The objective of this study was to determine whether immunologic anhydride-induced respiratory disease could be predicted on the basis of the level of specific immunoglobulin E (IgE) or immunoglobulin G (IgG) antibody. Eight-one anhydride-exposed employees in one plant were studied. Fourteen had disease and 67 did not. Immunologic studies were performed by enzyme-linked immunosorbent assay and expressed as titers. When optimal discriminant analysis was used, IgE < 1:5 and IgG < or = 1:10 were found to be the optimal titers for separating employees with and without immunologic respiratory disease caused by anhydrides. When IgG < or = 1:10 was used, 62 of 81 workers were correctly classified; the sensitivity was 100%, the positive predictive value was 45%, the specificity was 75%, and the negative predictive value was 100%. When IgE < 1:5 was used, 73 of 81 workers were correctly classified; the sensitivity was 86%, the positive predictive value was 67%, the specificity was 91%, and the negative predictive value was 97%. In conclusion, anhydride disease status can be predicted on the basis of specific IgG or IgE antibody level.