Food poisoning and factory farm animals: caveat lector

A new study of foodborne illness has just been published and has made quite a bit of news. The typical headline is something like: "Animals Farmed For Meat Are The No. 1 Source Of Food Poisoning Bug, Study Shows." That makes it sound like most cases of food poisoning are from a particular bug and that bug comes from farm animals. Although it is inaccurate, the stories beneath the headline don't do much to dispel it:

Researchers from Lancashire, England, and Chicago, IL, have discovered that animals farmed for meat are the main source of bacteria responsible for food poisoning.

They suggested that 97 pct of the foodbourne [sic] illnesses are caused by bacteria typically found in chicken and livestock. (ANI)

Or this:

Animals farmed for meat and poultry are the principle source of food poisoning in humans, according to new research published in PLoS Genetics.

Using a new method of evolution-based gene-typing, a group of U.K. and U.S. researchers found that 97 percent of the infections reviewed came through the food chain, from chicken, cattle and sheep. Previous studies had shown that the majority of cases were caused by wild animals or environmental sources.

Campylobacter jejuni causes more cases of gastro-enteritis in the developed world than any other bacteria, including E. coli, Salmonella, Clostridium and Listeria combined. The spread of the disease may be cut by improving food hygiene during preparation and by enforcing on-farm security measures, said Daniel Wilson of the University of Chicago, who led the research. (Bloomberg)

Science Daily is a little better, as it should be, since it is just harvesting press releases from the universities of the research labs where the work is done:

A study by researchers from Lancashire, England, and Chicago, IL, found that 97 percent of campylobacteriosis cases sampled in Lancashire were caused by bacteria typically found in chicken and livestock. The work is based on DNA-sequence comparison of thousands of bacteria collected from human patients and animal carriers.

Campylobacter jejuni causes more cases of gastroenteritis in the developed world than any other bacterial pathogen, including E. coli, Salmonella, Clostridium and Listeria combined. Wild and domestic animals act as natural reservoirs for the disease, which can also survive in water and soil. However, the relative importance of these sources is unclear, and recent work has suggested that livestock are not the main reservoir for human disease. (Science Daily)

A little better, but still not that great. All of these renditions imply that most food poisoning is due to a bug found in farm animals. In fact the correct statement is found in the paper itself:

Campylobacter is the most commonly identified cause of bacterial gastro-enteritis in the developed world. (Wilson et al.,"Tracing the Source of Campylobacteriosis, PLoS Genetics; Introduction [the abstract is less precise and less accurate])

Let's parse this. Campylobacter jejuni is a bacterium that is the most commonly identified cause of bacterial gastroenteritis. It is not necessarily the most common agent of food poisoning, just the bacterial cause most commonly identified. Only one third of outbreaks have an agent identified and the proportion is even less if you consider all food poisoning, including sporadic or isolated (non outbreak cases) like home food poisoning. The most commonly identified agent in outbreaks is Salmonella, not Campylobacter. The estimate that Campylobacter is the most common cause includes all the sporadic cases worldwide. Nor does this statement include the non-bacterial causes of food poisoning, which include viruses and parasites. Both are hard to diagnose and identify. Many of the two thirds of unidentified outbreak cases are undoubtedly viruses, like Norwalk agent (a calicivirus). So the statement may be true but it doesn't mean quite what it seems to on a quick read and certainly not what it was construed to mean in the press accounts (some of which confusion was caused by careless wording of the press release).

Having (I hope) cleared that up, campylobacteriosis is without doubt one of the most important kinds of food poisoning (producing diarrhea, vomiting, fever), so its source is important. It is commonly found in farm and wild animals and hangs out in soil, water, feces and lots of other places. There has been a strong suggestion that much campylobacteriosis may be waterborne. The purpose of the paper was to see if genetic analysis of C. jejuni isolated from ill patients could shed light on the source. Here's their technical description:

Here we report a systematic study of 1,231 cases of C. jejuni infection in Lancashire, England, which we have DNA-sequenced using multi-locus sequence typing (MLST). We infer the source of infection of each patient by comparison to 1,145 animal and environmental C. jejuni isolates collated from previous studies in livestock, wild animals and the environment, using a novel population genetics approach that models DNA sequence evolution and zoonotic transmission. We treat the animal and environmental reservoirs of C. jejuni as populations between which there may be gene flow (migration). Within these populations the bacteria evolve through de novo mutation and horizontal gene transfer (recombination). We estimate the amount of mutation, migration and recombination, and use these estimates to assign probabilistically each human case to one of the source populations. From these population assignments we estimate the total amount of human disease attributable to each source. [cites omitted]

I quoted this to highlight several things, which you might be able to see even if you don't understand the technical jargon. The first is that these researchers used a lot of relevant data. It is a unique dataset. But it is specific to one location, Lancashire, England and it pertains to sporadic cases, not outbreak cases. The source for an outbreak may well be different, for example, a drinking water supply. Second, the genetic landscape is quite complicated. The genetic information of Campylobacter is constantly changing in the bug population, in the environment, and in the human population. The genetic variation is both by mutation of genes and the transfer of chunks of genetic information between organisms. There is no unique fingerprint that connects the genetic pattern in humans with a genetic pattern in the various sources such as farm animal, wild animal or environmental sample. In order to figure this out they had to pioneer a new analytic method. This required a mathematical model of evolution and genetic variation.

I confess I haven't examined the model in detail (NB: this is a different kind of model than discussed in our long modeling series here). That would take more time than I have for something not germane to my own research work. But I looked at it enough to get the general idea. It appears that they looked at how much the sequences at specific parts of the genetic blueprint of the organism differed from each other, both within sources like sheep or the environment or patients; and between the sources. They did this by defining a distance that depended on the number of differences between the seven loci examined. They then used a Bayesian approach. Bayesian analyses differ from the conventional statistical approach by considering the unknown parameter as a random variable rather than considering the observed sample as a random variable. This requires assuming some prior probability distribution for the source probabilities and also a probability model for the process that generated the observed data (to get the likelihood). This is turn requires some assumptions about gene flow, evolutionary processes and other things. It also assumes there are no unknown reservoirs for Campylobacter other than the ones they considered. The results are from simulations of this scheme, run 100 times.

The conclusion of this work was that in 97% of the cases, chicken, cattle or sheep were the likely sources of infections in their isolates. Wild animals or the environment were rare sources. The implication is that the food chain is the source of almost all sporadic (non-outbreak) cases of campylobacteriosis in Lancashire, England, not wild animals or some environmental source like a water supply. There is as yet no evidence to support the proposition this is true everywhere or true of outbreak cases and certainly no evidence it is true of the bulk of food poisoning cases whose source is unidentified or not bacterial. It is very nice work, but still limited at this point.

So this is another case where the actual study is not nearly what it is portrayed in the news media (and the scientists and their press flacks must bear a large part of the blame for this); and even on its own terms, has many moving parts when we lift the hood; not only wheels within wheels, but many assumptions.

The lesson? As always, caveat lector.