Insights into Cow-Level Risk Factors for HPAI H5N1 clinical disease in Lactating Holsteins from a Single Outbreak in Colorado (US). An observational study

Highly Pathogenic Avian Influenza (HPAI) H5N1 (clade 2.3.4.4b) has spilled over into dairy cattle populations in the US. A year after the initial outbreak, many uncertainties remain about the virus´s epidemiology. The objective of this observational study was to examine cow-level factors that may influence cow´s susceptibility to manifest clinical signs during a HPAI H5N1 outbreak using on-farm herd records. The study was conducted on a commercial dairy farm in Colorado (US) that housed only mid-to-late lactation cows and experienced the peak of its outbreak in late May 2024. Herd records were used to identify clinical cases. The study population comprised 3,281 cows that were housed across 7 free-stalls and 5 open-lot pens. Pens were classified as “comparable” if they housed cows with similar individual-level characteristics, as determined using a network-based clustering approach based on ANOVA and Chi-squared tests. Univariate associations between case status and potential risk factors were evaluated in the full study population using t-tests and Chi-Square tests. The association between clinical disease and pregnancy was assessed using a 1:1 matched subset of pregnant and non-pregnant cows (n = 196), matched on parity, lactation stage, milk yield, and pen ID, and analyzed using conditional logistic regression. Other potential risk factors of clinical disease were evaluated in a subset of pregnant cows from comparable pens (n = 1,546) using mixed-effects logistic regression. The overall proportion of clinical cases reported was 14.0% (n = 458). Unadjusted analysis showed significant differences in pregnancy (6% in non-pregnant cows vs. 15% in pregnant cows) and parity (9% in 1st parity cows, 16% in 2nd and 3rd parity cows, and 18% in cows with ≥ 4 parities). Across all 12 pens, the proportion of clinical cases ranged from 7% to 27%, and when restricted to comparable pens (n = 5), the variation was similar (11% to 24%). The matched OR, using non-pregnant cows as the reference group, was 4.9 (95% CI: 1.6–14.9), indicating higher odds of being a case among pregnant cows. Among pregnant cows, the OR for multiparous cows vs. primiparous cows was 2.1 (95% CI: 1.5–2.8). The findings suggest that pregnancy and parity are 2 cow-level factors that contribute to HPAI H5N1 clinical disease. Expanding the number of herds studied is critical for a better understanding of these risk factors.