We evaluated two enzyme immunoassays (EIA) for detection of antinuclear ant
ibodies (ANA) which became recently available end which were designed for a
pplication on a fully automated system; (i) the EIA-ANA screen kit from Sig
ma Diagnostics (St. Louis, MO, USA) applied on APTUS(TM), an automated EIA
analyser from Sigma Diagnostics and (ii) the Cobas(R) Core Hep-2 EIA-ANA as
say applied on the fully automated Cobas Core immunochemistry analyzer from
Roche Diagnostics (Basel, Switzerland). The evaluation was done by an anal
ytic comparison of the automated systems to an established indirect immunof
luorescence (IF) method performed on SSA transfected human epitheloid cell
substrate slides. Three hundred and thirty six samples were tested with the
Sigma EIA-ANA assay and 603 samples with the Cobas(R) Core Hep-P EIA-ANA a
ssay. For both EIA systems, there was a trend of generally increasing signa
l from the assay system with increasing IF-ANA titers. With an IF-ANA refer
ence range of < 1:160, concordance between Sigma(R) EIA-ANA and IF-ANA was
86% and concordance between Roche(R) EIA-ANA and IF-ANA was 85%. When compa
red to IF-ANA with a reference range of < 1:160, sensitivity and specificit
y of the EIA-ANA screen was 0.65 and 0.92, respectively, for Sigma(R) EIA-A
NA and 0.61 and 0.91, respectively, for the Cobas(R) Core assay. The low se
nsitivity observed with both methods is a major concern. Sigma EIA-ANA scre
en revealed the presence of autoantibodies in 23 of the 29 samples containi
ng antibodies to extractable nuclear antigens (ENA) and/or double stranded
DNA and the Cobas(R) Core Hep-P EIA-ANA assay revealed the presence of auto
antibodies in 40 of the 43 samples containing antibodies to ENA and/or doub
le stranded DNA.