Key points

  • Maternal supplementation of omega-3 polyunsaturated fatty acids may reduce risk of atopic dermatitis (AD).

  • Fish intake during early childhood may decrease the risk of developing AD.

  • Early oral introduction of potentially allergenic foods in combination with improved skin barrier function may decrease the risk of developing food allergies in patients with AD.

  • Food allergy is common in patients with AD but usually does not exacerbate the condition directly.

  • There is favorable evidence for vitamin D supplementation as an adjunctive treatment in AD.


How does what we eat influence atopic dermatitis (AD)? Hundreds of studies of variable quality have attempted to answer this question. It is unclear whether the effects of diet on both AD risk and existing AD are mediated by pro- or antiinflammatory properties, skin barrier development, immune regulation, modulatory effects on the gut microbiome, or a combination of these. What is clear is that patients are gravitating toward more “natural” treatments or finding the “root cause” of AD, and dietary influences are one of the first places they may start.

The true relationship between nutrition and AD is difficult to tease out. One main reason is that high-quality dietary studies are difficult to perform, often resulting in inconclusive or inconsistent data. Observational studies rely heavily on patients’ reports of what foods they consumed and when, rather than direct observation of dietary intake. Depending on timing of retrospective studies, they may ask questions about food intake months prior to data collection, resulting in recall bias due to omissions or inaccurate recollections. These studies also rely on patients having sound knowledge of food ingredients and portion sizes, and they often assume that diets do not vary significantly during the study period. Questionnaires are limited by how specific the questions may be (e.g., “fatty fish” vs. the specific type of fish). Other interventional trials may exhibit bias or lack randomization. Even randomized controlled trials (RCTs) can have difficulty with blinding participants, patient adherence to dietary recommendations, and other methodologic challenges.

Other difficulties stem from the distinct phenotypes seen in AD. Often the questions to diagnose a child with AD are very broad. For example, “Has your child ever been diagnosed with eczema?” and “Has your child ever had a dry, itchy skin rash?” likely include many patients without AD. Eczema is often used as a catch-all term for any dry, itchy skin or scaly rash among patients and health care professionals alike. Even among those who are accurately classified as AD patients, there is variation in timing of onset, severity, responsiveness to therapy, associated comorbidities, and persistence over time. It is nearly impossible to control for this variability in large-scale population-based or cohort studies, though some have identified distinct AD phenotypes ( ).

Here we attempt to summarize the highest-quality evidence currently available regarding nutrition and AD. The chapter is divided in two sections: how diet may affect the risk of an individual developing AD, and how diet may affect patients who have already developed AD.

Diet and the risk of AD development

Intrapartum maternal diet

Western diets are typically high in proinflammatory omega-6 polyunsaturated fatty acid (PUFA) and low in antiinflammatory omega-3 PUFA. This may contribute to the higher incidence of AD in developed countries. One German study of over 2600 children found that maternal intake of margarine and vegetable oils (high in omega-6 PUFAs) in the last 4 weeks of pregnancy increased risk of eczema during the first 2 years of life, while high maternal fish intake (high in omega-3 PUFAs) decreased this risk ( ). A systematic review examined eight observational studies and four RCTs evaluating maternal intake of omega-3 PUFAs and its relationship to eczema in offspring. Half of the observational studies and half of the RCTs showed a protective effect of high omega-3 PUFAs/fish intake ( ). A meta-analysis of the RCTs examining patients with eczema and a positive skin prick test (SPT) of any type showed a statistically significant decreased relative risk of 0.53 with maternal omega-3 PUFA supplementation. This relationship disappeared when all cases of eczema were included ( ), suggesting this protection may only extend to specific AD phenotypes. A separate meta-analysis failed to show a relationship between omega-3 PUFA supplementation and AD ( ). A pooled analysis of 10 studies examining maternal fish intake only (rather than omega-3 PUFAs) failed to show any relationship with AD risk ( ).

Maternal vitamin intake may also have some impact. Increased maternal β-carotene consumption may result in lower AD risk at 16 to 24 months ( ). One study found that higher maternal folate intake from vitamin supplements, but not dietary folate, was associated with increased AD risk. This was unrelated to maternal serum folate levels ( ). Higher late pregnancy maternal nicotinamide levels were associated with decreased AD risk at 12 months of age, but not at 6 months ( ). Though nicotinamide is derived from dietary intake of vitamin B 3 and tryptophan, it remains unclear what diet is appropriate to achieve higher maternal nicotinamide levels, as additional nutrients and chemicals are related to this metabolic pathway ( ). There is no evidence that direct supplementation of vitamin B 3 or tryptophan in pregnancy decreases AD risk. A systematic review and meta-analysis showed inconclusive evidence for any effect of other maternal vitamins or zinc intake on infant AD risk ( ).

The effect of varying types of carbohydrate consumption remains unclear. The United Kingdom (UK)–based Avon Longitudinal Study of Parents and Children showed no relationship between eczema and maternal free/added sugar intake ( ). A multivariable analysis of over 600 mother-infant pairs demonstrated increased rates of AD in high-risk infants (first-degree family member with allergic disease) at 12 months of age in those whose mothers consumed higher levels of resistant starch and fiber from green vegetables in late pregnancy ( ). Finally, while several studies have not shown any relationship to maternal fruit and vegetable intake and AD, one study showed decreased AD risk at 16 to 24 months with increasing maternal consumption of citrus fruits ( ).

Both, a systematic review of 11 interventional studies and a Cochrane review failed to show a benefit to maternal dietary restriction of common food allergens during pregnancy ( ). Two separate studies showed no effect of maternal Mediterranean diet on preventing the development of AD in offspring ( ). One Japanese study of maternal intake of natto, a fermented soybean food, showed that AD risk at 6 months of age decreased with daily consumption ( ).

Breastfeeding and infant formula

The early infant diet is made up exclusively of human breast milk and/or an infant formula, with introduction of solid foods at around 4 to 6 months of age. Breast milk has beneficial effects on a child’s overall development and immune system, and may modulate risk of atopy later in life ( ). The effect of breastfeeding on AD remains controversial. One meta-analysis of 42 studies showed some low-quality evidence of decreased AD risk under 2 years of age with exclusive breastfeeding for 3 to 4 months ( ). A more recent meta-analysis of 27 prospective cohort studies showed no definitive relationship between either exclusive breastfeeding or total breastfeeding (any amount, whether combined or exclusive) and AD risk ( ). Family history of atopy appeared to influence the effect of exclusive breastfeeding, with decreased AD risk in those with a family history of atopy, and increased risk in those without ( ). Yet another study in a UK cohort of over 14,000 children showed that breastfeeding for 1 to 6 weeks or more than 6 months actually increased AD risk at 5 years of age ( ). Given conflicting reports and significant study heterogeneity, no recommendation can be made regarding breastfeeding as a preventive measure for AD at this time. A Cochrane review showed no benefit to maternal restriction of common food allergens while breastfeeding in reducing AD risk. The authors additionally point out that mothers on restricted diets gain less weight and may actually be putting themselves and their infants at risk of poor nutrition ( ).

Standard infant formula is cow’s milk based. Hydrolyzed formulas are composed of cow’s milk proteins casein and whey broken down to varying degrees. The German Infant Nutritional Intervention (GINI) study compared intervention and nonintervention arms (GINIplus) for high-risk infants. It showed significantly less AD up to 6 and possibly 10 years of age when using hydrolyzed formula in the first 4 to 6 months of life (both with and instead of breast milk) compared to cow’s milk formula ( ). Another study of high-risk nonexclusively breastfed infants in Singapore demonstrated decreased AD with partially hydrolyzed formulas ( ).

However, a Cochrane review found no evidence for hydrolyzed formulas preventing AD, either in comparison to exclusive breastfeeding or cow’s milk formulas ( ). A more recent birth cohort from France showed increased AD in high-risk infants who received partially hydrolyzed formula instead of nonhydrolyzed formula ( ). Based on these more recent studies, the recommendation to give hydrolyzed formula to high-risk infants if exclusive breastfeeding is not possible may need to be reconsidered.

Other types of formulas include amino acid–based (also known as elemental formula) and soy formulas. Amino acid–based formulas are not recommended for patients with AD in the absence of a documented cow’s milk protein allergy. Based on a Cochrane review, using soy formula instead of cow’s milk formula is not recommended to prevent food allergy or AD in at-risk infants ( ).

Dietary risk factors

There is concern that the Western diet (i.e., high in refined carbohydrates, saturated fat, and red meat, and low in fruits and vegetables) may increase AD risk ( ). Consumption of fast food three or more times per week was associated with an increased risk of severe eczema, whereas consumption of fresh fruits (once or twice per week for adolescents and three or more times per week for children) was protective against AD in an international study ( ). AD is less prevalent in areas with higher consumption of fish, vegetables, and cereal- and nut-based protein ( ). Rice, kimchi, and coffee decreased AD risk in Korean adults in another survey study of 17,497 participants, whereas AD risk increased with consumption of meats and processed foods such as instant noodles ( ).

Incorporating fish into the diet with the introduction of solid food may decrease AD risk. Eating fish at least once weekly decreased the odds of developing eczema by 28% by the age of 6 years in a questionnaire-based study of 4264 mother-infant pairs ( ). A meta-analysis also showed decreased AD risk with fish intake during the first 12 months of life ( ).

Though dairy has been reported as a trigger food for AD ( ), some evidence indicates that dairy products such as milk may actually be protective against AD. Consumption of milk and butter, as well as fresh fruit, pulses (beans, lentils, and peas), and rice three or more times per week decreased AD risk among 3209 children 6 to 7 years of age ( ). In a cross-sectional survey of 6- and 7-year-olds in Spain, milk, butter, and nuts were associated with a decreased AD risk when consumed three or more times per week ( ).

Avoidance of gluten is a commonly employed dietary tactic to minimize AD risk that is lacking in evidence. Several dietary associations have been drawn regarding AD from food frequency questionnaires. As part of the Nurses’ Health Study II (NHS-II), a prospective cohort study of women 25 to 42 years of age, 63,443 responses were analyzed for AD risk factors. Neither gluten intake nor a proinflammatory diet increased AD risk ( ).

Diet and existing AD

Skin barrier and food sensitization

There is an intimate relationship between AD and food allergy. While food allergies are more common in patients with AD, a common misconception is that food allergy frequently causes or exacerbates AD. In fact, the so-called atopic march starts with eczema, with an estimated 30% of patients subsequently developing food allergies, and only rarely do food allergies exacerbate AD ( ). There has been a recent international shift in dietary recommendations for avoidance of food allergy. Previously, parents were advised to consider withholding highly allergenic foods (e.g., peanuts, eggs) during early life to prevent allergy in high-risk children. However, the Learning Early About Peanut Allergy (LEAP) study, which showed that early introduction of peanuts decreased incidence of peanut allergy, has shifted the focus toward early introduction of allergenic foods to promote tolerance in both high- and low-risk infants ( ).

Mouse models were the first to show that removal of the stratum corneum followed by antigen exposure induces a Th2 inflammatory response and migration of sensitized Langerhans cell ( ). Additionally, in vitro human keratinocytes have shown increased epidermal permeability, decreased filaggrin expression, and improved integrity of the stratum corneum in the presence of histamine. These immune responses are important parts of the dual-allergen exposure hypothesis, which speculates that low-dose environmental exposure of a disrupted skin barrier to food allergens can lead to allergy (also known as transcutaneous sensitization), while high-dose exposure orally to those same allergens can induce tolerance ( ). If the stratum corneum is damaged by trauma, such as with scratching induced by active AD, histamine can increase the risk of transcutaneous sensitization ( ). Early-life environmental peanut exposure also has a higher risk of peanut allergy in patients with filaggrin mutations than those without, independent of AD ( ).

The relative balance of cutaneous versus oral exposure may ultimately be what determines food allergy development. This hypothesis could explain why eczema developed before 6 months of age (when oral solid food intake may be more restricted) is associated with higher risk of food allergy than eczema developed after 6 months of age ( ), and why patients with more severe AD and/or who apply peanut-containing topical products are more likely to develop peanut allergy ( ). The dual-allergen exposure hypothesis is changing the way that physicians approach food allergy, encouraging early oral exposure of potential culprit foods to promote tolerance. The way we approach AD treatment is also changing. Early and aggressive management of infantile AD with topical steroids and emollients to restore the skin barrier may have a lasting effect on patients and their ultimate development of food allergies.

Food allergy and sensitivity

Patients with AD are more likely to show sensitization to multiple foods ( ). Both patients and physicians commonly perceive that diet affects severity of AD as well as frequency of flares. The study of food-triggered AD has produced a vast array of literature, and it remains a somewhat controversial topic. Note that just because a patient has both AD and food allergy does not mean that exposure to their allergen will impact their AD. The two conditions can and frequently do coexist independently.

Immediate allergic reactions are what patients generally consider as food allergies and are mediated by food-specific immunoglobulin E (IgE). These reactions typically occur within minutes to a few hours of food ingestion and may cause urticaria, gastrointestinal (GI) symptoms, wheezing, and, at their most severe, anaphylaxis. AD can be indirectly aggravated by these immediate reactions with pruritus and subsequent scratching. However, direct exacerbation of AD is likely the result of a delayed eczematous reaction. This is an eczema flare due to a delayed hypersensitivity mediated by T cells rather than IgE. These occur 6 to 48 hours after exposure. Patients may not have immediate reactions to the same food, making a history of food-specific exacerbation unreliable.

Systemic allergic contact dermatitis is a type of delayed hypersensitivity to specific allergens found in a variety of foods, rather than specific foods themselves. This can be difficult to diagnose, as it may be mistaken for or superimposed on AD. Common culprits include nickel (beans, lentils, whole wheat bread, canned products, shellfish) and balsam of Peru (citrus fruits, tomatoes, chocolate, colas, cinnamon, and vanilla) ( ).

Food allergy testing

Food allergy testing in AD has unique challenges and high false-positive rates. The skin of AD patients is more reactive, and overall serum IgE levels can be elevated. For this reason antigen-specific IgE (sIgE) testing is preferred over radioallergosorbent testing (RAST) or SPT for diagnosing immediate allergic reactions, though all have relatively low positive predictive values in AD patients. SPT and sIgE are most helpful if they are negative, as these tests have a high negative predictive value in patients with AD and effectively exclude a food allergy ( ). In contrast, late reactions may be missed with IgE-specific testing alone, and even if this testing is positive, it has a low specificity for these types of reactions ( ). A meta-analysis examining atopy patch testing (APT) in the detection of food allergy in patients with AD found a pooled sensitivity of 53.6% and specificity of 88.6% ( ). Double-blind, placebo-controlled food challenges are considered the gold standard, but these are often expensive, time consuming, and may require medical intervention for the patient in the event of a severe reaction.


Hen’s egg allergy accounts for up to 70% of food allergies in AD patients, and up to 40% of patients with egg allergy will have AD ( ). The two proteins typically responsible for egg allergy are ovomucoid protein (most common, heat stable) and ovalbumin (heat labile), with the majority of egg allergens found in egg whites ( ). Approximately 85% of patients will have developed complete tolerance by preschool age ( ). The rate of reduction in egg white sIgE may predict which patients will achieve tolerance ( ). Egg-sensitive children have more allergic disease, including higher rates of severe and persistent AD, as well as concomitant asthma and rhinoconjunctivitis ( ). In infants with a suspected egg allergy and positive egg sIgE, there is some evidence supporting an egg-free diet to improve AD ( ). Though commonly practiced, there is little evidence to support egg exclusion in AD patients without suspected or documented egg allergy.


Peanut allergy is found in 6.6% of all patients with infantile AD ( ). Persistence of AD correlates with presence of peanut allergy, and though tolerance with age has been reported, it is uncommon and does not occur in patients with severe food allergy symptoms at onset ( ). Introducing peanuts early to children with AD and/or egg allergy decreases development of peanut allergy ( ). US guidelines recommend introduction of peanuts at 6 months of age for patients with mild to moderate AD, and consideration of preexposure testing in addition to early introduction for those with severe AD ( ). Since refined peanut oils remove and/or denature the allergenic proteins, they are generally safe for use in patients allergic to peanuts ( ).

Cow’s milk

Cow’s milk allergy is common in children with AD and may account for up to half of food allergies in this population ( ). Allergy to cow’s milk can cause both early and late skin reactions and may affect the skin alone without involvement of the GI tract ( ). Reactions in children with AD allergic to cow’s milk are more likely to involve urticaria than those without ( ). Cow’s milk allergy is often outgrown by late childhood, and presence or absence of AD does not impact timing or likelihood of achieving tolerance ( ). There is very little evidence regarding other nonhuman sources of milk in cow’s milk–allergic children with AD.

Wheat and gluten

Wheat allergy usually develops in infancy and can be associated with gluten as well as other various allergenic proteins. Wheat can cause both immediate and delayed reactions in AD patients ( ). Both types of reactions often occur without other systemic symptoms such as vomiting, diarrhea, rhinitis, or stridor ( ) and are considered a separate entity from celiac disease or nonceliac gluten sensitivity ( ). Children with wheat allergy often become tolerant by 3 to 5 years of age. A study of AD patients 14 years of age or older showed that AD was exacerbated by wheat in about 5% of patients and improves with an elimination diet ( ). In wheat-allergic AD patients, IgE antibodies to gliadin are typically seen in SPTs, RAST, and immunoblotting ( ), though as mentioned previously these are not the tests of choice for diagnosing food allergy in patients with AD.


Soy protein, derived from soybeans, is a popular nonmeat protein substitute found in many processed foods. Soy allergy in children with AD usually causes late or eczematous reactions, although rarely immediate reactions can be seen ( ). About 27% of AD patients may be sensitized to soy without any clinical relevance ( ). Tolerance typically develops by adolescence. However, older patients can develop a “late” soy allergy that may be associated with birch pollen or peanut allergy ( ).

Birch pollen–related foods

As part of the atopic march, allergic rhinoconjunctivitis often occurs in patients with AD. Allergy to birch pollen can result in oral allergy syndrome and late eczematous reactions due to cross-reactivity of birch pollen sIgE with proteins in a variety of fruits and vegetables, including celery, carrots, hazelnut, apples, and kiwi. This phenomenon plays an important role in areas where birch trees are native, including North America and Northern and Central Europe. Foods that were previously tolerated can suddenly lead to both immediate and late reactions in AD patients of all ages once sensitized to birch pollen ( ).

Elimination diets and other dietary interventions

Several other types of diets have been explored in AD. Elemental diets with an amino acid–based formula for 6 weeks were not beneficial for children under 3 years of age ( ). Elimination diets in which patients only consumed five to eight different foods in addition to whey or casein protein supplementation are not effective, and adherence to these strict diets is quite poor ( ).

In one survey of 169 adult AD patients, 87% had tried eliminating specific foods from their diet ( ). Multiple foods were reported to improve or exacerbate AD, both by their intake and by their elimination ( ) ( Table 8.1 ). Evidence does not support elimination diets in the management of AD, and in fact, strict elimination diets may increase the risk for nutrient deficiencies and obtunded growth and development ( ). Malnutrition can also retard skin barrier repair. Removal of specific foods based only on testing may also be harmful. In one study, 13% of those who eliminate foods they previously ate without reaction based on positive SPTs or sIgE levels subsequently developed intolerance and failed an oral food challenge ( ). Intolerance developed as early as 3 months after food elimination ( ). When special diets are attempted, duration should be limited to 4 to 8 weeks to evaluate the effect of dietary interventions before resuming a normal diet ( ).

Table 8.1

Patient-reported Exacerbating and alleviating foods in 169 patients with AD

Adapted from Nosrati, A., Afifi, L., Danesh, M. J., Lee, K., Yan, D., Beroukhim, K., et al. (2017). Dietary modifications in atopic dermatitis: Patient-reported outcomes. Journal of Dermatological Treatment , 28 (6), 523–538. .

Food %
Exacerbating foods
Dairy 24.8
Gluten 18.3
Alcohol 17.1
Sugar 16.5
Tomato 13.0
Citrus 10.0
Eggs 7.1
Meat, soda, spicy foods, processed foods, seafood <5.0
Improved when removed
White flour products 53.6
Gluten 51.4
Nightshades 51.4
Junk foods 51
Alcohol 50.8
Dairy 50.7
High fat foods 43.6
Tobacco 40.4
Shellfish 36.8
Caffeine 33.3
Pork 32.2
Red meat 24.1
Sodium/salt 23.2
Alleviating foods
Dietary supplements a 25.4
Water 14.2
Vegetables 11.8
Oils b 9.5
Fruits 6.5
Fish 1.7
Improved when added
Vegetables 47.6
Organic foods 39.5
Fish oil/omega-3 35.0
Fruits 34.6
Vitamin D 34.4
Probiotics 28.7

a Probiotics, vitamin D, vitamin C, zinc, or omega-3.

b Primrose, hempseed, cod liver, and coconut oils.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Jul 23, 2022 | Posted by in GENERAL & FAMILY MEDICINE | Comments Off on Nutrition

Full access? Get Clinical Tree

Get Clinical Tree app for offline access