MAP in the Environment

MAP in Food

Marshfield Clinic Retail Milk Study


It is a scientifically and generally accepted fact that foods derived from animals infected with Mycobacterium avium subspecies paratuberculosis (MAP) are continually entering the human food chain.  Cornell University recently confirmed the presence of MAP in retail beef and, as far as milk, multiple studies of pasteurization have been conducted.  With regard to milk, to date, ten studies have found that MAP can survive pasteurization standards used in the U.S., and one study found that MAP did not survive pasteurization.  However, controversy still surrounds this issue and final conclusions have not been drawn.  One must ask "Why not?"  The inability to resolve the controversy and arrive at conclusions appears to be a direct result of the methodologies used, i.e. simulation testing -- which opens the door for scientists to debate seemingly endlessly the adequacy and appropriateness of the methodologies used -- versus testing of the retail milk.

PARA has maintained from the beginning and to this day that the only way to answer the above question is to test supplies of milk and other retail foods.  Since the summer of 1997, PARA has been urging the FDA to conduct retail milk testing in the U.S., and the FDA has consistently ignored our concerns, refusing to test retail supermarket milk and dairy products for the presence of MAP.   Instead, they have relied upon one single, flawed and highly criticized "simulation study" to assure the American public that the milk and dairy products we give our children are free from contamination with live MAP

This critical research has now been undertaken...not by the FDA...but by the Marshfield Clinic in Marshfield, Wisconsin.  

Marshfield Clinic Retail Milk Study

The Marshfield Clinic Retail Milk Study is testing retail supermarket milk, taken from the shelves of supermarkets, for the presence of MAP. This study is being conducted by Dr. Jay L. E. Ellingson, Director of Marshfield Laboratories Food Safety Services.  In its July/August 2002 newsletter published by the Marshfield Clinic, Dr. Ellingson states:  "Mycobacterium paratuberculosis is considered a suspect zoonotic agent because it causes Johne's disease in cattle, and some studies in the UK say it causes Crohn's disease in humans.  The question is whether people get Crohn's disease by ingesting food product derived from infected cattle." 

This is the first and only retail supermarket milk testing program for the presence of MAP ever undertaken in the United States.  Results of the study will be published in the summer of 2003.  

With over 20 years of agricultural safety and health research experience, Marshfield Clinic's National Farm Medicine Center specializes in research, education, outreach, consulting and testing programs that deal with critical health issues in agriculture, where food safety begins.  For more information about Marshfield Clinic, visit their website at:

Recently, FDA was contacted by PARA for a comment on the Marshfield Study and clarification of their position about the safety of retail milk supplies in the U.S., in light of recent UK findings to the contrary.  (See MAP in Dairy Products.)   Not surprisingly, no one at FDA's CFSAN would speak with us, instead referred us to the Risk Assessment Department, to their attorneys.   Sadly, it appears that FDA considers PARA, the organization, more of a risk than PARA, the bacterium.  FDA's position remains the same: 

They will take no action! 
They will not exert the precautionary principle


The Case for Retail Testing

In this section we will discuss the scientific rationale for retail testing.

This document was first presented to the Food Safety Committee of the United States Animal Health Association in October 1998.  It has been updated to reflect important developments since that time.




Ever since Chiodini and Hermon-Taylor published in 1993 that MAP was capable of surviving commercial milk pasteurization, there has been much controversy about their results. On the one hand, some said that their results were sufficiently valid to initiate retail testing of milk, in order to determine whether or not viable MAP were present in retail milk. On the other hand, some people, including those responsible for ensuring the microbiological safety of milk, discounted Chiodini and Hermon-Taylor's results stating that the results were derived from laboratory simulations of pasteurization, and could not be extrapolated to the real world.

More laboratory simulations of commercial pasteurization were conducted, in three countries, the USA, the United Kingdom and Australia. These simulations were of varying degrees of sophistication, with some using commercial pasteurization equipment, some using laboratory scale pasteurization equipment, and some using test tubes.

The decision to study the subject by using simulations of pasteurization was based on the perception that testing of actual retail milk was impractical, due to the problems of studying MAP in the fertile biological environment present in retail milk, e.g. overgrowth by faster growing organisms.

These perceptions of the difficulty of testing the retail milk supply for the presence of viable MAP are no longer valid. Techniques recently developed in the United Kingdom make it easy, straightforward and relatively cheap to test retail milk, and other dairy products.

This document will discuss the rationale for conducting laboratory simulations of pasteurization, and how that rationale is flawed. It will also discuss how testing of actual retail milk is the only method by which we can be certain whether or not viable MAP bacteria are present in retail supplies of milk and dairy products.


MAP: an enigmatic organism.

MAP is one of the most enigmatic bacteria known. It is known to cause a chronic and fatal wasting disease, known as Johne's disease, in a wide range of animals, including cattle, sheep, goats, horses, deer, elk, rabbits, chickens, bison, dogs, and primates. However, the disease causing mechanisms of the organism are not known. There are no treatments for Johne's disease, and no effective vaccines: animals that contract the disease simply waste away until death.

What is known about MAP, a subspecies of Mycobacterium avium, is the following:

  • The organism is extremely slow growing, requiring a minimum of 24 hours to multiply. To culture a visible colony of the organism requires months, and when testing for the organism in animal feces to determine the animal's infection status, samples are usually held for 16 weeks before being declared negative.

  • The organism requires special environmental conditions for successful culturing. For example, the organism requires the presence of mycobactin in the culture medium before it will grow, since it is incapable of producing its own mycobactin. Hence it is labelled "mycobactin-dependent".

  • The organism apparently has the ability to suspend its metabolism. When it does so, it may remain inactive for periods of months or even years. This allows the organism to persist in the environment, for up to 2 years in water and 9 months in slurry. Since MAP is an obligate pathogen, requiring the environment inside an animal's body to replicate, it is not able to multiply in the environment, only to survive.

  • The organism is capable of existing without a cell wall. When it does so, it is said to exist in "spheroplast" (Cell-Wall-Deficient) form. After periods of months or even years of inactivity, the spheroplast may then revert to the "bacillary" (Cell-Wall-Intact) form, and begin multiplying again.

  • Because it has a thick and waxy (lipid-rich) cell wall, the organism has the ability to withstand treament with a wide variety of chemical decontaminants.

  • Data on the thermal tolerance of paratuberculosis show that it is more heat tolerant than its "parent" organism Mycobacterium avium, which itself can withstand temperatures 6 to 7 degrees higher than Mycobacterium bovis.


MAP in milk

Research has shown that MAP is present in milk taken from cattle infected with the organism. This is true of both clinically diseased animals and subclinically diseased animals. "Subclinical" animals are infected with the organism, but do not yet show any clinical symptoms of disease. An infected animal may remain subclinical for many months, or even years.

MAP may be present in milk through a variety of different routes.

  • In free suspension. Because paratuberculosis is a systemic infection, with clinically diseased animals shedding up to 5 x 1012 (5 trillion) bacteria per day, the organism is pervasive in tissue and body fluids.

  • Inside white blood cells. Paratuberculosis colonizes and multiplies in white blood cells (macrophages), which are an integral part of milk, i.e. the dam passing immunity to the calf.

  • In clumps. There is also the potential for faecal contamination of milk, since clinical cows suffer from severe liquid diarrhea, which may contaminate milk during the actual milking process. Bacteria that contaminate milk in this way will most likely be clumped, i.e. an entire colony of bacteria will literally stick together.

  • In fat droplets. Research shows that MAP has an affinity for the fat layer in milk. Thus, after milk has been taken from an infected cow, paratuberculosis bacteria in free suspension will migrate to fat droplets in the milk, and conceal themselves within those droplets.


The need for testing different a variety of milk products

Data on the thermal tolerance of MAP indicates (figure 1) that it is on the borderline when it comes to the question of eradication with pasteurization. Therefore, it is almost certain that factors such as the physical, chemical and biological environment in which it is pasteurized will make a difference between survival and killing, e.g. its presence inside macrophages or fat droplets.

Graph of pasteurization times for M. avium subsp. paratuberculosis, C. burnetti and M. bovis

Figure 1. Estimated pasteurization lines for clinical strains of M. paratuberculosis, when the initial concentration was 106 organisms/ml, and for comparable numbers of M. bovis and C. burnetti. Source: Sung and Collins, Appl & Env Microb 64(3),999-1005.

For example, the loss of a large part of the fat layer in skimmed milk might make the difference between whether the organism survives or not. Likewise, if milk is homogenized before pasteurization, this may also remove the fat layer protection for the organism, by breaking up protective fat droplets. However, if homogenization is carried out after pasteurization, then the bacteria will still enjoy fat protection.

The acidic environment of cheese and whey making processes may be enough to hinder or even stop the growth of the organism. Most likely the duration of immersion of the organism in an acidic environment may dictate the survival of the organism, i.e. cheeses matured for a year or more may be less likely to contain live paratuberculosis. Conversely, cheese matured over a short period may be more likely to contain live paratuberculosis. This is a most important consideration, since an estimated one-third of the cheese produced in the United States is made with raw milk.


Different Methods of Pasteurization

Currently, there are four main accepted methods of pasteurization. Each dictates the minimum time/temperature conditions which guarantee varying degrees of microbiological safety. Also, each is designed so as to minimize the amount of chemical and physical changes induced in the milk, so as to ensure its palatability to the buying public.

Low Temperature Holder (LTH, or Batch) pasteurization.

Milk pasteurized under the LTH method is heated to 63oC for 30 minutes. Although this process is designed so that every particle of milk is exposed to the required time/temperature conditions, this process is also designed to minimize damage to the fat in milk, in order to guarantee a full "creamy" taste. Products most often pasteurized with LTH pasteurization include cream, ice-cream and milk for local distribution from small dairies, who seek a "taste advantage" over their larger competitors.

Unlike all of the other pasteurization methods, LTH pasteurization does not involve the use of "holding tubes". All studies of MAP and LTH pasteurization, both simulations and thermal tolerance studies, find that MAP does indeed survive LTH pasteurization. This includes the simulation research conducted by the USDA Agricultural Research Service, which found that when 107 MAP per millilitre were added to sterile milk and LTH pasteurized, 102 bacteria survived the pasteurization process.

High Temperature Short Time (HTST) pasteurization.

HTST pasteurization is the most widely used method in the United States, and is used for the great majority of fluid milk sold, as well as several other dairy products. HTST pasteurization equipment is designed to guarantee that every single particle of milk is exposed to the minimum pasteurization conditions of 71.7°C for 15 seconds, through the use of a "holding tube" and a flow-diversion device which extracts product that is under-heated and resubmits it to the pasteurization process.

HTST is the most studied pasteurization process in relation to MAP. Simulation studies have been conducted in the U.S.A, the United Kingdom and Australia. All but two of those simulations have indicated that MAP may indeed survive HTST pasteurization, and thus may be present in retail milk supplies.

The most noteworthy research of HTST pasteurized milk was recently conducted in the United Kingdom. In Northern Ireland, in a preliminary study, MAP was cultured from 6 of 31 (19%) HTST pasteurized retail milk samples taken from operating commercial dairies. This research will be discussed in more detail in a later section.

High Heat Short Time (HHST) pasteurization.

This pasteurization process has not been studied in relation to MAP.

Ultra High Temperature (UHT, or Ultra) pasteurization.

UHT pasteurization involves the heating of milk to a minimum of 130°C for 1 or more seconds. While it is generally assumed that MAP does not survive UHT pasteurization, it is important to note that there are no data to support this assumption, and this assumption probably arose from comparison of the thermal tolerance of MAP with the thermal tolerance of other microbes, a comparison which may or may not be valid.

This overview shows that MAP is a common contaminant of raw milk. Are dairy product manufacturing practices, such as pasteurization, 100% effective at killing this microbe? The question remains open.


Laboratory Simulations of Milk Pasteurization

As all engineers and economists, meteorologists and mathematicians, car designers and computer scientists know, there are a number of problems faced when conducting simulations.

The most important problem is that it is usually impossible to simulate any given system in its entirety. There are many assumptions that must be made. Decisions must be made on which elements of the complex system are to be modeled, and which elements are not. One thing that must be certain is that factors that are omitted from the model must have little or no effect on the overall system.

With the factors that are chosen to be included in the model , it is impossible to be precise. The factors that are chosen to be included must be approximated, meaning that all of the parameters chosen for inclusion in the model must have a margin for error built into them.

When modeling a complex multi-element process, these margins for error are multiplied together, resulting in an even wider margin for error. When simulating a very large complex system, the margin for error of the entire simulation may be wider than the values being measured in the simulation.

The Pasteurized Milk Ordinance (PMO), the standard document by which most States set minimum legal requirements for milk pasteurization, is an extremely complex document. The PMO was first written in 1928, and has been revised numerous times in the last 60 years, by countless expert personnel, to take account of new knowledge in the fields of biology, chemistry and physics. The PMO takes into account an extremely wide range of biological, chemical, physical and legal factors which affect the production of milk and the public health. The PMO is 405 pages long.

The PMO is still not fully specific on all details of the milk pasteurization process. It does not include the precise specification for equipment used in the pasteurization process. It merely defines a range of operational parameters within which pasteurization equipment must operate. For example, there are four separate pasteurization processes described in the PMO, LTH (Batch), HTST, HHST and Ultra-pasteurization (UHT), the definition of each of which describes a range of operational parameters that are acceptable for each standard.

The PMO describes a process that is so complex that simulation of the entire system would certainly be an extremely difficult and horrendously expensive project. Due to the immense complexity of the process, giving rise to wide margins for error in the simulation, it may even be impossible to simulate the entire process with any degree of accuracy.

Use of sterile milk

Most simulations of pasteurization make use of sterile or near sterile milk (UHT) milk. The process of making milk sterile alters its physical and chemical and biological composition, thus introducing further margins for error to the simulation. For example, the fat layer of such mlk would be disrupted, depriving paratuberculosis bacteria that have an affinity for the fat layer of the natural protection they would enjoy in real world milk.

Use of bacteria in free suspension.

The normal method of contaminating sterile milk during simulations is to add quantities of MAP to that milk, i.e. the bacteria are free floating in the milk. This deprives the bacteria of the protection that they would experience in the real world, where they exist inside white blood cells (macrophages) in the milk. Also, once the bacteria have been added to milk, they are immediately submitted to pasteurization, thus depriving them of the chance to surround themselves with fat droplets, as they do in real world milk.

Summary of laboratory simulations.

The only times when simulation should be used is when it is impossible to study the actual system in operation, or prohibitively expensive to do so. We have discussed above some of the factors that influence simulations of milk pasteurization. It is clear that thorough and accurate simulations of commercial milk pasteurization would be extremely expensive, and may be futile, due to the extreme complexity of the system and the unavoidable wide margins for error that result from this complexity.

The reasons why all research has relied on simulations to date were based on the fact that up to now, testing of retail milk has been considered to be excessively difficult. This was due to the problems of microbial overgrowth of the sample, which results from the presence of organisms which grow much faster than paratuberculosis, combined with the difficulty of culturing paratuberculosis itself.

However, those constraints on retail testing no longer exist. As will be discussed in the next section, due to recent developments of novel testing methodologies, retail testing is now straightforward, effective and relatively cheap.


Retail Testing in the United Kingdom

Dr. I. Grant of The Queens Universtity, Belfast, has been developing techniques for use in testing for paratuberculosis for a number of years. The most important technique from that research, a technique that was  in development for 4 years, is known as ImmunoMagnetic Separation (IMS).

IMS is a technique for extracting paratuberculosis cells from liquid suspensions. IMS involves the following steps.

  1. Small steel spheres coated in rabbit antibodies to MAP are added the liquid to be tested.

  2. The anti-paratuberculosis antibodies bind to any MAP cells that are present in the liquid.

  3. The steel balls are removed from the test liquid by the use of a magnet.

  4. The material attached to the steel spheres is decontaminated.

  5. The resulting extract is cultured.

The current cost for testing of each liquid sample using IMS is £200 (US$320 approx). It is likely that this cost will fall as a result of improvements in the IMS testing methodology which will inevitably result as the process is utilized and improved upon.

On August 10th 1998, the UK Ministry of Agriculture, Fisheries and Food announced the results of a preliminary study, which was designed to prove and improve IMS technology. 31 samples of raw milk were tested, and 31 samples of commercially HTST pasteurized milk were tested. The raw milk samples were taken from milk before it was submitted for pasteurization at a commercial plant. The pasteurized samples were taken from cartons of retail milk as they rolled off the production line at the same commercial plant. IMS techniques were applied to the 62 samples. MAP was cultured from 10 of the 31 (32%) of the raw milk samples. MAP was cultured from 6 of the 31(19%) of the HTST pasteurized samples.

Based on the results of this preliminary study, the United Kingdom government ordered a more extensive sampling of British retail milk. Over the next 12 months, samples will be taken from up to 500 dairies across England, Wales and Northern Ireland, and submitted for testing. Overall, 1,000 samples will be tested, at an overall cost of £200,000. Types of milk to be sampled include full fat milk, skimmed milk, semi-skimmed milk and UHT milk. The study is expected to take 18 months to complete. The primary reason for this length of time is because the numbers of paratuberculosis bacteria shed by cattle vary throughout the year. Thus the numbers of paratuberculosis bacteria present in milk will vary seasonally, so the testing period must encompass a full cycle of the seasons.

The number of samples being studied in the United Kingdom, i.e. 1,000 samples, is certainly statistically large enough to give results which truly reflect the extent of MAP contamination of retail milk in that country. It will be vitally important to know this information if MAP and its contamination of dairy products ever becomes the subject of trade regulations intended to restrict the spread of the organism. The subject of trade barriers and MAP will be discussed in a later section.

UPDATE:  In May of 2002, results of the above study were published.  It is entitled "Incidence of Mycobacterium paratuberculosis in Bulk Raw and Commercially Pasteurized Cows' Milk from Approved Dairy Processing Establishments in the United Kingdom"  (Grant, I.R.; Ball, H J.; Rowe, M.T.  Applied and Environmental Biology, May 2002, p. 2428-2435, Vol. 68. No. 5).14   This study confirmed that MAP survives pasteurization standards used in the UK, the very same pasteurization standards used in the U.S. 


Required Laboratory Expertise

It has become clear in the last few years that extensive experience in two fields is necessary for the successful study of MAP in retail dairy products. These are

  1. Culturing the organism. MAP is notoriously difficult to culture in the laboratory. It is extremely slow-growing, and has extremely precise nutritional and environmental requirements. This is particularly important when attempting to culture paratuberculosis after pasteurization, since the organism will certainly be "sub-lethally injured" by the pasteurization process, if it has not been killed. Harsh treatment of the organism of the organism after pasteurization, such as overuse of decontaminants or antibiotics may kill the organism, being the "straw that broke the camels back". Also, MAP is known to have the ability to suspend its biological processes, and thus may require even more precise conditions to restart growth after the sub-lethal injury inflicted by pasteurization.

  2. Food microbiology. When studying an organism in relation to milk pasteurization, it is not sufficient to merely spike sterile milk with the organism. There is a wide range of physical, chemical and biological factors which may influence the outcome of the testing. Thus, extensive knowledge of these physical, chemical and biological factors is necessary, i.e. the testers should be Food Microbiologists. Also, the work of Dr Grant at Queens suggests that the decontamination methods used in veterinary medicine to culture MAP from fecal samples may be too harsh for these types of studies. While these methods do not damage the healthy robust organisms which are found in bovine fecal samples, they are harsh enough to destroy a large percentage of the sub-lethally injured organisms which are found in post-pasteurized milk.

Given the acute shortage of laboratories with experience in studying Mycobacterium paratubercuosis post-pasteurization (Grant et al in the UK, Sung and Collins in Wisconsin, Condron et al in Australia), it is important that any proposed study of retail testing include the involvement of several laboratories, each working independently but also working in close contact with established experts in the field.


MAP As A Trade Barrier

In 1997, the government of the Netherlands announced that paratuberculosis would be eradicated from Dutch herds of cattle by the year 2007, an aggressive and potentially expensive aim. The impetus to take this action was based on 3 primary concerns.

  1. Economic losses from paratuberculosis were mounting, since a substantial fraction of dairy herds in the Netherlands are infected with paratuberculosis.

  2. Mounting concerns for public health, and a strong will to exercise the precautionary principle. The hard-earned lessons of BSE have not been forgotten in Europe.

  3. Trade. The Netherlands is one of the world's largest dairy exporters, and has trade links with countries across the world, including a very close trading relationship with fellow European member states. If MAP is classified as a human pathogen, the Netherlands will be able to certify that their exported dairy products are free of MAP, and thus gain economic advantage over their international competitors.

The Australian government has begun the process of eradicating MAP from cattle herds and sheep flocks, for precisely the same reasons as the Netherlands. The Australian sheep and cattle industries have partnered with Australian State Governments, in New South Wales and Victoria, to institute control and eradication programs which will eventually eliminate Johne's Disease from herds of food animals in those, which are vital to economic health of those regions.

When these countries have reached their goal of paratuberculosis free status, they will be most vigilant to ensure that they do not become re-infected by imports from paratuberculosis-infected countries. Such trade barriers will certainly include restrictions on the movements of paratuberculosis-infected animals and breeding products, and may also extend to ban paratuberculosis-infected dairy foods.

With expected refinements of testing processes such as ImmunoMagnetic Separation, it may be possible to carry out on-the-spot testing for MAP contamination in as little as five years or less. Importing countries may then wish to test dairy imports at the point of entry.

Given that 20% of small US dairy herds are infected with MAP and that 40% of large US dairy herds are infected, it appears that the United States has the largest paratuberculosis problem in the world, at least in terms of bovine paratuberculosis.

Source:   Contact PARA:
Paratuberculosis Awareness & Research Association, 1999-2003.