Estimating herd prevalence of major mastitis pathogens in Canadian dairy herds using Bayesian latent class models

Mastitis remains a major challenge for dairy production worldwide, and understanding pathogen prevalence is essential for guiding control strategies. This study aimed to estimate the prevalence of major mastitis pathogens in bulk tank milk (BTM) from Canadian dairy herds in Alberta, Ontario, and Québec. A total of 340 herds were included, with BTM samples analyzed by qPCR for Staphylococcus aureus, Streptococcus uberis, Streptococcus dysgalactiae, Streptococcus agalactiae, Mycoplasma spp., Mycoplasmopsis bovis, Prototheca spp., Klebsiella spp., and Escherichia coli. We used a 2-stage Bayesian latent class model (BLCM) to estimate the provincial herd-level prevalence of each pathogen, along with 95% credible intervals (95% BCI). Samples classified as 'suspect' by qPCR were excluded from the BLCM analysis. We also classified pathogens as having a very low prevalence when there was a > 95% probability that the provincial prevalence was below 1%. The estimated prevalence of S. aureus was 56.4% in Québec (95% BCI: 48.5-63.9), 32.6% in Ontario (95% BCI: 23.7-42.2), and 35.0% in Alberta (95% BCI: 22.4-48.8). Streptococcus uberis was the second most prevalent pathogen, with estimated prevalence of 41.2% (95% BCI: 34.2-48.4) in Québec, 34.1% (95% BCI: 25.7-43.2) in Ontario, and 33.7% (95% BCI: 22.6-45.8) in Alberta. Streptococcus dysgalactiae followed, with estimated prevalence of 13.7% (95% BCI: 9.3-19.1), 8.2% (95% BCI: 4.0-14.2), and 8.5% (95% BCI: 3.1-17.3) in Québec, Ontario, and Alberta, respectively. The estimated prevalence for S. agalactiae, Mycoplasma spp., M. bovis, Prototheca spp., Klebsiella spp., and E. coli ranged from 0.3 to 5.8%. Specifically, S. agalactiae prevalence was 0.6% (95% BCI: 0.1-4.4) in Alberta, 1.8% (95%BCI: 0.3-5.5) in Ontario, and 2.1% (95% BCI: 0.7-4.8) in Québec. Mycoplasma spp. prevalence was 0.6% (95% BCI: 0.1-4.4), 2.7% (95% BCI: 0.7-6.9), and 1.6% (95%BCI: 0.4-4.1) in Alberta, Ontario, and Québec, respectively, while M. bovis prevalence was 0.5% (95% BCI: 0.1-3.8), 0.9% (95% BCI: 0.1-3.7), and 0.3% (95% BCI: 0.1-1.4). Prototheca spp. prevalence ranged from 0.4% (95% BCI: 0.1-2.5) in Ontario to 3.1% (95% BCI: 1.3-6.3) in Québec, and Klebsiella spp. prevalence ranged from 0.6% (95%BCI: 0.1-4.2) in Alberta to 1.8% (95% BCI: 0.3-5.5) in Ontario. Finally, for E. coli, the estimated prevalence was 4.1% (95% BCI: 0.8-11.0), 5.8% (95% BCI: 2.4-11.1), and 3.9% (95%BCI: 1.7-7.3) in Alberta, Ontario, and Québec, respectively. Our analyses suggested that all 3 provinces could be considered as having low prevalence of M. bovis. For microorganisms that mainly causes clinical mastitis, the lower prevalence estimates likely reflect, at least in part, the limitations of a single BTM sampling rather than true limited circulation. Positive predictive values were low for most pathogens, except for S. aureus, highlighting the need to interpret positive results with caution, whereas negative predictive values were consistently high for most pathogens (>0.98), indicating reliable negative results. Overall, these findings emphasize the importance of continuous surveillance, regular diagnostic testing, and implementation of biosecurity measures to reduce the prevalence of mastitis pathogens.