The competition ELISA method described in 1985 by Friguet and colleagues has frequently been used to determine dissociation constants (K-D) of antigen-antibody reactions. Subsequently Stevens suggested a correction for the bivalency of IgG. In either case, the K-D is assumed to vary only with the composition of competing fluid-phase antigen and consequently should not be affected by the solid-phase antigen. However, during the course of experiments defining epitopes to the V3 loop of human immunodeficiency virus type 1, both the composition and density of solid-phase antigen were capable of significantly influencing the calculated values. With one solid-phase antigen, the calculated K-D was a function of antigen density. With a second solid-phase antigen, the calculated K-D did not vary with the density. With the latter antigen or with low densities of the former, K(D)s calculated using the Stevens correction for bivalency were close to the estimate obtained by a radiolabeled peptide precipitation assay. Accordingly, since the density and composition of solid-phase antigen may alter the K-D calculated from competition ELISAs, such estimates should be confirmed by a more readily interpretable immunological method.
