Objective: To compare the measurement of antinuclear antibodies (ANA) by immunofluorescence and by 6 different commercial enzyme-linked immunosorbent assays (ELISAs) in clearly defined patient groups and serum samples.
Methods: Serum samples were derived from 3 sources: 1) patients with a known clinical diagnosis of systemic lupus erythematosus (SLE) (group 1), 2) sera with known monospecific ANA reactivity (group 2), and 3) sera from consecutive patients for whom ANA testing had been ordered (group 3). Each of these sera was tested for the presence of ANA by immunofluorescence on HEp-2 cells, and by using each of 6 commercially available ELISA ANA kits.
Results: In patients with known clinical SLE, 88% were ANA positive by immunofluorescence. Positivity with the different ELISAs ranged from 62% to 90%. While most ELISAs successfully detected antibodies to SS-A, RNP, and Sm, there were significant differences between assays in the detection of anticentromere antibodies and anti-DNA. Measurement of ANA in consecutive patients for whom an ANA test was requested showed that, generally, those assays with high sensitivity for detection of SLE had a high false-positive rate, whereas those assays with a low false-positive rate failed to identify some patients with a clinical diagnosis of SLE.
Conclusion: There are significant differences in the detection of ANA by immunofluorescence and different ELISA kits. Agreement between different assays is generally marginal. When ordering and interpreting an ANA test, the clinician must be familiar with the specific assay being used to measure ANA and the differences between the various ANA assays.