This review provides the results of 5 completed clinical trials and highlights 2 ongoing trials. The first of the completed trials, the BABYDIET study, was conducted in Germany and found that infants with a genetic risk for type 1 diabetes (which overlaps with celiac risk) did not have a significantly different prevalence of autoantibody production when gluten was introduced at 6 vs. 12 months. However, the second controlled trial found the exact opposite; although a smaller study (30 high-risk infants total), an increased incidence of celiac disease was observed when gluten was introduced at 6 months vs. 12 months, with significant differences observed between the microbiome of genetically susceptible infants and those without increased risk.
A much larger study (the CELIPREV study enrolled over 700 infants) published in NEJM found that delaying gluten introduction until 12 months (vs. 6 months) in those with genetic susceptibility did delay the onset of autoimmunity and celiac disease, but did not prevent either. Thus, while differences were observed at age 2, the incidence of both autoimmunity and celiac were not different by age 5. It’s important to keep in mind that all the infants in this study had a first-degree relative with celiac disease, and most of the infants had a high-risk genotype (which itself was a strong predictor of disease development), so extrapolating to infants without any heightened risk may not be applicable. Breastfeeding was also found to lack any impact on the study outcomes.
Another large study, PREVENT-CD, attempted to feed high-risk infants small amounts of gluten (100mg daily) during 4-6 months of age, yet no difference was detected in celiac disease prevalence by age 3 (and no effect of breastfeeding was observed), suggesting that early gluten introduction may not benefit high-risk infants.
The results of the largest study to date (the EAT study, published in JAMA Pediatrics) included over 1000 infants randomized at 13 weeks of age to an early introduction group that consumed 6 allergenic foods (cow's milk, hen's egg, peanut, sesame, cod fish, and wheat) in addition to breast milk, or to an exclusively breastfed group. It’s important to note that these infants were from the general population, i.e., they were not at high risk for the development of celiac disease or other autoimmune conditions. The early introduction group consumed approximately 3.2g of gluten per week, over 4.5-fold higher than the dose consumed in the PREVENT-CD trial.
This study did find a protective effect of this earlier introduction of a higher gluten dose. However, the number of total celiac cases was small, and the study was not intentionally powered to detect a difference in celiac prevalence. This last point is important, as it highlights the potential differences when enrolling high-risk vs. normal-risk infants, as the takeaway dietary messages may not be identical, i.e., early introduction of allergenic foods in infants that are not at high-risk may be protective (or at least not harmful), yet there may not be a benefit for early introduction to high-risk infants, and possibly some harm. This is precisely what was observed in a large prospective cohort (observational) trial published in JAMA, The Environmental Determinants of Diabetes in the Young (TEDDY) study. Enrolling over 6600 children at 6 centers, this study found that among children at high risk for type 1 diabetes or celiac disease, daily gluten intake was associated with a dose-dependent increase in celiac incidence.
The ongoing clinical trials may help to address some of the remaining questions, as both of the studies mentioned are enrolling infants at moderate to high genetic risk. The first of these studies (PreCiSe) will examine the effects of a gluten-free diet for the first 3 years of life on the incidence of celiac disease by age 7. Given that the peak incidence of celiac disease is between 2-3 years, this study has important implications for its prevention in high-risk infants. Additionally, this study will examine the effects of probiotic supplementation with Lactobacillus paracasei and plantarum, strains previously shown to favorably modulate the activity of regulatory T cells in children with celiac disease. The 2nd trial (GRAIN) is enrolling infants at moderate risk, and will evaluate the impact of a somewhat restricted gluten diet (<2g per day) during the first 5 years of life on the development of celiac disease by age 10.
Certainly, there are other variables that need consideration. For instance, infection with enterovirus between 1-2 years of age may interact with gluten exposure to influence risk of autoimmunity in susceptible children, as observed in the TEDDY study. Yet at this point, the message seems to be that while there may not be any benefit for restricting gluten in those with typical risk, there still may be benefit among those at high risk, the degree of which will hopefully be clarified in studies currently underway.