“Ecological changes, including those due to agricultural or economic development, are among the most frequently identified factors in emergence. They are especially frequent as factors in outbreaks of previously unrecognized diseases with high case-fatality rates, which often turn out to be zoonotic introductions…. “Agricultural development, one of the most common ways in which people alter and interpose themselves into the environment, is often a factor. Hantaan virus … causes over 100,000 cases a year in China …. Conversion of grassland to maize cultivation favored a rodent that was the natural host for this virus, and human cases increased in proportion with expansion of maize agriculture …. “Perhaps most surprisingly, pandemic influenza appears to have an agricultural origin, integrated pig-duck farming in China …. “Because humans are important agents of ecological and environmental change, many of these factors are anthropogenic. Of course, this is not always the case, and natural environmental changes, such as climate or weather anomalies, can have the same effect.”
Note that the original epidemiological focus on farms and the farm environment was not concerned with food safety. Joshua Lederberg led the team from the National Academy of Science’s Institute of Medicine whose report in 1992 was the foundation for developing institutional responses to emerging infections (7). I did not see any mention of foodborne diseases there either. However, before it was decided to start a new CDC journal, there was briefly a new section in the CDC’s Morbidity and Mortality Weekly Report (MMWR). The April 16, 1993, issue started with an introduction describing a new series in MMWR on emerging infectious diseases. The first report in the new section was on the West Coast O157:H7 hamburger outbreak (8), where Jack-in-the-Box is referred to anonymously as “chain A.” In practice, emerging foodborne diseases were an issue from the beginning. It is easy to forget that finding the human pathogen O157:H7 as a non-pathogenic commensal in cattle was a shock. It was even more surprising to find O157:H7 in or on plants, starting with the apple cider and apple juice cases that were occurring around the same time as the Jack-in-the-Box outbreak (9). The interaction between human pathogens and plants has continued to be even more surprising, up until today. A brief survey of the research projects and publications of only three labs can illustrate this: Fred Ausubel’s lab at Harvard/Massachusetts General Hospital (10); Maria Brandl’s lab at USDA Albany, CA (11); and Ariena H.C. van Bruggen’s lab at the University of Florida’s Emerging Pathogens Institute (12). There are many others. A new country doctor, or a Medical Extension Specialist, would have to integrate veterinary, epidemiological, and ecological issues of pathogens, as well as the microbial interactions of healthy farm environments. The transition from Burnet’s world to a view of dynamic pathogen evolution and emerging infectious diseases seems difficult enough. For some years, there has been an attempt to create a paradigm shift to integrate concepts of health and health disciplines “to obtain optimal health for people, animals and our environment.” The overall research program has often gone under the name of the One Health Initiative. The contributions of National Research Council and Institute of Medicine Reports on One Health are reviewed in the current issue of Emerging Infectious Diseases (December 2013) in an article by Carol Rubin, et al. (13). From a food safety perspective, “our environment” would include plants and crops, farms and the farm environment, and food production systems. The research seems to be ahead of the paradigm shift, but as a concept that might integrate at least some perspectives on farming and food safety with integrated epidemiological approaches, it seems like a welcome approach. It might at least provide a common language. I actually have higher hopes for integration of approaches in the farm environment and improving produce safety. The clinical diagnosis of the farm “It takes three generations to grow a farmer.” (14) Farming, in this country, is a highly skilled and complex integrated occupation. It’s also high-risk, both physically and economically. Mitigating economic risk is a constant challenge. The complexity of farming is reflected in the need for 80 to 100 years worth of historical memory to intelligently guide decision-making. “A year like this one,” just in terms of weather, may most closely resemble a year in the late 1940s for farmers I work with in Eastern Washington. A particular combination of events – like a devastating fire at grain harvest – in my county, Yolo County, CA, may require going back to the 1920s. You may only live once, but over three or four generations, many rare combinations of circumstances have often occurred before. Farmers have to integrate complex data into practical decision-making and action, all the time, year after year. In order to integrate decisions about human pathogens in the farm environment, they need well-founded and well-researched, reliable information. They need people to work with who can help apply that information to the specific individual conditions of their farms and fields, crops and rotations, animals, and rangelands, soil types and climatic variations, ecologies and water sources. Cooperative Extension has been one of the most effective research and technology transfer agencies in the history of the United States. It was one of the integral parts of the success of the Land Grant colleges. The training and culture within Cooperative Extension is both (neutral) research and solution-oriented: problem-solving. They are not an enforcement agency. They can do excellent research. They can work with farmers. In California, we have been dependent on a small group of Extension Specialists for key field research on produce safety. Because the culture of an institution and profession is so important, we thought that Cooperative Extension was the best place to house Medical Extension Specialists as they developed their own approaches. Extension has been repeatedly decimated over the past 20 years. In states such as California, there are fewer than half the professional staff positions (county- and university-based) remaining. It is surprising that the Land Grants have not been able to do more to protect this key part of their mission, nor to use it as a model for technology transfer in other areas besides agriculture (15). It seemed important not to further this decline. Additional funding for new positions could possibly come from non-traditional sources, including Health and Human Services, EPA, and DHS, while maintaining or improving funding for agricultural extension. A new profession does not have to be university-based, but much of the training for a new profession will probably have to be. David Acheson, when at FDA, wrote about the positive impact this approach could have in medical education. In general, we thought that the starting place should be Land Grants that had both medical and veterinary schools and a strong foundation of epidemiology, microbiology and molecular biology. There is more to disease and ecology than is reflected in my own focus, which has been mainly on emerging bacterial human pathogens, because of the food outbreak record, and on bacterial and fungal infections/symbioses of plants. Protozoan and viral diseases sourced from U.S. farms have seemed secondary. This is obviously a limited perspective. One of my friends in the 1970s studied the increase in schistosomiasis following the concrete lining of irrigation ditches in Sri Lanka, and there are many such counter-intuitive results. I tend to leave out parasitology, and– if you will pardon the older expression – fauna. In designing a broader program of training, these kind of gaps should be corrected. What farmers need to increase food safety, more than regulations, are professionals who can carry out a clinical diagnosis of their farm. The goal is to maintain a dynamic equilibrium of healthy ecological interactions, to prevent human pathogen transmission, and to avoid new pathogen emergence. Developing and maintaining a healthy dynamic equilibrium is going to be different in a complex and rich ecosystem, and more attainable, than in a sterile or semi-sterile environment (16). In California, it is striking how many key researchers on human pathogens in the farm environment are in Veterinary Medicine and Veterinary Medicine Extension. This must reflect the continuing influence of the late Dr. Calvin W. Schwabe, who founded the first department of epidemiology in a veterinary school at UC Davis and worked to develop the integration of animal and human diseases, the one medicine concept, and the one health concept discussed above (17). UCD also has run an internationally oriented master’s degree program in Preventive Veterinary Medicine (18). One can imagine a number of different types of professional training that would be good disciplinary backgrounds for this new profession. My guess is that the first new country doctor will be a vet. (1) Letter from Assistant Commissioner for Food Safety Dr. David Acheson to the author, received Aug. 23, 2007. Besides supporting the new positions proposed for UC Davis, he commented on curricula effects in medical education: “We believe that education in agriculture and human health, foodborne pathogens and zoonotic diseases are essential components of medical school programs….” (2) Quoted in James M. Hughes, Emerging Infectious Diseases: A CDC Perspective, 2001. Emerging Infectious Diseases, Volume 7, No. 3 Supplement, 494-496, June 2001. From: MacFarlane Burnet’s “Natural History of Infectious Diseases,” third edition, Cambridge University Press, 1962. (3) Burnet would probably have been appalled at the continuing devastation of then known and existing diseases 60 years later, including cholera, infant diarrhea and a plenitude of parasitical diseases. (4a) K.S. Bell, et al. 2004. Genome sequence of the enterobacterial phytopathogen Erwinia carotovora sussp. atroseptica and characterization of virulence factors. PNAS 101 (30): 11105-1110. Published online July 19, 2004. (4b) T.H.M. Smitz, et al. 2010. Complete genome sequence of the fire blight pathogen Erwynia amylovora CFBP 1430 and comparison to other Erwinia spp. Molecular Plant-Microbe Interactions. MPMI, 23 (4) pages 384-393, 2010. (5) David Satcher. 1995. Emerging infections: getting ahead of the curve. Emerging Infectious Diseases, Volume 1 (1), January-March 1995, pages 1-6. (6) Stephen S. Morse. 1995. Factors in the emergence of infectious diseases. Emerging Infectious Diseases, Volume 1 (1), January-March 1995, pages 7-15. (7) Joshua Lederberg, et al. Emerging Infections, Microbial Threats to Health in the United States. Institute of Medicine. National Academy Press, Washington D.C., 1992. (8) Multiple authors, starting with M. Davis and C. Osaki of the Seattle-King County Department of Public Health. Update: Multistate outbreak of Escherichia coli O157:H7 infections from hamburgers – Western United States, 1992-1993. MMWR, 42 (14), 258-262. April 19, 1993. (9) The hamburger outbreak overshadowed the May 1993 report in the Journal of the American Medical Association (JAMA) on hemolytic uremic syndrome (HUS) caused by drinking apple cider with O157:H7 in Massachusetts, and similar cases in New York and Connecticut. [R.E. Besser et al. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh pressed apple cider. JAMA 269 (17) 2217 – 2220. May 05, 1993.] With hindsight, similar cases are probably traceable to the 1980s, including in Sacramento, CA, where the linkage between apple juice and HUS was made but not with pathogenic O157:H7. Similarly the Listeria monocytogenes 1981 Canadian outbreak was actually the first demonstration of foodborne listeriosis. [E.T. Ryser and E.H. Marth (editors). Listeria, Listeriosis and Food Safety, Third Edition. CRC Press, 2007 (page 94)]. Cabbage grown with raw and composted sheep manure, from a herd which had had listeriosis deaths, was made into coleslaw. Outbreaks traced to sprouts and lettuce in the 1990s and early 2000s continued to implicate fresh produce, as did the Odwalla (unpasteurized) juice O157:H7 outbreak. Finally, with the 2006 spinach outbreak, FDA had had enough, and human pathogens on leafy greens and other raw produce were unquestionably a national and regulatory issue. [FDA Statement on Foodborne E. coli O157:H7 Outbreak in Spinach. Oct. 06, 2006. See “Next Steps” section. http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108761.htm ]. (10) http://ausubellab.mgh.harvard.edu (11) http://www.ars.usda.gov/pandp/people/people.htm?personid=10920 (12) http://www.epi.ufl.edu/?q=node/167 (13) Carol Rubin, et al., 2013. Review of Institute of Medicine and National Research Council Recommendations for One Health Initiative. Emerging Infectious Diseases, Volume 19, Number 12, pages 1913-1917. December 2013. (14) This common saying seems to evade Internet search terms. “It takes three generations to make a farmer” is quoted in “The Promised Land” by Leavitt Ashley Knight, “Everybody’s Magazine” Volume 28, page 788, 1913. But this was a play on the more common “It takes three generations to make a gentleman.” I’ve heard it in the rural West for more than 35 years, and it was quoted (with “make”) by Oklahoma Labor Commissioner Mark Costello in 2012. (15) As an observation, the decline of Extension in CA seems to be in an inverse function to the number of university positions at the vice chancellor level and above. (16) One of the most protected and controlled medical environments is the hospital intensive care unit (ICU). Yet the control of methicillin-resistant Staphylococcus aureus (MRSA) and other multi-drug resistant organisms in ICUs is not only difficult but also the subject of considerable controversy on even the research results of control methods. See: K.T. Kavanagh, et al., 2013. A perspective on how the United States fell behind Northern Europe in the battle against methicillin-resistant Stapylococcus aureus. Antimicrobial Agents and Chemotherapy, volume 57, number 12, pages 5789-5791. December 2013. Published ahead of print, Oct. 7, 2013. The analogy in food safety is sprout production, also under the most protected and sealed environment imaginable, in this case for crop production, where pathogen introduction and spread is facilitated by the absence of any ecological controls of pathogens in a semi-sterile environment. Similarly, control of human pathogens in the crop phyllosphere (the above-ground aerial surface layers of plants) seems to be more difficult because of the reduced richness of plant surface ecology relative to soil and roots (rhizosphere ecology). Control systems still may be possible, but any biocontrol or beneficial microbial treatment is going to be much less well-buffered. (17) R.S. Nolan, 2013. Legends: the accidental epidemiologist. Dr. Calvin W. Schwabe fathered a generation of veterinary epidemiologists. JAMA News, July 01, 2013. Posted June 19, 2013. https://www.avma.org/News/JAVMANews/Pages/130701m.aspx (18) http://www.vetmed.ucdavis.edu/mpvm/]]>
(14) Committee on Toxicants and Pathogens in Biosolids Applied to Land, National Research Council. Biosolids Applied to Land: Advancing Standards and Practices. The National Academies Press, National Academy of Science. 2002.
Human pathogens constantly flow from urban environments and animal production into farm environments, contaminating water and soil, and finding a home in wildlife. Then we ask farmers to deliver pathogen-free fruits and vegetables.
Produce is increasingly sold as centrally processed fresh-cut products, diced fruits and cut salads in wholesale and retail packages, facilitating cross-contamination without a kill step. Then it is shipped in closed plastic containers that can allow pathogen maintenance and growth, with use-by dates of about three weeks from harvest. A cold chain for processing and delivery, that prevents E. coli O157:H7, for example, from multiplying allows Listeria to survive or thrive.
On farm, two of the greatest hazards have nothing to do with actually growing the crop. Washing, cooling and storage — on-farm handling, and providing sanitary conditions for employees that do not lead to human-to-produce contamination. On farm regulation focuses on field practices.
Farmers are asked to cut down trees and shrubs, which provide habitat for birds; to fence off fields and water sources, which might attract deer and other large species; to bury poisoned bait stations around the perimeters of their field, to prevent squirrels, mice, shrews and other rodents from entering their fields; to remove the low-growing filter strips next to streams, which actually prevent fertilizer run-off and some pathogens from entering the streams, because they also attract wildlife.
They are supposed to have management practices that prevent flies and other insects, snails and other invertebrates, frogs and other amphibians, from carrying human pathogens onto their fields. All have been shown to be potential carriers. Since farm food safety is focused on growing operations, sources that provide human pathogens to contaminate the fauna of a farm ecosystem are not regulated.
I use produce outbreaks where there have been outbreak investigation reports as case studies that provide some, limited, evidence. They may or may not support this thesis. They provide a window into how produce safety is being framed and understood.
In this article two of the 2011 outbreaks are covered: the Jensen Farm Listeria-cantaloupe outbreak and the Jaquith Farms O157:H7 strawberry outbreak. The implications for food safety policy are covered on the national and then small-farm scales.
Using the normal rules of thumb that make sense of food safety, there should never have been a Jensen Farms lLsteria cantaloupe outbreak. We may be in a changed world.
The FDA environmental assessment Report (1) states:
“This is the first listeriosis outbreak associated with a whole fruit or vegetable raw agri- cultural commodity. Listeria monocytogenes contamination has historically been associ- ated with ready-to-eat and processed food products, such as deli meat, unpasteurized cheese, raw milk, fresh-cut fruit, and fresh-cut vegetables, and is typically thought of as an environmental contaminant of food plants. Known reservoirs for Listeria monocytogenes include ruminant animals (e.g., cattle, goats, and deer), decaying vegetation, and cold, wet, and difficult to clean environments.”
Fresh-cut fruits and fresh-cut vegetables are recent food additions to the list. The usual combination of concern is Salmonella and cantaloupes, not Listeria,
With 33 deaths out of 146 reported cases, the Listeria-cantaloupe outbreak was one of the worst foodborne outbreaks in over 80 years. The death rate is close to 25 percent.
In many outbreak investigation reports I have read, “absence of evidence” in a processing facility is taken to be “evidence of absence,” even when it was documented that management had spent up to a week sanitizing their plant post-outbreak and pre-inspection.
In the Jensen Farms case, the equipment design and facility design prevented sanitation, both before the outbreak and before the investigation. Outbreak strains were found within the packing facility and cold storage. It was a rare case of presence of evidence, and some equipment had to actually be cut out and replaced in order to be sanitized to show no Listeria.
When the FDA looked for a source of contamination on-farm, they took the more usual approach.
“All environmental samples collected in the growing fields were negative for Listeria monocytogenes.”
“FDA has determined that the growing environment cannot be eliminated as a potential contributor in the introduction of Listeria monocytogenes contamination…”
That is true, as far as it goes, but it is also true of pretty much every farm in the country, given the ubiquity of ruminants and of Listeria in the environment. It does not do much to explain the origins or root cause of this unique case (2).
Many observers, including Bill Marler, have characterized the on-farm handling as gross negligence; and condemned the lack of prevention by food safety auditors. Jensen Farms had passed its audits for safety. Its papers were in order.
Now Jensen Farms appeared to me to be a large or very large operation responsible for a national outbreak centered on Colorado and Texas. Based on the recalled canta-Jim loupes alone, it looked like it probably was one of the 5,600 largest farms in the country with farm gate sales of over $5 million.
Jim Prevor had outstanding coverage of the outbreak in his Perishable Pundit articles but with a very different perspective (3):
“If Wal-Mart really put food safety first, it would have never bought cantaloupes from this small producer, which produced in a six-week season what California ships in a day.”
A small farm with local production, and small farms are not safe? We do not seem to be talking about the same outbreak. When I asked him about this, it turned out we were using the same words with different meanings from different contexts (4).
Prevor’s context is the national produce market, dominated by the largest buyers and their major produce suppliers. The vocabulary of the buyers is different than farmers, and perhaps most of their customers. “Local” can be defined as “within a 7 hour drive” by Whole Foods. “Small” in this world of the top strata within the top 5,600 farmers (the number is for all crops, not just produce) includes Jensen Farms. He estimated that maybe 10 of the very large melon farms in the U.S. had the money, size and motivation for a particular kind of food-safety processing equipment.
His main point was that the largest buyers do not make food safety their first priority in buying decisions, when one looks at how they actually operate. For example, if they can label produce as “local,” in their terms, this can have top priority. Buying from a “small” farm like Jensen Farms may meet a marketing decision for “local” produce, despite there being major farms with better food safety capacity. Therefore “safe versus local” as a choice. He was using the words as industry terms; to people outside the industry, they read differently.
Prevor then makes a critical observation: that how major produce buyers are actually rewarded determines the effectiveness of food safety commitments. He wrote an interesting analysis of overall food safety issues, including changing liability so that food safety is not unrewarded (5). On a national policy level his
approach and mine have many points in common. Trying to use common language:
The food system as a whole should be evaluated by a hazard analysis. The worst hazards overall should have the highest priority for being dealt with.
To really have an effective outbreak, one has to have a way to spread and multiply the pathogens; incidental contamination won’t do the job.
Any centralized source for contaminating produce, which spreads, multiplies, and preserves pathogens is more dangerous than any incidental source. Handling and processing can be major centralized sources in handling, and fresh-cut is a special hazard in part because it is food processing without a kill step.
Widespread hazards from the farm environment also can create broadly contaminated produce, for example from contaminated surface water or ground water or flooding.
Within farming operations, handling operations, packing and cold storage as in Jensen Farms, might be the highest priority after water quality and direct exposure to feces. Texas may be the state that has gotten this right, with a new law focused on on-farm handling (and processing).
Incentives for individual decision makers also can determine over-all food safety. Company incentives for major buyers are also a critical control point for food safety. Business expansion and growth with increased contracted deliveries can be a hazard for two reasons: newly acquired equipment can be a hazard, and pressure to deliver can lead to pressure on safety standards (6).
However, right now, we are quite distant from this approach.
On October 06, 2006, as the spinach outbreak was ending, the FDA put out one of the bluntest statements issued by a bureaucracy, saying to the lettuce and spinach industry that enough was enough, clean up your act or we will do it for you; and inadvertently creating the category of “leafy greens” (7).
In my view, what happened next was a tactically brilliant takeover by the major processors and handlers of the produce safety issue, starting with the California LGMA. It left farmers under the control of processors, who somehow left out the food safety issues of handling and processing operations.
The fundamental flaw in this approach is that in the absence of an effective kill step, for fresh-cut in particular, they basically have to sterilize the farm environment instead, which is impossible. It can be an ecological, social and strategic disaster. [For detailed discussions of the impact on farm environments see the Wild Farm Alliance website (8).]
This “processor-favoring” approach dominates all discussions of food safety on farms.
A cold-hearted analysis of sprouts as a commodity, that takes food safety as the first priority, is that they should be cooked. A cold-hearted analysis of fresh-cut as a commodity is that it is a food safety failure. We could go back to local production consumed within a few days of harvest for salad greens; local food service doing their own cutting for “fresh-cut” fruits and vegetables. This eliminates the pathogen increase time from farm to retail to consumption, and other factors. As commodities, both sprouts and fresh-cut lack kill steps and provide humid sealed environments in plastic containers for pathogen maintenance and increase.
Food safety, looked at cold-heartedly, is not the top priority when fresh-cut processing is analyzed. Other factors, including market share, value-added, and consumer convenience are priorities. Priorities have to be balanced, perhaps. It is not a justification for displacing all food safety concerns onto farms.
That still leaves the whole produce outbreaks.
It’s curious. No one studies the effectiveness of small farms at achieving food safety. I once heard an FDA hearing commissioner explain to an audience of mainly large processors and grower-shippers: “what you call a spinach harvester, I call a pathogen inoculator.” Harvesting on small farms is completely different, but which techniques are better or worse for food safety in comparison?
Can small farms really have a better safety record, and if so why; due to which practices?
The 2011 E. coli O157:H7 outbreak in Oregon involving Jaquith Strawberry Farm, initially thought to be a small grower, looks like a fairly large producer, distributing through multiple channels. One of them, Ron Spada Farms, was reported to have recalled 4,800 flats still in inventory after the outbreak trace-back. Other strawberries went to roadside stands, farm stands and farmers markets as if they were self-produced (9).
This can make “knowing your farmer” problematic. The farmer wasn’t small, and the people who looked like small farmers weren’t farmers, or had not grown the crop. Consumers may have had the illusion they were buying from a small local farmers in many parts of Oregon. Only the epidemiological trace-back found the common source.
There was no complete outbreak investigation report, but the outbreak strain was found in deer feces in the actual sole-farm’s strawberry production field. If there is a widespread pathogen contamination of deer, it did not affect multiple producers.
UC Davis field studies showed that O157:H7 in deer in California was quite rare, challenging the frequently made association with some California outbreaks. Oregon might be different.
There are many actual small strawberry producers in Oregon. They all managed to avoid causing outbreaks. There also are very large producers (super-farms) including Driscoll’s, which appears to have an outstanding food safety record, who also grew and sold strawberries safely.
It looks like this outbreak could have been due to negligence, probably at harvest, on one “medium” sized farm; that ordinary precautions might have been enough to prevent contamination even with the presence of E. coli O157:H7 in local deer.
Is there a statistical distribution of food safety by farm size? We know that the pressure of growth can be a major hazard point, throughout the produce outbreak record. Food safety can slip under the pressure to meet increased sales.
Actual incentives are not just a hazard point for the largest national produce buyers; they can compromise safety priorities for individuals or operations anywhere in the supply chain.
This outbreak involved multiple illnesses and one death. It was a relatively local, perhaps “regional” and “limited” outbreak. That is cold or no comfort to the family of the person who died or to those who became ill.
One can imagine a consensus on a set of national priorities focused on fixing the worst hazards first, an FDA that puts more emphasis on causal analysis of outbreaks as Michael Taylor has called for, and co-management of food safety and the preservation of habitats and the farm environment, as both the California LGMA and the Wild Farm Alliance have called for. Overall food safety would be greatly improved.
That is not where we are, right now.
What can I say to a small organic farmer who asked me about the impact of the three major outbreaks in 2011, and the oncoming regulations under the Food Safety Modernization Act (FSMA)?
By acreage, she and her husband are small farmers under any U.S. definition, owning and farming 20 acres. By farm gate sales they are in the top 20 percent of U.S. farmers.
Something like this:
Most of the work being done on food safety is not for their benefit and the regulatory schemes in process may be mostly harmful, imposing burdens without improving overall food safety because it avoids dealing with major hazards in the overall system of food production.
Regulators have a compulsive belief in the efficacy of complex paperwork, almost for its own sake. It could be called faith-based paperwork. The same money could be spent on solving problems, but they assume the solutions are all known.
In contrast, they should try and align their ordinary commercial records with food safety, an approach also favored by some major food safety consultants for the largest firms.
They need to know how to approach a recall, and be able to trace-up deliveries quickly.
Regulation will be based on the largest farms, even when these have been the source of multiple recalls or even outbreaks. Systemic failures such as the Listeria-cantaloupe case will not lead to new systemic approaches.
Expect to pay for third-party audits that lack meaning; and Extension services, which might have been helpful in actually solving safety problems, have been cut in half.
Reduce known hazards. There may be some crops, and some ways of growing and harvesting crops, that have to be dropped. Look carefully at washing, handling and packing.
Make food safety an over-riding contingency of sale in delivery contracts. Make food safety a part of any plan to grow production and sales, from the start. Consider increasingliability insurance and buying business interruption insurance.
Improved safety, and meeting all imposed regulations, provides no legal protections. In- cidental contamination beyond farmer control still comes under strict liability.
They can improve their odds.
I can guess that the ethical standards they use in all their approaches to farming do have a beneficial impact on food safety as well, for their size of farm, because they frame their own personal incentives, and oppose temptations to cut safety to meet a dellivery.
“Don’t screw up.” Know the common important hazards and have a systematic plan to control their risks.
You are always,farming on the brink of disaster.
(1) Environmental Assessment: Factors Potentially Contributing to the Contamination of Fresh Whole Cantaloupe Implicated in a Multi-State Outbreak of Listeriosis October 19, 2011
(2) A lot of specific information is missing from the report that could be useful in analyzing the outbreak and preventing future outbreaks. Jensen Farms used equipment that could not be cleaned, that came from a non-cantaloupe use (supposedly for potatoes). Was the previous operation contaminated with the outbreak strains? Similarly, a truck was used to bring culls to a cattle operation as feed and could have been a contaminant source for the cleaning and packing facility. Were outbreak strains found in the cattle operation? A neighboring field was reported to have received municipal sewage sludge, of unreported treatment, and Listeria bacteria are resistant to degradation in sludge. It would be important to know if the outbreak strains were found in the neighboring field.
I read ” All environmental samples collected in the growing fields were negative…” in a lawyerly fashion, and wonder why environmental samples collected in the larger environment were not reported on.
For a comprehensive review of Listeria and food safety and Listeria in the agricultural environment see, for example: Listeria, Listeriosis, and Food Safety, Third Edition
Edited by: Elliot T. Ryser, Michigan State University, East Lansing, USA; Elmer H. Marth, University of Wisconsin, Madison, USA March 27, 2007 CRC Press
(3) Perishable Pundit: The Cantaloupe Crisis: The Truth That Dare Not Speak Its Name: The Priority Can Be Safe Or The Priority Can Be Local, But It Cannot Be Both , October 4, 2011
(4) Perishable Pundit: “A Choice Had To Be Made: Which Was The Top Priority: Buying Cheap, Buying Regional Or Buying Safe?”
(5) Jim Prevor “How to Improve Food Safety” The New Atlantis, May 21, 2010
(6) For example, that is my reading of the outbreak reports for two major cases: Odwalla juice, and Dole spinach, as well as others.
(7) FDA Statement on Foodborne E. coli O157:H7 Outbreak in Spinach, October 06, 2006.
See “Next Steps” section.
(9) Deer Confirmed as Source of Strawberry Outbreak August 18, 2011, Food Safety Website
Daniel B. Cohen is the owner of Maccabee Seed Co., an agricultural research-and-development and consulting company in Davis, CA. See also “Sizing Food Safety Regulations to Fit the Farm” and “An Outbreak Like Germany’s Could Happen Here”]]>
When I was asked by a small organic farmer in California what the implications were of the 2011 O104:H4 European outbreak, I said it could happen here. It also could have worse consequences here.
Let’s review the O104:H4 outbreak conclusions:
In Germany there were nearly 4,000 cases, 54 deaths and 845 cases of hemolytic uremic syndrome (HUS) — unusually, mainly in adults (1) (2).
One hundred percent of the German O104:H4 cases were associated with the production of fenugreek sprouts on a single organic farm in Lower Saxony, with regional distribution centered on Hamburg. The organic farm was found to have satisfactory sanitation and proce- dures and is back in operation.
The epidemiological evidence indicates that a particular lot of fenugreek seed, sold from or produced in Egypt, had become contaminated with this highly unusual outbreak strain. The same outbreak strain was associated with a secondary outbreak in France, also linked to fenugreek sprouts from the same seed lot, but produced as children’s projects and served at a local festival.
Where it starts getting more complicated is that sampling for O104:H4 found no positive results from fenugreek seed lots. There also were no positive results from tens of thousands of environmental samples taken during the post-outbreak investigation, although contaminated sewage and streams were found during the outbreak period. The conclusion was that much better culturing methods were needed for tracking the environmental presence of this pathogen.
Partially confirming data on O104:H4 having a regional origin comes from European tourists, who have become infected with related, but not identical O104:H4 strains after trips to Turkey, Egypt, Tunesia and Afghanistan (3). This supports the idea of regional endemic sources of O104:H4, but not where and how the fenugreek seeds were actually contaminated.
The Egyptian Ministry of Agriculture strongly objected to the identification of Egyptian seeds as the outbreak source in the absence of positive detection of the outbreak strain.
That is where things stand now, with clear identification of the food carrier and the source farm, but with some ambiguity as to the ultimate root cause of the outbreak and the specific mechanism of contamination.
The primary role for epidemiologists and public health officials during a foodborne outbreak, like the O104:H4 outbreak in Germany and France, is to stop the epidemic. If an intervention stops the outbreak, then the identification of the cause is verified. Identification is based on epidemiological evidence; and direct detection is a kind of luxury of evidence. The tools they use have been developed primarily for this purpose, and may be less useful in determining other aspects, such as the search for confirmation of root causes and mechanisms of contamination.
How did the context in Germany shape the investigation and outcome in ways that might be different from the U.S.?
When O104:H4 infections hit Germany, the first “vegetables of interest” were E. coli-contaminated organic cucumbers from Spain, revealed publicly before there was confirmation that the E. coli did not match the outbreak strain. This was curious because of the lack of any outbreak — O104:H4 or other — due to the same two suppliers in Spain itself. Not that any pathogen finding is good, but at least one of the sources was told his cucumbers were seen spilled on the ground at the Hamburg Central Produce market. Perhaps even the non-outbreak pathogens were not due to the farm source. Spanish farmers and the Spanish government saw this as lingering prejudice against Spain.
European Union and German payments were later made to partially compensate Spanish growers.
In contrast, as far as I can tell, there have been no EU lawsuits against the farm for the actual outbreak. So what would be a strict liability case in the U.S. appears to be neither negligence nor foreseeable under the German equivalent law.
There were at least two cultural presumptions, also different in Germany from the U.S. that may reinforce this. Organic appears to be regarded as naturally safer, to the point where the Health Minister was quoted as saying “Of course we know that organic is not really safer.” And fenugreek sprouts were part of a common elaborated sprout cuisine in which sprouts are perceived as a natural healthy food.
U.S. epidemiologists seemed stunned that raw sprouts were not on the top of the list for food surveys in the case/control studies, given the outbreak histories for sprouts. German epidemiologists initially thought including sprouts would only create false positives in case/control studies because sprouts, in general, are so common a food item.
A CDC-FDA-state investigation of a similar outbreak here in the United States would focus more quickly on sprouts given the long history of sprout outbreaks. But perhaps the greatest difference in attitudes is toward organic farming.
Why should Europeans perceive organic farming as safer than conventional agriculture? A substantive reason is that the rules elaborated for organic animal production may provide better protections against transmission of human pathogens, both in the U.S. and the EU. Here I will use U.S. examples.
The Organic Foods Production Act (OFPA), passed in 1990, created the National Organic Program (NOP). Labeling was the basis of the law, that food be correctly identified as organic, but that was the hook for protecting the word “organic” and defining the farming practices.
The U.S. rules for organic compost, in an abundance of caution, were based on EPA guidelines for the use of human waste. This was supposed to provide a wide margin of safety for the use of animal manures. In many studies of E. coli O157:H7 or surrogate pathogen survival under these rules, undertaken since the spinach crisis, it appears that there is a margin of safety even for these STEC but it is not particularly wide.
Organic rules prevented the re-feeding of waste meats from slaughter, including brains and nervous tissue, long before mad cow (BSE) emerged, They ban the non-therapeutic use of antibiotics. They ban the use of dried cow manure as dairy bedding. They require more adequate space for animals and birds so that, for example, de-beaking of chickens is not allowed, or needed due to the insanity produced in birds from overcrowding in a stimulus- deprived environment. The overcrowding of birds, or cattle in CAFOs, may facilitate disease transmission and therefore, in some CAFOs, the routine use of antibiotics, but there also are ranchers who buy antibiotics by the multi-gallon tote without confining their cattle. Most antibiotic sales (by volume) are for animal use in agriculture, between 70 – 80 percent (4).
Organic rules ban the re-use of manures as animal feed.
Last summer, when I was working with a “medium-sized” organic vegetable producer, the price of composted chicken litter (for organic fertilizer) had gone through the roof because of increased demand for pelleted chicken litter in cattle feeding. Thanks in part to the California Leafy Green Marketing Agreement rules, at least there is now research interest in how exactly to p
asteurize different manures for conventional agriculture [for Salmonella, see (5)].
Since it is not an obvious step to feed brains or manure back to cattle, where did this come from? As far as I can tell, many of the changes in conventional animal production after World War II were driven by economists looking for efficiencies in production, without considering biology. They looked at the nitrogen content of manure, for example, and saw inefficiencies in cattle digestion (!) which could be re-captured, rather than fertilizer for soils (6).
All of these factors had to be specifically banned in organic animal production either be- cause they were once routine, or still are routine, in U.S. conventional agriculture. They all could play a role not only in the transmission of human pathogens but in the selection and evolution of more dangerous pathogens: by multiplying and spreading pathogens in closed biological systems, by acting as selection systems for antibiotic resistance and increased tolerance of acidity and other stresses, and by favoring selection for lateral transmission of groups of pathogenicity traits between pathogen species.
Even when there is an underlying substantive basis, EU positive attitudes toward organics may be due more to complex cultural attitudes than to substance.
Food safety criticisms of organic that I hear or read in the U.S. are often based on a failure to understand that most agricultural manures go into conventional agriculture, and for decades only organic agriculture had rules that were specifically designed to prevent human pathogen transmission (7). There is nothing to prevent conventional agriculture from using similar approaches, and sometimes there is a convergence toward safety.
What are some contrasts we could expect in a similar outbreak in the U.S.?
U.S. epidemiologists might identify a sprout-caused outbreak more quickly, and limit the duration of the outbreak.
There would be great differences in the cost of paying for health care, and who is liable for payment.
There would be no compensation to farmers who were misidentified as growing an outbreak crop. There would be no exemption from liability for the farm that caused the outbreak, because they have strict liability for putting out adulterated food.
There would be no presumption that “organic is safer.” There are still major constituencies for whom organic production is an ambiguous, minor but often irritating fact of life in farm and food production industries. They could condemn an outbreak as due to being organic (8), rather than despite being organic, as in the EU.
A final question from the O104:H4 outbreak: what can be done to improve the safety of raw sprouts, whether produced on an organic farm or a conventional farm?
For the actual production process itself there were only two factors: seeds and water. Sprouts were produced with clean equipment in a well-sanitized manner. There is a minimal ecology to sprout production, which may be part of the problem.
Changing from sprouts to greenhouse production of baby vegetables in soil might seem to be a minor change, (let the sprouts grow some true leaves!), but the research on both greenhouse and field production indicates that human pathogen uptake from the soil to plants is highly inhibited, and this can include seed-borne human pathogens.
In a practical sense, sprout safety requires either human pathogen-free seeds specifically produced for sprouting, a kill step that can be used after seeds are sprouted and pathogens are exposed for treatment, or both. People could also cook their sprouts, the original use for soy sprouts from the very beginning of soy as a major food in China (9).
The rational basis for organics includes an ecological approach to farming and an ap- preciation for the biological opportunities and constraints on and around their farms. Conventional farmers can use the same approaches and often do.
I also see sprout production, and the history of sprout outbreaks in this country, as a metaphor for the current approaches to regulating food safety on farm. There could not be a more simplified system for contained and controlled food production than sprouts,
yet it fails repeatedly. Applying the same approaches used in sprout production to the complexity of farms and the farm environment seems both irrational and doomed to fail.
1) Epidemic Profile of Shiga-Toxin-Producing Escherichia coli O104:H4 Outbreak in Germany N Engl J Med 2011; 365:1771-1780 November 10, 2011
(2) STEC Workshop Reporting Group. Experiences from the Shiga toxin-producing Escherichia coli O104:H4 outbreak in Germany and research needs in the field, Berlin, 28-29 November 2011 . Euro Surveill. 2012;17(7):pii=20091. Available online:
Date of submission: 10 February 2012
(3) Eurosurveillance, Volume 17, Issue 4, 26 January 2012 Rapid communications OUTBREAK OF HAEMOLYTIC URAEMIC SYNDROME DUE TO SHIGA TOXIN-PRODUCING ESCHERICHIA COLI O104:H4 AMONG FRENCH TOURISTS RETURNING FROM TURKEY, SEPTEMBER 2011 N Jourdan-da Silva1, M Watrin2, F X Weill3, L A King1, M Gouali3, A Mailles1, D van Cauteren1, M Bataille4, S Guettier4, C Castrale5, P Henry5, P Mariani6, V Vaillant1, H de Valk1
-antibiotics-sold-in-us/”>-antibiotics-sold-in-us/ (5) Appl Environ Microbiol. 2012 Feb;78(4):1302-7. Epub 2011 Dec 16. Validating Thermal Inactivation of Salmonella spp. in Fresh and Aged Chicken Litter. Kim J, Diao J, Shepherd MW Jr, Singh R, Heringa SD, Gong C, Jiang X. http://aem.asm.org/content/early/2011/12/12/AEM.06671-11.short?rss=1 (6) The greatest contribution to food safety in the future might be to require biology and microbiology courses for economics students. And there is an interesting possibility that the ethical considerations used in designing organic rules for animals will be shown to also have had long term economic value. (7) For example, I happened to be at one of the organizing meetings for the CA LGMA when CFSAN investigators revealed that the farm associated with the spinach outbreak was organic in transition. The chief scientist of a major produce organization exclaimed “We were so close!” Perhaps he was under stress. The same attitude came from state senators at a California hearing on the same crisis. “Suppose I don’t want my food grown organic with manures.” Most outbreaks and recalls are in conventional produce and it looks to me like there are fewer outbreaks due to organics than a proportional expectation would predict. (8) GMO biotechnology companies sometimes have an ambiguous relationship to organic food production. When they want to avoid labeling GMO crops and foods they can point consumers to organics, where GMO is prohibited. When organic farmers block release of crops like GMO herbicide-resistant alfalfa, they get incensed. The USDA’s official position for the last few years has been “people; we can all get along.” (9) See page 296 of H.T. Huang’s magisterial Fermentations and Food Science; Volume VI: 5 of the series Science and Civilization in China (Joseph Needham). Cambridge University Press, 2000. The first written documentation of soy sprouts in Chinese history includes the steps: wash after three days and fry in oil.
(5) Appl Environ Microbiol. 2012 Feb;78(4):1302-7. Epub 2011 Dec 16.
Validating Thermal Inactivation of Salmonella spp. in Fresh and Aged Chicken Litter.
Kim J, Diao J, Shepherd MW Jr, Singh R, Heringa SD, Gong C, Jiang X. http://aem.asm.org/content/early/2011/12/12/AEM.06671-11.short?rss=1
(6) The greatest contribution to food safety in the future might be to require biology and microbiology courses for economics students. And there is an interesting possibility that the ethical considerations used in designing organic rules for animals will be shown to also have had long term economic value.
(7) For example, I happened to be at one of the organizing meetings for the CA LGMA when CFSAN investigators revealed that the farm associated with the spinach outbreak was organic in transition. The chief scientist of a major produce organization exclaimed “We were so close!” Perhaps he was under stress. The same attitude came from state senators at a California hearing on the same crisis. “Suppose I don’t want my food grown organic with manures.” Most outbreaks and recalls are in conventional produce and it looks to me like there are fewer outbreaks due to organics than a proportional expectation would predict.
(8) GMO biotechnology companies sometimes have an ambiguous relationship to organic food production. When they want to avoid labeling GMO crops and foods they can point consumers to organics, where GMO is prohibited. When organic farmers block release of crops like GMO herbicide-resistant alfalfa, they get incensed. The USDA’s official position for the last few years has been “people; we can all get along.”
(9) See page 296 of H.T. Huang’s magisterial Fermentations and Food Science; Volume VI: 5 of the series Science and Civilization in China (Joseph Needham). Cambridge University Press, 2000. The first written documentation of soy sprouts in Chinese history includes the steps: wash after three days and fry in oil.]]>
What exactly is a small farm?
The last Census of Agriculture data is from 2007, although a new census will be made this year. In 2007, there were 2.2 million “farms.” About 1.3 million of these had farmgate gross sales of less than $10,000. Only 57,000 of them had farmgate sales of $1 million or more (sales, not profit). (1)
This is about the sales volume of a medium-large urban food co-op in the 1970s, 40 years ago, when a million was worth a million, you might say. All of agriculture in the U.S. had farm sales of around $300 billion, roughly equivalent to a single major multi-national oil company.
Almost all the farmers I work with are part of the 57,000. But the 57,000 are as stratified as the whole farm population. About 10 percent of them (5,600) have farmgate gross sales of $5 million or more. Some of the farms I work with are among those 5,600. The USDA doesn’t break down the 5,600 any further, but there are also a wide range of farm sizes within the category “$5 million or more.”
Some are truly ‘superfarms’ with significant share of production of an entire commodity. When I analyzed California leafy greens using the 2002 Census of Agriculture data, for example, 59 lettuce farms grew 69 percent of the state lettuce crop on 1,000 acres or more. In 2006, USA Today reported that a single spinach grower produced almost all of the spinach for Fresh Express, at that time the largest fresh-cut processor in the country.
What is the definition of a “superfarm”? There are both organic and conventional vegetable operations in California that farm more than 20,000 acres — almost 30 square miles. They could certainly qualify, but where the cutoff would be is not clear. Some are simply outstanding operations, including for food safety. Perhaps 10 percent of the farms with sales above $5 million, or about 500 farms, have very large roles in many produce categories.
In 2011, the USDA’s Economic Research Service published “Financial Characteristics of Vegetable and Melon Farms” by Mir Ali and Gary Lucier, which shows continuing concentration across all vegetable and melon types (2).
The fact that the same individuals may control more than one farm “entity” complicates matters, but in general this means that farm statistics understate the degree of concentration and disparity between farm sizes.
The term “farms” includes farms and ranches, both crops and animal production, including large CAFOs (concentrated animal feeding operations), as well as woodland and pasture farms.
It’s pretty easy to guess that the 1.3 million farms with less than $10,000 in sales should be in the “very small” farm category. But where this category ends is not clear.
The Tester Amendment to the Food Safety Modernization Act, under which the FDA is writing new farm food safety regulations, establishes a kind of definition of small as less than $500,000 in (farmgate) sales and selling most production “locally” within a state, or within 275 miles if sales cross state lines. From my perspective, and looking at the O104:H4 outbreak in Germany, this sounds more regional than local, but regional is not a defined category.
The FDA briefly describes the FSMA as follows:
“FSMA is the most sweeping reform of FDA’s food safety authority in more than 70 years, and a law of this scope and complexity often comes with direction from Congress for the federal agency responsible for implementing it to go through a process called rulemaking…”
Under the new food safety law, FDA will be issuing a number of rules including a preventive controls rule in food facilities, a foreign supplier verification rule, and a produce safety rule.”
For farmers, the question of how a “food facility” is defined — because it can include on-farm basic operations — and how rules for produce are developed, and where federal regulation stops and state regulation takes over, are all crucial questions, which are partly dependent on the size of the farm being “small” or not. Very different language is used by different parts of the industry for both “small” and “local.”
One could say that a small farm has average annual sales somewhere in the range of from $10,000 to about $500,000, but less than $1 million dollars, recognizing that a lot of jockeying is going on because of the legal implications of the categories for food safety. By this very inclusive definition, there are about 840,000 “small farms,” if by “small” you include what might be more reasonably called “medium” or “medium-large” farms.
There are no “average” farms except as an inappropriate statistical fiction.
Crops are not evenly distributed geographically. For example, California grows about 1/2 of U.S.-produced fruits, vegetables and nut crops, and for produce like leafy greens, the Salinas valley and the desert production of the Imperial Valley and Arizona dominate U.S. production in very specialized sub-regions, which are often dominated in turn by the largest farms.
It might make more sense to write leafy green regulations geared toward the largest farms in these areas rather than trying to impose a one-size-fits-all on smaller farmers growing a variety of crops in varying climates and geography. Many fruits and vegetables to be regulated as “produce” have similar concentrations, with large farms producing much of the crop in specific regions.
Just using very broad statistical indicators, it should be obvious that U.S. agriculture is highly heterogeneous. At an ecological, geographical, geological, biological, social, and farm or farm-field level the heterogeneity is infinitely more complex.
It does not seem like a rationally conceivable project to write national regulations for food safety on-farm, although it may be legally mandated.
Not, that is, if by regulation one means detailed instructions for every farm and farm size and crop and diversity of production and integration of production between crops and animals and ecosystem interactions and farm benefits to the environment. Even looking just at what I look at (mostly) in food safety, the production of “specialty crops” (fresh vegetables, fruits, nuts and other raw products), U.S. farm complexity defeats centralized rule-making that could function.
The knowledge needed, if anywhere, is in the fields.
One could get far more impact for every regulatory buck spent if the initial goal were the regulation of the 6,000 largest farm operations. These truly approach what we think of as “industrial agriculture,” with very large field sizes — often in large monocultures, and often collectively dominating the overall production of a particular commodity.
They have many characteristics in common and look like the place one could start implementing national regulations. One could make sure these 6,000 farms had good food safety procedures and then perhaps move on to the next 50,000 farms with sales over $1 million — a tougher nut to crack but conceivable. If we had well-validated food safety procedures f
or the very largest farms, that is. Do we?
Apparently not. Otherwise hundreds of millions of dollars of ongoing and new research projects — federal, state, and private — would not be going into fundamental and operational (or validational) research on farm food safety, in particular aimed at these kinds of farms. A several-year study headed by the University of Maryland will for the first time
have some access to the private data of the largest produce farms and processors, for example.
There is a further problem in fitting on-farm food safety to the scale of farming. If one looks at some of the metrics and “‘supermetrics” developed by groups like the California Leafy Green Marketing Board and private buyers, respectively, it looks to me like the very largest farms and superfarms have the best fit to the metrics and face the least economic damage from applying them.
One of the metrics commonly used is to buffer production away from natural habitat. This can be a reasonable issue for a 1,000 acre field situated on a multi-thousand acre production farm. But it gets more expensive the smaller the field and the closer the natural habitat. At the limit of difficulty would be a small field, close to habitat, where the requirements for buffering (set on a national standard) leave no field left to farm.
But my impression is that economic constraints start with even quite large farms.
There is quite a lot of well-validated data, not on how to farm safely, but on how to farm unsafely in particular ways, on how to specifically screw up. Some of these are general principles that appear truly scale neutral. Some are commodity-specific and one hopes reflected in parts of the commodity-specific guidances by the FDA and others, and are mostly scale neutral.
The human pathogens of major concern in food safety have their individual ecologies and preferences. Situations where each pathogen thrives can be looked for. These include Shiga toxin-producing E. coli (such as O157:H7), Listeria, Salmonella, Shigella and Campylobacter. There are also common pairings between a particular crop and a particular pathogen found to be of greatest concern for that crop. Salmonella on cantaloupes is a common pairing.
So I could imagine several thousand pages of regulations being written for food safety on “small” and “very small” farms being replaced, instead, by one sentence: “Don’t screw up.”
For example: “Don’t harvest strawberries, produced near ground level, next to wild animal droppings. And if rain or sprinkler irrigation could have splash-dispersed potential pathogens, expand the area where there is no harvest.”
This is a strategy of known defect reduction, or elimination, often used in sensory control of food quality. Defect reduction is an achievable process of improvement on any farm. It would seem to be more effectively approached through education and cooperative efforts, leaving farmers in control — as they should be. They know their farms. It could have dramatic impact, over time, in reducing food safety hazards on-farm. This is a statistical reduction of harm, effective on average, and not a guarantee. There are fairly simple steps that farmers have control over that would increase food safety.
On the farm, water quality, presence of excrement that can contact a crop, worker health, and crop “handling” — preparation for sale off farm — seem to be the main hazard points under control. Yet most regulations try and comprehensively detail every aspect of growing a crop. Growing crops up until harvest might be the least important factor in increasing overall food safety. The steps from harvest to the consumer are the places that seem to have the greatest impact on food safety. Those steps include handling, processing, commingling, storing, packaging, labeling (such as use-by dates, use-instructions), transportation (and consumer or retail behavior), which can have far more destructive effects. Except insofar as a grower is involved in all of these, they remain out of the grower’s control.
In 2007, we found that processor/handlers of whole produce, and fresh-cut products, were the known sources of many multistate produce outbreaks and recalls over the previous 20 years. We also found that fresh-cut produce, compared to whole produce, looked like the cause of more outbreaks than would be the case if on-farm contamination was all that mattered.
For E. coli O157:H7 and many produce categories like leafy greens, we came up with the notion that the introduction of a new pathogen into a new kind of processing and food category without a kill step — fresh-cut for both retail and food-service/wholesale — combined with certain commercial realities to create the majority of the outbreaks and recalls in the U.S. (3). Consequently, the greatest impact on food safety would come from better and specific regulation of the fresh-cut industry, or self-regulation for that matter, rather than regulating “farms.”
One type of regulation that, from the record, could have great impact is simply to have use-by dates on bagged fresh-cut produce for consumers, or containerized fresh-cut for food service, be determined with food safety criteria, as well as by “freshness” and, in fact, to be under some regulation of any kind.
The three outbreaks in 2011 mentioned at the beginning of this commentary show the limitations of this approach as the sole analysis — it is only true as far as it goes.
(1) 2011 Census of Agriculture
(2) Financial Characteristics of Vegetable and Melon Farms
(3) September 22, 2009 testimony of David Runsten, Director of Policy and Programs, Community Alliance with Family Farmers, Davis, California, at the USDA hearing on the proposed National Leafy Green Marketing Agreement, Monterey, California
Daniel B. Cohen is the owner of Maccabee Seed Co., an agricultural research-and-development and consulting company in Davis, CA.]]>
There is another STEC group that parallels the unusual pattern of the current outbreak in Germany. It as characterized by a form of Shiga toxin that becomes more deadly after it is attacked by the patient’s body’s defense mechanisms.
Most STEC like O147:H7 have a class of Shiga toxin known as Stx2. But there are a variety of subtypes of Stx2 such as Stx2a, Stx2b, Stx2c Stx 2d and others. There is a form of Shiga toxin Stx2d that becomes more virulent following attack by the mucosal enzyme elastase, whether from mice or humans.
Elastase attacks the toxin by cleaving off two amino acids from one of its components, at the C-terminal end of the A2 peptide (short protein) of Stx 2d. In lab tests with mouse models in 2002, as well as other tests, conducted by Dr. Alison O’Brien’s group at the Uniformed Services University of the Health Sciences (Bethesda, Maryland), these E. coli with toxins activated by exposure to mucosal preparations or the specific enzyme elastase, were more virulent than any other (STEC) Stx E.coli strains .
This type of Shiga toxin was then named “Stx 2d activatable” because instead of damaging the toxin by splitting it, the elastase, or crude mucosal preparation containing elastase, made it worse — “activating” it.
Professor Dr. Helge Karch and associated researchers at the National Consulting Laboratory on Hemolytic Uremic Syndrome, University of Muenster and the Interdisciplinary Center for Clinical Research, Muenster, wanted to study what the clinical implications of this kind of strain were for humans, rather than mice. They did a stratified analysis of 922 human STEC strains from patients with HUS, bloody or nonbloody diarrhea and asymptomatic carriers, looking at different kinds of STEC strains and disease course outcomes.
It takes more than Shiga toxin to have a virulent STEC pathogen. In particular, in “normal” STEC strains, tissue invasion is aided by an adhesive protein called intimin, encoded by the eae gene. In most STEC descriptions, having the eae gene for intimin is a crucial component for pathogenicity.
The Muenster researchers found that Stx2d activatable strains were also highly virulent in human cases, not just in mice or the in vitro model of Vero cells (the basis of calling Shiga toxins “verotoxins” is based on this kind of test). This was despite being eae (-), lacking the production of intimin. They also found the unusual pattern of adults and older patients with bloody diarrhea and HUS, as in the O104:H4 outbreak [2,3].
There were several alternative hypotheses to explain this. One was that the increased virulence of the activated toxin was a compensation for lack of intimin, or that toxin delivery into tissue was actually aided by the peptide editing caused by human mucosal elastase. A second possibility was that there was overexpression of the Stx gene to compensate for lack of intimin. A third possibility was that Stx2d activatable STEC strains had an unknown virulence factor that replaced the intimin functionality.
The first characterization by the Robert Koch Institute of the German O104:H4 outbreak strain included the notation that it was also eae (-), lacking intimin. Rather like the third hypothesis for Stx2d activatable strains, the outbreak strain had different adhesion genes characteristic of the entero-aggregative types of pathogenic E. coli, not seen in a “normal” STEC outbreak, whether in O157:H7 or pathogenic non-O157s.
Here the close parallel ends.
The O104:H4 outbreak strain, has no need for intimin, one might say, since it has other adhesion and invasion factors from its enteroaggregative genetic background. O104:H4 appears to be an enteroaggregative type of E. coli with Shiga toxin functionality from normal STEC. It is in a sense, a Shiga toxin containing enteroaggregative strain type. It could be
described as a STEC “armed” enteroaggregative E. coli or as a hybrid of STEC and enteraggregative types.
Virulence and adult bloody diarrhea and adult HUS at a higher rate of conversion may be due entirely to these enteroaggregative derived genes, not due to increased virulence of the Shiga toxin subtype, or increased virulence due to peptide editing by elastase, or some other enzyme.
The DNA sequence of the toxin from O104:H4 is supposed to be the same as Stx2, not Stx2d-activatable.
However, the 5-day time course from bloody diarrhea to HUS is somewhat suggestive of a pathogen-human-pathogen interaction like that of mucosal elastase and Stx2d activatable strains. Indicators of patient status slightly improve on the third day, and some patients think they are recovering just before plunging into a rapid development of HUS.
It is possible that this reflects a similar trigger, where toxin is functionally enabled for increased virulence and rapidly deployed in response to the same human defenses that give patients or physicians the illusion of pos- sible improvement.
STEC strains can change and evolve even during the infection course of a single patient (3), losing or gaining plasmids and phages, with their associated functionalities for Shiga toxin or intimin or drug resistance, for example. The parallel unusual clinical patterns between the O104:H4 outbreak strain and STEC Stx2d activatable pathogenic strains — severity of adult clinical symptoms — may be an example of convergent evolution, getting to the same results by completely different mechanisms. The mechanism could also be analogous or very similar, if not identical.
Regardless of mechanism, two parallel types of cases, with similar clinical pattern and outcomes, can still give clinicians and researchers sugges- tions for approaches, context and comparisons for treating the disease course.
Daniel B. Cohen
Maccabee Seed Company
 Martina Bielaszewska, Alexander W. Friedrich, Thomas Aldick, Robin Schürk-Bulgrin, and Helge Karch Shiga Toxin Activatable by Intestinal Mucus in Escherichia coli Isolated from Hu- mans: Predictor for a Severe Clinical Outcome Clin Infect Dis. (2006) 43(9): 1160-1167 doi:10.1086/508195
 “A subset of eae-negative STEC strains (i.e., STEC that produce Stx2dactivatable) differ from other eae-negative STEC strains by their strong association with severe disease… By its association with severe outcome of the infection caused by eae-negative STEC, Stx2dactivatable parallels the pathogenetic significance of Stx2 among eae-positive STEC … However, in contrast to Stx2, which is associated with HUS development in children <5 years old…, most of the patients with HUS from whom we isolated Stx2dactivatable-producing STEC were adults.” From Bielaszewska et al., next citation.
 Activation of Shiga toxin type 2d (Stx2d) by elastase involves cleavage of the C-terminal two amino acids of the A2 peptide in the context of the appropriate B pentamer. Melton-Celsa AR, Kokai-Kun JF, O’Brien AD Mol Microbiol. 2002 Jan;43(1):207-15.
 Alexander W. Friedrich, Wenlan Zhang, Martina Bielaszewska, Alexander Mellmann, Robin Köck, Angelika Fruth, Helmut Tschäpe, and Helge Karch Prevalence, Virulence Profiles, and Clinical Significance of Shiga Toxin-Negative Variants of Enterohemorrhagic Escherichia coli O157 Infection in Humans. Clin Infect Dis. (2007) 45(1): 39-45 doi:10.1086/518573]]>
German medical and public health authorities are burdened by the equivalent of more than 10 normal years worth of case loads occurring in a single month, the continuing need for secondary infection prevention, the outbreak source investigation and perhaps other issues as well. Despite this, a major study conducted as the outbreak was peaking was published in the New England Journal of Medicine . This included a prospective study on a subset of patients.
An article  and editorial  published in Eurosurveillance also presented some key information. The specific content of these reports and their implications for medical and public health response to possible new clusters or outbreaks remains largely unreported, two weeks later. Here are some of the vignettes and possible lessons learned based on these three reports, news articles during the development of the outbreak — largely from Germany — reports from the French outbreak, and other background sources.
One of the important findings of the prospective study in the NEJM article  was that daily laboratory testing for platelet counts, creatinine and lactase dehydrogenase (LDH) were crucial for monitoring the approach of HUS in patients.
It took a median of 5 days, from known (bloody) diarrhea to HUS onset, when this occurred. For the first three days levels were fairly steady, and in fact platelet counts and LDH, if anything, improved. This corresponded with some patients re-porting that they felt they had begun to recover from bloody diarrhea even as HUS was about to rapidly develop.
On day four creatine and lactate dehydrogenase rose modestly, while platelet levels began a somewhat steeper corresponding decline. During day five creatinine levels had tripled over the baseline; lactase dehydrogenase had tripled or quintupled and platelet levels had fallen to about one fifth of baseline (50-60,000 platelets per cubic mm from the chart).
These were the accurate predictors for HUS; indicating kidney damage response (creatinine), tissue damage response (lactase dehydrogenase) and the consequence of blood cell damage (platelet decrease) or hemolysis.
The Eurosurveillance article  presented one of the first highly probable human-to-human transmissions in this outbreak, in this case mother to infant.
Returning from northern Germany to the Netherlands, the mother became ill 8 days later, and the child 15 days later. The mother developed severe HUS 4 days after admission to the hospital for bloody diarrhea. Her child developed “blood abnormalities compatible with HUS,” fever — which appears unusual for these cases — and severe neurological symptoms. Three weeks after admission, as of the reporting date, the child remained on dialysis and had unknown neurological consequences.
This fits with a reported pattern of approximately 8 days for incubation of O104:H4 , about twice as long as for O157:H7. The timing of HUS onset also fits with the NEJM description of timing of HUS approximately 5 days after initial admitting symptoms — here bloody diarrhea and abdominal pain for two days for the mother.
Crucially: “The child had spent much time with her mother during her stay at the first hospital, when the mother developed diarrhea. Before admission of the child to (the hospital) the mother had not been advised to take any specific hygienic measures.”
One of the most important findings reported in the Eurosurveillance editorial  was that roughly one half of patients with HUS developed severe neurological symptoms. These included: disorientation, dissociation, stupor, and life-threatening seizures. Neurological symptoms, present in other Shiga toxin-producing E. coli (STEC) outbreaks, are a signature severe partner to HUS and gastroenteritis in outbreaks with this strain of O104:H4 (See also ).
Clinical presentations from this outbreak are quite surprising compared to past experience both with O157:H7 and non-O157 STEC E. coli cases over the last 30 years.
Some of this can be explained, perhaps, by the presumed food carrier being sprouts. Most cases are adults, who may be presumed to be more likely to eat sprouts than those 17 years or younger. More cases are adult women, in particular, than would be expected at random. This also could be explained by dietary choice. The vast majority of deaths are in older and elderly patients. This could be explained by the observation that there still are no effective treatments for HUS in particular, and that the elderly may be less successfully treated by ameliorative or palliative care, or may in fact be more susceptible to harm from some of the treatments.
However, in comparing adults and children, the O104:H4 outbreak is not only different from expectations but backward from previous experience. In the past, HUS has been a disease outcome overwhelmingly of infants and children, but here it is overwhelmingly an adult outcome. Bloody diarrhea is more common in adults while vomiting is more common in children.
The disease course, once started by dietary or human-to-human transmission, is statistically very different for adults and children compared to other STEC.
Another, much smaller O104:H4 outbreak occurred in the Gironde region of southwest France centered on Begles . This outbreak has 15 cases versus more than 4,000 in the German outbreak. The outbreak strain is consistent with the May-June outbreak strain in Germany.
“Preliminary data indicate that this outbreak shares the same novel epidemiological, clinical and microbiological features identified in the E. coli O104:H4 outbreak in Germany , including a predominance of adult women among the cases, an unusually high proportion of HUS cases among identified possible outbreak cases, a longer median incubation period than expected for cases of Shiga toxin-producing E. coli infection, and a genetically related E. coli O104:H4 producing a CTX-M ESBL.”
In this case three kinds of sprouts were produced for a local event. The first food surveys came back with little in common, but further epidemiological questioning showed the initial patients had in fact consumed cold foods decorated with the sprouts at the same event.
Because there is an overlap of a seed supplier with the German sprout producer/farm, fenugreek seeds purchased from Egypt from two different years’ stocks are leading candidates for possible seed-borne transmission through sprouts.
Because of the mult
possibilities of routes of contamination, however, European health authorities are being much more reserved in their comments than earlier. The UK seed company involved in distribution questioned how seeds sold for sprouts to many locations could cause an outbreak in only one location if they were, in fact, the source.
The French case also illustrates a problem raised by U.S. epidemiologists: how to distinguish different reporting criteria when describing the outbreaks. Using the French data, it would look like a 50% transition rate from bloody diarrhea to HUS vs a 20% rate in Germany.
Medical treatment and follow-up need to be planned before patients see a doctor or are admitted to a hospital until after patients are no longer contagious, if that can be determined.
The mother to child transmission case from the Netherlands shows the terrible individual consequences of the failure to have a system in place for comprehensive care, including family care. In a mass patient event, pre-planning may be critical to limit secondary transmission.
Patients reported in the German press going to doctors or hospitals with preliminary symptoms including diarrhea and stomach cramps, but basically being discharged to return home without treatment or follow-up and told to come back if they developed bloody diarrhea. When this did develop and they returned, they found that while there was segregation of potential STEC patients, the conditions were horrendous, including one bathroom for more 100 patients. Sanitation efforts were quickly overwhelmed. One patient commented that if she wasn’t sick before, she would be now.
German press reports also showed there was great concern and confusion among families of patients about how to treat their family members who were returning home from the hospital. There was apparently a lack of follow up or information provided.
There may need to be special training for both conventional medical staff, social workers or other paraprofessionals to provide accurate care before admission, at admission, and to provide follow-up care and information after returning home in a mass event.
German press reports showed pro-active actions by public health authorities to prevent child to child transmission in school settings. This included segregating or keeping home children who were ill, shutting down and sanitizing classrooms, and in one case where both food worker employees and children became ill, closing an entire school. It is not clear where children who may have been ex- posed to carriers actually went.
The presumption from O157:H7 experience is that children would be slower to clear STEC from their system than adults, taking a median time of something like 30 days. There also is a reasonable presumption that children may have more difficulty understanding and carrying out basic sanitation, either because of im- maturity or lack of adequate facilities.
O104:H4 may act differently in this regards as well, however. It infects children and adults differently than in O157:H7. If it acts more like some enteroaggrega- tive strains, then adult food handlers may be more of a concern (5), although the complexity of this group makes any generalization difficult (6).
The reports of probable person to person transmission from visitors to Germany to other countries include at least one case in the United States, and possibly two, where the second case was fatal. The impression is that most of these cases, relatively few so far, were adult to adult. Adult transmission should be considered in food handling, indirect hand to mouth transmission through different media such as some cosmetics, or intimate contact, including sex. But there has been little discussion of any of these potential routes.
There have been no reports of multiple transmissions from passengers returning from Germany by air, despite closed, shared facilities.
Most of the lessons on acute care of O104:H4 won’t be known until the German public health and medical personnel have the time to collate, evaluate and un-derstand their experiences.
There were reports that some physicians and hospitals were trying aggressive early treatment with antibiotics. Normally this is not recommended because of the potential for releasing even more Shiga toxin, but the press reports were that they found normal procedures were not working. These would presumably be antibiotics that O104:H4 does not have resistance markers to. The employment, timing, follow-up, efficacy and consequences for further resistance selection to new classes of antibiotics all have to be a concern.
In previous STEC outbreaks patients with central nervous system (CNS) dys- function as well as HUS had worse outcomes and mortality. In a 2010 review F. Obata suggested the evidence points to a weakening of both the blood-brain barrier and the blood-cerebrospinal barrier, possibly with the direct involvement of Shiga toxin Stx-2 subtypes involved .
One of the features of clinical treatment in the O104:H4 outbreak has been greater deployment of monoclonal antibodies (MAb’s) targeted at different com- ponents or actions of Stx.
One private company made their MAb drug freely available. This was targeted at disrupting one of two mechanisms of hemolysis, the disruptive triggering of apo- posis, normally a genetically controlled coordinated induction of cell death (9). A second method of triggering hemolysis involving Stx is the disruption of ribo- somal structures in cells preventing protein synthesis and causing cell death .
Alison D. O’Brien (10) and her lab at the Uniformed Services University of the Health Sciences in Bethesda MD has conducted some of the key research on monoclonal antibodies to target shiga toxin function and STEC pathogenicity for many years. Some of the commercial products came from earlier research at this lab. A number of animal models have been used by different labs with results that seem consistent with clinical experience (see ).
The overall efficacy of MAb’s and passive antibody therapy in the outbreak will be worth learning. There was one point made in the Eurosurveillance editorial :
“Worryingly, especially patients with seizures seemed to respond only weakly to standard antibody-based treatment regimes.”
There could be a role for vaccination against STEC E. coli, particularly for limited use to prevent li
kely secondary infections alongside other methods .
For O157:H7 there are two commercial vaccines developed for cattle. These were tested in herds in both Canada and the United States to see if they would cause a reduction in O157:H7 numbers and shedding, helping protect beef against being a carrier. Apparently they are not being commercially used.
This is a similar situation to West Nile Virus vaccines, where at least three commercial vaccines are available and in use for horses, but none for humans. Vaccine development to prevent STEC infections in people may require a different business or nonprofit model.
Reports on household members of primary cases who remained healthy have shown different levels of antibodies to Stx and to O157:H7 in both IgG and IgM classes. If an IgM was effective at all, oral vaccination might be effective and a more acceptable means of prevention of spread including mass inoculation.
I am quite aware of my limitations in preparing this kind of review, but I also note that communication between medical subspecialists and the general public are difficult. Sometimes the general public includes other medical professionals in differing specialties who may be impacted by a new outbreak of O104:H4 yet not have followed the literature from Europe on this issue.
To reflect my lack of expertise, I borrowed the presentation format of the late columnist Herb Caen (San Francisco Chronicle for most of his long career). This was known as “three dot journalism” where he mixed gossip, reporting, social commentary, culture, politics, tragedy and profundity into a city-wide swirl. I suppose one could call this “three dot medical journalism.”
Food Safety Website has space for comments, and in this case I would especially invite medical and epidemiological comments to correct errors of content or interpretation or new information.
Daniel B. Cohen
Maccabee Seed Company
 Epidemic Profile of Shiga-Toxin-Producing Escherichia coli O104:H4 Outbreak in Germany — Preliminary Report
Christina Frank, Ph.D., Dirk Werber, D.V.M., Jakob P. Cramer, M.D., Mona Askar, M.D., Mirko Faber, M.D., Matthias an der Heiden, Ph.D., Helen Bernard, M.D., Angelika Fruth, Ph.D., Rita Prager, Ph.D., Anke Spode, M.D., Maria Wadl, D.V.M., Alexander Zoufaly, M.D., Sabine Jordan, M.D., Klaus Stark, M.D., Ph.D., and Gérard Krause, M.D., Ph.D. for the HUS Investigation Team
June 22, 2011 (10.1056/NEJMoa1106483)
 Kuijper EJ, Soonawala D, Vermont C, van Dissel JT. Household transmission of haemolytic uraemic syndrome associated with Escherichia coli O104:H4 in the
Netherlands, May 2011. Euro Surveill. 2011;16(25):pii=19897. Available online: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19897
(3) Jansen A, Kielstein JT. The new face of enterohaemorrhagic Escherichia coli infections. Euro Surveill. 2011;16(25):pii=19898. Available online:
 Gault G, Weill FX, Mariani-Kurkdjian , Jourdan-da Silva N, King L, Aldabe B, Charron M, Ong N, Castor C, Macé M, Bingen E, Noël H, Vaillant V, Bone A, Vendrely B, Delmas Y, Combe C, Bercion R, d’Andigné E, Desjardin M, de Valk H, Rolland P. Outbreak of haemolytic uraemic syndrome and bloody diarrhoea due to Escherichia coli O104:H4, south-west France, June 2011. Euro Surveill. 2011;16(26):pii=19905. Available online: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19905
 Oundo, J. O., Kariuki, S. M., Boga, H. I., Muli, F. W. and Iijima, Y. (2008), High Incidence of Enteroaggregative Escherichia coli Among Food Handlers in Three Areas of Kenya: A Possible Transmission Route of Travelers’ Diarrhea. Journal of Travel Medicine, 15: 31-38. doi: 10.1111/j.1708-8305.2007.00174.x
 Q&A: Dr. Robert Tauxe on the Outbreak in Germany, by Daniel B. Cohen, June 22, 2011, Food Safety Website.
 “Doctors Shaken By Outbreak’s Neurological Devastation; At first, doctors were most concerned about the kidneys of patients with enterohaemorrhagic E. coli, or EHEC. In the past week, howver, it has become apparent that neurologi- cal side effects of the bacterial infection could be even worse”. By SPIEGEL Staff. Spiegel Online 06/09/2011.
 Adv Appl Microbiol. 2010;71:1-19. Epub 2010 Feb 20. Influence of Escherichia coli shiga toxin on the mammalian central nervous sys-
tem. Obata F
(9) Induction of apoptosis by Shiga toxins. Vernon L Tesh. Future Microbiol. Author manuscript; available in PMC 2011 January 1.Published in final edited form as:Future Microbiol. 2010 March; 5: 431-453. doi: 10.2217/fmb.10.4.
 Mouse Models of Escherichia coli O157:H7 Infection and Shiga Toxin Injec- tion. Krystle L. Mohawk and Alison D. O’Brien. J Biomed Biotechnol. 2011; 2011: 258185. Published online 2011 January 3. doi: 10.1155/2011/258185.
 Stx2- but not Stx1-specific human monoclonal antibody protects piglets chal- lenged with enterohemorrhagic Escherichia coli producing Stx1 and Stx2. Kwang-il Jeong, Saul Tzipori, and Abhineet S. Sheoran. Published in final edited form as: J Infect Dis. 2010 April 1; 201(7): 1081-1083. doi:10.1086/651198
 Strategies to Reduce Person-to-Person Transmission during Widespread Escherichia coli O157:H7 Outbreak. Edmund Y.W. Seto,* Jeffrey A. Soller,† and John M. Colford Jr. Emerging Infectious Diseases, Vol 13, No. 6, June 2007.]]>