Design: multisite randomised controlled trial (PrEggNut)
Population: 2,137 pregnant women whose unborn child had at least two biological family members with medically diagnosed allergic disease
Primary outcome: IgE-mediated egg or peanut allergy in the infant at 1 year
Effect: 7.8% (83/1066) vs 8.4% (89/1064); RR 0.93 (95% CI 0.69-1.26; P = 0.65)

PrEggNut randomised 2,137 pregnant women whose unborn child had at least two biological family members with medically diagnosed allergic disease — a deliberately high-risk group — to a diet with at least six eggs and sixty peanuts a week, or no more than three eggs and thirty peanuts, from before 23 weeks' gestation until four months postnatally.

IgE-mediated egg or peanut allergy at one year occurred in 7.8% of infants in the high-intake group and 8.4% in the standard group. Safety measures were similar.

The hypothesis followed logically from what is already established: introducing peanut early to the infant reduces peanut allergy, so pushing exposure earlier still, through the placenta and breast milk, looked like a reasonable extension. It does not work that way. Whatever tolerance is being induced by early introduction appears to require the infant's own gut, not the mother's diet.

What this settles is a question asked in every antenatal and postnatal clinic. The answer is now evidence-based and symmetrical: no reason to increase these foods in pregnancy, and no reason to avoid them either. The prevention that does work still happens after birth, at weaning.

Tell women with a family history of allergy to eat normally in pregnancy and lactation
Keep the message that has evidence: introduce egg and peanut to the infant at weaning, not late
Record the family allergy history in the newborn notes so the weaning conversation happens on time
Do not let a null result for maternal loading be heard as support for maternal avoidance
Adapt the advice to how these foods are actually eaten at home rather than to trial quantities

Don't overread it

A null result for increasing maternal intake is not evidence for avoiding these foods in pregnancy; avoidance was never tested here and has no support.

The statistics, in plain English

A relative risk of 0.93 with a 95% confidence interval of 0.69 to 1.26 includes 1.0, and in a trial of more than 2,000 women that interval is narrow enough to rule out any large protective effect — this is a well-powered null result, not an inconclusive one. It does not exclude a small benefit: the lower bound of 0.69 leaves room for a relative reduction of up to about 31%, which against a background risk of 8% would be about 2.5 percentage points in absolute terms.