Serodiagnosis of hepatitis C virus. Effect of new evaluation of cutoff values for enzyme-linked immunosorbent assay in Brazilian patients.

With the goal of reducing false-positive results in enzyme-linked immunosorbent assay (ELISA) serodiagnosis of hepatitis C virus in clinical practice, a study was undertaken to establish better cutoff values. We examined 277 serum samples from patients with hepatitis (non-A, non-B; B; autoimmune); subjects with antinuclear antibodies or rheumatoid factor, anticytomegalovirus or Epstein-Barr virus IgG or IgM antibodies, or parasitic disease (Chagas disease, leishmaniasis); and healthy volunteers. Concordant positive results in 2 different immunoblot assays in 250 samples were taken as indicative of true-positive, and when negative, of absence of infection. Reactivity in 3 ELISA tests were evaluated for the manufacturer recommended cutoff (CO) and for 2CO, 3CO, and 4CO; and corresponding sensitivity and specificity were calculated for single or combined pairs of ELISA tests. Although CO is adequate for blood bank screening, because it provides maximal sensitivity, the frequently observed false-positive results could be significantly reduced by increasing the cutoff value to 2CO, with no significant loss in sensitivity either in relation to pairs of immunoenzymatic tests or to a single ELISA.