Monovariate and multivariate analyses including logistic regression were performed to determine associations among predicting variables [age, gender, immunofluorescence pattern, and anti-nuclear antibody (ANA) titer] and anti-extractable nuclear antigen (ENA) and -dsDNA antibodies (Abs) in 1089 patients with positive fluorescent ANA (FANA) test results. Samples with high titer ANAs had an increased frequency of anti-ENA and -dsDNA Abs. The receiver operating (ROC) curves of the ANA titer for anti-ENA Abs had a larger under the curve area compared to the ROC curve for anti-dsDNA Abs, indicating that ANA titer is better for predicting anti-ENA Abs than anti-dsDNA Abs. There was no relation noticed between immunofluorescence patterns and anti-ENA and -dsDNA Abs except an increased frequency of anti-dsDNA Abs found in samples with a homogeneous pattern. Probability calculations on the basis of the ANA pattern and the titer showed that samples with low titer ANAs (1:160 or less) had low probabilities for anti-ENA Abs (0.002-0.009) regardless of immunofluorescence patterns. However, samples with a homogeneous pattern at any titers including low titers had high probabilities for anti-dsDNA Abs. A decreased frequency of anti-dsDNA Abs as measured by Crithidia assay was noticed in samples from patients aged 50 or older. In contrast, no association was noticed between age and anti-ENA Abs. There was no female preponderance found in the presence of anti-ENA and -dsDNA Abs. In conclusion, our study shows that the ANA titer but not the immunofluorescence pattern is useful in predicting anti-ENA Abs. In contrast, both the ANA titer and the immunofluorescence pattern help in predicting anti-dsDNA Abs. Samples with low titer ANAs (1:160 or less) may not need a further test for anti-ENA Abs unless an ANA-associated disease is highly suspected. However, a test for anti-dsDNA Abs should be considered in samples with a homogeneous pattern at any titer including low titers.