Brucellosis (caused by the bacterium Brucella abortus) is a zoonotic disease endemic in wild elk (Cervus canadensis) of the Greater Yellowstone Ecosystem, US. Because livestock and humans working with elk or livestock are at risk, validated tests to detect the B. abortus antibody in elk are needed. Using the κ-statistic, we evaluated the buffered, acidified plate antigen (BAPA) assay for agreement with the results of the four serologic tests (card test [card], complement fixation test [CF], rivanol precipitation plate agglutination test [RIV], standard plate agglutination test [SPT]) that are approved by the US Department of Agriculture for the detection of the B. abortus antibody in elk. From 2006 to 2010, serum samples collected from elk within B. abortus-endemic areas (n = 604) and nonendemic areas (n = 707) and from elk culture-positive for B. abortus (n = 36) were split and blind tested by four elk serum diagnostic laboratories. κ-Values showed a high degree of agreement for the card (0.876), RIV (0.84), and CF (0.774) test pairings and moderate agreement for the SPT (0.578). Sensitivities for the BAPA, card, RIV, CF, and SPT were 0.859, 0.839, 0.899, 1.00, and 0.813, whereas specificities were 0.986, 0.993, 0.986, 0.98, and 0.968, respectively. The positive predictive values and the negative predictive values were calculated for 2.6%, 8.8%, and 16.2% prevalence levels. These findings suggest the BAPA test is a suitable screening test for the B. abortus antibodies in elk.
Keywords: Brucella abortus; buffered antigen plate agglutination (BAPA); elk; negative predictive value; positive predictive value; sensitivity; specificity; κ-statistic.