The exclusion of oats from the diet in coeliac disease is controversial.
To study the long-term safety of oats in the treatment of children with coeliac disease.
Altogether 32 children with coeliac disease were enrolled in a 2-year controlled trial. Twenty-three children in remission were randomized either to oats or gluten challenge; when small bowel histological relapse was evident after gluten challenge, a gluten-free diet including oats was started. Furthermore, nine newly detected coeliac patients adopted an oat-containing gluten-free diet. Small bowel mucosal morphology, CD3+, αβ+ and γδ+ intraepithelial lymphocytes, human leucocyte antigen (HLA) DR expression and coeliac serology were determined. After the trial, the children were allowed to eat oats freely; follow-up was extended up to 7 years.
In coeliac children in remission, oats had no detrimental effect on intestinal histology or serology during the 2-year trial. In contrast, the gluten-challenge group relapsed after 3–12 months. Complete recovery from the disease was accomplished in all relapsed and newly detected patients on an oat-containing gluten-free diet. After the trial, 86% of the children preferred to consume oats and they all remained in remission.
In most children with coeliac disease, long-term consumption of oats is well tolerated, and it does not result in small bowel mucosal deterioration or immune activation.
The treatment of coeliac disease (CD) is based on the life-long exclusion of wheat, rye and barley prolamins. The issue whether oats can be safely consumed by coeliac patients has been debated as the gluten-free diet (GFD) was advocated over 50 years ago. Early short-term reports showing controversial results on oats toxicity were based on the observation of symptoms or fat malabsorption without small bowel mucosal biopsies, and it was not ruled out that the oat products used in these trials were free of contamination with other cereals1, 2 In 1995, the possible toxicity of oats was re-evaluated in a large controlled study in coeliac adults,3 and since then several additional studies have shown that most adults with CD and dermatitis herpetiformis tolerate uncontaminated oats,4–9 even in long-term use.10
Despite the fact that controlled studies3, 5, 11, 12 have so far given evidence about the safety of oats in the majority of patients with CD, doubt remains as to whether oats is eventually non-toxic and should really be included in the coeliac diet. Firstly, there is concern about the possible contamination of commercial oat products by other gluten-containing cereals.13, 14 Secondly, it has been noted that a subgroup of CD patients experience more abdominal symptoms while consuming an oat-containing GFD.7, 8 Thirdly, it was recently demonstrated by the Oslo group that all three oats intolerant coeliac patients had oat avenin-reactive T-cell lines in their small bowel mucosa,15 and furthermore it has been found that oats may contain also some wheat-like sequences in its protein structure.16 A fourth concern has been that there are only few studies about oats in children with CD.17 Recently, a randomized double-blind study in newly diagnosed coeliac children was published showing that an oat-containing GFD did not prevent clinical, serological or small bowel mucosal recovery in 1 year.12 However, 26% of children in the oat-containing GFD group withdrew from the study, which necessitates caution.12 Furthermore, long-term oats studies in coeliac children are totally lacking.
Our aim was to investigate the long-term safety of an oat-containing GFD in 32 children suffering from CD. We especially focused on minor small bowel mucosal inflammatory changes as measured by detailed morphometry, and on humoral response. Firstly, we conducted a 2-year prospective, randomized follow-up study, where children with CD in remission were randomized either to oats or gluten (oats, wheat, barely and rye) challenge. When small bowel histological relapse was verified during gluten challenge, the patients excluded wheat, barely and rye from their diet but continued to consume oats for 2 years. Similarly, newly detected CD children adopted an oat-containing GFD. After the 2-year trial, patients were allowed to eat oats freely in conjunction with an otherwise GFD and the clinical follow-up was extended up to 7 years.
Patients and methods
Subjects and study design
Altogether 36 consecutive over 7-year-old children with either previously diagnosed or newly detected CD were recruited for the study during 1996–1997 at the Department of Paediatrics, Tampere University Hospital; 32 of the children consented and four refused because they found the protocol too laborious. In all patients, the diagnosis of CD was based on the presence of small bowel mucosal severe partial or subtotal villous atrophy with crypt hyperplasia, and initially all had been serum endomysial antibody (EmA) positive. Twenty-three out of the 32 children were previously diagnosed CD patients and had been treated with a conventional (avoiding wheat, rye, barley and oats) strict GFD for at least 2 years before the present study and all had evinced a good clinical and serological response. At the baseline of the study, these 23 patients in remission were randomized either to undergo open oats challenge or a gluten challenge allowing in addition to oats the consumption of wheat, rye and barley. When clear small bowel mucosal relapse was verified during the gluten challenge, patients reverted to a GFD, this time avoiding wheat, rye and barley but with consumption of oats allowed to continue. Similarly, nine consecutive newly diagnosed CD children were advised to adopt a GFD with oats (Figure 1). Currently, in Finland, gluten challenge is usually not needed during the clinical work-up of CD.
During the first 2 years on an oat-containing diet clinical, nutritional and serological assessments were carried out at 0, 1, 3, 6, 12, 18, 24 months. Small bowel mucosal biopsies were evaluated at baseline and after six and 24 months on oats. In patients undergoing a gluten challenge, follow-up examinations were carried out as in the oats group until small bowel mucosal histological relapse was evident. After the relapse and commencement of an oat-containing GFD follow-up, examinations were carried out in the same way as in the oats challenge group (see above).
After the 2-year trial, patients were allowed to eat oats freely with an otherwise GFD and the clinical follow-up visits including nutritional and serological assessment were carried out once a year or every other year until March 2004. Small bowel mucosal biopsies were only considered if patients’ clinical condition or serology implied a relapse of the disease.
Oat product and dietary assessment
At the beginning of the study, children and their parents were given instructions by a trained dietitian on the oat-containing GFD and possible gluten challenge. A detailed dietary analysis and a history of occasional or regular consumption of gluten-containing products or oats were assessed by means of an interview and a 4-day record of food intake at baseline and after 3, 6, 12 and 24 months from entering the study. Based on the data obtained, the daily consumption of gluten or oats in grams was assessed as previously.18 Patients in the gluten-challenge group were advised to eat 20 g gluten per day as wheat, rye, barley and oats products – meaning approximately 8–10 slices of normal bread daily. The goal for daily intake of oats was 50 g as porridge or home-baked bread. The rolled oats (one single batch; Melia Ltd, Raisio, Finland) given to the patients without charge for 2 years were free of gluten contamination as confirmed by enzyme-linked immunosorbent assay (ELISA, Ridascreen Gluten Kit; Biopharm, Darmstedt, Germany) and polymerase chain reaction techniques.19 After the 2-year trial, patients continued with commercially available oats products from the general market. The purity of these products has been previously analysed: 29 out of 30 tested samples had gliadin levels below 28 mg/kg (=p.p.m.) and only one organic cultured oat product was clearly wheat-contaminated in excess of 200 p.p.m. gluten,19 the proposed level for a new Codex Standard for gluten-free products.20
Small bowel mucosal morphology
Small bowel mucosal biopsies were taken by an adult-size Watson capsule from the proximal jejunum, and all specimens were evaluated blindly by the same investigator, who had no prior knowledge of disease history or laboratory values. One part of each specimen was processed and stained with haematoxylin-eosin and studied under light microscopy. Morphometric analysis covering villous height and crypt depth ratio (Vh/CrD) and the height of enterocytes (μm) was performed in well-oriented biopsy samples, as previously described.21 Poorly oriented sections were not accepted, and the samples were re-dissected until they were of good quality.
For immunostaining, the other part of the small bowel biopsy specimen was freshly embedded in optimal cutting temperature compound (Tissue-Tek; Miles Inc, Elkhart, IN, USA), snap-frozen in liquid nitrogen and stored at −70 °C until used. Immunohistochemical studies using the avidin-biotin immunoperoxidase system were carried out on 5-μm-thick frozen sections. CD3+ intraepithelial lymphocytes (IELs) were stained with monoclonal antibody Leu-4 (Becton Dickinson, San Jose, CA, USA), αβ+ IELs with monoclonal βF1 antibody (Endogen, Woburn, MA, USA) and γδ+ IELs with TCRγ antibody (Endogen). The numbers of positively stained IELs were counted with a ×100 flat-field light microscope objective from randomly selected surface epithelium; at least 30 fields were counted along the epithelium and IEL density expressed as cells/millimetre epithelium as described in detail elsewhere.22 The reference values were set at 37 cells/mm for CD3+, at 25 cells/mm for αβ+ and at 4.3 for γδ+ IELs, and for Vh/CrD for the ratio 2.8.22 Small bowel mucosal human leucocyte antigen (HLA) DR expression was examined by staining frozen biopsy specimen with monoclonal antibody HLA DR (Becton Dickinson) in a dilution of 1:1500. The expression was considered enhanced when strong in the villous epithelium or present in crypts. Negative crypt and only slight-to-moderate villous epithelium expression was considered negative.23
Serum IgA-class EmA was investigated by an indirect immunofluorescence method using human umbilical cord as substrate; a dilution of 1:≥5 was considered positive.24 Serum IgA-class tissue transglutaminase antibodies (tTG-ab) were determined by ELISA (Celikey; Pharmacia Diagnostics, GmbH, Freiburg, Germany) using human recombinant tissue transglutaminase as antigen, a unit value (U) ≥5 being considered positive. Serum gliadin antibodies (AGA) were measured by in-house ELISA and the lower limit of positivity for IgA-class was 0.2 ELISA units (EU) per millilitre.24, 25 None of the patients studied had selective IgA-deficiency.
Clinical and nutritional evaluation
Possible symptoms were enquired at each follow-up visit. Weight (kg) and height (cm) were measured and progression of puberty evaluated. A growth chart was drawn up for each child. Blood haemoglobin level was measured using a routine laboratory method; values 11–16 g/dL for children 7–13 years of age and 12–16 g/dL (girls) and 13–16 g/dL (boys) over 13 years of age were considered normal.
Fisher's exact test was used in cross-tabulations and a two-tailed t-test was used to compare the parameters in different time-points; a P-value of <0.05 was considered significant. Quantitative data are given as means with 95% CI; when the confidence intervals do not overlap, the difference between the groups is significant.
The study protocol was approved by the Ethical Committee of Tampere University Hospital. All parents and children (when applicable) gave written informed consent.
Thirteen children with CD in clinical remission were randomized to oats challenge and 10 to gluten (in addition to oats, also wheat, rye and barley) challenge (Figure 1); the groups were similar with respect to age, duration of GFD and history of symptoms. Nor did age differ between these two groups and the newly-detected CD group (Table 1).
|CD in remission||Newly detected CD, GFD with oats (n = 9)|
|Oats challenge (n = 13)||Gluten challenge* (n = 10)|
|Median age at baseline of the study (range), years||11 (9–17)||13 (7–15)||12 (8–14)|
|Median duration of gluten-free diet (range), years||6 (2–16)||5 (2–10)||0|
|Original symptoms or signs leading to the diagnosis of CD, n|
|Anaemia and malabsorption||1||1||0|
|Screening in risk group‡||2||2||2|
Diet during a 2-year controlled trial
During a 2-year controlled trial, the median daily consumption of oats was 45 g/day (range 13–81 g/day) in coeliac patients in clinical remission randomized to oats challenge. At the same time, in coeliac children in remission randomized to gluten challenge, the median daily intake of gluten was 14 g (range 7–19 g/day); after histological relapse of the disease, the patients adopted a GFD containing oats for 2 years (median intake of oats 41 g/day, range 24–59 g/day). In newly detected coeliac children, the median daily consumption of oats was 43 g (range 19–64 g/day) during the study.
The overall compliance to a GFD was good during the study. Only one previously detected patient in the oats challenge group had transient dietary lapses after 6 months from the beginning of the study; all the rest adhered to a strict GFD throughout the 2-year study.
Withdrawals and symptoms during a 2-year trial
Two CD children in clinical remission experienced intolerable abdominal symptoms (abdominal pain and vomiting) immediately after intake of oats, and within 1 month both discontinued the study (Figure 1). At the follow-up examination, coeliac serology was negative, and small bowel biopsies taken instantly after appearance of the symptoms and cessation of oats showed that mucosal villous morphology was normal and the densities of IELs were even lower than at the beginning of the study, and mucosal HLA DR expression had become negative in both; furthermore, no signs of oats allergy were verified in them by prick, patch and RAST tests (data not shown). A third patient, from the gluten-challenge group, discontinued the study shortly after gluten-induced histological relapse and commencement of an oat-containing GFD; she was asymptomatic but found the study protocol too laborious (Figure 1).
Nine out of 13 previously diagnosed patients in the oats challenge group were totally asymptomatic throughout the study; two temporarily experienced slight abdominal distension after the first 3 months of the study, but in both the symptoms resolved while continuing the consumption of oats. In the gluten-challenge group, four out of 10 patients developed abdominal symptoms at the time of small bowel histological deterioration; they all, as well as the newly diagnosed untreated coeliac patients, became asymptomatic during an oat-containing GFD.
Small bowel biopsy findings during a 2-year trial
At the beginning of the study, all CD children in remission had normal small bowel mucosal villous morphology, and there were no significant changes in the Vh/CrD and enterocyte heights during the 2-year oats challenge (Figure 2, Table 2). In parallel densities of CD3+, αβ+ and γβ+ IELs decreased (P = 0.02, P = 0.03 and P = 0.02, respectively) and enhanced mucosal HLA DR expression disappeared in all on an oat-containing GFD (Figure 2, Table 2).
|Variable||CD in remission||Relapsed CD patients*|
GFD with oats
|Newly detected CD|
GFD with oats
|Oats challenge||Gluten challenge*|
|Enhanced mucosal HLA DR expression; n (%)|
|Baseline||10/13 (80%)||6/10 (60%)||10/10 (100%)||8/9 (89%)|
|6 months||8/11 (73%)||10/10 (100%)‡||6/8 (75%)||5/7 (71%)|
|24 months||0/11 (0%)†§||1/7 (14%)†||3/8 (38%)‡|
|Enterocyte height; mean (95% CI)|
|Baseline||39 (38–40)||42 (39–46)||30 (25–34)||31 (28–33)|
|6 months||41 (39–42)||30 (25–34)||38 (36–40)||36 (33–38)|
|24 months||41 (39–43)||38 (37–40)||37 (34–40)|
By comparison, in all coeliac patients randomized to gluten-challenge, small bowel mucosal deterioration was evident within 3–12 months; the mean Vh/CrD and the enterocyte height decreased and the densities of IELs increased statistically significantly (Figure 2, Table 2). All these parameters in relapsed coeliac patients were similar to those with newly detected untreated CD. Commencement of oat-containing GFD did not prevent small bowel mucosal recovery in relapsed and newly detected coeliac patients; after 6 months on oats, the mean Vh/CrD and enterocyte heights were significantly higher and the densities of CD3+ and αβ+ IELs significantly lower than those during the gluten-containing diet (Figure 2, Table 2). After 2 years on an oat-containing GFD, small bowel mucosal villous morphology, enterocyte heights and the mean densities of IELs did not differ significantly from the baseline values of treated coeliac children who had been for years on a traditional GFD, avoiding also oats (Figure 2, Table 2). Enhanced small bowel mucosal HLA DR expression was detected less frequently in patients using a long-term oat-containing GFD than in patients on a long-term traditional GFD and active CD (Table 2).
Serology during a 2-year trial
During the oats challenge, serum IgA-class EmA and tTG-ab remained negative in all coeliac children in remission adhering otherwise to a strict GFD. One patient admitting to passing dietary lapses had positive EmA-titre at 6 months; after adopting a strict GFD, serum EmA became negative even though the patient continued to consume oats (Table 3).
|Variable||Reference||CD in remission||Relapsed CD patients *|
GFD with oats n = 10
|Newly detected CD|
GFD with oats n = 9
|Oats challenge n = 13||Gluten challenge*n = 10|
|Endomysial antibodies, median (range)||>1:5|
|Baseline||0 (0–0)||0 (0–0)||1:500 (1:100–1:4000)||1:1000 (1:5–1:2000)|
|1 month||0 (0–0)||1:50 (0–1:1000)||1:200 (1:5–1:500)||1:200 (1:5–1:1000)|
|3 months||0 (0–0)||1:500 (0–1:8000)||1:5 (0–1:200)||1:50 (0–1:200)|
|6 months||0 (0–1:5)||1:200 (1:5–1:1000)||0 (0–1:50)||1:5 (0–1:200)|
|12 months||0 (0–0)||1:500 (1:500–1:500)||0 (0–0)||1:5 (0–1:50)|
|18 months||0 (0–0)||0 (0–0)||0 (0–1:5)|
|24 months||0 (0–0)||0 (0–0)||0 (0–0)|
|Transglutaminase antibodies, median (range)||≤5.0 U/L|
|Baseline||0.3 (0–1)||0.3 (0–1)||93 (16.3 to >100)||90 (3.0 to >100)|
|1 month||0.2 (0–1)||6 (0.1 to >100)||13 (4–91)||71 (1 to >100)|
|3 months||0.5 (0–1)||22 (1 to >100)||4 (1–33)||9 (1–27)|
|6 months||1 (0–3.1)||57 (7 to >100)||2 (1–11)||4 (1–11)|
|12 months||0.2 (0–2)||>100 (16 to >100)||1 (0.2–1)||4 (1–7)|
|18 months||0.4 (0–1)||1 (0.2–1)||1 (1–6)|
|24 months||1 (0–1)||1 (0.1–2)||1 (0.4–3)|
|Gliadin antibodies, median (range)||≤0.2 EU/L|
|Baseline||0 (0–0.2)||0 (0–0)||1 (0–1)||0.4 (0–20)|
|1 month||0 (0–0)||0 (0–4)||0 (0–0.4)||0.3 (0–7)|
|3 months||0 (0–0.2)||1 (0–1)||0 (0–1)||0 (0–1)|
|6 months||0 (0–0)||1 (0–1)||0 (0–1)||0 (0–0.3)|
|12 months||0 (0–0)||0 (0–0)||0 (0–0.3)|
|18 months||0 (0–0)||0 (0–0)||0 (0–0.3)|
|24 months||0 (0–0.3)||0 (0–0)||0 (0–0.3)|
In the gluten-challenge group, serum EmA and tTG-ab became positive within a month in five out of nine patients, and these antibodies were positive in all 10 at the time of small bowel mucosal histological relapse. In all relapsed CD patients, as well as in all newly detected untreated patients, serum EMA and tTG-ab titres decreased over time with an oat-containing GFD, being negative in all after 2 years. Changes in serum IgA-class AGA levels were less prominent; two patients had borderline-positive values after 2 years on an oat-containing GFD (Table 3).
Long-term clinical follow-up
After the 2-year controlled trial, all 29 CD children completing the study were allowed to eat commercially available oats freely (Figure 1), and the clinical follow-up was extended up to 7 years (median 5 years). Seven were lost to follow-up as moving away from the Tampere University Hospital area. Three (14%) of the remaining 22 children with CD stopped the consumption of oats, because they did not like the taste of it. The remaining 19 (86%) decided to continue the oat-containing GFD. Detailed food recordings were no longer carried out, but at clinical control visits the patients reported that they consumed oats, only less than during the challenge study. Fourteen out of the 19 patients were keeping an otherwise strict GFD; serum EMA and tTG-ab remained negative in all. Five adolescents consuming an oat-containing diet and one on a traditional GFD reported transient dietary transgression, and during these lapses serum EmA and tTG-ab also became temporarily positive in two of the patients. All coeliac children consuming oats remained asymptomatic, and their development in height, weight and puberty was normal (data not shown). At the end of the follow-up, median haemoglobin levels were 13 g/dL (range 12–14 g/dL) in girls and 14 g/dL (range 14–17 g/dL) in boys. None was suspected of relapse of CD on clinical or serological grounds and therefore no additional small bowel biopsies were carried out.
In the present study, we showed for the first time in children with CD that consumption of considerable amounts of oats is well tolerated, and it does not result in small bowel mucosal damage or humoral response even when taken for many years. These findings are in accordance with earlier results from clinical trials in adults with CD and dermatitis herpetiformis,3–6, 9 and in vitro immunologic studies in organ culture systems.26, 27 Even more reassuringly, in the current study after long-term consumption of an oat-containing GFD, all CD children were found to have normal densities of αβ+ IELs (Figure 2c). In CD, the most sensitive indicator in small bowel mucosal response to gluten has been an increase in IELs,28 and especially αβ+ IELs are regarded as highly gluten-dependent.22, 29 It is known that in CD the density of γδ+ IELs remains high over a long period on a GFD,30, 31 but interestingly, even the densities of these cells decreased in our patients over time with oats (Figure 2d). It is noteworthy that the small bowel mucosal response was totally opposite when the children started to eat, in addition to oats, also wheat, rye and barley. During such gluten-challenge, coeliac serology quickly became positive and small bowel mucosal histological relapse was evident within 12 months in all (Table 3, Figure 2). Furthermore, in these relapsed patients, the consumption of oats in an otherwise GFD did not prevent small bowel mucosal recovery, this again suggesting that oats is neither toxic nor immunogenic in CD.
According to the current literature, some coeliac patients experience gastrointestinal symptoms more often during an oat-containing than during a traditional GFD.7, 8 In general, such symptoms have been mild, and the appearance of abdominal distension and flatulence has previously been explained by an increased intake of fibre from oat products.9 In the current study, children were encouraged to consume considerable large quantities of oats daily, and it is possible that the transient abdominal symptoms at the beginning of the study were because of such increased fibre intake. However, in two children evincing immediate dramatic abdominal symptoms after oat intake (vomiting and pain), the aetiology of the symptoms remained obscure. Interestingly, in the follow-up examination, these two patients were found to evince no signs of gluten-induced mucosal inflammation or cereal allergy. In our recent study in adults, oat-containing GFD caused more intestinal symptoms than the traditional diet.8 In that study, we did not find any difference in small bowel mucosal morphology between the diet groups after 1-year follow-up, whereas the densities of IELs were significantly higher in the oats group. The existence of symptoms was not associated with the increase in the density of IELs.8 Moreover, it must be noted that the biopsies were not taken from all patients, and at the baseline the densities of IELs were already higher in patients randomized to oats-group than in those randomized to non-oats group. Different study designs and different oat products in our earlier study in adults and current study in children disturb the comparison between these studies. Still recently, a Norwegian research group demonstrated that some oat-intolerant patients with CD might have avenin-reactive T cells in the small bowel mucosa;15 the clinical relevance of such findings is, however, unknown. Nonetheless, it would seem that even if most CD patients tolerate oats, there are some who have to avoid it to remain in remission.
Fear of wheat contamination originating in the harvesting and milling process in commercially available oat products has led to a reluctance in recommending oats as part of a GFD in many countries.13, 14 However, it must be noted that oat is by no means different from other cereals; in recent studies, gluten contamination was also found in naturally gluten-free products, i.e. maize, rice and buckwheat.32, 33 As there is a market demand for oat products among patients with CD, coeliac societies and the industry should make an effort also to promote the development of safe oat products free of gluten contamination.
In oats, avenin accounts for only 5–15% of the total protein, whereas in wheat, barley and rye prolamins constitute 30–50% of the total protein. Moreover, the putative toxic amino acid sequences are less frequent in avenin than in other prolamins, which would explain the apparent non-toxic nature of oats.34 It has therefore been argued that large quantities of oats might still be harmful to coeliac patients. In earlier short-term studies in children with CD, the daily intake of oats was fairly low (median of 15 and 24 g/day),12, 17 but in the present series even ingestion of a median of 43 g (up to 81 g/day) oats daily seemed not to be harmful. In fact, this is well in accord with earlier results showing that adults with CD do also tolerate large amounts of oats.9
We conclude that uncontaminated oats can be safely included in a GFD also in the majority of children suffering from CD. Oats diversifies a GFD, and when allowed most coeliac children prefer to consume some oats in their diet. As there is a clear market demand for oats, coeliac societies and industry should make an effort to produce oat products free of wheat contamination.
The present study and the Coeliac Disease Study Group are supported by grants from the Medical Research Fund of Tampere University Hospital, the Foundation of the Friends of the University Children's Hospitals in Finland, the Foundation for Paediatric Research, the Finnish Medical Foundation, the Yrjö Jahnsson Foundation, the Cultural Foundation and the Academy of Finland Research Council for Health.