Objective: To evaluate a commercial microtiter enzyme immunoassay for antinuclear antibodies (ANA) by comparing the results of tests performed with this assay to an established indirect immunofluorescence method performed on human epitheliod cell substrate slides.
Design: Both analytical methods were used to test for the presence and levels of ANA in stored sera from 313 patients previously shown to have detectable ANA and from 102 healthy control subjects. Follow-up tests for specific autoantibodies (anti-dsDNA antibodies and antibodies to extractable nuclear antigens [ENA]) were performed on all sera from patients. The medical histories of all patients were reviewed to determine the presence of systemic rheumatic diseases (SRDs). Different cut-off levels of positivity were examined to determine the sensitivity and predictive values of positive results on the enzyme immunoassay for detecting patients with SRDs or sera with positive tests for specific autoantibodies.
Results: Among patients with clinically diagnosed SRDs (n = 197), the enzyme immunoassay was positive for ANA (> or =1 U) in 100% and the indirect immunofluorescence method was positive (titer > or =40) in 95.4% of cases. Among ANA-positive patients with no SRDs (n = 116), testing by enzyme immunoassay and indirect immunofluorescence yielded positive results in 97.6% and 75.6% of cases, respectively. Among healthy control subjects, each of the two methods was positive in 15% of cases. As expected, most patients with SRDs had higher levels of ANA than did ANA-positive patients with other clinical diagnoses. A cut-off level of > or =3 U on the enzyme immunoassay correctly classified 77% of patients with a SRD as "positive" and 88% of patients with other clinical diagnoses as "negative." The probability of detecting a positive result for specific autoantibodies on second-order testing increased directly with the level of ANA. A cut-off level of > or =3 U had a sensitivity of 92% for identifying sera with positive specific autoantibodies, and results > or =3 U had a predictive value of 52% for a positive second-order test result.
Conclusion: Enzyme immunoassay is substantially equivalent to indirect immunofluorescence for detecting clinically important ANA. Cut-off levels for positive results on the enzyme immunoassay can be established that optimize the usefulness of this method in diagnostic algorithms for specific autoantibodies.