Performance characteristics of three Brucella canis serological assays in the United States

<i>Brucella canis</i> is a zoonotic pathogen of dogs that poses diagnostic challenges. While direct detection of <i>B. canis</i> by PCR or culture is ideal, serologic diagnosis is necessary for identification of carrier animals and can support a clinical diagnosis of brucellosis. Prior to 2022, <i>B. canis</i> seroscreening in the United States was primarily performed using a commercially available rapid slide agglutination test. However, the kit was discontinued by the manufacturer in early 2022, leaving a gap in the availability of commercial <i>B. canis</i> seroassays. The goal of this study was to compare the performance of three <i>B. canis</i> serologic tests that are currently available: VMRD <i>Brucella ovis</i> ELISA, Bionote Anigen Rapid C.Brucella Ab immunochromatographic lateral flow assay, and VMRD <i>B. canis</i> indirect fluorescent antibody (IFA) assay. A panel of 56 banked serum specimens originally submitted to the Cornell University Animal Health Diagnostic Center (the study reference laboratory) for <i>B. canis</i> seroscreening was distributed to 12 testing laboratories. Each sample was run on three assays developed at the reference lab: rapid slide agglutination test with 2-mercaptoethanol (2-ME RSAT), agar gel immunodiffusion test using cytoplasmic antigen (AGID II), and Canine Brucella Multiplex. Five testing labs ran the ELISA, six ran the lateral flow, and six ran the IFA. When evaluated as a screening assay, we compared the assays to the 2ME-RSAT. The ELISA had the highest sensitivity (96.8, 95%CI 83.8-99.9) but the lowest specificity (79.3, 95%CI 57.9-92.9). The sensitivity of the lateral flow was 90.6% (95%CI 75-98%) and the IFA was 87.5% (95%CI 71-96.5). Specificity for the lateral flow was 95.8% (95%CI 78.9-99.9) and IFA was 97.5% (95%CI 67.6-97.3). When compared to AGID II and Canine Brucella Multiplex, the test assays were all highly sensitive, but specificity was <90%. Interrater reliability was highest for IFA (<i>Κ</i> = 0.92) and lowest for the lateral flow (<i>Κ</i> = 0.82). Serial testing of positive samples with a more specific test, such as AGID II, will continue to be necessary when using any of the three assays tested in this study.