Mycobacterium paratuberculosis (Mptb) is the causative agent of Johne's disease of ruminant animals including cattle, goats, and sheep. It has been suggested that this organism is associated with Crohn's disease in humans, and milk is a potential source of human exposure to this organism. A total of 18, including 7 regular batch and 11 high temperature short time (HTST) pasteurization experiments, were conducted in this study. Raw milk or ultra-high temperature pasteurized milk samples were spiked at levels of 10(3), 10(5), and 10(7) cfu of Mptb/ml. Escherichia coli and Mycobacterium bovis BCG strains at 10(7) cfu/ml were used as controls. Pasteurization experiments were conducted using time and temperature standards specified in the Canadian National Dairy Code: regular batch pasteurization method: 63 degrees C for 30 min, and HTST method: 72 degrees C for 15 s. The death curve of this organism was assessed at 63 degrees C. No survivors were detected after 15 min. Each spiked sample was cultured in Middlebrook 7H9 culture broth and Middlebrook 7H11 agar slants. Samples selected from 15 experiments were also subjected to BACTEC culture procedure. Survival of Mptb was confirmed by IS900-based PCR of colonies recovered on slants. No survivors were detected from any of the slants or broths corresponding to the seven regular batch pasteurization trials. Mptb survivors were detected in two of the 11 HTST experiments. One was by both slant and broth culture for the sample spiked to 10(7) cfu/ml of Mptb, while the other was detected by BACTEC for the sample spiked to 10(5) cfu/ml. These results indicate that Mptb may survive HTST pasteurization when present at > or = 10(5) cfu/ml in milk. A total of 710 retail milk samples collected from retail store and dairy plants in southwest Ontario were tested by nested IS900 PCR for the presence of Mptb. Fifteen percent of these samples (n = 110) were positive. However, no survivors were isolated from the broth and agar cultures of 44 PCR positive and 200 PCR negative retail milk samples. The lack of recovery of live Mptb from the retail milk samples tested may be due to either the absence of live Mptb in the retail milk samples tested or the presence of low number of viable Mptb which were undetected by the culture method used in this study.