A balanced look at gluten sensitivity
The Monash paper,1 which was previously discussed on SBM, suggested that it might be the fructans in wheat and not gluten that is responsible for symptoms in IBS sufferers who feel better on a gluten-free diet. Fructans belong to a group of short-chain carbohydrates known as FODMAPs that are readily fermented by bacteria in the intestine. If fructans were really to blame for wheat-induced gastrointestinal symptoms, this would be good news for IBS sufferers currently on a gluten-free diet — for some, a diet low in FODMAPs would be less restrictive than one without gluten, making it less prone to nutritional deficiencies.
As it stands, the existence of NCGS has neither been proven nor disproven by anyone. But gluten sensitivity sits at the intersection of several dilemmas in medicine today and, unlike how it’s portrayed in the media, is hardly an all-or-nothing affair. Proving it wrong will not instantly heal the people who have prescribed themselves a gluten-free diet. Proving it wrong will not produce a cure for IBS, a shorter time to a celiac diagnosis, or the correct way to handle potential celiac disease. Neither will proving it right. In the eyes of one gluten avoider, “Modern medicine is really good at crisis intervention… they don’t do well with chronic issues”.2
With these issues in mind, it’s time to move past the media debates and fad dieters and take a balanced look at NCGS. This overview will use four recent articles by the Monash group1, 3, 4, 5 as a framework to uncover some of the factors contributing to the gluten sensitivity phenomenon. Even though it has been gone over many times, a discussion of the FODMAPs study is still in order — in fact, essential — to appreciate the central importance of the elimination diet in diagnosing a food sensitivity. Along the way, we’ll get an idea of who the gluten sensitive might be, how successful gluten-free diets really are, and how challenging it is to pursue a celiac diagnosis.
To refresh, here are some of the key ideas discussed:
Gluten: A protein found in relatively large amounts in wheat, but also found in other grains.
Celiac disease: A genetically-linked autoimmune disorder caused by the small intestine reacting to gluten. Causes pain, altered bowel movements and malnutrition.
Irritable bowel syndrome: A diagnosis made purely on the basis of symptoms (bloating, pain and either diarrhea, constipation, or both), with no known current cause.
Elimination diet: A diet which removes specific foods or classes of foods from the diet in an effort to identify the trigger to an allergic reaction or symptom(s).
A food sensitivity is born
In 2011, an international panel of experts defined NCGS as “a non-allergic and non-autoimmune condition in which the consumption of gluten can lead to symptoms similar to those seen in celiac disease”.6 Non-intestinal symptoms can include behavioral changes, bone or joint pain, muscle cramps, leg numbness, weight loss, and chronic fatigue.7 But this is only a working definition, and there is still a lot of uncertainty as to what NCGS actually represents, even for its proponents.
So far, NCGS appears to be a heterogeneous phenomenon, with at least three different and potentially-overlapping subtypes6, 8:
A type reminiscent of celiac disease — a family history of celiac disease or the celiac HLA genes but no increase in intestinal permeability and no auto-immune response.
A type reminiscent of irritable bowel syndrome — intestinal motility issues but a greater frequency non-intestinal symptoms.
A type reminiscent of food allergies — prior history of allergy or concurrent food sensitivities and some immunological markers for celiac disease, but no activation of mucosal basophils by gliadin.
Multiple subtypes imply multiple mechanisms, and this too needs to be ironed out for NCGS. Sapone et al.9 described NCGS as “an inflammatory condition mostly supported by innate immune mechanisms,” which differentiates it from celiac disease and wheat allergy, and in vitro evidence suggests that other poorly-digested wheat proteins known as amylase-trypsin inhibitors might trigger the innate immune response.6 Changes in intestinal microbiota may also play a role in NCGS.10 And, of course, the fructans have their own special magic.
When NCGS was officially created by consensus, things weren’t proceeding according to the normal order. Celiac researchers, mainly centered at the University of Maryland, were pretty convinced that they had seen non-celiac patients whose symptoms had improved on a gluten-free diet. They searched for a set of biomarkers that would shed light on these cases, but found nothing consistent except for the absence of two celiac hallmarks — an autoimmune response and increased intestinal permeability.6, 8
From a food sensitivity perspective, however, searching for biomarkers is putting the cart before the horse. Food sensitivities, like cow’s milk protein intolerance or certain food additive intolerances, are only recognized after they have been demonstrated in studies using a double-blind, placebo-controlled food challenge (DBPCFC). The DBPCFC works regardless of the mechanism behind the reaction, so it can be used before any biomarkers have been identified.
Back in 2011, only one study had been completed on gluten sensitivity using a DBPCFC,3 and this first bit of positive evidence came from the Monash group. The study looked at a small group of IBS sufferers who identified themselves as gluten sensitive and who tested negative for celiac disease. They could only enter the study if their symptoms — abdominal pain, bloating, gas, constipation, diarrhea, or tiredness — were currently well controlled by a gluten-free diet.
Most DBPCFC studies use a crossover design, but the first Monash study randomly divided the participants into test and control groups. All participants continued on their usual gluten-free diet during the course of the study and were challenged with either gluten-free or gluten-containing low-FODMAP bread and muffins. The gluten group experienced greater gastrointestinal symptoms and tiredness compared to the control group, so it looked like gluten could induce GI symptoms on its own.
Shortly thereafter, a group at the University of Palermo11 used a DBPCFC to see whether any of their IBS patients were affected by wheat. In the end, they reclassified 30% of the 920 patients as suffering from wheat sensitivity. Notice I said wheat sensitivity — this study used capsules of wheat flour as the challenge instead of gluten. It also differed from the Monash study in that it included a stricter elimination diet before and during the challenges, not just relying on participants to maintain their own gluten-free diets.
The elimination diet — or the baseline diet — just may be the most important part of a DBPCFC. It puts everyone on a level playing field, stabilizing symptoms and doing away with any effects that might carry over from the participants’ normal diets. An elimination diet excludes the food under investigation and, if the study is thorough, all other foods or food additives that might produce a reaction. If symptoms don’t improve by the end of the elimination diet, then the study is on the wrong track.
The Palermo researchers were looking at a broader set of food sensitivities than just wheat. Gluten-free participants were asked to resume a normal diet for 2-4 weeks before beginning the elimination diet, which excluded wheat plus cow’s milk, eggs, tomato, and chocolate — foods known to aggravate IBS — and any other triggers identified by the participants. A challenge was considered positive if symptoms returned after having disappeared during the elimination diet. Of those testing positive for wheat sensitivity, the majority were also positive for cow’s milk protein intolerance when challenged separately; half were also sensitive to eggs and tomato.
Since many different foods are commonly, although not universally, reported to trigger or worsen IBS symptoms, the Palermo study didn’t go far enough to determine whether NCGS (or wheat sensitivity) is really its own entity outside of IBS or to explain the possible overlap between the two conditions. On the other hand, the Monash results were significant because gluten alone seemed to have caused an adverse reaction in IBS sufferers. But this result needed to be replicated under stricter conditions, so 2013 brought a new paper from the Monash group1 on a second gluten trial.
In IBS, the enteric (intestinal) nervous system is easily overstimulated, leading to abnormal contractions and an exaggerated perception of pain. Food triggers IBS symptoms by stimulating the enteric nervous system directly through various chemical receptors or indirectly by pressure.12 FODMAPs — which include fructans, galactans, polyols, fructose, and lactose — lead to more pressure in the intestine than other foods because they are poorly absorbed, rapidly fermented, and osmotically active.13 In other words, they cause gas and loose stools. The low-FODMAP diet, developed in 2008 by the same group at Monash University that we have been discussing, is the first comprehensive diet plan shown to reduce symptoms in the majority of IBS sufferers.14
Before the low-FODMAP diet, avoiding single foods like wheat or milk wasn’t always effective in treating IBS14; eliminating more than one carbohydrate type, like fructans and fructose, seems to have an additive effect that controls IBS symptoms better.15 That said, not everyone with IBS is sensitive to all of the FODMAPs, and the diet is highly customizable. The effectiveness of the low-FODMAP diet means that any diet study involving IBS sufferers must control for the effects of high-FODMAP foods.
The second DBCPFC study from the Monash group — the one that has received all the media attention — was a randomized crossover trial using stricter diets and stricter testing to make sure that participants did not have latent celiac disease. In this trial, meals were provided and additional snacks were logged. Here they ran two experiments — one where the elimination diet excluded high-FODMAP foods and the other where the elimination diet excluded high-FODMAP foods, dairy, and known triggers of pharmacological food intolerance reactions.
In the first arm of the study, participants were given challenge foods containing gluten, gluten plus whey, and pure whey, each separated by a washout period. Overall, symptoms improved during the low-FODMAP elimination diet but worsened for each of the challenges. Individually, though, only 3 out of 37 showed a significant response attributable to gluten when all challenges were considered.
In the second arm, 22 people from the first group were rechallenged with foods that contained gluten, whey, or no additional protein. This phase, which was meant to verify the results of the first using an even stricter elimination diet, could not reproduce any of the previous reactions. For example, only 2 out of 22 responded to gluten in the second arm, and these were not the same people who responded to gluten in the first.
One blogger took issue with the fact that many of the participants reacted to whey: “If some of the ‘placebo’ and low-gluten patients were, in fact, sensitive to whey, then that would invalidate the results of the study.” But that’s not true. The point to remember is that most people did not react to the same challenge food twice, and only one person consistently reacted to whey. Furthermore, it isn’t a catastrophe in a study like this if someone reacts to the placebo — this can happen for many reasons, including just having a bad day. What matters is how the reaction to the real challenge compares to the placebo reaction; if it is significantly greater, then the challenge is positive.
What did the researchers conclude? Considering both arms of the study and the immunological testing that also took place, no evidence of gluten-specific effects were found in patients on a low-FODMAP diet. They did not conclude that gluten sensitivity does not exist, only that “gluten may not be a specific trigger of functional gut symptoms once dietary FODMAPs are reduced.” That, however, does not settle the debate over gluten sensitivity. What the two Monash trials have shown us is the correct elimination diet to be used when looking for NCGS.
Gluten free with symptoms
The story doesn’t end there, though. From the Monash work we also learned that a good number of gluten-free dieters continue with the diet even though their symptoms do not completely resolve. This came out in two ways. To be included in the two Monash studies, participants needed to have well-controlled symptoms on a gluten-free diet. For the second study, for example, 147 volunteers were screened using several questions, one being, “Do you currently feel in control of your symptoms?”4 To this, 22% answered “no,” 3% answered “sometimes,” 16% answered “mostly,” and 59% answered “yes.” Then, when participants rated their symptoms at the beginning of the study, 11 of 37 were under significant distress, and 22% of the group significantly improved while on the low-FODMAP elimination diet.1
It is tempting to think that the people who did not see complete improvement on their own gluten-free diet were not really sensitive to gluten in the first place, but we don’t have the data to support this — only 58% of the larger group believed that they were strictly gluten free, and it is not clear how many of the gluten ‘cheaters’ still had symptoms. As the researchers pointed out, we also do not know how the symptomatic group felt before they began avoiding gluten — perhaps they had actually improved quite a bit on a gluten-free diet, but just not completely. Another explanation is that FODMAPs were really to blame for their symptoms, and avoiding only wheat wasn’t enough to provide complete relief.
It isn’t surprising that a person could be mistaken about having gluten sensitivity. According to someone like Dr. Oz, gluten sensitivity can be diagnosed by avoiding gluten for a while — say two weeks or a month — and then trying some to see what happens. But avoiding a staple food like wheat amounts to a major change of eating habits, and it is likely that wheat will not be the only food that is eliminated. For those who aren’t accustomed to performing experiments on themselves, these diet instructions don’t explain that any and all foods excluded during an elimination diet must be excluded for the entire diet and then used as a challenge later on. On top of this, placebo challenges really are necessary, and one round of challenges may not be enough — when symptoms are subjective or when a strong nocebo effect is expected, the active and placebo challenges should be randomized and repeated three times.6, 7
Oddly, many participants from the second Monash study opted to continue their gluten-free diet after the study ended because the diet made them “feel” better.5 This intrigued the researchers, so they invited the participants back for a third study to find out whether gluten was affecting their mental state.5 This trial proceeded along the same lines as the second, and mental state was assessed using the Spielberger State Trait Personality Inventory. Gluten still had no impact on gastrointestinal symptoms, but it did induce feelings of depression. This is a very interesting result because it could explain the lure of the gluten-free diet; however, researchers are still a long way off from uncovering a definite link between gluten and depression, transient or otherwise.
The missing celiacs and other sufferers
The final bit of information from the Monash researchers comes from the 147 people who applied to participate in their gluten studies.4 Of these, 44% had initiated their gluten-free diet themselves; the rest were following the advice of an alternative health professional (21%), dietitian (19%), or general practitioner (16%). In the surveys that followed, the researchers determined that two-thirds had not adequately excluded celiac disease, meaning that 15% had not been tested at all for celiac disease and that 29% of those who had undergone duodenal biopsies did not consume enough gluten beforehand for the results to be valid . This includes half of the tests initiated by general practitioners.
Considering that one in 56 people with symptoms will have celiac disease,19 it is important to know why people feel confident diagnosing themselves with gluten sensitivity without first ruling out the more serious problem. Even if they do maintain a strict gluten-free diet for their entire lives, without a formal diagnosis, these people with possible celiac miss out on follow-up care, accommodations by schools and employers, and tax breaks or food subsidies in many countries. The truth is, many have pursued a celiac diagnosis or wish that they could have pursued one. The decision to go it alone is not taken lightly — consider these situations pulled from stories posted online:
They met the serologic for celiac disease but had a negative biopsy.
Their doctor won’t test them even though they have asked.
Their doctor says that they cannot have celiac disease because they are overweight.
Their doctor told them that they are too old to have celiac disease.
They were not told to or told not to do a gluten challenge before the endoscopy.
Their doctor ordered a colonoscopy to check for celiac disease.
They already believe that they have celiac disease because of their symptoms and family history.
An alternative practitioner told them that they have celiac disease because their reaction to gluten is severe.
They are already gluten free and don’t want to “get glutened” again in order to be tested.
Having a celiac diagnosis won’t affect their commitment to a gluten-free diet.
They believe that the Cyrex test is more thorough than conventional celiac testing from a doctor.
They believe that celiac disease is only a minor part of the spectrum of gluten-related disorders.
They are worried that their health and life insurance premiums will increase.
They don’t want a preexisting condition on their the health insurance records.
They do not have enough money for an endoscopy.
Their health insurance wouldn’t pay for an endoscopy.
Regardless of what we might think about these reasons, they paint the picture of a larger problem. Roughly 80% of celiac cases go undiagnosed,20 a situation that The Lancet has described as a “public health farce”21 in light of the millions of healthy people who choose to be gluten free. Undiagnosed celiac disease can lead to osteoporosis, anemia, infertility, certain intestinal cancers, other autoimmune disorders, and an increased risk of mortality, although the latter is still under debate.22, 23 Still, the average time to receive a celiac diagnosis has been around 11 years.24
Right now we don’t know how many people with self-diagnosed gluten sensitivity actually have celiac disease, but the Monash data tells us that this is a significant concern. Efforts are being made to cut down the time it takes for a celiac diagnosis, but progress is slow.24 Increasing celiac awareness among physicians is part myth busting and part reeducation, and in the absence of mass screening, doctors must be able to spot non-GI or “atypical” symptoms and associated conditions that suggest the presence of the disease.25 At the same time, the protocol for diagnosing celiac disease is being reexamined26 — in Europe, the duodenal biopsy, currently regarded as the gold standard for celiac diagnosis, can be omitted in certain circumstances.27
We also don’t know how often potential celiac disease is actually mislabeled as gluten sensitivity. Potential celiac disease describes cases where serologic and genetic markers for celiac disease are positive but the biopsy is negative.24 There is no consensus on whether to treat potential celiac disease, nor, for that matter, whether it is a real problem or just a false-positive result.24, 28 Some evidence suggests that a gluten-free diet is beneficial,28 while another report describes patients whose symptoms spontaneously improved even though they continued to eat gluten.29 Until the issues surrounding celiac and potential celiac disease have been resolved, non-celiac gluten sensitivity — or at least one of its subtypes — cannot be clearly defined.
Where do we go from here?
In a large Australian survey from 2013, roughly 7% of respondents had diagnosed themselves with gluten sensitivity.30 The gluten free fad is just as popular in Australia as it is in the US,30, 31 so this figure might also estimate the prevalence of gluten sensitivity here, even though a 2009 estimate was much lower at 0.55%.32
Who might be in the 7%? Some with IBS, some with a botched celiac diagnosis, some with potential celiac disease, some who are intolerant to foods that they inadvertently eliminated because of their new diet, some who get only a psychological benefit from avoiding gluten, some whose symptoms continue despite their best efforts, some who really are hypochondriacs, and perhaps some with NCGS as its own entity. It doesn’t even matter whether non-celiac gluten sensitivity exists — these people will remain.
But it’s going to be a while until gluten sensitivity is understood. According to Dr. Alessio Fasano, director of the Center for Celiac Research at the Massachusetts General Hospital, NCGS stands where celiac disease stood 40 years ago.8 Even so, all we need are a few good DBPCFC trials to answer the basic question as to whether gluten sensitivity exists. This may be easier said than done, however, as the Monash group found such a strong nocebo effect in their trials that they now believe that future research should involve IBS sufferers who have not tried a gluten-free diet before.33
In the meantime, more people will suspect that they are sensitive to gluten, and if they are not to be lost to self-diagnosis or alternative medicine, the medical community must be able to lead them through the process of sorting out their suspicions. The Monash researchers have suggested an interim pathway for diagnosing gluten sensitivity, which includes:4
The adequate exclusion of celiac disease.
The exclusion of other dietary triggers, like FODMAPs or other foods suspected by the individual.
A gluten-free diet, if symptoms did not resolve or improve in Step 2.
Blinded gluten challenges, if symptoms did improve in Step 3.
Rechallenges with gluten to establish their gluten threshold.
As a clinical approach, this moves us away from thinking about gluten sensitivity in all or nothing terms and addresses the issues one person at a time.