Validation of a polymerase chain reaction method for the detection of rendered bovine-derived materials in feedstuffs

This study validated a polymerase chain reaction-based method for the detection of a specific bovine mitochondrial gene derived from rendered bovine tissues and admixed with complete animal feed. Four laboratories participated in this effort: one state laboratory and three Food and Drug Administration (FDA) laboratories, including one FDA field laboratory. The protocol used a statistical approach of 90% probability, with a 95% confidence interval for determining acceptable rates of false-positive and false-negative samples. Each participating laboratory analyzed 30 samples of feed each containing 0, 0.125, and 2.0% bovine meat and bone meal (BMBM), for a total of 90 feed samples. The samples were randomized such that the analysts were unaware of the true identity of the test samples. The results demonstrated that all laboratories met the acceptance criteria established for this protocol. The overall rates of false-negative results were 0.83% (1/120) at the level of 0.125% BMBM and 1.67% (2/120) at the level of 2% BMBM. The overall rate of false-negative results for all levels of BMBM was 1.25% (3/240). The rate for false-positive results was 0.83%.