Sensitive and efficient methods for detecting anti-erythropoietin (anti-EPO) antibodies are needed for analysis and, above all, for large scale screening of human serum samples. ELISA is an attractive alternative to labor-intensive radioimmunoprecipitation assays but apparently conflicting reports question its sensitivity. We sought to resolve this issue by directly comparing different reported ELISA approaches to determine whether rhEPO-coating methods affect detection of anti-EPO antibodies. Investigators reporting low sensitivity had used ELISAs in which rhEPO was directly coated to microtiter plates while the high sensitivity ELISA used plate-bound streptavidin to bind biotinylated rhEPO. Using anti-EPO positive human sera, our results confirmed a large (100- to 300-fold) difference in sensitivity between the ELISAs and suggested that the inferiority of the low sensitivity ELISA was caused by the direct coating of rhEPO which may disrupt epitopes by masking recognition sites or introducing conformational changes. Thus, a bridging ELISA can be an appropriate and effective system for antibody analysis and screening of human sera with high sensitivity and specificity but only if performed with streptavidin binding of biotinylated antigen. This finding may also be more generally applicable to the detection of antibodies against other protein antigens.