Much of the disagreement about products on the lay-Celiac List, and generally within the Celiac community, is ultimately due to confusion about the maximum tolerable daily level of gluten.
Can anyone help? Even though this issue is not fully settled (and probably varies amongst individuals), discussing even vague boundaries would still be very useful to Celiacs who are confronted with daily decisions on which foods to avoid.
This resulted in a lengthy discussion on the concept of how much gluten a typical Celiac can tolerate in a day without damage. The listowners have edited and compiled the discussions. They appear below.
The Listowner's summary of (and commentary on) the discussion follow:
What follows are excerpts from the discussion, with names removed:
Eleven patients were involved. The study seems well done to me, included counting of intraepithelial lymphocytes, and the only possible defect I can see is that the patients were in the age range of 12-15 years when examined. It is possible that increases in hormone levels during puberty might have made them more resistant to damage at that time, relative to a younger or older group of patients.
I am sometimes involved in experiments related to wheat breeding and genetics in which rigorous statistical planning, protocols, and analysis are used. Elaborate statistical analysis is justified and necessary in such work. I also work in areas of biochemistry where elaborate statistical analysis is not usually warranted or necessary, such as molecular structure analysis. When we carry out the amino acid sequencing of a protein, we do not apply statistical analysis to the results--except in the most rudimentary sense--nor do we consult a statistician before beginning the experimental work.
I think there can be little hope of using the approaches [#2] is insisting upon to advance our knowledge in the area of celiac disease. The situation is enormously complex, involving so many factors (many of which may not even be recognized!) that definitive experiments are for the most part impossible to do. With regard to numbers of patients, obtaining 11 patients to participate in a study borders on the miraculous as there are not that many definitively diagnosed celiac patients available and those are rarely willing to involve themselves in experimentation. There is no scientific definition of what a "highly sensitive celiac" is. I could go on and on. Should we close up shop and quit trying even though we cannot meet the standards that statisticians consider necessary? Can no valid conclusions be drawn from experiments that do not involve hundreds of patients from a random pool?
Our knowledge about celiac disease increases in small increments and incomplete knowledge is a fact of life. This means that people must make arbitrary judgements based on incomplete information. As I mentioned, I know of no such experiments on rice or corn that even approach the sophistication of the Ejderhamn et al. work (even though there are amino acid sequences in corn prolamins that are not all that different from those in wheat). Are all celiac patients going to give up rice and corn because there are no carefully designed, statistically controlled experiments on these grains? I doubt it. I would say there is a reasonable body of evidence, developed without benefit of rigorous statistical design or analysis, indicating that rice and corn are safe.
I am merely trying to point out some of the complexities involved in doing research in the area of celiac disease. Despite these complexities, I think that reasonable conclusions can be drawn from experience and the analysis of many different experiments,even though the results of any single experiment might be questioned on rigorous statistical grounds.
It is sometimes quite difficult to balance the safety of a strict gluten free diet recommendation with the practical issues facing the patient as well as the emotional, social and psychological impact that the diagnosis has on patients.
I individualize the level of detail that I go into with patients according to their level of understanding and what follow up they will get.
If Patients are compliant with careful follow up which might include periodic biopsies (maybe once every 5 years) which are normal and the patient remains asymptomatic then I consider that they are as gluten restricted as needs be. In patients who may not have as close a follow up, then I think its a lot harder to predict how senstive they are going to be. Unfortunately many patients will play down their symptoms, rather than question their diets, in others there may a great discrepancy between the amount of small bowel damage and the severity of the symptoms. For these patients I emphasize the as little as reasonably acheivable for the patients.
If patients are going to have little or no follow up, then I err on the conservative side and suggest that a strict diet is the best advice. Of course it is likely that the sensitivity to ingested gluten may change with time, age, amount of damage, state of the host immunity etc. so it not a static environment.
[#4] and I have disagreed before over the appropriate standards for making judgments about what is safe to eat. One source of our disagreement may be our respective scientific training and disciplinary perspectives. [#4] seems to be looking for fundamental, universal laws, as bench scientists often do. As a psychologist, I am more concerned with studying the range of individual and developmental differences in reactions to experiences. We may not have a generally accepted definition of what a "highly sensitive celiac" is, but do any of you doubt that there is a substantial range of sensitivity in the population?
Like [#4], I believe that statistical procedures need not always be part of scientific investigation. However, the concern I expressed about Ejdermann et al. and similar studies was more fundamental; it was about the basic principles of research design that underly appropriate use of both statistics and nonstatistical methods. To what extent do the methods of an experiment permit the generalized conclusions one wishes to draw?
[#4] suggests that "definitively diagnosed celiac patients are rarely willing to involve themselves in experimentation." He suggests giving up on the effort to study representative samples. To me, this statement highlights why we should be concerned about bias in the results that we have. [#4] also acknowledges that "the situation is enormously complex, involving many factors, (many of which may not be recognized)." This is exactly the kind of research situation in which careful sampling, extensive measurement, and thorough analysis of findings is most needed. Good sampling in a study of response to a toxin cannot always be random, but it should at least represent those populations hypothesized to be at greatest risk, if one plans to generalize to that group.
[#4] suggests that the arguments in my previous post could be extended to the admonition that celiacs also should avoid rice and corn, which have not been proven "safe" by the methods I proposed. I disagree. There is theoretical and empirical/clinical history to warrant concern about wheat, rye, and barley. (Oats may remain a gray area in terms of their biochemistry.) A weak finding of no effect from an agent theoretically presumed to have toxic effects is different from a similar finding obtained when one had no particular reason to hypothesize an effect.
Although I have expressed my arguments in abstract scientific terms, my ultimate concerns, like [#4]'s, are practical. During a recent stay in Australia, I had trouble finding products that were as gluten free as those I eat in the U.S., and my clearcut adverse reactions included temporary loss of my hard-won lactose tolerance. When I talked with owners of health food stores or contacted manufacturers about suspect products, the message I got was along the lines of "it says gluten free and meets our standard, what kind of ridiculously unreasonable person are you to expect MORE?" I felt rather like little David Copperfield asking for more (GF?) porridge.
In the absence of definitive information about thresholds of safety (which [#4] suggests may be impossible to obtain), strict regulations, manufacturing standards, and dietary guidelines create a climate in which the largest possible number of celiacs can eat safely. As [#5] and [#6] have suggested, individual patients and their physicians may choose a less conservative standard, which may be appropriate in many cases. However, it is much easier for an individual celiac to decide to eat a little wheat starch than it is for him or her to have to find gluten free products that meet more than a lax regulatory standard or convince hospital and institutional dietitians that a diet that contains some gluten, but has become officially accepted as okay, is NOT something they can tolerate.
I am not sure if non invasive monitoring alone will suffice to detect damage from oats. The antibodies may not be incited by exposure to oats in the absence of wheat. Are the connective tissue antibodies produced specifically following the exposure to wheat proteins only and if they are then they may not detect a response to gliadins. I am not aware if any of the studies on oats in past looked at antobodies. The d-xylose test also may not be very sensitive for lesser degrees of damage. ?significance
Certainly careful follow up of patients seems justified. The "how to" will need to be tailored to the clinical situation. a counsel of perfection may be to rebiopsy every few years. I recognize that this is easier to do in adults than in children.
I thank [#2] for [the] thoughtful and well put comments and agree that part of the problem is that we are speaking to different concerns. I admit that the following discussion of a few points has been shifted somewhat to my own concerns. One point concerns the worth of the 1988 study of Ejderhamn et al. (I don't know the authors and I approach the study at face value as published).
The question Bill Elkus asked was: is there any minimal amount of gluten ingestion that is not harmful to celiac patients? The study of Catassi et. al. ["Dose dependent effects of protracted ingestion of small amounts of gliadin in coeliac disease children; a clinical and jejunal morphometric study". Gut 34: 1515-1519, 1993] presented evidence that 100 mg per day of gliadin is potentially harmful mainly on the basis of patients showing infiltration of intraepithelial lymphocytes. (Perhaps [Dr #5] would question the validity of these results because low level effects are difficult to evaluate. I haven't any experience with lymphocyte counting.) However, if we accept the Catassi et al. results as valid, we can proceed to consider studies involving lesser amounts of gluten protein intake. I mentioned the study of Ejderhamn et al. because it is the only one I know that deals directly with the question at hand. (Although, in the past, I have brought up the possibility that the studies of Holmes and coworkers on the incidence of cancer in celiac patients provide indirect evidence for a lack of harm from long term ingestion of wheat starch. This is on the basis that the people in the control group were presumably including wheat starch products in their diets, yet showed no significant increase in cancer incidence.
If the study of Ejderhamn is all we have, we should give the authors due credit for an inportant first step in filling a major gap, try to extract whatever useful information we can from it and then proceed to the next experiments. In a workup that included the most sensitive test currently known, infiltration of intraepithelial lymphocytes, no evidence of harm was noted for all 11 patients in the study, all of whom had been eating wheat starch products for some time (years). Of course, it would have been good to have 100 patients in the study. Of course, it would have been good to have adults as well as children (I pointed out the potential problem of dealing with adolescents who sometimes seem to have a holiday from celiac disease, presumably because of increased hormone levels). I think, however, that useful information can be obtained from study of even a single well chosen patient--although journal referees would unfortunately probably never agree to its publication. And obviously, limitations must be considered.
The near impossibility of obtaining large samples of celiac patients to participate in studies that involve biopsies must be kept in mind. Almost everything we know (think we know?) about the harmful grains and proteins in relation to celiac disease has been obtained from experiments that could be criticized for the number of patients involved and the sampling of the overall patient population represented by those patients (and frequently for other weak points)
Now, for the individual celiac patient, is the information from the study of Ejderhamn et al. of any value? It is far from the final word, but I don't think it is valid simply to dismiss the results. Many celiac patients in the US are avoiding anything made with grain alcohol, shampoos with wheat germ extracts, white vinegar (which is mostly made from corn-based alcohol), and so on. Some are concerned to the point of wondering whether they should induce vomiting or use a purgative upon accidentally ingesting gluten or ingesting what they think is "hidden gluten." Is this justified? I doubt it--while allowing that for a few people, probably a small sampling of all celiac patients, such concerns may be worthwhile because of individual complications that are considerably less than general in the patient population as a whole. I think the question coming up here is something like this: if you have no obvious response to gluten, are you doing yourself harm to, say, use vanilla flavoring that comes in a (wheat) grain alcohol base. I think the chances of harm from this are not worth considering, yet apparently large numbers of celiac patients in the US wouldn't think of using such an alcohol-containing flavoring.
Certainly, I would like to see a major study in which a large number of carefully chosen celiac patients participate and in which a well-characterized purified gliadin preparation is fed at a level of, say, 10 mg per day, with the results followed up at least as carefully as those in the study of Ejderhamn et al. Unfortunately, I don't expect to see any such study in the near future. In the meantime, if some celiac patients choose to consider the results of Ejderhamn et al. as sufficiently encouraging to allow them to partake of an occasional indulgence in MacDonalds' french fries, despite the hash browned potatoes (with some wheat in them) having been fried in the same oil, without fearing that they will be struck down by fatal cancer at an early age, I would say that decision has as much validity, perhaps more validity, than the decision of the person who chooses to avoid them.
From my standpoint, as a biochemist trying in small ways to further our understanding of the proteins and peptides responsible for the manifestations of celiac disease, I do not feel comfortable with the pronouncements of patients or physicians that are not backed up by publication of results in scientific journals. It has been stated that there is a wide range of sensitivity among celiac patients. There probably is. But what published direct evidence do we have for this range of sensitivity? Nothing much that I know of--although Van de Kamer, Weijers, and Dicke (1953) described a patient with a hypersensitive reaction that occurred within 3-6 hours of eating wheat, used this patient in screening of grain fractions, thereby I think invalidating most of their results. With regard to sensitivity, I have no doubt that there are strong psychological effects involved in some individuals. Furthermore, I speculate that there may well be significant differences in mechanism of response among patients--with highly sensitive patients perhaps exhibiting a Type I immediate hypersenstivity response, involving IgE, and others with a Type IV delayed hypersensitivity response mediated by T cells. People with an IgE mediated response might even in some cases be reacting to non-glut en proteins found in wheat rather than to gluten proteins or peptides. Those with immediate hypersensitivity are difficult to convince that all celiac patients don't respond to small amounts of wheat gluten the way they do. There is a tendency to think that even those patients who don't have an immediate response are doing themselves irreparable harm. There is no significant evidence to indicate they are. There is a tendency to think all responses are just part of the same spectrum, that it is all just celiac disease. If you consider the complete human immune system as comprising the spectrum, I guess that is true. As scientists, however, if more than one mechanism is involved in producing a reaction to wheat ingestion, should we lump all types of response together as celiac disease? In the interest of scientific progress, I don't think that would be helpful. We need to classify and to reduce problems to units that can be dealt with experimentally.
Obviously, we don't have enough information. I think we need to scrutinize carefully what we think we know from a solid scientific standpoint, separate out what is mere opinion, and move forward as best we can with new experiments to provide solid information where we need it. Opinions can be helpful in providing the basis for new experiments, but must be considered suspect. As [#5] said, this is where art meets science. Deconstructivism interfacing with logical positivism.
My apologies to physicians who have to deal in the real world with helping and advising their patients on a daily basis. The abstract approach doesn't provide much of value to them in the short run. Nevertheless, I hope they are supportive of scientific efforts that in the long run would provide a better basis for advising their patients.
Finally, I emphasize that I am not a physician and am not recommending any particular course of action to patients or physicians. This exposition is for the purpose of discussion among professionals and no other.
[#4] has put in sharp focus the difference between opinion and what is now termed "evidence based" decisions. With the current level of information we are limited to making 1.broad generalizations about Max tolerable for the celiac patients at large or 2. appraoching it one patient at a time and allowing the pateints to establish what the threshold for damage is. The former is based an very shaky( scientifically speaking) data, the latter depends on the patient and the doctor both being compliant with careful follow up. I agree completely with [Dr #4] that studies based on one or a few seemingly very sensitive celiacs may not generalisable to the celiac population at large.
How then can we generate data that would be helpful to the patients and the doctor advising them? Should we look at the effects of low doses of a defined peptide given in double blind fashion over a prolonged period of time? What population should be studied. ? Do as the Finns and study uncomplicated celiac disease? What are the tests to be applied: biopsies, serology, tests of absorption? Is there a risk that challenging someone with a dose of gluten that proves high enough to incite damage will alter the natural history of the illness?
Another approach which may not be as tight scientifically but more clinically relevant would be the ongoing comparison of 2 different dietary strategies. randomise patients to receive either the rigid absolutely gluten free diet instruction and the other a more liberal gluten restriction diet (ignoring all of the things [Dr #4] mentioned and maybe even other trace contaminants. Similar measures could be used to assess damage and one could also include quality of life measures in addition. This type of study while not as scientifically solid as the the first would be much more clinically relevant to patients, dietitians and doctors.
I suppose a 3rd way would be to do a comparison of calculated dietary intake and compare this with the subjective and objective measures to assess outcome. The latter 2 types of studies may be easier to get patients to partake in.
I think [the] third option, reproduced below, is the most realistic way to assess response to low levels of gluten, both in terms of ethics and relative to the possibility of recruiting/maintaining a sample varying in range of sensitivity. I suppose a 3rd way would be to do a comparison of calculated dietary intake and compare this with the subjective and objective measures to assess outcome. The latter 2 types of studies may be easier to get patients to partake in.
Any such correlational design probably should include repeated measurements on all key variables beginning at time of diagnosis. It also should allow for the strong possibility of some rational self-selection of dietary strategies. As [#12] has suggested, some people probably minimize their symptoms or have gut damage without overt distress, but others probably adapt toward a level of compulsiveness that allows them to feel good and/or get a normal followup biopsy. These adaptations would minimize differences in outcomes between those who adopt strict and lax criteria in managing their own diets, but could be figured into a design if anticipated and measured.
My compliments to you for a very thoughtful response and excellent summary. I do think that those of us who counsel patients in this arena will continue to be frustrated with some of the perpetual myth that surrounds this difficult issue. Perhaps I'm fortunate to some degree in that invariably I've found my families with celiac disease educated and motivated. To this degree I think the families need to understand some of the difficulties we face in trying to make suggestions based upon fact. The problem is that I do feel that my personal experience suggests that there is not only differential responses between patients, but I might go so far as to suggest that there may be differential responses of the same patient depending upon "other factors".
Given that the process of antigen presentation from the intestine depends in large part on mucosal integrity and normal function of M cells and lymphoid aggregates, it wouldn't be surprising to imagine how a secondary viral process, e.g. rotavirus, could transiently affect antigen presentation in a manner which could result in a dynamic state of exposure of food antigen. I have seen on several occasions infants with classical milk protein intolerance who were successfully transitioned from elemental formula back to cow's milk after time, only to "reactivate" their milk intolerance with abnormal biopsies and colitis after an episode of viral gastroenteritis. Perhaps a similar mechanism could account for some of the variability of why patients may tolerate limited gluten for a while (even with normal biopsies) only to "reactivate" later. It's only speculation on my part, but certainly there is a parallel example with cow's milk intolerance.
In the meantime, I try to discuss these issues with the parents and insert some common sense into the equation. At times it's difficult due to a zealous attitude of some individuals. Thanks for your comments,
Patients with gluten challenge have reactive -gliadin or other prolamine sensitive- lymphocytes in the peripheral blood. There are several methods to measure the reactions. One of them I used extensively in Hungary in the mid-eighties is the leukocyte migration inhibition test. After incubation the peripheral lymphocytes of patients with the supposed antigen the sensitized lymphocytes porduce several lymphokines. One of them is the leukocyte migration inhibition factor, which inhibits the migration of polymorphonuclear leukocytes form an agarose microdroplet. By measuring the area of migration compared with control cells give an index which tells whether was any lymhokine production for the antigen. This can be a less invasive way to decide whether a patient has cellular immune reaction (by the way, which is responsible for the intestinal damage). to the given antigen.
I have tested kids who -usually under grandparents "supervision"- ingested pancakes, and a week later their leukocyte migration inhibition test became pathologic, although they had negative test before. Usually it became negative within 3-6 months on a gluten-free diet, and apparently a load of gluten enough to increase the number of sensitive lymphocytes in the peripheral blood.
By measuring the cellular immune reaction with this "old-fashioned' way or other methods helps us to answer questions, like is there a small amount of gluten which a patient can tolerate (measure the cellular immune reaction before and after a low dose of gluten challenge) or whether oat proteins are dangerous (challenge volunteers and measure the cellular immune reaction). I feel that studies using peripheral blood cells are easier and less invasive than intestinal biopsies. Furthermore, it is proved that the lymphokine production is present in both the intestine (after incubating biopsy specimens with gluten the supernatant inhibits leukocyte migration; see studies published by Ferguson and others). There are several other ways. I have published a paper in 1989, which showed that a single challenge of gluten increases the intestinal permeability within 5 hours, or there is an increase in the leukocyte chemotactic factor level in the serum. Also there is a degranulation of mast cells, which means that any measurable mast cells-derived factor can be used to test oat proteins or the lowest level of tolerable gluten. All this test requires only urine collection or blood drawing.
To perform these studies we need volunteers who are willing to participate. Evidently, this kind of studies are experimental but may help us to develop future clinical methods which help to take care of celiac patients much better.
The prevalence of gluten intolerance in [Finland] is 1:360. I think that celiac patients tolerate wheat starch or small, accidentally taken, amounts of gluten. An early diagnosis and treatment are much more important issues for celiac patients.
An adult CD patient in Finland may consume perhaps 200 g/day of gluten-free flours. Assuming that these flours contain max 200 ppm of gluten (= 20 mg of gluten/100 g dry weight) we come to figures that the daily intake of gluten is 40 mg (20 mg glidin).
Our children with celiac disease are followed routinely at the outpatient clinic. On our strict gluten-free diet (but allowing wheat starch) symptoms disappear, endomysium, reticulin human umbilical cord and gliadin antibodies disappear also and the mucosa recoveres. I have had cases where I do not see negative seroconversion for example in endomysium antibodies. I do not remember any case where 100% gluten-free diet was needed (no wheat or barley starch-based products) to correct this. All patients were ingesting gluten deliberately or by accident perhaps some grams per day. Buckwheat or millet could be wheat contaminated similar to oats.
When I perform a gluten challenge to an adolescent for example (see Arch Dis Child 1989;64:1604-07), I want to give at least 10 g of gluten per day (equals perhaps four slices of wheat bread) and the mucosa will relapse after one month to two - 7 years (my longest challenge). Stenhammar in Sweden had to challenge one child for 14 years before he could prove the child to be a celiac. If the challenge would have been performed with gluten-free flours containing wheat starch only, 50 kg per day should have been ingested (=10 g of gluten).
If amounts of gluten ingested as wheat starch is not tolerated, we of course can prescribe a 100% gluten-free diet to our patients. We need more studies where ill-effects of small amounts gliadin are looked for. The problem is what to use as a gold standard (clinical outcome for 80 years or normal jejunal mucosal morphology or small changes in the 'normal' mucosa as measured by computers or immunohistochemistry or organ culture inflammation by some peptides or some aminoacid long peptide immunological T cell reactions for example). We will find many nonceliacs to be reactive, they are perhaps later celiacs or never according to the present criteria? In some years biopsy will not be the gold standard for celiac disease?
Q. How can I determine the amount of gluten from the weight of the wheat in food?
A. For 100 units of whole grain wheat, about 70 units of white flour results from the milling process. The rest is separately sold as wheat bran or wheat germ. Those 70 units of flour are about 10%- 15% protein, thus about 7 to 10 units of protein for 100 units of whole wheat. The protein is about 80% gluten, thus about 6 to 8 units of gluten for 100 units of whole wheat. Since one typically sees wheat flour as an ingredient, appying the 70% factor implies 8 to 12 units of gluten per 100 units of wheat flour.
Q. Sometimes I see references to the amount of prolamin or gliadin instead of gluten.
A. To cereal scientists, gluten is the same as prolamin, but in some older terminology only the gliadin fraction is termed prolamin. Gliadin makes up about half of the gluten. The other half is often called glutenin, but it is very similar to the gliadin half in composition and structure and I suspect that it is toxic to a large extent. It would be simplest to say that gluten equals gliadin equals prolamin as far as toxicity is concerned.
Q. How is wheat starch made? Why is some starch said to be more toxic than others?
A. Most of the wheat grain and of white flour is made up of starch granules. Starch granules make up about 75% of grain or of white flour. In the processes used to make wheat starch, a small amount of the gluten protein (actually mostly the gliadin fraction, but not entirely), sticks to the surface of the starch granules. The amount depends on the washing method, how many times the granules are washed, and factors like that. Wheat starch can be made very low in surface protein and it is only the surface protein that is of concern (there are some internal granule proteins, but we are pretty sure that they are not gluten proteins).
Q. How much gluten is in wheat germ and wheat bran?
A. I don't know how much gluten they contain, but my feeling is that it is likely to be more than you would find in wheat starch and so best avoided by celiac patients.
The purpose of this copyright is to protect your right to make free copies of this paper for your friends and colleagues, to prevent publishers from using it for commercial advantage, and to prevent ill-meaning people from altering the meaning of the document by changing or removing a few paragraphs.