Antibody-Based Prediction Rules for Connective Tissue Disease: Analysis of 12,555 Patients with Antinuclear Antibody Testing

Ryo Rokutanda¹, Mitsumasa Kishimoto¹, Yasuharu Tokuda², Ken-ichi Yamaguchi¹, Hisanori Shimizu¹, Yasuhiro Suyama¹, Yuri Ohara¹, Yoichiro Haji¹, Chisun Min¹, Akira Takeda¹, Yukio Matsui³ and Masato Okada¹, ¹Division of Allergy and Rheumatology, St. Luke's International Hospital, Tokyo, Japan, ²Institute of Clinical Medicine, University of Tsukuba, Ibaraki, Japan, ³Immuno-Rheumatology Center, St. Luke's International Hospital, Tokyo, Japan

Meeting: 2012 ACR/ARHP Annual Meeting

Keywords: Antibodies, Antinuclear antibodies (ANA), diagnosis, mixed connective tissue disease (MCTD) and systemic lupus erythematosus (SLE)

Session Information

Title: Epidemiology and Health Services Research: Epidemiology and Outcomes of Rheumatic Disease II

Session Type: Abstract Submissions (ACR)

Background/Purpose: Antinuclear antibody (ANA) is widely used as a screening test for connective tissue diseases (CTDs). The sensitivity of this test is high in 3 CTDs–systemic lupus erythematosus (SLE), systemic sclerosis (SSc), and mixed connective tissue disease (MCTD). ANA-associated specific autoantibodies are also identifiable in 3 additional CTDs including polymyositis and dermatomyositis (PM/DM), as well as primary Sjögren syndrome (SS). However, appropriate cutoff points for ANA titers or in combination with other CTD-specific antibodies remains unclear.

Methods: We reviewed records for all patients (n=12,555) with ANA immunofluorescence assay testing from January 2003 to June 2012. Among 3302 patients with positive ANA (³a40), 94 were excluded due to lack of information about ANA titers. Data collection also included specific CTD diagnoses, ANA titers, and results of specific antibody tests for DNA, Sm, U1RNP, Ro, La, centromere, Scl-70, and RNA polymerase III. We calculated sensitivity, specificity, and positive (PPV) and negative (NPV) predictive values for diagnosing each CTD at different ANA cutoffs. We then performed classification and regression tree (CART) analysis to develop prediction rules for each CTD. Potential predictors included ANA titers, ANA staining patterns, and specific antibody test results.

Results: Of 12,461 patients with ANA testing, 665 patients (5.3%) were diagnosed as having at least one of the 6 CTDs of interest. Area under the ROC curves of ANA testing were 0.943, 0.977, 0.926, 0.840, and 0.740 for SLE, MCTD, SSc, SS, and PM/DM, respectively. The sensitivities of ANA (³a40) for SLE, SSC, MCTD, SS, PM/DM were 98.6%, 95.9%, 100%, 83.0%, and 69.6%, while PPV was 6.7%, 5.1%, 0.7%, 6.3, and 1.7%, respectively. Prediction rules developed by CART were as follows: (1) ANA (³a40) with anti-DNA antibody (PPV: 70.3% for SLE), (2) ANA (³a40) with anti-U1RNP antibody (PPV: 26.8% for MCTD), (3) ANA (³a320) with anti-centromere antibody (PPV: 44.7% for SSc), and (4) ANA (³a320) and anti-Ro without anti-DNA (PPV: 49% for SS).

Conclusion: ANA testing demonstrates substantial sensitivity, but has low PPV for CTDs. Nonetheless, in combination with other antibody testing, our prediction rules reveal a high specificity and PPV for diagnosing CTDs.

Disclosure:

R. Rokutanda,
None;

M. Kishimoto,
None;

Y. Tokuda,
None;

K. I. Yamaguchi,
None;

H. Shimizu,
None;

Y. Suyama,
None;

Y. Ohara,
None;

Y. Haji,
None;

C. Min,
None;

A. Takeda,
None;

Y. Matsui,
None;

M. Okada,
None.

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/antibody-based-prediction-rules-for-connective-tissue-disease-analysis-of-12555-patients-with-antinuclear-antibody-testing/