• DOI:10.7326/0003-4819-146-11-200706050-00008
• Corpus ID: 6640507

Meta-analysis: Diagnostic Accuracy of AntiCyclic Citrullinated Peptide Antibody and Rheumatoid Factor for Rheumatoid Arthritis

@article{Nishimura2007MetaanalysisDA,
  title={Meta-analysis: Diagnostic Accuracy of AntiCyclic Citrullinated Peptide Antibody and Rheumatoid Factor for Rheumatoid Arthritis},
  author={Kunihiro Nishimura and Daisuke Sugiyama and Yoshinori Kogata and Goh Tsuji and Takashi Nakazawa and Seiji Kawano and Katsuyasu Saigo and Akio Morinobu and Masahiro Koshiba and Karen M Kuntz and Isao Kamae and Shunichi Kumagai},
  journal={Annals of Internal Medicine},
  year={2007},
  volume={146},
  pages={797-808}
}

• K. Nishimura, D. Sugiyama, +9 authors S. Kumagai
• Published 5 June 2007
• Medicine
• Annals of Internal Medicine

Context Are autoantibodies against cyclic citrullinated peptide (CCP) better serum markers for rheumatoid arthritis than rheumatoid factor (RF)? Contribution This meta-analysis of 86 studies found that the positive likelihood ratio for anti-CCP antibody was greater than that for IgM RF for identifying patients with rheumatoid arthritis (12.5 vs. 4.9). Sensitivity was similar for the 2 tests, although specificity of anti-CCP antibody (95%) was higher than specificity of IgM RF (85%). Cautions Fewer studies evaluated anti-CCP antibody than RF. There was possible publication bias for reporting positive findings regarding anti-CCP antibody. Implication Anti-CCP antibody positivity seems to be more specific than IgM RF positivity for identifying patients with rheumatoid arthritis. The Editors Rheumatoid arthritis is the most common autoimmune disease, affecting approximately 1% of the world's population (1). It causes persistent synovitis, pain, joint destruction, and functional disability. Because irreversible joint destruction can be prevented by intervention during the first months of disease, early diagnosis of rheumatoid arthritis is important (24). Rheumatoid factor (RF) is an antibody directed against the Fc region of IgG that has been used as a diagnostic marker for rheumatoid arthritis. However, it is nonspecific and may be present in healthy elderly persons or in patients with other autoimmune and infectious diseases (5). Other rheumatoid arthritisassociated autoantibodies known to be specific for rheumatoid arthritis include antiperinuclear factor and antikeratin antibodies (6, 7). Because of rigorous technical requirements for their detection, antiperinuclear factor and antikeratin antibodies have never been widely used as markers for rheumatoid arthritis, despite their high specificity. The epitopes of their antigens are arginyl residues citrullinated by peptidyl arginine deiminase (810). Some enzyme-linked immunosorbent assays (ELISAs) use linear citrulline-containing peptides that have similar sensitivity to and higher specificity than RF for diagnosing rheumatoid arthritis (11). To improve sensitivity, assays that use cyclic citrullinated peptide (CCP) were developed to detect anti-CCP antibody (12). In this systematic review, we summarize published data on the sensitivity, specificity, and likelihood ratios of RF and anti-CCP antibodies for diagnosing rheumatoid arthritis. We also summarize results of studies that assessed the associations of these markers with development and radiographic progression of rheumatoid arthritis. Methods Data Sources and Searches We developed a protocol for the review and followed standard reporting guidelines (13, 14). We searched MEDLINE for studies published in any language through September 2006 that examined autoantibodies against citrullinated proteins, rheumatoid factor, or both for the diagnosis of rheumatoid arthritis. Our searches (available on request) were based on combinations of the following index terms: rheumatoid arthritis, antiperinuclear factor, antikeratin antibody, citrullinated protein, anticyclic citrullinated peptide, rheumatoid factor, sensitivity, specificity, mass screening, predictive value of tests, receiver-operating characteristic curve, and accuracy. We also reviewed reference lists of retrieved studies and review articles. Study Selection Two reviewers independently scanned abstracts that met the inclusion criteria. We included studies that evaluated the utility of assaying anti-CCP antibody or RF for diagnosis of known or suspected rheumatoid arthritis, enrolled at least 10 participants, were published after 1987, and provided enough data to allow calculation of sensitivity and specificity for diagnosis of rheumatoid arthritis. We used the 1987 revised American College of Rheumatology (ACR) criteria as the reference standard of rheumatoid arthritis (15). In general, we regarded reports of patients with symptom duration of less than 1 year as studies of early rheumatoid arthritis, although we also used the researchers' definitions of early rheumatoid arthritis. Data Extraction and Study Quality Assessment We extracted data by using a standard form that included the demographic characteristics of the participants, inclusion and exclusion criteria, number of participants who were evaluated with the index test, and methods of antibody testing. Two investigators independently assessed the design of the studies by using previously developed quality criteria for studies of diagnostic tests (1618). These assessments addressed the technical quality of the anti-CCP antibody test, technical quality of the RF test, application of the reference or index test, blinding of observers, description of the study sample, and cohort assembly. We used coefficients to examine interrater agreement for our initial overall quality score (19) and resolved any item discrepancies through discussion. Data Analysis We used a random-effects model to combine estimates of sensitivity, specificity, and positive and negative likelihood ratios (1921). We planned analyses that were stratified by generation of anti-CCP antibody assay (first [anti-CCP1] second [anti-CCP2]) and by RF subtype (IgA, IgG, and IgM). We analyzed subgroups of relevant studies that included patients with early rheumatoid arthritis and that evaluated combination testing for anti-CCP antibody and RF. We conducted a stratified analysis for different threshold and measurement methods when we suspected heterogeneity among studies. We also conducted threshold analyses and metaregression to assess whether the threshold effect and heterogeneity among studies existed (22). We examined funnel plots for diagnostic odds ratios to explore the possibility of publication bias (23). For analyses, we used MetaDiSc, version 1.1.4 (Hospital Universitario Ramn y Cajal, Madrid, Spain); Stata, version 8.2 (Stata Corp., College Station, Texas); and R, version 2.21 (R Foundation for Statistical Computing, Vienna, Austria). Role of the Funding Sources This study was supported in part by a Grant-in-Aid for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and a grant for Research on Allergic Disease and Immunology from the Ministry of Health, Labor and Welfare, Japan. The funding sources had no role in the collection, analysis, or interpretation of the data or in the decision to submit the manuscript for publication. Results Search Results and Characteristics of Studies We identified 302 reports, of which 86 met the inclusion criteria (11, 12, 24106) (Appendix Figure). Thirty-seven studies in 14949 patients (11, 12, 24, 26, 2938, 4042, 44, 45, 47, 48, 50, 52, 54, 56, 58, 6062, 64, 66, 67, 70, 74, 76, 97, 99, 100) reported on the diagnostic accuracy of anti-CCP antibody, whereas 50 studies in 15286 patients 12, 24, 27, 29, 30, 3237, 39, 40, 4244, 47, 48, 50, 52, 54, 55, 6062, 64, 66, 70, 7274, 76, 8085, 8898, 100) reported on the diagnostic accuracy of RF. Appendix Figure. Study flow diagram. Appendix Table 1 (11, 24, 26, 2938, 4042, 44, 45, 47, 48, 50, 52, 54, 56, 58, 6062, 6467, 74, 76, 97, 99, 100) and Appendix Table 2 (12, 24, 27, 29, 3335, 37, 3943, 45, 47, 48, 52, 65, 66, 7274, 76, 80, 81, 88, 9092, 9498, 100) summarize the characteristics of the included studies. In anti-CCP antibody and IgM RF studies, respectively, the median numbers of participants were 404 and 226, their median ages were 57 years and 53 years, and the median proportions of women were 59% and 68%. Studies of anti-CCP antibody that were published after 2000 usually addressed anti-CCP2 assays. Characteristics of control groups varied. Among the anti-CCP antibody studies, 5 used patients with undifferentiated arthritis, 13 used patients with other rheumatic diseases, 1 used healthy persons, 1 used hepatitis C carriers, and 17 used a mix of healthy persons and patients with other diseases. Among the IgM RF studies, 5 used patients with undifferentiated arthritis, 16 used patients with other rheumatic diseases, 2 used healthy persons, 1 used hepatitis C carriers, 1 used patients with polymyalgia rheumatica, and 22 used a mix of healthy persons and patients with other diseases. Three studies did not report details on the control group. Appendix Table 1. Characteristics of Studies of Autoantibodies against Cyclic Citrullinated Peptide* Appendix Table 2. Characteristics of Studies of Rheumatoid Factor* Study Quality Only 1 study satisfied all criteria on our quality checklist. Twenty-two studies (30%) met at least 70% of the criteria, and 9 studies (10%) met fewer than 50% of the criteria. The coefficient for interrater agreement was 0.92 on the quality score. Most studies adequately described the technical aspects of assaying anti-CCP antibody and RF. In 86% (32 of 37) of anti-CCP antibody studies and 82% (41 of 50) of RF studies, the 1987 revised ACR criteria were used as the reference standard for rheumatoid arthritis. Most studies did not explicitly mention blinding of investigators to the clinical assessment or to the reference standard. Most studies (90%) enrolled patients with known or suspected rheumatoid arthritis. Characteristics of these patients were fully described in just over half of the studies. Enrollment was prospective in 18 of 37 anti-CCP antibody studies and 25 of 50 RF studies. Studies of RF showed a wide range of sensitivities and specificities (Appendix Table 1). One study (35) reported very low sensitivity and specificity. In this study, 57% of control patients had conditions that can present with RF-positive arthritis (primarily the Sjgren syndrome or Wegener granulomatosis). Laboratory techniques for measuring RF varied across studies. Fifteen studies used nephelometry, 16 used latex agglutination, and 16 used ELISA. Twenty-two studies used less than 20 U/mL as the cutoff value for negative test results, 11 used less than 40 U/mL as the cutoff value, and 17 did not report cutoff valu