The COBAS Core HEp2 ANA enzyme immune assay (EIA) was evaluated in a precision and a clinical sample study in comparison to indirect immunofluorescence assay (IFA) on HEp2-cells. In the precision study the COBAS Core EIA yielded intraassay coefficient variations (CVs) mostly below 9%, and interassay CVs between 4.7% and 10.4%. When comparing the COBAS Core EIA to IFA, the results corresponded well in healthy subjects, systemic lupus erythematosus, mixed connective tissue disease and rheumatoid arthritis. In the case of Sjogren's syndrome and scleroderma patients the COBAS Core EIA yielded a lower rate of positive results compared to IFA. This discrepancy may be explained by the lack of detection of autoantibodies to nuclear antigens that can be identified only by IFA due to their compartmentalization and higher localized antigen density in HEp2 cells. The discrepancies in the group of dermato/polymyositis patients are due to the fact that the EIA contains mainly nuclear antigens and was able to detect only antibodies against the cytoplasmic Jo1 antigen that was added to the HEp2 nuclear extract. Routine sera were also evaluated; good agreement was found in sera from patients attending tertiary reference centres for autoimmune diseases but a higher number of discrepancies was reported in sera from unselected populations.