Modeling the Contributions of Cross-Reactant Chemicals in Immunoassays Available to Purchase

Immunoassay techniques often involve reagents that react with constituents of samples other than the target compound. The phenomenon is called cross-reactivity. This paper describes a study with three objectives aimed at overcoming some of the problems related to crossreactivity: (1) to investigate the effects of cross-reactivity on the accuracy and precision of analytical measurements; (2) to develop mathematical techniques to “deconvolute” analytical measurements that contain cross-reactions; (3) to define the minimum analytical and statistical experimental design required to implement the mathematical techniques developed.

It was found that cross-reactivity affects method accuracy, but not necessarily precision; an additive bias occurs because what is measured is actually the sum of the reactions of the target analyte plus nontarget cross-reactions. The study was successful in developing a stochastic model by which the desired “deconvolution” could always be accomplished under certain experimental conditions. The model and estimation procedures are algorithmically simple and extendable to any number of cross-reactants. The minimum experimental design requires that (1) the sample mixture be qualitatively prescreened for the identity of potential cross-reactants, (2) associative rate constants be determined for the target analyte and cross-reactants, (3) the associative rate constants not be equal for the specific reagent used, and (4) the analytical runs produce a time series of measurements over the period during which binding occurs.