Abstract #W158

# W158
Enterobacteriaceae bacteria counts vary for US commercial dairy total mixed rations fed during summer months.
J. Goeser*1,2, J. Becker2, K. Bryan3, S. Bascom4, C. Wacek-Driver5, R. Schmidt6, C. Stoffel7, N. Michael8, 1Rock River Laboratory Inc, Watertown, WI, 2Univer, Madison, WI, 3Chr. Hansen Inc, Milwaukee, WI, 4Phibro Animal Health Corp, Teaneck, NJ, 5Forage Innovations LLC, Bay City, WI, 6Lallemand Animal Nutrition, Milwaukee, WI, 7Papillon-Ag, Easton, MD, 8Arm & Hammer Animal Nutrition/Church & Dwight Inc, Milwaukee, WI.

Feed Enterobacteriaceae colony count per g (cfu/g) can be a bacterial contamination measure however population data for commercial dairy TMR are unknown. The objective here was to determine if a food safety assay (3M Petrifilm Enterobacteriaceae count plate) was capable of culturing Enterobacteriaceae in TMR and determine if counts varied for commercial TMR. Samples were collected (n = 370), from April through September 2018, by US dairy industry professionals and submitted to Rock River Laboratory (Watertown, WI) for analysis. Samples were processed according to 3M Petrifilm instructions (2017, reference 6420/6421). In brief, 10g of TMR was blended into 90 mL Butterfield buffer, shaken and diluted 10, 100, 1000 and 10000-fold. 1mL of diluted solution was plated and incubated at 30C for 24h. Plates counts were manually counted and TMR cfu/g determined by multiplying plate count by plate dilution factor. A separate subsample (approximately 150 g) was dried, ground and analyzed for nutritive parameters by NIR to evaluate correlations with Enterobacteriaceae counts. The bacterial count data were found to be not normally distributed, thus data were transformed using log10 transformation. The log10 transformed data were found to be normal using the continuous fit - normal function in JMP v14.0. The resulting population statistics (log10 cfu/g) in TMR were then as follows: mean = 2.75, standard deviation = 1.18, coefficient of variation (CV) = 42.9%, minimum = zero, maximum = 5.02, and 15th and 85th percentiles = 1.67 and 3.95, respectively. Correlations were evaluated using the response screening and multivariate methods functions in JMP v14.0. Independent pairwise correlations were significant (P < 0.05) for 11 NIR predicted nutritive parameters, with the largest r-values being water soluble carbohydrate (r = 0.21), dry matter (r = 0.18), and in situ rumen starch digestibility (3h; r = −0.18). These nutrition and bacterial count correlations warrant further investigation. The 3M Petrifilm proved capable of culturing Enterobacteriaceae colonies with TMR samples and log10 transformed data population statistics, with a CV greater than 40%, suggest variation exists in Enterobacteriaceae counts in TMR on commercial dairies.

Key Words: feed contamination, enterobacteria