Incidence and specificity of interference in two-site immunoassays.

Using a modified "two-site" immunoradiometric assay, termed an "interference assay," we have demonstrated the occurrence of non-analyte antibody-binding substances in approximately 40% of serum samples. These substances multivalently bind antibodies from any of several species at a site other than the antigen-binding site. Using a two-site immunoradiometric assay for human choriogonadotropin, we have investigated their effect on analyte quantification. In this system these antibody-binding substances mimic the presence of analyte; when analyte is actually present, they can also cause over- or underestimates. Addition of excess nonspecific immunoglobulin from an appropriate species eliminates this interference. However, the amount of nonspecific immunoglobulin added to an assay system may not always be enough to block interference in all samples.