Detection of Mycobacterium avium subspecies paratuberculosis in environmental samples from infected Czech dairy herds

The objective of the present study was to evaluate the suitability of environmental sampling to screen Czech dairy herds to detect Mycobacterium avium ssp. paratuberculosis (MAP) and to find the most convenient location for the MAP detection in the lactating cow area. Environmental samples (ES, n = 72) from milking parlour holding pens (n = 19), milking alleyways (n = 19) and free-stall alleyways (n = 34) from 19 herds were simultaneously tested to detect MAP by a quantitative PCR (qPCR) and bacterial culture. Eight and thirteen samples from the milking parlour holding pens, twelve and eleven samples from the milking alleyways and eleven and eighteen samples from the free-stall alleyways were qPCR and culture positive, respectively. A 4.6 times higher probability of being culture positive than qPCR positive was detected for the assessable MAP detection results from the free-stall alleyways [P = 0.008 6, odds ration (OR) = 4.572 8)] and no association was found between the results from the milking parlour holding pens (P = 0.191 4) and the milking alleyways (P > 0.999 9) and the diagnostic method used. The percentage of qPCR-positive samples in the tested locations was detected for the milking alleyways (63.2%), free-stall alleyways and milking parlour holding pens. The herd infectious status was in agreement with 16 (84.2%), 14 (73.7%) and 12 (63.2%) qPCR results from the milking alleyways, free-stall alleyways (32.4%) and milking parlour holding pens (42.1%), respectively. No statistically significant differences were detected for these results (P = 0.396 1). MAP was detected by the qPCR and bacterial culture in all three locations where the ES were collected. We suggest an environmental sampling followed by MAP detection by qPCR as an easy-to-perform time-saving protocol for MAP screening in Czech dairy herds. Although the milking alleyways seem to be the most convenient location for the environmental sampling, this assumption was not statistically supported.