Two types of flexible, sensitive and rapid competitive flow injection enzyme immunoassay were developed and evaluated for their potential use in the bioanalysis of steroids. Instead of the more typical approach where the signal is generated from antibody-bound hapten-enzyme conjugate, the non-bound fraction passing through the affinity column was allowed to react with an enzyme substrate from a merging channel in a post-column reaction system. The enzyme product (p-aminophenol or 4-methyl umbelliferol) was amperometrically or fluorometrically detected. Several parameters known to affect signal generation in the immunoassay were evaluated, including flow rate through the affinity column and through the reaction coil, the length of the reaction coil and of the affinity column. In the pre-incubation approach, where samples were mixed with enzyme conjugate and antibodies before injection, a sample throughput as high as 20 h(-1) was possible. The signal precision was about 1% (RSD) for cortisol (0.6-80 pmol) and 2% (RSD) for budesonide (0.02-12.5 pmol). In the displacement assay for cortisol, enzyme-labelled analyte was displaced from the affinity column when the sample was injected into the flow. A standard curve was obtained with a signal precision of 4-20% for 12.5-1250 pmol injected. The same instrumental set-up was used in both types of immunoassay, and thus a highly flexible system was obtained. A simple replacement of the affinity column from protein G in the pre-incubation approach to a column containing primary antibodies in the displacement assay was needed.