Skip to main content

Celiac disease – Genetic testing and clinical utility

Celiac disease is a digestive disease that damages the small intestine and interferes with absorption of nutrients from food. People who have celiac disease cannot tolerate gluten, a protein in wheat, rye, and barley. Gluten is found mainly in foods but may also be found in everyday products such as medicines, vitamins, and lip balms. When people with celiac disease eat foods or use products containing gluten, their immune system responds by damaging or destroying villi—the tiny, fingerlike protrusions lining the small intestine. Villi normally allow nutrients from food to be absorbed through the walls of the small intestine into the bloodstream. Without healthy villi, a person becomes malnourished, no matter how much food one eats. Celiac disease is both a disease of malabsorption—meaning nutrients are not absorbed properly—and an abnormal immune reaction to gluten. Celiac disease is also known as celiac sprue, nontropical sprue, and gluten-sensitive enteropathy. Celiac disease is genetic, meaning it runs in families. Sometimes the disease is triggered—or becomes active for the first time—after surgery, pregnancy, childbirth, viral infection, or severe emotional stress (NIDDK).

CD is caused by a reaction to gliadin, a prolamin (gluten protein) found in wheat, and similar proteins found in the crops of the tribe Triticeae (which includes other cultivars such as barley and rye). Upon exposure to gliadin, and certain other prolamins, the enzyme tissue transglutaminase modifies the protein, and the immune system cross-reacts with the small-bowel tissue, causing an inflammatory reaction. That leads to a truncating of the villi lining the small intestine (called villous atrophy). This interferes with the absorption of nutrients, because the intestinal villi are responsible for absorption. The only known effective treatment is a lifelong gluten-free diet.

From a genetic testing point of view CD is an interesting case, the majority (>97%) of people with CD have either HLA DQ2, DQ8 or both and the number of alleles also influences the risk of developing the disease. Having said that, roughly 30% of the population carry one or more of these alleles but the prevalence of CD is only around 1-2%, so the HLA alleles are necessary, but not sufficient, for the development of celiac disease. They are not particularly predictive either (although homozygous DQ2 and DQ2/DQ8 have a much higher risk, of the order of 1/7-1/10 – Megiorni et al, 2009) so is there any use in testing for these alleles in asymptomatic individuals?  The interesting thing is that a genetic test based on HLA alleles has an almost perfect negative predictive value, if the HLA alleles are not there then the chances of developing, or of any symptoms being due to, CD are minimal, move on to the next candidate for the diagnosis. There are also arguments in favour of testing based on some interesting facts about CD:

1. The majority of people with CD don’t even know that they have it, only about 10-20% of cases have been diagnosed, the rest remain undiagnosed for a variety of reasons, mainly because symptoms are not yet so severe that they lead to diagnostic testing. The diagnosis itself is not so easy either, the full diagnosis requires a biopsy of the intestine and histological examination of the villi to establish that they are damaged (JAMA commentary).

2. The biopsy is the final diagnosis if the individual tests positive for anti-TG (transglutaminase) and/or anti-endomysium antibodies. In most countries (at least in Europe) the biopsy is required to be positive before there is any reimbursement of gluten free foods, which are quite expensive. This is not really a good situation, often it is the case that gluten sensitivity is suspected and the sources are eliminated leading to some recovery – at this point, if the patient wants reimbursement he/she has to become ill again, and not trivially since it involves actual physical damage to the small intestine.

3. CD is on the increase, in the USA levels are 4 fold higher than 50 years ago, and this is not because of better diagnosis as the samples were compared to actual stored blood samples, the data reflect a true increase in the prevalence of the disease (Rubio-Tapia, et. Al, 2009)

4. Approx 50% of people with type 1 diabetes showed some immunological reaction to gluten and other wheat proteins and about 10% are actually diagnosed as CD (Mojibian, 2009)

5. CD sufferers have higher mortality, as described in a recent study of >40,000 individuals (Ludvigsson, 2009), but the surprising finding was that even higher mortality was seen in the antibody positive but biopsy negative groups:


  HR for increased mortality
Celiac disease


(no villous atrophy)


Latent Celiac
(antibodies but normal mucosa)


In Sweden only those with celiac disease are treated with a gluten free diet while very few with inflammation or latent disease are. The hypothesis is that the continued consumption has health risks. Note that the study looked at mortality rates but of course this is a marker for overall poorer long term health and increased incidence of chronic disease. Many European countries will only reimburse gluten free foods for those with full celiac disease – these data make the diagnostic situation even less acceptable.

6. CD may be preventable? There is evidence that infant feeding can affect onset of CD. Interesting data from the Swedish epidemic when CD levels went up 3-4-fold during 1984-1996 whereas there was no change in neighbouring countries. The rise in CD coincided with a change in the instructions given to mothers (for which compliance is >90% in Sweden), leading to abrupt gluten introduction and cessation of breast feeding. In 1996 the instructions were changed to encourage longer breast feeding and to introduce gluten gradually between months 4-6 and while still breast feeding – the CD rate dropped 4-fold in the ensuing years (Olsson, 2008; Myléus, 2009).


7. A study by Norris et al showed that feeding between 4-6 months was linked to the lowest risk of developing CD (Norris 2005).


The combined evidence lead to the ESPGHAN (European Society for Pediatric Gastroenterology, Hepatology, and Nutrition, Agostoni et al 2008) to issue a statement advising avoidance of early (<4 months) and late (>6 months) introduction of gluten, and to do so gradually while still breast feeding. Although the evidence is not from clinical trials, babies have to be fed and a choice needs to be made, in the circumstances the evidence points to this being the best advice (there are some clinical trials in progress or recruiting but they seem to be comparing introduction at 6 months and 12 months, except for 1 Finnish trial which does not specify - As always, clinical trials and nutrition is a difficult subject – given the evidence would you want your child to be in the late introduction arm?

Arguments for genetic testing: In newborns it would be helpful to identify the genetically predisposed so that the correct feeding advice can be given (it’s not just celiac disease but other autoimmune diseases as well, with shared HLA alleles). Why test though, why not give the same advice to all? The feeding program needs to be carefully followed and not all populations are as compliant as the Swedish (certainly not the Italians!) so it makes sense to target just the at risk groups, and if we don’t start somewhere we’ll never get to personalised medicine!

Are some people squeamish about the idea of genetically testing newborns for disease risk? It shouldn’t be the case here, there is a clear prevention advantage and in any case, if it were proposed to serologically HLA type them then I bet there would not be protests – it’s genetic exceptionalism…

For older children and adults – at least it can be useful in relatives of CD patients, excluding those who are not predisposed and identifying those who should be screened for auto-antibodies and then the biopsy if necessary. As a general screen? Maybe not an economical test yet for the state or insurers to pay for but it could be encouraged, the statistic that only 10-20% of CD sufferers are diagnosed is highly significant. If the true numbers were correctly diagnosed, and also if the bar was lowered to include antibody positives then all of a sudden there would be 10-20 fold more gluten sensitive individuals and this would have a positive effect on the gluten free food market, prices would drop, choice would increase and the lot of the CD suffer would improve. After all it’s not really necessary for it to be a disease, “just” avoid the gluten. Unfortunately it’s not so easy to do so, gluten is everywhere, and it’s use has increased enormously over the last few decades. It’s used as an ingredient in most processed food, to give texture - I am based in Italy, in a salumeria recently I was choosing some prosciutto, there were two types (actually being Italy there were probably about 20 types, just using licence here) and I asked the difference: one was free of phosphates while the other was also free of lactose and gluten – what the hell was gluten doing in ham in the first place, in Italy of all places?!

Some testing options:

Prometheus have a fairly expensive test at $329
Genelex at even more - $445
Decode and 23andme do not have SNPs that comprehensively type the HLA alleles but do have some other slight risk alleles discovered by GWAS. Not sure about Navigenics and Pathway as I couldn’t find the SNPs on their sites.

A promising recent development was published in PLOS One (Monsuur et al,  2009) where they describe 6 tagging SNPs that comprehensively genotype the relevant HLA alleles – any test based on these SNPs would be a lot cheaper than the PCR based methods used by Prometheus – e.g. 23andme could add them to their scan (as long as there are no exorbitant licence fees since the authors have applied for a patent on the findings … that’s another story for another day…)


  1. Excellent post, Keith.
    Keep up the good work.

  2. Excellent review - I hadn't realised that the genetics of any complex disease had this level of (negative) predictive power.

    Are you going to blog on the patenting issue?

    My reading of the docs in your last link is that the patenting office accept the novelty and industrial applicability of the SNP tagging approach to HLA typing, but reject the notion that there is any invention in it - largely on the basis of the author's previously published work, amusingly.

    This matters to my group, as we use some of these SNPs (the assays for which can be readily purchased, as described in the paper) for tagging HLA for T1D risk. The patent applied for only covers CD, but even so it would be miserable if the assays became unavailable ...

  3. Thanks for the comments Neil.

    I hadn't thought of posting on the patenting issue, it's something I follow fairly closely, I'm not a lawyer and find patent applications excruciating to read - but it is an important issue.

    I just briefly lookeed over the reviewers report and I'm happy to see that 0/17 claims are considered innovative! It's good logic. A tagging SNP to determine HLA type? Obvious and done before. Use tagging SNPs to determine CD risk? Obvious of course since HLA involvement is long established.

    I'm not so happy about patenting work that has been carried out in publicly funded labs by people paid salaries and grants by the public (through tax and charity). These people are rewarded for their work by salary and publication and they are not investing any risk. It would be a shame if this patent application is delaying the development of a cheap CD test for this very important and underdiagnosed disease. Also particularly important in this case because the alternative to the SNPs is a very expensive assay. Currently there are a few serological screening programs to assess the ture incidence of the disease - with a cheap SNP assay it would be possible to increase the coverage by eliminating the 70% or so who don't have the alleles.

    On the other hand, if there was no access to patent protection maybe there would be delay anyway, maybe the publication of the details would be delayed while a commercial test is developed. I don't believe this would happen though, most scientists that I know and have known would consider the publication more important and by holding back they risk being beaten by another group.

    Patenting is a nuisance! I don't think though that it would have any effect on academic research, there are not many cases that I know of where patent holders demand licences from academics - it's not good PR!

    Best wishes for your work on T1D, I have a strong interest in that, and complements on the website -

  4. very helpful. Thanks, I think it could be very useful.

  5. Enterolab does genome testing for CD and non-celiac gluten sensitivity. Something like $169 including FedEx shipping. They don't test for HLA-DQA.

    Gluten Free Society also does genome testing for CD and non-celiac gluten sensitivity. They test for both HLA-DQA and HLA-DQB for something like $350.

    In v3 23andMe tests for 5 of the 6 SNPs referenced in the Monsuur et. al article you cite above. DQ 2.5 = rs2187668 (T). DQ7 = rs4639334 (A). DQ8 = rs7454108 (G). DQ 2.2 requires a test of 3 genes, two for inclusion, rs2395182 (T) and rs7775228 (C). The third one, that would exclude a person from being DQ 2.2 and make them DQ 4 instead, is to not be rs4713586(C). 23andMe does not test for the last one. Someone meeting all three criteria for DQ 2.2 has genoset 221.

  6. I once went to a clinical trial for celiac disease. I'm so glad that I did because ever since, I've gotten my eating habits underway and controlled. I don't feel sick often anymore.

  7. This was a really interesting read. One of my best friends had celiac disease. She was very paranoid about eating gluten, but I understood that she was trying to eat safe. She would get really sick if she accidentally ate something with it. She passed away a year ago from other health issues, but whenever I see gluten-free food at the grocery store it reminds me of her!
    Claudia Rosenburg |

  8. Thanks for this article. I truly hope more of the genetic interpretation sites add in more of the celiac markers. We are pretty concerned that our daughter has Celiac and she does have some major genetic markers but it's difficult to get a confirmation since I keep her off gluten. She had some once recently and she spiked a 105 fever within hours and had abdominal pain. Seemed more severe than typical intolerance. Don't really want to reintroduce it to get a diagnosis, though, and not sending my 4yo in for biopsy.

  9. A very interesting read having just taken a FamilyTreeDNA Family Finder DNA test which covers 3 out of the 6 coeliac genetic markers. However, reading the last response I find it, well, sad really doesn't express it properly, but that a parent wouldn't have a 4yo child not tested simply because it involves a 'biopsy'. Given that this is considered the 'Gold Standard' by which CD is properly diagnosed (although I have a cousins whose 2 biopsy's have been negative but genetic and blood tests both positive), the child will continue to live her life with a 'I'm on a GF diet because my mum says I should be and it hurts if I eat gluten' and will, most likely, get tagged as she gets older with simply being on a fad diet!

    Biopsy's are not the evil many people seem to believe they are. Yes, it involves some surgery which, like everything we do every single day has risk, but the risk surely outweighs, by far, the consequences of having a proper diagnosis.

    I wasn't diagnosed until I was 37 and have had 2 further biopsy's in the intervening years because of other issues. My symptoms didn't appear until 6 months before the first biopsy so I was lucky that I made the decision to have it. But, I would be extremely upset (to use a mild word) if I had discovered that my mother had decided, when I was aged 4, to not have a biopsy because of HER beliefs, which could potentially have found the truth at an early age.

    I was told when I had the first biopsy that now that CD had been diagnosed there would be, in all likelihood, secondary allergies that had been masked by the CD and to expect these to show (which did happen - allergies to amines being the predominant one) but the biopsy did not hurt, I can't even remember it (unlike some of the tests I did have to go through back in the 80's) and it turned my life around from the hell it was before I was diagnosed. Now my DNA test has, on the three markers contained within FTDNA's results, helped show and confirm what I already knew.

    Good article and good read.


Post a Comment

Popular posts from this blog

Genetic testing and potential harm: DTC or trust me I’m a doctor?

Recently at a couple of conferences ( European Human Genetics conference and Consumer Genetics Conf ) there have been various speakers questioning DTC genetics and calling for all health related personal genetics to be delivered through medical practitioners. I argued in the past that unregulated tests delivered through practitioners actually have the potential for more harm, not less. By coincidence last week some discrepancies in a DTC and a via MD test were pointed out to me – and they seem topical. Breast feeding has many benefits one of which appears to be increased IQ scores – however not all studies agree, some indicating that results may be confounded by maternal intelligence (see Wikipedia ). Sometimes inconsistencies in associating an action with an outcome can be resolved by looking at genetic variation (which tends to increase the error bars when not accounted for). So in 2007 some headlines were made when a study was published by Caspi’s group ( PNAS, open access )

The Great Health Data Deficit: Are Environmental causes for Disease a Mirage?

Ever since I was a child the message has been eat your greens (we got caned at primary school if we didn’t), you need to exercise, carrots will make you see in the dark, an apple a day keeps the doctor away. It continues today with the “5 a day”, all the healthy food pyramids, anti-junk foods, and so on. Yes the environment can be dangerous, it can kill us. There are some indisputable examples, a bullet in the head for example, or getting hit by a car is an environmental impact that almost always results in injury and death. There are others on which we more or less agree, like smoking is harmful (evidence is very strong, although it has never actually been proven, who knows, it could be the regular movement of hand to mouth that causes all the damage, we need to do the proper controlled trial…) For the rest we have always sort of somehow known that eating badly can cause disease – it seemed so obvious, and all the little research studies just confirmed what we knew and lead to the

Breast feeding: Personal genetics is in it’s infancy…

There is a lot in the new today about breast feeding – should it be exclusive for at least 6 months ( WHO ) or should it include some earlier introduction of solid foods ( BMJ article )? A lot of the reporting is confusing the issue, claiming or implying that the BMJ authors are saying the 6 months is too long (and pointing out that they have been consultants to the baby food industry). The BMJ authors are not saying that, they simply say that it might be prudent to introduce gradually some solid foods, while still breast feeding before 6 months. There is a useful account at the Marion Nestle blog and I left a comment there which I have expanded upon here. This could be an area where genetic testing can help – blanket advice is never the best for all but confusion (as we have now) is even worse. I blogged about celiac disease and genetic testing a while ago and part of the post addressed exactly the issue of breastfeeding and weaning. ...babies differ! Probably, very likely, one