Associations between bovine viral diarrhea virus, bovine leukemia virus, and Mycobacterium avium subspecies paratuberculosis exposure on H5N1 clinical signs and shedding among dairy cows in the United States.

Highly pathogenic avian influenza virus (HPAI) of the H5N1 subtype was first detected in dairy cattle in the United States in March 2024 and subsequently spread throughout the US dairy industry. Although early research has focused on the epidemiology and detection of H5N1, the influence of concurrent endemic infections, such as bovine viral diarrhea virus (BVDV), bovine leukemia virus (BLV), and Mycobacterium avium ssp. paratuberculosis (MAP) on H5N1 disease in dairy cattle has not been evaluated. The objective of this observational study using a convenience sample of dairy cattle from H5N1-affected farms was to evaluate exposure to these endemic dairy cattle diseases and their association with H5N1 clinical signs, viral shedding, and antibody production. Serum samples, along with cow characteristics and H5N1 clinical status, were collected from a convenience sample of cattle on 11 commercial dairy farms in 3 states (California, Colorado, and Michigan) between April and December 2024. A total of 423 cows were tested for BVDV antigen, 459 for BLV antibodies, 108 for BLV DNA by PCR, and 459 for MAP antibodies. No cattle tested positive for BVDV antigens, but 72.7% of farms (9/11) had cattle that tested positive for MAP antibodies. All farms had cattle test positive for BLV antibodies. Multivariable mixed-effects logistic regression models with farm as a random effect were used to identify predictors of H5N1 clinical status, shedding, combined clinical and shedding status, and antibody status. Bovine viral diarrhea was not included in the modeling because the disease was not detected in any cows in this study. Potential confounders, including pregnancy status, lactation number, and days in milk, were evaluated in the modeling process. Cows positive for MAP antibodies had 2.66 (95% CI: 1.12-6.29) times higher odds of being clinical and 2.21 (95% CI: 1.05-4.67) times higher odds of shedding H5N1. Although no comorbidities were found in the final shedding and clinical model, pregnant cows were 2.21 (95% CI: 1.11-4.39) times more likely to be shedding and clinical compared with nonpregnant cows. Regarding H5N1 serum antibody modeling, cows in their fourth lactation or greater were 2.83 (95% CI: 1.43-5.63) times more likely to have antibodies against H5N1 compared with cows in their second lactation. Interpretation of these findings should consider that sampling occurred during an active outbreak and relied on convenience sampling and producer-reported clinical status. Nonetheless, these results suggest that MAP exposure and pregnancy status are associated with clinical signs and potential shedding of H5N1 in dairy cattle and should be considered in future control strategies.