Eczema in a baby is not evidence of a food allergy, and UK guidance does not treat the two as one question. NICE sets out when food allergy should be suspected in a child with atopic eczema, one time-limited formula trial for bottle-fed babies under six months, and an eight-week limit on a cow's milk-free diet before specialist dietary advice is required. The NHS puts it plainly to parents: "do not change your diet unless you are told to by a doctor (this includes if you are breastfeeding a baby with eczema)" (NHS, Atopic eczema).
Advice on allergenic foods has also moved in the opposite direction from what parents were told a decade ago, on the strength of two UK trials, LEAP and EAT. Both screened and supervised their participants in ways a kitchen cannot reproduce, which is why every decision below has a named person attached: a GP, a health visitor, a dietitian or an allergy clinic.
When food allergy is suspected at all
NICE CG57 asks clinicians to "consider a diagnosis of food allergy in: children with atopic eczema who have had immediate symptoms from eating a particular food; babies and young children with moderate or severe atopic eczema that has not been controlled by optimum management, particularly if associated with gut dysmotility (colic, vomiting, altered bowel habit) or failure to thrive" (NICE CG57, 1.4.1.2).
The second trigger is narrower than it looks. It is not eczema by itself but moderate or severe eczema that optimum management has failed to control, usually alongside a gut symptom or poor weight gain. Care that has been intermittent rather than optimum has not tested anything yet, and what optimum looks like in the first year is covered in baby eczema in the first year.
For milder disease the guideline points the other way, asking clinicians to "reassure children with mild atopic eczema and their parents or carers that most children with mild atopic eczema do not need to have tests for allergies" (1.4.1.5). On tests sold direct to parents it is blunter: "advise children with atopic eczema and their parents or carers not to use high street or internet allergy tests, because there is no evidence of their value in managing atopic eczema" (1.4.1.6). Skin prick and specific IgE testing are separate things, arranged and read by a clinician, and they follow the suspicion rather than replacing it.
The formula trial, and the eight-week limit
For one group of babies the guideline names a trial with a start and an end. "For bottle-fed babies aged under 6 months with moderate or severe atopic eczema that has not been controlled by optimal treatment with emollients and mild topical corticosteroids, offer a 6- to 8-week trial of an extensively hydrolysed protein formula or amino acid formula in place of cow's milk formula" (1.4.1.7).
Four conditions sit in that sentence: bottle-fed, under six months, moderate or severe disease, and an emollient and mild steroid routine that has already failed. The formula is offered by a clinician rather than picked off a shelf, and six to eight weeks is the whole of it.
What follows is equally specific. "Refer children with atopic eczema for specialist dietary advice if they have followed a cow's milk-free diet for longer than 8 weeks" (1.4.1.8). That is a referral trigger, not a signal to carry on alone. Two further recommendations close off the substitutions parents make themselves: "do not use diets based on unmodified proteins of other species' milk (for example, goat's milk or sheep's milk) or partially hydrolysed formulas to manage suspected cow's milk allergy in children with atopic eczema" (1.4.1.9), and "offer diets that include soya protein along with specialist dietary advice for children aged 6 months or over" (1.4.1.10).
If the baby is breastfed
Here the guideline states the uncertainty rather than resolving it. "For children who are being breast fed, explain that it is not known whether changing the mother's diet will reduce the severity of the atopic eczema. If food allergy is strongly suspected, consider a trial of an allergen-specific exclusion diet" (1.4.1.11).
So a mother cutting dairy on the strength of a baby's eczema is acting on something the guideline says is not known, and where an exclusion is tried it is allergen-specific and prompted by a strong suspicion, not a broad removal of several foods. The NHS names breastfeeding mothers directly in its instruction not to change diet without a doctor's say-so.
Why the avoidance advice was reversed
LEAP randomised 640 infants aged at least 4 months and under 11 months, all of whom "had to have severe eczema, egg allergy, or both", to consume or avoid peanut until 60 months of age (Du Toit et al, 2015). The primary outcome was the proportion with peanut allergy at 60 months. Among infants with a negative skin prick test at entry, peanut allergy was present in 13.7% of the avoidance group and 1.9% of the consumption group. Among those with a positive test at entry, the figures were 35.3% and 10.6%.
EAT asked a different question of a different population: 1,303 exclusively breastfed infants from the general population, randomised at 3 months to early introduction of six allergenic foods (peanut, cooked egg, cow's milk, sesame, whitefish and wheat) or to exclusive breastfeeding to around six months, then UK practice (Perkin et al, 2016). The primary outcome was food allergy to one or more of the six foods between 1 year and 3 years of age. In the intention-to-treat analysis it developed in 7.1% of the standard-introduction group and 5.6% of the early-introduction group, which did not reach significance (P=0.32). Only in the per-protocol analysis, among families who managed the full schedule, was prevalence lower with early introduction, 2.4% against 7.3%. The trial's own summary is the useful line for a parent: "the early introduction of all six foods was not easily achieved but was safe".
Avoidance did not protect, and in high-risk infants it looked worse. The other prevention question parents ask, whether a supplement in pregnancy or infancy changes the odds, has its own thin evidence base in probiotics.
The condition that came with the reversal
LEAP did not license peanut at home. Every infant was screened by skin prick test before entry, and those with a wheal larger than 4 mm were excluded from the trial altogether. Infants randomised to consumption then underwent a baseline open-label food challenge, negative-test infants receiving 2 g of peanut protein and positive-test infants incremental doses. Infants who reacted at that challenge "were instructed to avoid peanuts", and stayed in the intention-to-treat analysis but not the per-protocol one. A positive result at the clinic changed what the family was told to do, before any peanut was eaten at home.
UK guidance carries that forward. The British Society for Allergy and Clinical Immunology states that "infants with moderate to severe eczema and/or eczema which began in the first 3 months of life are at greatest risk of reacting to egg and peanut", and that "it is recommended that parents and carers with infants with pre-existing food allergy and/or severe eczema speak to a healthcare professional for advice on allergen introduction" (BSACI). It adds that the benefit of testing such babies first "needs to be balanced against the risk this could cause a delay", so whether a test happens is a clinical judgement, not a fixed rule.
The NHS weaning page says the same in parental language. Where a baby already has eczema or a diagnosed food allergy, or there is a family history, the page says "talk to your GP or health visitor first", and describes introducing allergenic foods "one at a time so that you can spot any reaction" (NHS, Food allergies in babies and young children). It also warns that "delaying the introduction of peanut and hen's eggs beyond 6 to 12 months may increase the risk of developing an allergy to these foods". Allergy UK adds that parents "should be actively discouraged from offering allergenic foods merely to test for allergy if there is no intention to keep the food regularly in the diet".
Who decides, and in what order
The route starts in primary care. A GP or health visitor assesses whether the CG57 criteria are met, arranges the formula trial where the baby fits 1.4.1.7, and refers on: "refer children with moderate or severe atopic eczema and suspected food allergy for specialist investigation and management" (1.7.1.5). That sits apart from the general dermatology criteria, set out in when to see a dermatologist. A dietitian takes the nutritional side once a food is excluded beyond eight weeks, and an allergy clinic handles testing and any supervised introduction.
None of this maps onto a self-directed exclusion and reintroduction sequence. That approach exists, it has a method, and it is written for adults: see how to do an eczema elimination diet properly. Run at home for an infant, the same steps strip out the screening, supervision and nutritional oversight that make the evidence above hold.
References
- National Institute for Health and Care Excellence. Atopic eczema in under 12s: diagnosis and management. NICE Clinical Guideline 57. https://www.ncbi.nlm.nih.gov/books/NBK571009/
- NHS. Atopic eczema. https://www.nhs.uk/conditions/atopic-eczema/
- NHS. Food allergies in babies and young children. https://www.nhs.uk/conditions/baby/weaning-and-feeding/food-allergies-in-babies-and-young-children/
- Du Toit G, Roberts G, Sayre PH, et al. Randomized trial of peanut consumption in infants at risk for peanut allergy. New England Journal of Medicine, 2015, 372:803-813. https://doi.org/10.1056/NEJMoa1414850
- Perkin MR, Logan K, Tseng A, et al. Randomized trial of introduction of allergenic foods in breast-fed infants. New England Journal of Medicine, 2016, 374:1733-1743. https://doi.org/10.1056/NEJMoa1514210
- British Society for Allergy and Clinical Immunology. Early introduction of food allergens. https://www.bsaci.org/resources/allergy-management/food-allergy/allergy-prevention/early-introduction-of-food-allergens/
- Allergy UK. Identifying high-risk infants and reducing food allergy risk in primary care. https://www.allergyuk.org/hcp-resources/reducing-food-allergy-risk/