Evaluation of estrogenic activity in animal diets using in vitro assay

A yeast estrogen bioassay (RIKILT REA) was in-house validated for feed on the 5 μg 17β-estradiol-equivalents per kg level according to EC Decision 2002/657/EC. All the performance characteristics met the criteria as defined in the Decision and the REA is able to detect 17β-estradiol in animal feed at a low level of 1.15–2 μg kg⁻¹. Subsequently, the developed and validated procedure was applied to determine the estrogenic activity in 24 feed samples intended for food producing animals, pets and laboratory animals. Two batches of rodent diet Murigran and one dog feed have been presented as a suspect, i.e. gave responses above the determined decision limit (CCα) and detection capability (CCβ). In assessing the performance of the estrogenic activity in these diets evaluated by comparison with the 17β-estradiol calibration curve, 17β-estradiol-equivalence levels of 7.07 μg EEQ kg⁻¹ and 9.54 μg EEQ kg⁻¹ in two batches of rodent diet and 5.3 μg EEQ kg⁻¹ in dog feed have been established. The activities observed in the rodent feed could be explained by chemical analysis, revealing high amounts of genistein, daidzein and trace amounts of zearalenone. In addition, the estrogenic activity in one of rodent feed was above the established CCα, but below the CCβ values established and all other samples showed no estrogenic activity with responses below the CCα value, which corresponds to levels below 2 μg EEQ kg⁻¹.