To construct a competitive ELISA standard curve for the detection of glucose-6-phosphate debydrogenase (G6PD), we used highly purified native G6PD (nG6PD) as both immobilized and soluble antigens and anti-G6PD serum raised against nG6PD as antibody. The polystyrene cuvettes coated with nG6PD were challenged with a mixture of a limiting amount of anti-G6PD serum and various doses of nG6PD as competitors followed by incubation with alkaline phosphatase-anti-IgG conjugate. The competitive ELISA did not exhibit the typical sigmoidal dose-response curve characteristic of competition immunoassays under the optimal concentrations of antigen and antibody. The soluble nG6PD used as competitor failed to effectively inhibit the binding of antibodies to the immobilized nG6PD. The addition of NADP, a cofactor of G6PD enzyme, to coating buffer used for immobilizing nG6PD to the cuvettes and PBS-Tween-BSA buffer for diluting competitors did not improve the inhibition of antibody binding to immobilized nG6PD by soluble n/G6PD. The addition of BSA to coating buffer did not increase inhibition, either. Surprisingly, when partially active G6PD (paG6PD), obtained by repeated freeze-thawing, was used as competitor, the antibody binding to either immobilized nG6PD or immobilized paG6PD was inhibited 49-58%. We conclude that an effective competitive ELISA system with nG6PD enzyme and anti-G6PD serum for the detection of G6PD may not be established due to the poor inhibition of antibody binding to immobilized nG6PD by soluble nG6PD under the present assay conditions and that the inhibition may be improved by using an inactivated enzyme as competitor regardless of the type of immobilized antigen used. These results imply that the immobilized nG6PD may undergo denaturation upon binding to the polystyrene cuvettes and that our anti-G6PD serum may recognize denatured enzyme better than active enzyme.