with established RA: results from the South Swedish Arthritis Treatment Group Register. Rheumatology (Oxford) 47 (4), 495–9.
877
Copyright of International Journal of Rheumatic Diseases is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
