Where does E. coli O157:H7 come from? Food Inc. and cookie dough versions

Is E. coli O157:H7 associated with things other than feedlot cattle?

I had a few people call me recently, saying, I saw that movie, Food, Inc., which says that E. coli O157:H7 is predominately in feedlot cattle because of the grain they are fed, and that’s how the bug came to exist. So how did it get into Nestle cookie dough?

It’s sort of a mantra of raw milk enthusiasts and wannabe food safety types that E. coli O157:H7 is a product of feedlot cattle, and that grass-fed creatures are benign entities for the dinner plate.

A blogger yesterday wrote, “… hamburger tainted by e-coli, a virus that breeds in a cow’s stomach when it is feed grain instead of grass (which, of course, most cows are nowadays in order to fatten them quickly and cheaply).”

It’s a bacterium, not a virus.

Nicholas Kristof, a columnist for the N.Y. Times, wrote yesterday, on Sunday, June 21/09, that, “There is some evidence that pathogens, including E. coli, become much more common in factory farming operations. Move feedlot cattle out to a pasture for five days, and they will lose 80 percent of the E. coli in their gut, the film says.”

That evidence is about as strong as the whisps of evidence compiled by Danny Sugarman that The Doors’ frontman Jim Morrisson is still alive and didn’t die from excess in a Paris bathtub in 1971. But, every teenager goes through their Doors phase (I can only find the clip below in Spanish, but Canada’s The Guess Who stands up much better with the hindsight of time; they know they are drunken buffoons, and not a drunken buffoon trying to be a poet).

Scientific uncertainty can easily be exploited by the certainty of filmmakers, who cherry pick facts and flourish on rhetoric. And I guess if it’s repeated ad nauseum for 11 years by writers from the N.Y. Times to your-favorite-bullshit blogger it becomes fact.

That line, “Move feedlot cattle out to a pasture for five days, and they will lose 80 percent of the E. coli in their gut,” comes from a 1998 paper published in the journal Science by Diez-Gonzalez of Cornell University, and colleagues.

I had one of my colleagues, Rena Orr, write a review of the controversy back in Nov. 2000.

Since September 1998, there has been conflicting information on the effect of diet on E. coli shedding from cattle. The conflict arises in part from the effect of diet on the ability of E. coli to develop acid resistance. … Diez-Gonzalez et. al demonstrated that feeding a high-grain diet to cattle results in an acidic environment in the colon. Because the animals incompletely digested the starch in grains, some starch was able to reach the colon where it fermented, producing fermentation acids. The researchers believe an acidic environment selects for or induces acid resistance among the Escherichia coli population. … Diez-Gonzalez et al. concluded that if cattle were given hay for a brief period (five days) immediately before slaughter, the risk of foodborne E. coli infection would be significantly reduced because the acidity in the colon is greatly reduced. "Our studies indicate that cattle could be given hay for a brief period immediately before slaughter to significantly reduce the risk of food-borne E. coli infection."

The Science article received mainstream media attention, and was covered by the Associated Press and The New York Times, as well as scientific releases and reports. In the Irish Times, it was cited as the basis for concluding that because Irish cattle are fed a grass-based diet rather than grain, Ireland has a low incidence of E. coli O157:H7. Hancock et al. contend that this conclusion is unsupported or contradicted by several lines of evidence. The E. coli that contaminate beef typically originate from the hide, the hooves, or the equipment used in slaughter and processing rather than directly from the colon, and likely replicate in environments unlike the colon. Therefore, the induced acid resistance of E. coli contaminating beef is likely to be unrelated to the pH of its ancestral colonic environment. The E. coli O157:H7 bacterium uses several mechanisms to survive acid environments, some of which are innate and are not influenced by environment . Although acid resistance is likely a factor in an infective dose, induced acid resistance has not been shown to be a factor in E. coli O157:H7 infectivity by experimental (dose-inoculation) or observational (epidemiological) data . Therefore, acid resistance induced by exposure to weak acid may not influence the virulence of this pathogen.

Published data on E. coli O157:H7 tends to contradict or does not support the effects of the dietary change proposed by Diez-Gonzalez et al. In a recent study on three different grain diets (85% cracked corn, 15% whole cottonseed and 70% barley, or 85% barley), the fecal pH of the animals fed the corn diet was significantly lower (P < 0.05) than the fecal pH of the animals fed the cottonseed and barley and barley diets, likely resulting in a less suitable environment for E. coli O157:H7 in the hindgut of the corn fed animals (2000, Buchko et al). In the Journal of Food Protection, researchers concluded that changing from grain to a high roughage diet did not produce a change in the E. coli concentration that was large enough to deliver a drastic improvement in beef carcass hygiene. Sheep experiencing an abrupt diet change have higher concentrations and increased shedding of fecal E. coli O157:H7 for longer periods than sheep fed a consistent high-grain diet. Another study compared the duration of shedding E. coli O157:H7 isolates by hay-fed and grain-fed steers experimentally inoculated with E. coli O157:H7 as well as the acid resistance of the bacteria. The hay-fed animals shed E. coli O157:H7 longer than the grain-fed animals, and irrespective of diet, these bacteria were equally acid resistant.

These results suggest that the proposed dietary change would actually increase contamination with E. coli O157:H7. Also, the 1,000-fold reductions in total fecal E. coli demonstrated by Diez-Gonzales et al. are far greater than those recorded in cattle experiencing similar ration changes . Finally, extensive surveys show that grain-fed feedlot cattle have no higher E. coli O157:H7 infection prevalence than similarly aged dairy cattle fed forage (hay) diets. Abrupt feed change immediately before slaughter could have unexpected deleterious effects. The proposed diet change has the potential to increase the risk of bovine salmonella infections, a potential source of food poisoning. The dietary change results in sharply reduced volatile fatty acid concentrations in the large intestine as well as changes in the bacteria, allowing for colonization of Salmonella.

See, that’s a really long explanation. It’s not as soothing as, change cattle diet, disease prevented. And that was written nine years ago.

Mike Osterholm, director of the Center for Infectious Disease Research and Policy and professor in the School of Public Health at the University of Minnesota wrote a cleaner critique in 2007 in the Minneapolis-St. Paul Star Tribune:

"Russo cited conclusions from a 1998 study from Cornell University that cattle fed a diet of grass, not grain, had very few E. coli, and that those bacteria that survived in the cattle feces would not survive in the human when eaten in undercooked meat, particularly hamburger. This statement is based on a study of only three cows rotated on different diets and for which the researchers did not even test for E. coli O157:H7. Unfortunately, the authors extrapolated these incredibly sparse results to the entire cattle industry. The Cornell study is uncorroborated in numerous published scientific papers from renowned research groups around the world. Finally, work conducted by the Minnesota Department of Health as part of a national study on foodborne disease recently showed that eating red meat from local farms was a significant risk factor for E. coli infection. …

And as my colleague David Renter wrote in Sept. 2006,

"Cattle raised on diets of ‘grass, hay and other fibrous forage’ do contain E. coli O157:H7 bacteria in their feces as do other animals including deer, sheep, goats, bison, opossum, raccoons, birds, and many others.

"Cattle diet can affect levels of  E. coli O157:H7, but this is a complex issue that has been and continues to be studied by many scientists.  To suggest switching cattle from grain to forage based on a small piece of the scientific evidence is inappropriate and irresponsible.  Several pieces of evidence suggest that such a change would not eliminate and may even increase E. coli O157:H7 in cattle.

"The current spinach outbreak may be traced back to cattle manure, but there are many other potential sources.  Simplistically attacking one facet of livestock production may be politically expedient, but instead provides a false sense of security and ignores the biological realities of E. coli O157:H7. In 1999, for example, 90 children were felled by E. coli O157:H7 at a fair in London, Ont. The source? A goat at a petting zoo, hardly an intensively farmed animal."

Last time I looked, E. coli O157:H7 and about 60 other shiga-producing E. coli that are known to cause illness in humans are present in about 10 per cent of all ruminants – cattle, sheep, goats, deer, elk -– and I can point to outbreaks associated with all of those species. Pigs, chickens, humans, birds and rodents have all been shown to be carriers of shiga-producing E. coli but the resevoir appears to be ruminants. The final report of the fall 2006 spinach outbreak identifies nearby grass-fed beef cattle as the likely source of the E. coli O157:H7 that sickened 200 and killed 4.

How the E. coli O157:H7 got into the cookie dough remains to be seen. Biology is complex and constantly changing – even at farmer’s markets, which was the big solution of Food, Inc. But it’s only a movie.

That Cornell paper can be found here:

Diez-Gonzalez, Francisco, Todd R. Callaway, Menas G. Kizoulis, James B. Russell. Grain Feeding and the Dissemination of Acid-Resistance Escherichia coli from Cattle. Science: Sept 11, 1998. Volume 281, Number 5383, pages 1666-1668.