600 - Proposal for a Reliable and Feasible Algorithm for the Identification of Antinuclear Antibodies

Sunday, November 6, 2011: 9:00 AM-6:00 PM
Hall F2 - Poster Hall (McCormick Place West)
Kaline M.C. Pereira, Alessandra Dellavance and Luis E. C. Andrade, Universidade Federal de São Paulo and Fleury Health and Medicine Laboratories, Sao Paulo, Brazil
Presentation Number: 600

Background/Purpose: Antibodies to Sm, RNP, SS-A/Ro, SS-B/La, Scl-70 and Jo-1 (known as extractable nuclear antigens - ENA) are particularly relevant in the diagnosis of systemic autoimmune rheumatic diseases (SARD). The known clinical associations of anti-ENA antibodies were originally defined on the basis of double immunodiffusion (DID) assay, which is cost-effective and extremely specific, but has limited sensitivity, is time-consuming, needs qualified personnel and is not appropriate for automation. Solid phase assays (ELISA and similar), in turn, are highly sensitive and ready for automation, but show variable diagnostic performance and often yield unexpected positive results. The present study analyzed the diagnostic performance of six ELISA kits and DID for determination of anti-ENA antibodies and proposes an algorithm combining ELISA and DID for efficient high throughput.

Method: 290 serum samples from patients with well characterized autoimmune (ACR criteria) and non-autoimmune rheumatic diseases, chronic viral hepatitis, and healthy controls were tested for anti-ENA antibodies by DID and six ELISA kits according to manufacturer’s instructions. Clinical diagnosis was the gold standard for determining the diagnostic performance of each assay. An algorithm combining a screening step by ELISA and a confirmatory step by DID was applied to 16,485 samples for which anti-ENA antibodies had been ordered in a large clinical laboratory.

Result: The sensitivity of ELISA tests for the detection of anti-ENA was higher than DID for all ELISA kits. The sensitivity of anti-Sm for the diagnosis of systemic lupus erythematosus (SLE) was 6.7% in DID and ranged from 18.6% to 44.2% in ELISA kits. The same was observed with anti-Scl-70 regarding the diagnosis of systemic sclerosis (SSc) (13.8% sensitivity by DID; 28.6 to 37.9% by ELISA) and anti-Jo-1 for polymyositis (PM) (5.9% sensitivity by DID; 9.5 to 17.7% by ELISA). In contrast the positive predictive value (PPV) of antibodies to Sm, Scl-70, and Jo-1, was 100% for SLE, SSc and PM, respectively, when assayed by DID, but had an average of 52.2% (ranging from 0 to 100%), when assayed by ELISA. ELISA, but not DID, yielded several anti-ENA positive results in non-SARD patients, such as those with chronic hepatitis, ankylosing spondylitis and osteoarthritis, and even in healthy individuals. No sample with a positive result in DID was negative in ELISA, what allowed the proposal of a two-step algorithm based on ELISA screening and confirmation by DID. With such algorithm only 17% of 16,485 samples were positive by ELISA and required confirmation by DID. In comparison to the standard one-step DID operation, this strategy resulted in a 60% reduction in the operation cost, and a significant reduction in the turn-around time from 96h to 24h in over 80% of the samples.

Conclusion: The proposed algorithm, whose effectiveness was demonstrated in the autoantibody routine of a large clinical laboratory, takes advantage of the best features of each method, ie, speed, automation and higher sensitivity of the EIA and the high specificity and positive predictive value of DID, allowing the report of highly reliable positive results with genuine value to the clinician.

Keywords: Enzyme-Linked Immunoabsorbant Assays (ELISA), antinuclear antibodies (ANA), autoantibodies, autoimmune diseases and diagnosis

Disclosure: K. M. C. Pereira, None; A. Dellavance, None; L. E. C. Andrade, None.