Accuracy in naturally occurring anabolic steroid assays in cattle and first approach to quality control in Italy.

This study assessed the accuracy of 16 commercial and three self-produced kits and drew the basis for using an external quality control (EQC) system. The commercial kits were mainly developed for blood sex steroid determination in humans but also have been used in cattle. Parallelism, recovery and precision tests were performed for progesterone (P4), testosterone (T) and oestradiol (E2) assays. Moreover, anonymous QC samples were sent to be analysed to some Italian laboratories. All kits showed a fair degree of parallelism (P < 0.01), even though 2/7 kits for T and 1/6 kits for E2 determination showed a regression coefficient (r2) lower than 0.98. For P4, an acceptable range of accuracy was achieved in the recovery test only by 1/6 kits; two kits showed fair or great overestimation and two kits considerable underestimation. For T, an acceptable range of accuracy was achieved only by 1/7 kits. For E2, 4/6 kits presented a variable degree of underestimation and two kits showed great overestimation. In the intra-assay precision test quite good repeatability was achieved only using samples with high hormone concentrations. While assaying samples with low concentrations we found a number of RSD > 10%. Moreover, the laboratories participating in the EQC produced statistically different (P < 0.05) results, particularly for high and medium concentrations. In conclusion, the use of commercial kits for screening naturally occurring sex steroid concentrations in cattle blood, in the case of suspected illegal treatments, requires preventive validation procedures and the development of an opportune EQC system.