Feeding infants peanut early prevents allergy. The old advice to avoid it backfired.

For infants at high risk of peanut allergy, introducing peanut early sharply reduces the chance they develop it — the opposite of the avoidance advice parents were given for years. In the landmark LEAP trial, high-risk infants who ate peanut regularly from infancy had a peanut allergy rate of 1.9% at five years, against 13.7% in those who avoided it, and the guidelines were rewritten as a result [s1] [s3].

What the LEAP trial found

The Learning Early About Peanut Allergy (LEAP) trial randomly assigned 640 infants with severe eczema, egg allergy, or both — the group at highest risk — to either consume or avoid peanut until 60 months of age [s1]. The infants were between 4 and 11 months old at randomisation and were split into two cohorts based on a skin-prick test done at the start, one with no measurable reaction and one with a small (1-to-4-mm) wheal [s1].

Among the 530 infants who initially had negative skin-prick tests, the prevalence of peanut allergy at 60 months was 13.7% in the avoidance group and 1.9% in the consumption group (P<0.001) [s1]. Among the 98 infants who already had positive tests — a more sensitised group — the effect held: 35.3% in the avoidance group developed peanut allergy versus 10.6% in the consumption group (P=0.004) [s1]. There was no significant difference between the groups in serious adverse events, meaning early feeding did not come at a measurable safety cost in this trial [s1].

This was a genuine reversal. For much of the 2000s, families with allergy history were advised to keep peanut away from young children, on the intuition that early exposure would sensitise them. LEAP showed the intuition was backwards for high-risk infants: early, sustained exposure trained tolerance.

What the EAT trial complicated

A second trial tested whether the same logic extends to the general population and to a wider set of foods. The Enquiring About Tolerance (EAT) study recruited 1,303 exclusively breast-fed three-month-olds from the general population and randomly assigned them either to early introduction of six allergenic foods — peanut, cooked egg, cow's milk, sesame, whitefish, and wheat — or to the standard UK practice of exclusive breastfeeding to about six months [s2].

Here the headline result was null. In the intention-to-treat analysis — which counts every child as assigned, regardless of what they actually ate — food allergy to one or more of the six foods developed in 7.1% of the standard-introduction group (42 of 595) and 5.6% of the early-introduction group (32 of 567), a difference that was not statistically significant (P=0.32) [s2].

But the trial's per-protocol analysis, restricted to the children who actually consumed the assigned amounts, told a different story: any food allergy was significantly lower with early introduction (2.4% vs 7.3%, P=0.01), as was peanut allergy (0% vs 2.5%, P=0.003) and egg allergy (1.4% vs 5.5%, P=0.009) [s2]. The catch is in why the intention-to-treat result was null: the trial reported that early introduction of all six foods "was not easily achieved" — many families could not get young infants to eat the required quantities — though the approach was safe [s2].

The honest reading of EAT is therefore two-sided. Early introduction appears to work when a family manages it, but achieving it in practice, across six foods in a three-month-old, is hard — and a benefit that depends on adherence is a weaker public-health tool than one that shows up regardless [s2].

What the guidelines changed to

On the strength of LEAP in particular, a National Institute of Allergy and Infectious Diseases expert panel issued addendum guidelines for the prevention of peanut allergy [s3]. Rather than a single rule, the addendum provides three separate guidelines for infants at different risk levels, covering how to define risk, when to test, and how and when to introduce peanut-containing foods, whether in a clinician's office or at home [s3]. Its conclusion is explicit: guidelines were developed for the early introduction of peanut-containing foods into infants' diets across risk levels [s3]. This is close to a full reversal of the earlier avoidance-based advice.

What is settled and what is not

What the trials establish is strongest and clearest for peanut in high-risk infants: early, regular exposure lowers the allergy rate substantially, and the LEAP data are the anchor for that claim [s1]. For the broader population and for the full menu of allergenic foods, EAT shows the biological effect is plausibly real but that adherence — getting a small infant to reliably eat the food — is the practical bottleneck, and the intention-to-treat result did not reach significance [s2].

What the evidence does not do is set a single universal age or amount that fits every child. The risk categories, the role of testing before introduction, and the timing are exactly the individualised judgments the NIAID addendum reserves for a clinician and family together, particularly for an infant with severe eczema or a known egg allergy [s3].

This article is informational and is not medical advice. For an infant with eczema, a known food allergy, or a strong family history, decisions about how and when to introduce allergenic foods should be made with a clinician who can assess that child's risk.