Antinuclear antibodies (ANAs) have become routine laboratory parameters in clinical hospitals. However, ANA testing by indirect immunofluorescence (IIF) assays is not an automated laboratory test. Efforts are being made to develop easy and semi- or automated methods to screen for ANAs. We evaluated the clinical performance of a new ELISA developed to screen for connective tissue disease related ANAs. The presence of serum ANA was studied with a commercial ELISA (Varelisa ANA CTD Screen) in 472 patients (202 SLE, 41 Sjögren syndrome, 11 CREST, 59 rheumatoid arthritis, 30 seronegative spondyloarthropaties, 77 inflammatory bowel disease, 13 reactive arthritis, 11 giant cell arteritis, 28 ankylosing spondilitis). A hundred and five sera from healthy subjects were used as controls. Receiver operator characteristics (ROC) analysis was carried out in order to optimize the cutoff. At target specificities of 80/90%, sensitivities of 80.8/ 73.9% were achieved. At the manufacturer's cutoff (ratio >or=1.0) sensitivity/specificity of 71.4/91.2% was found. At that cutoff, a positive likelihood ratio of 8.11 was found. For helping in the diagnosis of connective tissue diseases a test employing a subset of the most prevalent specificities reveals a good compromise as indicated by a high-positive likelihood ratio. However, the presence of ANAs in pathologies other than connective tissue diseases, such as SLE or Sjögren syndrome, may be of clinical significance as well. In these cases an IIF assay test is still mandatory, especially in autoimmune laboratories.