Elucidation of non-parallel EIA curves.

ABSTRACT

Quantitative determinations by EIA can be only obtained by reverse regression when linear portions of sample and standard curves are parallel. However, analysis of complex biological fluids often yields sigmoid curves displaying lower slopes, thus invalidating any quantitative interpretation. We hypothesized that this phenomenon was due to a competition effect between the target (for example an antigen) and related molecules for the binding sites (for example a capture antibody) immobilized onto the solid phase. This has been confirmed experimentally using various target-to-competitor ratios and formulated as a mathematical model. The slope decrease in target detection was related to the proportion of competitor, not in a linear, but in an exponential manner. This mathematical model has been computerized and can be used to correct aberrant sample curves provided the relevant parameters have been previously determined in the same systems.