Factors influencing adherence in a trial of early introduction of allergenic food

Abstract

Background: The Enquiring About Tolerance (EAT) study examined whether the early introduction of 6 allergenic foods from 3 months of age in exclusively breastfed infants prevented the development of food allergy. The intervention was effective in the per-protocol analysis for allergy to 1 or more foods and for egg and peanut individually, but only 42% of early introduction group (EIG) children met the per-protocol criteria.

Objective: We sought to identify which factors were responsible for nonadherence in the EAT study.

Methods: Factors influencing adherence within the key early introduction period in the EIG (up to 6 months of age) were divided into enrollment and postenrollment factors, and their association with nonadherence was explored.

Results: In an adjusted analysis, at enrollment, increased maternal age, nonwhite ethnicity, and lower maternal quality of life were independently and significantly associated with overall nonadherence in the EIG. Enrollment eczema and enrollment serum allergen-specific IgE sensitization to 1 or more foods (≥0.1 kU/L) were not related to overall nonadherence. After enrollment, 2 factors were significantly related to EIG overall nonadherence: parent-reported IgE-type symptoms with infant allergenic food consumption by 6 months of age and reported feeding difficulties by 4 months of age.

Conclusion: If early introduction of allergenic foods were to be considered a strategy to prevent food allergy, families of nonwhite ethnicity, those with older mothers, and those with infants with reported feeding difficulties or early-onset eczema would benefit from support to promote early and sustained consumption.

Keywords: Food allergy; adherence; allergens; breastfeeding; diet; infancy; randomized controlled trial.

Graphical abstract

Fig 1

Adherence with introduction of allergenic foods in the EIG up to 1 year of age. The figure presents the relative proportions of the EIG consuming 100%, 75% (the per-protocol threshold), 50%, or 25% or not having started consuming each of the 6 early introduction foods from enrollment through 12 months of age. The food-specific per-protocol adherence percentage (among those whose food-specific adherence status was evaluable) is shown in parentheses.

Fig 2

EIG enrollment IgE sensitization and overall and food-specific per-protocol adherence. Penalized logistic regression of the association between enrollment IgE sensitization (≥0.1 kU/L) to specific foods or to 1 or more of the 6 early introduction foods and the association with food-specific and overall nonadherence are shown.

Fig 3

Reporting in the key early introduction period (up to 6 months) of IgE-type symptoms to specific foods, IgE-type or non–IgE-type symptoms to any of the early introduction foods, and the association with food-specific and overall per-protocol adherence. Penalized logistic regression of the association between symptoms with consumption of the 6 allergenic foods and food-specific and overall nonadherence. Symptoms manifesting by 6 months of age are presented for IgE-type symptoms for each specific food, IgE-type symptoms to 1 or more of the 6 early introduction foods, and non–IgE-type symptoms to 1 or more of the 6 early introduction foods.

Fig 4

Contributions of subgroups to the proportion of food allergy cases in the SIG. A,Bar charts provide prevalence calculations used to estimate the reduction in total allergy burden and number needed to treat. Per-protocol adherence rates are shown for those EIG participants whose adherence rates were evaluable and also as a proportion of the whole EIG (percentages in parentheses). B, An 80% treatment effect and 85% adherence across all risk factors for allergy is assumed. For example, infants in EAT with visible eczema comprised 61.9% of the total food allergy burden. Hypothetically, if per-protocol adherence could be achieved in 85% of this subgroup, then 52.6% (61.9%*85%) of the allergic burden would experience the intervention. Moreover, if an intervention effect of 80% is assumed, then the total reduction in food allergy that would be realized from intervening on this subgroup would be 42.1% (52.6%*80%).