A statistical approach for the analysis of complex samples by immunoassay is proposed in this article. Two enzyme-linked immunosorbent assays (ELISAs), one of them in the conjugate-coated format and the other in the antibody-coated format, were evaluated for their suitability to the analysis of mixtures of three organodithiophosphorus pesticides: azinphos-methyl, azinphos-ethyl and phosmet. It was found that the apparent affinity of the antibody to each analyte changed in the presence of a cross-reacting compound in the antibody-coated ELISA format, but not when the conjugate-coated ELISA format was used. The assays were thereafter applied to the analysis of mixtures of the three recognized pesticides. With the conjugate-coated ELISA format, accurate and precise determinations of mixtures could be performed if an azinphos-methyl standard curve was employed, with recoveries between 71% and 130% and with coefficients of variation lower than 12.7%. Neither accurate nor precise measurements could be accomplished with the enzyme immunoassay using the antibody-coated ELISA format, independently of the standard curve used. It is thought that the study presented here will have applicability in a variety of cases where the analytical goal is semiquantitative screening based on the total quantity of an unknown mixture of related compounds.
