This Week in Allergy & Immunology — Sep 20, 2026

Welcome to This Week in Allergy and Immunology. This week we're covering 10 notable papers spanning the early-life origins and prevention of food allergy, the diagnosis and immunotherapy of nut allergy, and a cluster of advances in inborn errors of immunity — plus real-world biologic data in severe asthma and a provocative position paper on bacterial allergens. Let's dive in.

We start with prevention, and with a large, clean negative trial in the New England Journal of Medicine. The PrEggNut investigators, led by Palmer, randomised more than two thousand pregnant women whose unborn child had at least two close family members with allergic disease, assigning them either to a diet rich in egg and peanut — at least six eggs and sixty peanuts per week — or to a standard diet capped at three eggs and thirty peanuts, from before twenty-three weeks' gestation through four months of lactation [1]. At one year, IgE-mediated egg or peanut allergy affected about eight percent of infants in each arm, with no significant difference between the groups and no safety signal either way. So maternal allergen loading during pregnancy and breastfeeding does not prevent food allergy in high-risk infants. The practical message is one of reassurance rather than prescription: we continue to tell pregnant women not to restrict egg or peanut, but we should stop implying that eating more of them will protect the baby. The prevention window that matters remains the infant's own diet.

Two papers in the Journal of Allergy and Clinical Immunology In Practice sharpen where else that window might lie. Nilsson and colleagues followed 563 high-risk infants born at term with prospectively collected feeding data from the first postnatal days [2]. Just over a fifth of infants received commercial formula within twenty-four hours of birth, and those infants had roughly three and a half times the odds of IgE-mediated egg allergy at one year, and about two and a half times the odds of egg sensitisation — despite timely egg introduction at six months. No association was seen for other foods, though the case numbers were small. The proposed mechanism is displaced colostrum and its effect on the neonatal gut microbiome and immune development. This is observational, from a single high-risk cohort, and the authors are explicit that randomised confirmation is needed. But it is a reasonable addition to the antenatal conversation with allergy-prone families: if formula supplementation in the first day is avoidable, avoid it.

In the same journal, Bodén and colleagues used the NorthPop birth cohort in Sweden, nearly four and a half thousand children, to ask whether where infant eczema sits on the body predicts food allergy [3]. Parent-reported eczema distribution at nine months was related to physician-diagnosed food allergy and to measured sensitisation at eighteen months. Eczema confined to the trunk and limbs carried no increased risk. Eczema on the head, neck or hands — the sites in constant contact with food allergens — roughly doubled the odds of food allergy, and widespread disease involving both those sites and the body raised the odds more than fourfold. That maps neatly onto the dual-allergen-exposure hypothesis, and it gives you a quick bedside risk stratifier at the nine-month check: an infant with facial and hand eczema, particularly if it is widespread, deserves a more deliberate conversation about early allergen introduction and closer follow-up.

Moving from prevention to diagnosis, and to what may be the most immediately actionable paper of the week. Brettig and colleagues, again in the Journal of Allergy and Clinical Immunology In Practice, report the Australian ADAPT programme — the first nationally standardised model of care for infant peanut oral immunotherapy, across ten public hospital sites [4]. Infants under twelve months with a prior mild-to-moderate reaction to peanut plus documented sensitisation underwent a threshold oral food challenge, seven escalating doses from 15 up to 1,300 milligrams of peanut protein, to both confirm the diagnosis and set the starting immunotherapy dose. Of 778 infants challenged, about two thirds reacted, with a median eliciting dose of 120 milligrams. The striking figure is the other third: close to thirty-five percent OF INFANTS had a negative challenge despite a convincing history and positive testing, and were delabelled on the spot. Skin prick and Ara h 2 results were higher in the reactors but predicted neither threshold nor severity. Anaphylaxis occurred in about twelve percent of positive challenges — roughly eight percent of all challenges — and all events were managed safely. Among those who went on to immunotherapy, nearly all started below their measured threshold. The conclusion is hard to argue with: in infant peanut allergy, the challenge earns its place whether your plan is immunotherapy or avoidance, because a third of these children do not have the disease you are about to treat them for.

Staying with immunotherapy, the Nut CRACKER study from Elizur and colleagues, also in the Journal of Allergy and Clinical Immunology In Practice, addresses a question that comes up constantly in tree nut clinics: does treating one nut cover another [5]. In a prospective cohort of walnut-and-hazelnut co-allergic patients aged four and up, hazelnut cross-desensitisation occurred in just over half OF PATIENTS who started walnut oral immunotherapy, and in about two thirds OF THOSE who reached walnut desensitisation and were then challenged to hazelnut — against a natural resolution rate of roughly one in six in observational controls. Immunologic markers moved accordingly, with falling hazelnut sensitisation and basophil reactivity and rising specific IgG4. Persistent hazelnut allergy was predicted by a larger baseline hazelnut skin prick wheal, a baseline reaction dose of sixty milligrams or less, and higher IgE to Cor a 14. So for most co-allergic patients, a single walnut protocol may spare a second course — and component testing at baseline helps you tell the patient in advance whether that is likely.

Turning to immunodeficiency, where the Journal of Allergy and Clinical Immunology carries three papers worth your attention. ABACHAI, from Krausz and colleagues, is the first prospective interventional trial of abatacept in CTLA-4 insufficiency and LRBA deficiency — twenty adults, eighteen with CTLA-4 insufficiency, given 125 milligrams subcutaneously weekly for twelve months [6]. The primary endpoint was safety defined by failed infection control, and severe infections occurred in three patients at a rate comparable to the year before enrolment, with no EBV or CMV reactivation. Lymphoproliferation and immune activation fell, and quality of life and gastrointestinal symptoms improved. Importantly, cytopenias and renal involvement were not adequately controlled, and a fifth of patients discontinued, mostly for unrelated serious adverse events. This is a small, single-arm study, but it converts years of off-label use into prospective evidence, with an honest map of where abatacept does and does not work.

Alongside it, the proceedings of the First International KREC Consortium Meeting make the case for adding kappa-deleting recombination excision circles to newborn screening, so that infants with severe B-cell lymphopenia are found before recurrent infection or irreversible damage rather than after [7]. The priorities identified are health-economic justification, assay harmonisation, minimising false positives, and defined follow-up pathways — this is an implementation and advocacy document, not new outcome data, but it is the roadmap worth knowing if screening policy in your region is in play. And complementing it, Ljung Sass and colleagues report a systematic review and meta-analysis of paediatric CD19-positive B-cell reference values, pooling twenty-nine publications and more than ten thousand healthy children [8]. They propose five distinct age bands with corresponding lower limits of normal, and confirm that children under six differ consistently from adults. With CD19-directed therapies now reaching paediatric patients, having age-stratified normals to judge B-cell recovery against is a genuinely useful piece of infrastructure.

Two final papers. In Annals of Allergy, Asthma and Immunology, DeMartino and colleagues report CROSSROADS-3, a self-controlled claims analysis of 321 United States patients who started tezepelumab in the first nine months after approval [9]. Nearly half had been on another biologic in the preceding year, and among those with data available, about two thirds had blood eosinophils under 300. Annualised exacerbations fell by roughly sixty percent after initiation, cumulative oral corticosteroid exposure dropped by about forty percent, and exacerbation-related inpatient, emergency and outpatient visits all declined — with benefit in both biologic-naive and biologic-experienced patients. This is a single-arm, before-and-after design, so regression to the mean and channelling of sicker patients are real caveats, but the low-eosinophil and prior-biologic-failure signal is clinically useful. And in Allergy, an EAACI position paper led by Jappe catalogues sixty-six IgE-binding components across twenty bacterial species — staphylococcal superantigens, bacterial enzymes, Chlamydia components — against a database that currently lists just one official bacterial allergen [10]. Some may serve as markers of disease severity, and the authors argue that conditions we currently call non-allergic may eventually be reclassified.

If you only have time for one paper this week, make it the ADAPT programme evaluation in the Journal of Allergy and Clinical Immunology In Practice [4]. It tells you that roughly one in three infants you might have started on peanut immunotherapy on history and testing alone is not actually allergic — and that changes what you do at your next new-patient visit.

Here are the key takeaways from this week in Allergy and Immunology. First, loading maternal diet with egg and peanut during pregnancy and lactation does not prevent infant food allergy — reassure, but do not prescribe. Second, formula in the first twenty-four hours of life was associated with roughly triple the odds of egg allergy in a high-risk cohort, an observational finding worth raising antenatally. Third, infant eczema on the face, neck and hands, especially when widespread, marks substantially higher food allergy risk, while trunk-only eczema does not. Fourth, confirm infant peanut allergy with an oral food challenge before committing to immunotherapy or to lifelong avoidance. Fifth, walnut immunotherapy cross-desensitises most walnut-hazelnut co-allergic patients, and baseline hazelnut wheal size, threshold dose and Cor a 14 IgE tell you who will be the exception. And finally, abatacept now has prospective safety and efficacy data in CTLA-4 insufficiency and LRBA deficiency, with clear limits around cytopenias and renal disease.

That's your roundup for This Week in Allergy and Immunology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And think of one colleague — in any specialty — who never has time to keep up with the literature. Tell them about AudioScholar: a free ten-minute weekly for every specialty, to listen to in any podcast app, or to read at audioscholar dot C C.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.