1 2 3 4 5 6 7 8 9 Does Feeding Distillers Grains in Rations Increase E. coli O157:H7? Terry Klopfenstein, Dave Smith, Galen Erickson and Rod Moxley University of Nebraska-Lincoln Introduction In 1997 Hudson Foods in Columbus, NE, recalled 25 million pounds of ground beef 10 because it was positive for E. coli O157:H7. A few months later Beef America, a regional packer 11 in Norfolk, NE, had a ground beef recall. Both recalls bankrupted the companies. That got our 12 attention in Nebraska especially, but across the national beef industry as well. Meat and Poultry 13 Magazine calculated that the ten-year cost of E. coli O157:H7 from ’93 to ’03 was about $2.7 14 billion. The public health cost was estimated at $989 million per year. We have learned a 15 considerable amount about E. coli O157:H7 in the past 15 years. 16 Background on E. coli O157:H7 17 Prior to 1998, it was believed that the prevalence of E. coli O157:H7 in feedlot cattle was 18 low ─ perhaps 2 to 3% of all cattle being positive. New analytical techniques essentially 19 increased this prevalence by a factor of about 10. This procedure uses an IMS technique which 20 effectively ”finds” the E. coli O157:H7 in cattle feces. Feces have been the most common 21 material measured to determine if cattle are positive for E. coli O157:H7. Cattle feces contain 22 hundreds of types of bacteria measured in millions or more of bacteria per gram. Some of these 23 are coliform bacteria, including generic E. coli, which are not typically pathogens. So finding the 24 E. coli O157:H7 is like “finding the needle in the haystack”. 25 Using the IMS and other more sensitive analytical techniques, we followed 100 steers 26 through a summer feeding period ─ sampling them weekly. Prevalence was low (Figure 1) until 27 mid July and then increased from less than 10% to 80% over a two-week period. Prevalence then 1 28 declined to about 30%. We did not identify reasons for the dramatic change (Khaitsa et al., 29 2003). We sampled cattle in 74 pens in five commercial feedlots (Figure 2). Prevalence averaged 30 30% in the summer and each of the 44 pens sampled in the summer had at least one positive 31 animal (Smith et al., 2001). The range was from 1 to 80% positive animals ─ the variation 32 largely unexplained (Smith et al., 2001). Prevalence was only 6.1% in winter months and 14 of 33 30 pens had no positive animals. Still the range in prevalence was 0 to 56%. We concluded that 34 probably all feedlots have E. coli O157:H7 positive cattle and prevalence is higher in the summer 35 than winter and prevalence is highly variable and the variation is largely unexplained. 36 Elder et al. (2000) sampled cattle entering the packing plant and then the carcasses after 37 slaughter. They showed a positive relationship between E. coli O157:H7 positive cattle entering 38 the plant and the number of positive carcasses. It is believed the carcass gets contaminated from 39 the hide as it is being removed. Because the contamination is on the surface of the carcass, the 40 trim has the E. coli O157:H7 and is mixed into the ground beef as it is processed. The packing 41 industry has done an excellent job in intervention (Koohmaraie et al., 2005). This includes 42 carcass washes, steam pasteurization, organic acid rinses and “test and hold”. Test and hold is a 43 procedure where trimmings are sampled and tested for E coli O157:H7 and “held” before 44 processing. Only trimmings testing negative are ground. The small number of positives are heat 45 processed to destroy the organism. With these excellent interventions, the percentage of positive 46 samples declined from about 0.8% in 2000 to about 0.2% the past four years. 47 Unfortunately, 2007 was not a good year for ground beef safety and recalls. There were 48 eight recalls in 2006 and all of them were initiated because of company sampling. In 2007, there 49 were 20 recalls and nine of theses recalls resulted from illness investigation. This was(is) a 50 problem for the cattle industry and health officials looked for reasons why E. coli O157:H7 2 51 seemed to be a greater problem in 2007 compared to the previous four years. Because the ethanol 52 industry grew in 2007 and had attracted attention because of the effect on corn prices, some 53 theorized that feeding ethanol byproducts was the cause of the E. coli O157:H7 recalls. Late in 54 2007, research from Kansas State showing a connection between distillers grains (DG) feeding 55 and E. coli O157:H7 shedding was reported. The media, looking for a good story, made the 56 Kansas State research findings look like a food safety crisis. Let’s look at the data from Kansas 57 State and Nebraska and see if it confirms the media hype. 58 59 Distillers Grains and E. coli O157:H7 Shedding The Kansas State researchers have reported four studies. Jacob et al. (2008a) reported a 60 study using 370 feedlot cattle sampled at 122 and 136 days on feed. Prevalence overall was fairly 61 low (under 10%). On day 122 cattle were statistically more likely to shed E. coli O157:H7 when 62 fed 25% DG in the diet. On day 136 there was no effect of feeding DG. Jacob et al. (2008b) 63 sampled cattle for 12 weeks during the feeding period. Samples were fecal samples collected 64 from the pen floor. Feeding DG significantly increased E. coli O157:H7 shedding although there 65 was no difference on five of the 12 sampling periods. 66 The third Kansas State study was a challenge experiment where calves were inoculated 67 with nalidixic acid resistant E. coli O157:H7. This allowed the researchers to estimate the 68 number of E. coli O157:H7 shed. Fecal samples were collected for 42 days. E. coli O157:H7 69 shedding was not different for calves fed DG during the first five weeks but was statistically 70 greater during the last week of sampling. Based on these three studies, the Kansas State 71 researchers concluded that DG feeding increased E. coli O157:H7 shedding. In each of the three 72 experiments there were sampling times when DG statistically increased shedding, however, as 3 73 with most results in E. coli O157:H7 research, the results were somewhat inconsistent making 74 interpretation of the results somewhat difficult. 75 Recently, the Kansas State researchers have reported to the Kansas Livestock Assoc. 76 results of an experiment supported by the Kansas Beef Council and NCBA (Nagaraja et al., 77 2008). Seven hundred cattle were fed for 150 days ─ half were fed DG. Pen floor samples were 78 collected weekly or every two weeks and a total of 3,560 samples were collected and analyzed. 79 This is a large scale study with good statistical power. Overall prevalence was fairly low (5.1%). 80 “Although prevalence of E. coli O157:H7 in pen floor fecal samples was numerically higher on 81 some sampling weeks in cattle fed DDGS, there was no significant effect of DDGS (P = 0.2)”. 82 All of the Kansas State studies were conducted with steam-flaked corn (SFC) diets with 83 or without 25% (DM basis) DG. This may be important as we compare the Nebraska research to 84 the Kansas State research. Corrigan (2007) has shown that DG do not respond the same in SFC 85 diets compared to dry-rolled corn (DRC) or high moisture corn diets (HMC) (Figure 3). If cattle 86 gains and efficiencies respond differently to DG levels in SFC and DRC or HMC diets, then it is 87 possible that any effect on E. coli O157:H7 might respond differently as well. The Nebraska E. 88 coli O157:H7 research is with DRC or HMC only. 89 It is logical that the diet fed to cattle could influence the growth of E. coli O157:H7 in the 90 hindgut. Research has shown that the primary reservoir of E. coli O157:H7 is the hindgut and 91 that the E. coli O157:H7 attach to the intestinal wall of the hindgut. Interestingly, the E. coli 92 O157:H7 have no effect on cattle performance. There are two opposing theories on how the diet 93 affects E. coli O157:H7 in the hindgut. The first theory is that starch escaping digestion in the 94 rumen and small intestine is fermented in the hindgut producing volatile fatty acids and lowering 95 pH. This is theorized to inhibit growth of the E. coli O157:H7. Fox et al. (2007) showed support 4 96 for this theory. Steam flaking reduced starch in the hindgut and increased E. coli O157:H7 97 shedding. However, Depenbusch et al. (2008) said “E. coli O157:H7 was not related to fecal pH 98 or starch”. We reanalyzed the data of Peterson et al. (2007a) where diets with decreasing 99 amounts of corn were fed ─ decreasing the amount of starch in the diet. Amount of starch in the 100 101 diet was not related to E. coli O157:H7 shedding (P = .22). The opposing theory is that starch in the hindgut is the substrate for E. coli O157:H7 so 102 by reducing the amount of starch getting to the hindgut, E. coli O157:H7 would be reduced. 103 Peterson et al. (2007a) and Folmer et al. (2003) showed this did not work. While it is logical that 104 diet affects E. coli O157:H7 growth in the hindgut, clearly neither of the two opposing “starch 105 theories” have been “proven” so we look for other possible theories. 106 The Nebraska research on the effect of DG on E. coli O157:H7 shedding was reported by 107 Peterson et al. (2007b). The research was primarily focused on vaccination as an E. coli 108 O157:H7 intervention. Because the study was superimposed on a nutrition study, we reanalyzed 109 the data (Figure 4). Wet DG were fed as 0, 10, 20, 30, 40 and 50% of diet dry matter replacing 110 DRC and HMC. In this experiment samples of the hindgut mucosa were analyzed as well as 111 fecal samples. Results were similar but more consistent for the mucosal samples (Figure 4). 112 There was a significant effect of level of DG on E. coli O157:H7 shedding, however, it was not a 113 linear relationship. None of the levels of DG feeding were statistically different from the control 114 (ODG). The 10, 20, and 30% DG levels numerically decreased the shedding of E. coli O157:H7. 115 Interesting, this is within the range of feeding (25%) the Kansas State researchers used. Our 116 research is with DRC and HMC while their research is with SFC which may well make a 117 difference. 5 118 At the 40 and 50% DG feeding level, E. coli O157:H7 shedding numerically increased 119 compared to the control. Note that the statistical difference is between the 10, 20, and 30% DG 120 levels and the 40 and 50% levels. So does DG decrease or increase E. coli O157:H7 shedding? 121 The media (Des Moines Register, January 27, 2008) chose the negative approach ─ “studies at 122 the University of Nebraska suggest feeding DG increases levels of a DEADLY form of E. coli 123 bacteria”. 124 Peterson et al. (2007b) were studying vaccination. The pattern of E. coli O157:H7 in 125 hindgut mucosa for unvaccinated cattle is similar to that discussed previously (Figure 5). 126 However, there was only one positive steer among the vaccinated cattle and that was at the 50% 127 DG feeding level. So which is more important, finger pointing at feed ingredients such as DG or 128 looking for interventions? Over four studies involving 1,784 cattle, vaccination has reduced E. 129 coli O157:H7 shedding by 65%. This is equivalent to the effect of winter versus summer on 130 shedding. Feeding a direct-fed microbial (Peterson et al. (2007a) reduced shedding over two 131 years by 35%. These two interventions plus others being researched have considerable merit. 132 Conclusions 133 1. It is reasonable to think that what we feed cattle might affect the bacterial population of 134 the hindgut. This has already been demonstrated in different studies. 135 2. Research at Kansas State and University of Nebraska-Lincoln both suggest that under 136 some feeding levels and some other, as yet unknown, conditions, DG may increase E. 137 coli O157:H7 shedding. 138 3. Results of E. coli O157:H7 research in general and specifically with DG feeding are 139 inconsistent. To date, there has been no demonstrably consistent effect of DG feeding on 140 E. coli O157:H7 shedding. 6 141 4. Response in E. coli shedding to DG feeding may be affected by DG level and other 142 dietary ingredients such as the corn type. The issue is highly complex and more research 143 is needed to address the many factors, including those feed components given 144 concurrently with DG, that could be involved. 145 5. Interventions and research on interventions is much more important than “finger 146 pointing” at different feedstuffs, especially when data are inconsistent and more research 147 is needed. 148 6. At this point, there is no scientific evidence that feeding DG, at least at levels being used 149 commercially, is the cause of a food safety crisis! Additionally, there is no scientific 150 evidence to suggest that the feeding of DGs is the cause of the 2007 recalls. 151 152 References Corrigan, Mark C., Galen E. Erickson, Terry J. Klopfenstein, Kyle J. Vander Pol, Matthew A. 153 Greenquist, and Matthew K. Luebbe. 2007. Effect of corn processing and wet distillers grains 154 inclusion level in finishing diets. Nebraska Beef Cattle Report. MP-90:33-35. 155 Depenbusch, B. E., T. G. Nagaraja, J. M. Sargeant. J. S. Drouillard, E. R. Loe, and M. E. 156 Corrigan. 2008. Influence of processed grains on fecal pH, starch concentration and shedding 157 of E. coli O157:H7 in feedlot cattle. J. Anim. Sci. 86:632-639. 158 Elder, R. O., Keen, J. E., Siragusa, G. R., Barkocy-Gallagher, G. A. Koohmaraie, M., and 159 Laegreid, W. W. (2000). Correlation of enterohemorrhagic Escherichia coli O157 prevalence 160 in feces, hides, and carcasses of beef cattle during processing. Proceedings of the National 161 Academy of Sciences of the United States of America, 97-2999-3003. 7 162 Folmer, Jeffrey, Casey Macken, Rod Moxley, David Smith, Mindy Brashears, Suzanne Hinkley, 163 Galen Erickson, and Terry Klopfenstein. 2003. Intervention strategies for reduction of E. coli 164 O157:H7 in feedlot steers. Nebraska Beef Report. MP 80-A:22-23. 165 Fox, J. T., B. E. Depenbusch, J. S. Drouillard, and T. G. Nagaraja. 2007. Dry-rolled or steam- 166 flaked grain-based diets and fecal shedding of E. coli O157:H7 in feedlot cattle. J. Anim. Sci. 167 85:1207-1212. 168 Jacob, M. E., J. T. Fox, S. K. Narayanan, J. S. Drouillard, D. G. Renter, and T. G. Nagaraja. 169 2008a. Effects of feeding wet corn distiller’s grains with solubles with or without monensin 170 and tylosin on the prevalence and antimicrobial susceptibilities of fecal food-borne 171 pathogenic and commensal bacteria in feedlot cattle. J. Anim. Sci. 2008. In press Eprint 172 available. doi:10.2527. 173 Jacob, M. E., J. T. Fox, J. S. Drouillard, D. G. Renter, and T. G. Nagaraja. 2008b. Effects of 174 dried distillers’ grains on fecal prevalence and growth of Escherichia coli O157 in batch 175 culture fermentations from cattle. Appl. Environ. Microbial. 74:38-43. 176 Jacob, M. E., G. L. Parsons, M. K. Shelor, J. T. Fox, J. S. Drouillard, D. U. Thomson, D. G. 177 Renter, and T. G. Nagaraja. 2008c. Feeding supplemental dried distiller’s grains increases 178 fecal shedding Escherichia coli O157 in experimentally inoculated calves. Zoonoses Publ. 179 Hlth (Accepted for Publication). 180 Khaitsa, M. L., D. R. Smith, J. A. Stoner, A. M. Parkhurst, S. Hinkley, T. J. Klopfenstein, and R. 181 A. Moxley. 2003. Incidence, duration and prevalence of Escherichia coli O157:H7 fecal 182 shedding by feedlot cattle during the finishing period. J. of Food Protection, Vol. 66, No. 11, 183 1972-1977. 8 184 Koohmaraie, M., T. M. Arthur, J. M. Bosilevac, M. Guerini, S. D. Shackelford, and T. L. 185 Wheeler. 2005. Post-harvest interventions to reduce/eliminate pathogens in beef. Meat 186 Science 71:79-91. 187 Nagaraja, T. G., J. Drouillard, D. Renter, S. Narayanan. 2008. Distillers grains and food-born 188 pathogens in cattle: Interaction and intervention. KLA News and market Report Vol. 33, No. 189 35. 190 Peterson, R. E., T. J. Klopfenstein, G. E. Erickson, J. Folmer, S. Hinkley, R. A. Moxley, and D. 191 R. Smith. 2007a. Effect of Lactobacillus Strain NP51 on Escherichia coli O157:H7 fecal 192 shedding and finishing performance in beef feedlot cattle. J. Food Prot. 70, No. 2:287-291. 193 Peterson, R. E., T. J. Klopfenstein, R. A. Moxley, G. E. Erickson. S. Hinkley, G. Bretschneider, 194 E. M. Berberov, D. Rogan, and D. R. Smith. 2007b. Effect of a vaccine product containing 195 type III secreted proteins on the probability of E. coli O157:H7 fecal shedding and mucosal 196 colonization in feedlot cattle. J. Food Protection. Vol. 70, No. 11:2568-2577. 197 Smith, D. R., M. P. Blackford, S. M. Younts, R. A. Moxley, J. T. Gray, L. L. Hungerford, C. T. 198 Milton, and T. J. Klopfenstein. 2001. Ecological relationships between the prevalence of 199 cattle shedding Escherichia coli O157:H7 and characteristics of the cattle or conditions of the 200 feedlot pen. J. Food Prot. 64 (12) 1899-1903. 9 Longitudinal study Research Feedyard Khaitsa et al. 2003. J Food Prot 66 (11) 1972-1977. 100% 90% Conditions of natural O157:H7 exposure 70% Prevalence Feces from100 steers (10 pens of 10) cultured each week 80% 60% 50% 40% 30% E. coli O157:H7 recovered every week and at least once from every animal 20% 10% 0% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Week Conclusions E. coli O157:H7 is ubiquitous to cattle populations Variable in prevalence by TIME and PLACE Figure 1. This project was supported by the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service, grant number #0002501. Prevalence of cattle shedding E. coli O157:H7 Summer 1999 / 2002 44 pens of cattle Prevalence Summer E. coli O157:H7 • 4,952 cattle, 44 pens – 1,501 culture positive (30%) – EVERY pen (100%) – Variable prevalence (1(1-80%) Winter E. coli O157:H7 • 2,941 cattle, 30 pens – 179 culture positive (6.1%) – 16/30 pens (53%) – Variable prevalence (0(0-56%) • Significant difference by season Smith DR, et al. 2001. J Food Prot 64 (12) 1899-1903 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 pens ranked by prevalence Prevalence of cattle shedding E. coli O157:H7 Winter 2000 / 2001 / 2002 30 pens of cattle Prevalence Seasonal prevalence of fecal shedding in commercial feedyards 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 pens ranked by prevalence Figure 2. 10 WDGS & Grain Processing 6.5 Feed/Gain 6.0 5.5 5.0 y = -0.0186x + 6.1219 R2 = 0.9617 4.5 DRC HMC SFC y = -0.0085x + 5.4226 4.0 R2 = 0.767 3.5 y = -0.0003x + 5.4653 R2 = 0.0078 3.0 0 10 20 30 40 Level of diet DM (WDG) Corrigan et al., 2007 Nebraska Beef Rep. Figure 3. Effect of level of distiller’s grains on E. coli O157:H7 colonization by cattle b 0.25 0.2 Probability Wet distiller’s grains plus solubles replaced corn in each treatment diet b 0.15 0.1 0.05 ab a a a 0 00DG 10DG 20DG 30DG 40DG 50DG Diet Figure 4. 11 Colonization of non-vaccinated cattle Colonization of vaccinated cattle Figure 5. 12