Epicutaneous sensitization in the development of food allergy: what is the evidence and how can this be prevented?

Abstract

There is increasing evidence regarding the importance of allergic sensitization through the skin. In this review, we provide an overview of the atopic march and immune mechanism underlying the sensitization and effector phase of food allergy. We present experimental models and human data that supports the concept of epicutaneous sensitization and how this forms one half of the dual allergen exposure hypothesis. We discuss specific important elements in the skin (FLG and other skin barrier gene mutations, Langerhans cells, Type 2 innate lymphoid cells, IL-33, TSLP) that have important roles in the development of allergic responses as well as the body of evidence on environmental allergen exposure and how this can sensitize an individual.

Given the link between skin barrier impairment, atopic dermatitis, food allergy, allergic asthma and allergic rhinitis, it is logical that restoring the skin barrier and prevention or treating atopic dermatitis, would have beneficial effects on prevention of related allergic diseases, particularly food allergy. We present the experimental and human studies that have evaluated this approach and discuss various factors which may influence the success of these approaches, such as the type of emollient chosen for the intervention, the role of managing skin inflammation and differences between primary and secondary prevention of atopic dermatitis to achieve the desired outcome.

INTRODUCTION

Allergic diseases are the most common chronic diseases in Westernized countries and bear a substantial health and socio-economic burden. Atopic dermatitis (AD) affects approximately 20% to 30% of children.^1–3 Asthma affects 9–18%^{4, 5} and allergic rhinitis (AR) affects up to 40% of children.⁶ Food allergy (FA), is an epidemic among children in westernized countries,^7–9 and significantly impairs quality of life.^{10, 11} Between 15–40% of children with FA have experienced a severe, life-threatening allergic reaction and 30% report allergies to multiple foods.^{12, 13} The rate of confirmed FA is much higher in children with AD.^14–16 FA is expensive - resulting in $24.8 billion dollars/year in expenditures to the healthcare system and US families,¹⁷ and in Europe mean annual household costs were found to be €791 higher amongst households with a food-allergic child than those without food-allergic children.¹⁸

PATHOGENESIS OF FOOD ALLERGY

The genesis of FA is a complex process, influenced by genes, host immune responses, epithelial function and environment factors. Increased antibiotic use, non-vaginal births, ultra-sanitary lifestyles, less time spent outdoors and the resulting “modern” microbial community structures of the gut and the skin, beginning at birth, have been implicated in aberrant immune system maturation, and development of atopy, including FA (Figure 1).¹⁹ For years, it was thought that children became sensitized to food allergens by exposure through the gut (via breastfeeding or early consumption);²⁰ there is now increasing evidence that early life allergen exposure through the skin causes FA (Figure 2),^{21, 22} whereas, early oral exposure causes tolerance (known as the dual allergen exposure hypothesis: Figure 3). In the last 5 years, several studies have demonstrated allergen specific oral tolerance induction to allergenic foods in high risk children.^23–27

Figure 1:

Factors leading to skin barrier dysfunction: The skin is constantly exposed to environmental factors, both natural (eg, bacteria, viruses, fungi, food and aero allergens) and man-made (e.g., detergents, high pH creams and lotions). In those genetically predisposed to allergic disease (e.g., filaggrin, SPINK5, and loricrin mutations), these factors lead to skin barrier dysfunction, epicutaneous damage, and allergic sensitization.

Figure 2:

Pathways for skin barrier dysfunction and eczema leading to food allergy: Food allergy is manifested through the skin (1) the initial step towards food allergy is skin barrier impairment caused by environmental pollutants, detergents, infections, and genetics. (2) Skin barrier impairment leads to skin inflammation and clinical atopic dermatitis. (3) Exposure of allergens through skin that has an impaired barrier (dry) or has clinical atopic dermatitis leads to sensitization and food allergy.

Figure 3:

Dual-allergen Exposure: Increasing evidence suggests that early life allergen exposure through the skin causes T cell deviation towards a Th2 allergenic type and subsequent food allergy whereas early oral exposure causes T cell deviation towards tolerogenic Th1 and Treg subtypes (dual allergen exposure hypothesis).

LIMITATIONS OF ORAL TOLERANCE INDUCTION

The effect of oral tolerance induction appears to be allergen specific. For example, in the LEAP study, early peanut consumption did not lead to a prevention of tree nut or sesame seed allergy.²⁸ Since FA develops early in life,²⁹ there is a narrow window of opportunity to induce tolerance with oral introduction of multiple foods. In the HealthNuts population based study, 3.1% of children already had challenge proven peanut allergy (PA) by 12 months of age, and 9% were egg allergic.¹ In addition, introduction of multiple allergenic foods into the diet of young infants is challenging. In the EAT study, the adherence rate for introducing 6 allergenic foods into the diet was only 42%.³⁰ Thus, there is a need for an alternative approach to prevent FA.

There is evidence that diet diversity in infancy reduces food sensitization and allergic asthma (AA), but results for AD and AR are mixed.³¹ A systematic review found that breastfeeding was protective for AA; however, the evidence for AD and AR was weaker with no effect on FA.³² Other methods such as Vitamin D supplementation has produced mixed results.^{33, 34} There is little evidence for the role of prebiotics, and whilst meta-analyses and 2 individual studies show some benefit from probiotics, these studies have issues such as objective evaluation of outcomes, design and blinding.^{35, 36}

ATOPIC MARCH

AD is often associated with the subsequent development of allergic disease, including FA, AA, or AR, as characterized by the concept of the ‘atopic march’.^37–39 In the ALSPAC study, an oozing, crusting skin rash was an independent risk factor for the development of PA.⁴⁰ Furthermore, early onset (particularly within the first 3 months), severe AD markedly increases the risk of FA.^{16, 41, 42} In the HealthNuts study, children with early onset severe AD had a 50% rate of challenge proven egg, peanut or sesame seed allergy by 12 months of age. In the LEAP screening study,⁴³ there was a dose dependent increase in food sensitizations with increasing SCORing Atopic Dermatitis (SCORAD) levels in children between 4–11 months of age. Duration of infantile AD prior to inducing remission with proactive topical steroid treatment has also been shown to increase the risk of FA for each month that passes.⁴⁴ This evidence suggests that shortening the duration of AD, by reducing inflammation, reduces the opportunity for epicutaneous exposure to environmental food allergens thereby preventing sensitization and subsequent FA.

THE IMMUNE RESPONSE BEHIND FOOD ALLERGY

A dysregulated immune system involving a Type 2 T helper (Th2) cell mediated inflammatory response is associated with FA.^{45, 46} This involves the formation of specific IgE antibodies against environmental and food allergens, as well as the production of several cytokines, including interleukin (IL)-3, IL-4, IL-5, IL-9, IL-13, IL-25, IL-31, IL-33 and Thymic stromal lymphopoietin (TSLP).^47–50

The immune response in FA includes two phases.^{51, 52} The sensitization phase (Figure 4)⁵³ is initiated when specific resident dendritic cell (DC) subsets capture allergens in the skin (also applicable to airways or gut) and transport the allergens to draining lymph nodes (LNs), where they are processed and presented to naïve CD4+ T cells. Within the LNs, in the presence of IL-4 and endothelial cell-derived Type 2 cytokines, the T cells differentiate into allergen-specific CD4+ T cells producing high levels of IL-4 and IL-13. IL-4 and IL-13 favor B cell isotype class switching to specific IgE and drives the production of IgE memory B cells (Figure 4).⁴⁷ Through facilitated antigen presentation, a very low concentration of allergen can drive complex formation between sIgE, allergen, and low-affinity IgE receptor on antigen-presenting B cells (CD23+ cells);⁵⁴ which then further drives Th2 cell proliferation, leading to further B cell isotype switching and IgE production.⁵⁵ As the B cells mature, they differentiate into plasma cells and produce large amounts of allergen-specific IgE antibodies (sIgE) which bind to high-affinity FcεRI receptors on the surface of mast cells and basophils. During this phase, a memory pool of allergen-specific B cells and allergen-specific CD4 positive T helper 2 cells are generated. More recently, a subset of Th2 cells (Th2A cells) have been suggested to play an important role in the allergic immune response.⁵⁶ Group 2 innate lymphoid cells (ILC2), which are found interspersed throughout barrier surfaces in the lung, gut, and skin, are also emerging as key regulators and effectors in type 2 immunity and tissue repair. They are enriched in human skin lesions, and ILC2 isolated from AD lesions, are activated by IL-33. They secrete proallergic cytokines, including IL-5 and IL-13 (Figure 5). IL-5 triggers the recruitment of eosinophils. IL-13 promotes recruitment of inflammatory cells, alterations in the skin microbiome, and decreases in epidermal skin barrier.^57–59

Figure 4:

Mechanism of sensitization phase in allergic disease. During the allergic sensitization phase, in the setting of an impaired barrier, specific resident dendritic cell (DC) subsets capture allergens in the skin and transport the allergens to draining lymph nodes where they are processed and presented to naïve CD4+ T cells. Within the lymph nodes, the presence of IL-4 and IL-13 favors B cell isotype switching to specific IgE cells that differentiate into plasma cells and produce large amounts of allergen-specific IgE antibodies (sIgE). The sIgE bind to high-affinity FcεRI receptors on the surface of mast cells and basophils. During the sensitization phase, a memory pool of allergen-specific B cells and allergen-specific CD4 positive T helper 2 cells are generated.

Figure 5:

Mechanism of effector phase in allergic disease. During the effector, which follows the sensitization phase, subsequent encounters with a previously sensitized allergen leads to IgE-cross linking and activation of basophils and mast cells. Degranulation of mast cells and basophils leads to the release of preformed and de novo mediators which cause the symptoms of the immediate phase reaction as well as the subsequent late-phase reaction by activating memory allergen-specific Th2 cells.

The effector phase (Figure 5), follows the sensitization phase and is triggered when the host subsequently encounters a previously sensitizing allergen and causes cross-linking of FcεRI receptor-bound sIgE on sensitized mast cells and basophils leading to the release of preformed and de novo inflammatory mediators. These processes lead to the immediate phase reaction as well as subsequent late-phase allergic reaction through the activation of memory allergen-specific Th2 cells.⁴⁷ As these cytokines accumulate, allergen specific Th2 cells are activated and produce IL-4, IL-5, and IL-13 among other cytokines. Recent evidence suggests that IL-13 is a key cytokine that drives peripheral inflammation in AD and is overexpressed locally, whereas IL-4 has a more central effect.⁶⁰ The cytokines maintain allergen-specific IgE levels, eosinophilia, mucus production, and recruitment of inflammatory cells to inflamed tissues leading to tissue damage.

COMPONENTS OF IMMUNE SYSTEM LEADING TO EPICUTANEOUS SENSITIZATION

Thymic Stromal Lymphopoietin (TSLP)

An important cytokine believed to play a role in epicutaneous sensitization through an impaired skin barrier is TSLP.⁶¹ TSLP is increased in the stratum corneum (SC) of patients with AD and is positively correlated with AD severity (using SCORAD).⁶² TSLP is highly expressed by the keratinocytes of children with AD⁶³ and by the epithelial cells of asthmatics.⁶⁴ TSLP deficient mice are protected from developing allergic skin and airway inflammation following antigen exposure,^65–67 as well as intestinal food allergy,⁶⁸ which highlights the importance of this cytokine in allergic sensitization. Pro-inflammatory cytokines Tumour necrosis factor-alpha (TNF-α) and IL1-α, secreted in response to skin stripping,⁶⁹ or scratching, induce TSLP secretion in vivo from human keratinocytes.⁷⁰ TSLP levels increase following topical ovalbumin (OVA) application onto tape stripped skin.⁷¹ TSLP expression correlates with Langerhans cell (LC) maturation, upregulation of the TSLP receptor on LCs, migration of LCs to skin draining LNs where they promote the differentiation of naïve Th cells to Th2 cells and Th2 proliferation.^{63, 66, 72} TSLP induces migration of Th2 skewed antigen presenting cells (APCs) to the mesenteric LNs following tape stripping, providing evidence of a skin to gut migration.⁷³ These findings link TSLP to the early stages of epicutaneous sensitization: increased secretion of TSLP by skin disruption or epicutaneous allergen exposure interaction with LCs to prime T helper cells.

Interleukin 33 (IL-33)

IL-33 is another epithelial cytokine that is released in response to defects in skin barrier surfaces characterized by AD.⁷⁴ IL-33 is a member of the IL-1 cytokine family and is expressed in epithelial barrier tissues as well as lymphoid organs and plays an important role in triggering allergic inflammation after exposure to allergens.⁷⁵ IL-33 levels are increased in the skin lesions and serum of individuals with AD,^{76, 77} and in mouse skin following epicutaneous sensitization with OVA.⁷⁶ IL-33 polarizes skin-derived DCs and innate lymphoid cells to drive Type 2 cytokine production following epicutaneous sensitization to peanut extract,⁷⁸ promotes increased secretion of IL-5 and IL-13 by in vitro polarized Th2 lymphocytes,⁷⁹ and contributes to elevated serum IgE levels and eosinophilia when injected in vivo.⁸⁰

A study exploring the role of IL-33 in epicutaneously sensitized FA, found IL-33 to be essential for inducing IgE-dependent anaphylaxis; IL-33 deficient mice and mice treated with soluble IL-33 receptor blocker were protected from oral-challenge induced anaphylaxis.⁸¹ Another study by Galand et al, demonstrated increased local and systemic release of IL-33 following mechanical skin injury induced by tape stripping in mice;⁸² IL-33R (ST2) deficiency or ST2 blockade prior to oral challenge, significantly reduced the severity of food induced oral anaphylaxis. The existing evidence suggests that IL-33 plays a key role in epicutaneous sensitization followed by subsequent food-induced anaphylaxis. Neutralization of IL-33 has also been shown as a promising strategy for the treatment for FA and AD.^{83, 84}

EXPERIMENTAL MODELS OF EPICUTANEOUS SENSITISATION

Several animal models have shown that epicutaneous exposure to specific food allergens induce a potent allergic Type 2 immune response, associated with IL-4 secretion by T cells from draining LNs and high levels of allergen specific IgE by B cells.^85–87 In addition, epicutaneous sensitization to food allergens has been shown to lead to systemic food allergic reactions, including anaphylaxis, on subsequent oral exposure.⁷¹ Bartnikas et al. demonstrated that epicutaneous OVA and/or peanut application onto tape stripped skin of BALB/c mice led to intestinal mast cell expansion, increased serum IL-4 and food-induced anaphylaxis.⁷¹ Following epicutaneous sensitization, systemic anaphylaxis was induced by a single oral challenge with allergen. Conversely, oral immunization with the cholera toxin (CT) adjuvant failed to induce anaphylaxis on subsequent oral challenge.⁷¹ In other experimental models of IgE-mediated anaphylaxis where mice were sensitized by intraperitoneal (IP) OVA injection with alum adjuvant, anaphylaxis was much harder to elicit on oral challenge, only after multiple oral challenges.^{88, 89} This demonstrated that epicutaneous sensitization primes the gut in a more effective manner for subsequent allergic reactions than oral or IP sensitization. When Bartnikas et al. performed epicutaneous sensitization in IgE-deficient mice, there was no intestinal expansion of mast cells, no rise in serum IL-4 and no features of anaphylaxis on intragastric challenge, which suggested that the effects arising from epicutaneous sensitization are IgE dependent.⁷¹ Th2 cytokines IL-4 and IL-13 have also been investigated in experimental food allergy models of epicutaneous sensitization; whereas mice deficient in IL-4 (genetically bred or in the presence of anti IL-4-blocking antibodies) were still able to mount Th2 responses following epicutaneous exposure,^{90, 91} however, IL-13 deficient mice (IL-13−/−) were not.⁹²

In a more recent study by Geha et al.⁹³ mechanical skin injury alone not only promoted cutaneous sensitization to foods, but also expanded and activated intestinal mast cells, thus promoting anaphylaxis in mouse models. Mechanical skin injury causes systemic release of IL-33, which interact directly with mast cells through the ST2 receptor,⁹⁴ and leads to expansion of both mucosal and submucosal mast cells in the jejunum.⁹³ Additionally, cutaneous injury led to increased intestinal permeability. The increase in intestinal mast cells as well as permeability that is elicited by mechanical skin injury or scratching, may play a significant role in promoting anaphylaxis in patients with FA and AD. In another study by Kawasaki et al.⁹⁵ the authors demonstrated that inflammation from skin barrier disruption alone can worsen symptoms of FA even when the exposure to the allergen is removed. BALB/c mice were sensitized epicutaneously by repeated application of OVA through disrupted skin. After the first oral challenge, the skin barrier was further disrupted, and a second oral challenge was performed. Skin barrier disruption amplified the allergic reactions induced by the oral challenge, whereas topical pretreatment with dexamethasone reduced allergic reactions.⁹⁵ These studies highlight the role that a disrupted skin barrier plays in the development of FA, and which may also contribute towards the severity of reactions.

Experimental models have carefully evaluated whether the skin barrier needs to be disrupted to achieve epicutaneous sensitization. In most models, mice only developed epicutaneous sensitization with the application of peanut to abraded skin.^{85–87, 90, 96} Conversely, application of peanut or OVA onto intact skin, or bypassing the SC (via subcutaneous skin layer) led to non-allergenic responses.⁸⁵ In a recent experimental model, repeated applications of roasted peanut extract onto intact skin induced sensitization to the major peanut allergens Ara h 1 and Ara h 2 and anaphylaxis upon IP peanut re-challenge.⁷⁸ The authors postulated that peanut protein was able to sensitize through the skin without the need for SC disruption or an adjuvant, because peanut protein itself had adjuvant properties. In support of this, they showed that peanut antigen led to bystander sensitization for the milk allergen α-lactalbumin. Having previously shown that application of milk allergen α-lactalbumin onto intact skin did not lead to sensitization or allergy on re-challenge, the authors then showed that concomitant application of peanut protein in addition to milk allergen α-lactalbumin onto intact skin led to anaphylaxis on subsequent oral exposure to milk allergen α-lactalbumin.⁷⁸ In contrast, Walker et al. showed that, in filaggrin (FLG) deficient mice, epicutaneous sensitization to peanut or OVA required costimulation with Alternaria alternata or house dust mite extract.⁹⁷

BARRIER DYSFUNCTION AND ITS ROLE IN AD AND FA

The discovery that FLG loss-of-function mutations, which have an important role in formation and integrity of the SC, are highly associated with AD and xerosis (dryness), was a major milestone in understanding AD as a ‘barrier’ disease.^{98, 99} Experimental models have shown that filaggrin protein deficiency in the skin leads to facilitated entry of irritants, pathogens and allergens, thereby eliciting an inflammatory response (Figure 1).^{100, 101} Similarly, decreased epithelial nasal barrier function has been shown to facilitate transepithelial allergen passage, allergic sensitization and allergen induced mast cell degranulation.¹⁰² Thus skin barrier impairment may itself lead to a maladaptive immune system with a predisposition towards atopy.

The local cytokine milieu has also been shown to affect FLG expression in the skin.^{103, 104} Filaggrin protein expression was assessed in the skin of 69 patients with at least one FLG loss-of-function mutation; filaggrin protein was reduced in non-inflamed skin, and even further in acute AD lesions.. Keratinocytes cultured in the presence of IL-4, IL-13 and /or IL-22 had reduced FLG expression, whereas those cultured in presence of IFNγ had increased FLG expression.^{105, 106} Taken together these studies show that skin barrier abnormality due to FLG loss-of-function mutations can enhance allergen penetration thus favoring Th2 inflammation. On the other hand, Th2 inflammation can impair skin barrier function via a reduction in filaggrin protein expression, leading to a vicious cycle of inflammation and skin barrier disruption.

There has been conflicting research as to whether FLG mutations alone increase the risk of FA, in the absence of AD or other factors. Brown et al. showed a significant association between FLG loss-of-function and challenge proven IgE-mediated PA, even after adjusting for AD in cohorts from the UK, Netherlands, Ireland and Canada.¹⁰⁷ However, in infants, both the HealthNuts study and the EAT study showed that although FLG null mutations increased the odds of food sensitization, it did not significantly increase the odds of FA.^{108, 109} Venkataram et al.¹¹⁰ explored the longitudinal relationship between FLG mutations and FA; path analysis suggested an association between FLG mutations and AD in younger children and the progression to food allergic sensitization and FA in older children.

Beyond filaggrin, there are multiple other components of the skin barrier that may have a role in AD and FA.¹¹¹ Mutations in the corneodesmosin gene leads to Peeling Skin Disease associated with severe skin barrier defect, pruritus and food allergies.¹¹² Peeling Skin Disease shares clinical features with Netherton syndrome which is caused by SPINK5 mutations as well as patients with mutations in the transglutaminase 5 gene and desmoglein gene.^{113, 114} Importantly, Type 2 immune activation is associated with global downregulation of skin barrier genes including filaggrin, loricrin and involucrin thereby allowing food allergens to penetrate through the skin.¹¹⁵

Environmental factors, such as detergents and microplastics have an effect on epithelial barrier function. Anionic surfactants and commercial detergents have been shown to decrease tight junction barrier integrity in human keratinocytes¹¹⁶ and human bronchial epithelial cells,¹¹⁷ potentially increasing the risk of atopic disease. Surfactants have been shown to disrupt the lamellar architecture in the SC.¹¹⁸ In a mouse model, Ochi et al. showed that detergent-mediated skin damage promoted IgE/IgG1 responses and Th2 differentiation when epicutaneously sensitized to papain, a protease allergen. They also found that TSLP and IL-33 contributed to epicutaneous papain sensitization and the atopic march upon subsequent airway challenge with the protease allergen.¹¹⁹ In humans, application of traditional alkaline soaps increased transepidermal water loss (TEWL) and skin erythema, which are signs of prolonged damage to the skin barrier.¹²⁰ The environmental damage caused by plastics are well documented; however, the role of microplastics, which have become ubiquitous in the environment, on human health is poorly understood. It is in our air, our water, and in the foods we eat. It is within us and has been detected in human lungs and feces. Ingestion of microplastics has been implicated in gut dysbiosis through mechanical disruption and microbiome alterations leading to inflammatory responses.¹²¹ An in vitro study of peripheral blood mononuclear cells found that high exposure to polypropylene microplastics (~20 μm and 25–200  μm) increased histamine levels.¹²² In an in vitro model of human epithelial colorectal cells, researchers found that exposure to microplastics upregulated transcriptional factors involved in cell inflammation and proliferation. The chronic exposure to low doses of microplastics has the potential to induce epithelial cell injury and alterations to intestinal barrier function.¹²³ In a study of human lung epithelial cells, researchers found that exposure to inhaled polystyrene microplastics cause inflammatory and oxidative injury along with the disruption of intercellular junction proteins in the lung, potentially leading to pulmonary barrier dysfunction.¹²⁴

IN VITRO MEASURES OF EPICUTANEOUS SENSITIZATION

In vitro evidence has shown that PA correlates with peanut-specific proliferation in the skin-homing memory T cell subset, suggesting that sensitization occurs though the skin.^{125, 126} The differential peanut-specific Th cell proliferation in peanut allergic versus peanut tolerant children was assessed using homing markers on CD4+ Th cells for the skin (Cutaneous Lymphocyte Antigen: CLA+) and gut-associated-lymphoid tissue (integrin α4β7+) as markers for the route of initial sensitization to peanut.¹²⁵ Peanut specific Th cell proliferative responses were higher in CLA+ Th cells in peanut allergic children than peanut tolerant children. Cytokine responses from CLA+ peanut-specific Th cells in peanut allergic children showed a trend towards Th2-polarization (IL-13 and IL-4 secretion) whereas responses from α4β7+ peanut-specific Th cells in peanut tolerant children trended towards Th1 (IFNγ and TGFβ), suggesting that the origin of peanut sensitization occurs in the skin (Figure 3). Sensitization through the skin may be explained by the route of exposure to food allergens via the environment (dust and surfaces) or through transfer of the food allergen through hand-body contact, which is explored further below.

ENVIRONMENTAL EXPOSURE TO ALLERGENS IN HUMANS

In humans, several observational studies have shown an association between epicutaneous and environmental exposure to peanut allergens and the development of peanut sensitization,^{22, 127, 128} and challenge proven PA.^{40, 128, 129} One of the first studies to highlight the risk of epicutaneous peanut exposure was the ALSPAC birth cohort, where topical creams containing peanut (‘Arachis’) oil applied onto the skin of children with AD was an independent risk factor for the development of PA (peanut skin prick test (SPT) >8mm or sIgE >15kU/L).⁴⁰ Among PA children with AD in the ALSPAC study, 90% had been topically exposed to creams containing Arachis oil in the first 6 months of life.⁴⁰ In another study, household peanut consumption was used as an indirect marker of environmental peanut exposure.¹²⁹ Peanut allergic children had a 10-fold higher level of household peanut protein consumption (18.8 grams/week) than high-risk egg allergic children without PA (1.9 grams/week) (p<0.0001) and 3-fold higher level than non-atopic children without PA (6.9 grams/week) (p<0.0001).^129. Peanut butter consumption was more highly associated with PA than covered forms of peanut foods (e.g. chocolate peanuts); the authors postulated that environmental peanut exposure was exerting its’ impact through the skin, as peanut butter is sticky and therefore more likely to be transferred onto the skin of infants. In support of the dual allergen exposure hypothesis, infants who had eaten peanut in the first year of life were not affected by environmental peanut exposure. Subsequent studies showed that household peanut consumption is highly correlated with the concentration of peanut in the dust of an infants bed-sheet and play-area, even if the infant does not eat peanut.¹³⁰

Further work analyzing peanut dust concentration in infants with AD, milk and/or egg allergy, showed that, for every unit increase in peanut dust concentration from the living room floor, there was a 1.97-fold increase in peanut sensitization and 2.34-fold increase in PA.^22.In a population based study, children carrying at least one FLG loss-of-function mutation, were at increased risk of developing PA if they had high environmental peanut exposure in their bed-dust around the time of birth;¹²⁸ for each unit increase in peanut dust concentration during infancy, there was a 6-fold increase in school-age peanut sensitization and a 3.3-fold increase in school-aged PA.¹²⁸ In another population based Swedish study, environmental peanut exposure in high risk children (based on parental atopy or egg sensitization), increased risk of school-aged peanut sensitization.¹²⁷ Based on this research, one could postulate that reducing environmental exposure to peanut (and potentially other allergenic foods), in high-risk children might reduce the risk of sensitization and allergy; however, this has not been proven. Additionally, peanut levels in areas outside the home are also high (for example in schools),¹³¹ thus an environmental allergen reduction approach may be difficult to implement. In summary, only high-risk children (those with an impaired skin barrier, egg allergy or predisposition towards atopy) seem to be at risk of peanut sensitization from high environmental peanut exposure, which has important ramifications for future research in this field. Interestingly, associations between impaired skin barrier and allergy may also be pertinent for inhalant allergies; environmental cat exposure around the time of birth was found to increase the risk of indoor allergen sensitization in children with at least one FLG loss-of-function mutations.¹³²

THE ROLE OF THE MICROBIOME IN EPICUTANEOUS SENSITIZATION

With growing evidence that food sensitization can occur through the skin, more attention has been paid to dysbiosis of the skin microbiome as a factor in AD and FA.¹³³ Skin commensals have been found to be essential for resident T cells and educate keratinocytes to become effective in combating skin pathogens. Conversely, commensals also appear to be required so that the host does not attack normal colonizing flora. Microbial dysbiosis has been observed in infants with AD,^{134, 135} particularly with S. aureus colonisation.¹³⁶ In a study using minimally invasive skin tape stripping, shotgun metagenomic studies revealed that the skin of AD FA+ children was colonized with Staphylococcus aureus.¹³⁷ In contrast, topical transplantation of commensal bacteria from healthy volunteers to patients with AD improved AD severity and reduced topical steroid use; this forms the basis for further work in this area.^{138, 139}

Schwartz et al. identified a critical role for the microbiome in mice with an impaired skin barrier. The authors reported on their observation of skin inflammation in FLG-mutant (FLG^ft/ft) mice triggered by the interplay between the microbiome; specifically they reported a shift towards pathogenic staphylococci, IL-1β, and mast cells.¹⁴⁰ Further experimental models have shown that applying Staphylococcus enterotoxin B (SEB) on skin enhances allergic responses to food antigens.^141–144

In humans, S. aureus colonisation is significantly associated with AD severity, persistence and subsequent deterioration of AD.¹⁴⁵ Tsilochristou et al. showed the importance of S. aureus in driving AD severity, and specific IgE to cow’s milk, hen’s egg and peanut, independent of AD severity in the LEAP study.¹⁴⁵ Additionally, S. aureus colonization at any time was associated with prevention of acquisition of natural tolerance to hen’s egg, and the disruption of oral tolerance induction to peanut.¹⁴⁵ Taken together, these studies highlight the role of S. aureus dysbiosis in children with AD, on AD severity and risk of AD deterioration, in addition to increasing risk of sensitization to food, enhanced allergic responses, persistence of FA and inhibition of oral tolerance induction.

PHENOTYPIC MEASURE OF SKIN BARRIER DYSFUNCTION

Functional assessment of the integrity of the epidermis using TEWL has been shown to predict subsequent development of AD regardless of filaggrin mutation.^{146, 147} It has also been associated with food sensitization (even after adjusting for AD and FLG mutation status) in young infants.¹⁴⁸ Increased TEWL may reflect the spectrum of genetic and environmental factors leading to barrier impairment, such as water hardness,^{149, 150} frequency of washing and commercial detergents (known to decrease tight junction barrier integrity in human keratinocytes).^{116, 117} These findings are being assessed in randomized controlled trials (RCTs) to prevent AD such as the SOFTER (Softened Water for Eczema Prevention) trial.¹⁵¹

If the skin barrier is impaired, more water evaporates from the skin enabling allergens, irritants and bacteria to penetrate the skin.¹⁵² Carriage of a FLG loss-of-function mutation is associated with increased TEWL at three months, even in the absence of AD,¹⁴⁸ suggesting this already shows a predilection towards skin barrier impairment which could subsequently lead to clinical AD. TEWL at 2 days and 2 months was shown to predict AD at 12 months,¹⁵³ and subsequently the same group showed that raised TEWL predicts FA at 2 years of age.¹⁵⁴ In a recently published study, it was shown that TEWL after skin tape stripping separated AD with FA vs AD without FA.¹³⁷

IDENTIFYING SKIN BARRIER MARKERS TO PREDICT FOOD ALLERGY

A minimally invasive skin-tape-stripping measure of the SC in combination with a comprehensive multi-omics approach was used to determine whether AD with FA (AD FA+) children have skin biomarkers which distinguish them from AD without FA and non-atopic children.¹³⁷ Despite similar skin disease severity, the SC integrity and FLG content were significantly lower in children who were AD FA+ vs. AD FA−. Lipidomics of the SC in the AD FA+ group revealed a relative reduction in esterified ω-hydroxy fatty acid (EO) sphingosine (S) ceramides (CER) which are ultra long chained ceramides (EOS CER) required for normal skin barrier function. At the same time, a significant increase in nonhydroxy fatty acid sphingosine ceramide levels was observed in AD FA+ skin samples, resulting in a disproportionate decrease in EOS CER in the skin. Transcriptome studies of the skin tape also revealed the AD FA+ skin samples expressed the highest Type 2 immune expression levels.

Interestingly, STS proteomics revealed an immature keratin profile consistent with keratinocyte hyperproliferation in the SC of AD FA+ participants. A network analysis demonstrated that keratin 5, 14 and 16 expression, and reduced filaggrin breakdown products (urocanic acid), were strongly correlated with AD FA+, with increased TEWL, and were the most important predictors of AD FA+.¹³⁷ These data support the importance of skin barrier dysfunction in the pathogenesis of epicutaneous sensitization to environmental foods and may contribute to the persistence or severity of FA by chronically stimulating Type 2 immune responses in the skin.

THE ROLE OF EMOLLIENTS TO PREVENT THE DEVELOPMENT OF FOOD ALLERGY

Based on the body of evidence that sensitization occurrs through the skin, it would be logical to assume that reducing both severity and duration of AD could potentially reduce the incidence of FA. Pilot studies showed that standard moisturizers could reduce AD by up to 50%, primarily by enhancing skin hydration^{155, 156} The summary of the different pilot studies is presented in Supplementary Table A. These pilot studies were then followed up by larger RCTs trials; these are presented in Tables 1 (non-lipid emollients), 2 (emollients containing at least one ceramide) and 3 (emollients containing a combination of emollients and topical steroids). A more detailed version of the trials is presented in Supplementary Table B. Supplementary Table C lists emollients that have been tested in clinical trials. An RCT by Yonezawa et al. found statistically lower TEWL, diaper dermatitis, and risk of dry facial or body skin in the interventional group (1–2 applications per day for 12 weeks with a moisturizer containing either paraffin or alcohol/ceramide, n=114) than in the control group (n=113).¹⁵⁷ The BEEP (Barrier Enhancement for Eczema Prevention) and PreventADALL (Preventing Atopic Dermatitis and ALLergies in Children) RCTs (Table 1) evaluated whether including petrolatum based oils in baths or applying petrolatum-based moisturizers from the first few weeks of life (and small ‘tastes’ of allergenic foods in PreventADALL) could prevent AD and FA. Both studies showed no significant reduction in the incidence of AD, and in the BEEP study, there was an increased rate of infections and a trend towards increased food allergy in the intervention group.^{158, 159} Similarly, an RCT by Dissanayake et al. compared the incidence of AD and FA in the first year of life in infants treated with synbiotics and an emollient (alone or in combination) and controls and found no statistical difference.¹⁶⁰ The negative findings in these studies were surprising given the pilot study findings, and the authors discussed various reasons as to why this might have occurred including low adherence rate with the intervention (especially PreventADALL), contamination of the control arm (BEEP), and the type of emollient used. The PEBBLES (Prevention of Eczema By a Barrier Lipid Equilibrium Strategy) pilot study (Table 2) used a ceramide-containing trilipid cream (Supplementary Table A) for 6 months and demonstrated a reduction in investigator observed AD at 12 months, and food sensitization at 6 and 12 months of age (Figure 6).¹⁶¹ Per protocol analyses showed that in infants who received ≥5 days per week of treatment (twice daily trilipid cream), 0% (0 of 21) of children developed food sensitization compared to 19% (7 of 36) in the control group. Miyaji et al.’s retrospective cohort study⁴⁴ (Supplementary Table A) showed that in infants with moderate-severe AD, proactive topical steroid treatment commenced by 4 months of age versus after 4 months of age was associated with an almost 2-fold reduction in FA by 24 months (46.3% in the delayed steroid treatment group versus 25.3% in the early steroid treatment group (p=0.01). A number of emollient studies for prevention of atopic diseases are ongoing (e.g., NCT02906475, NCT03871998, NCT03409367, and NCT03667651) and results have not been published (Tables 1 and 2).

Table 1.

Randomized Trials Assessing Non-Lipid based Emollients for Atopic Dermatitis

Prevention type	RCT Author & Study Name	Participant Number & Study Arms	Population	Emollient	Duration, Frequency, & Location	Outcome measures	Results
1ry	Blume-Peytavi et al. Atopic Dermatitis in Atopy Predispos ed Infants (ADAPI) NCT02906475	160 2 arms	0–14 days of age Term newborns 18–45 year old moms Birthweight 2.5 to 4.5 Kg I, II, III, IV Fitzpatrick classification Parent or sibling with physician diagnosed AD, AR, or AA Must be breastfed and/or on cow-milk based formula	Milk Lotion (no details about brand(s) or ingredients provided)	12 months Applied once daily to entire body and face + Bathing addendum if needed	1ry: Cumulative AD incidence at 12 months 2ndry: Cumulative AD incidence at 24 months EASI at 12 and 24 months IDQoL at 12 and 24 months TEWL, skin surface pH, and Stratum Corneum Hydration on mid-alveolar forearm at 14 days, 1, 3, 6, 12, and 24 months	Study ongoing- no results published to date.
1ry	Jonathan Hourihane et al. Short-term Topical Application to Prevent Atopic Dermatitis (STOP AD) NCT03871998	242 2 arms	Full term healthy infant Up to 5 days of age 1st degree relative with AD, AR, AA, or FA	Commercially available moisturizer (no details about brand(s) or ingredients provided)	2 months 2x daily full-body application by intervention group standard skin care advice with no moisturizer application in control group	1ry: Cumulative incidence of AD at 12 months Cumulative incidence of IgE-mediated food allergy at 24 months 2ndry: TEWL and Natural Moisturizing Factor measured at birth, 2, 4 and 8 weeks and at 6 and 12 months Microbial community and skin biomarker analysis in a subset of study participants, comparison of changes over 1st year between cheek and antecubital fossa and between intervention and control group	Study ongoing- no results published to date.
1ry	Chalmers, J.R. et al. Daily emollient during infancy for prevention of eczema: the BEEP randomised controlled trial.159 ISRCTN21528841	1394 Control group (skin care advice only):701 Intervention group (skin care advice+emollient): 693	Up to 3 weeks of age 1st degree relative with physician-diagnosed AD, FA, AR	Choice between Doublebase Gel® (Dermal Laboratori es Ltd.) & Diprobase Cream® (Merck Sharp & Dohme Ltd.)-paraffin based without alcohol or lipids	12 months >=1x daily	1ry: Blinded assessment of AD at 24 months of age using UKWP diagnostic criteria 2ndry: Time to onset of AD Severity of AD based on EASI & POEM Confirmed diagnosis of FA at 24 months to milk, egg, and/or peanut (parental report, allergic sensitization, food challenge- if required) Parental report of clinical diagnosis of any FA between 12 and 24 months Safety & cost-effectiveness of prevention strategy Role of FLG genotyping as possible stratifier of response to emollient intervention	No statistically significant difference in AD: present at age 2 (within the past year) in 150 (25%) of 612 in control group and in 139 (23%) of 598 in intervention group (adjusted RR 095 [95% CI 0·78 to 1·16]; p=061) Subgroup analyses (number of 1st-degree relatives with AD, AA, or AR, FLG genotype, season of birth, water hardness, and probiotic use) also showed no statistically significant differences Adherence was considered satisfactory in only 51% of intervention group (all of those who did not complete any questionnaires were assumed as not using emollient). Adherence was considered satisfactory at 70% (311 of 442) in those who completed all questionnaires Adherence in intervention group decreased with each time-point: 466 (88%) of 532 had satisfactory adherence at 3 months, 427 (82%) of 519 at 6 months, and 375 (74%) of 506 at 12 months
1ry	Skjerven HO et al. Skin emollient and early complement ary feeding to prevent infant atopic dermatitis (PreventAD ALL): a factorial, multicentre, cluster-randomised trial158 NCT02449850	2701 1 of 4 similar sized groups: No intervention Skin care (oil-bath & Ceridal face cream) Consecutive introduction (3–4 months of age) of peanut, milk, wheat, & egg >= 4 days/week complementary to breastfeeding Both interventions	Infants General population	Ceridal face cream (alcohol & polyethelyne glycol) in skin care group and dual intervention group	9 months >= 5 days/week	1ry: AD assessed first at 12 months FA to intervention food assessed first at 36 months 2ndry: Recurrent wheeze or AA, allergic sensitization, allergy to other foods, anaphylaxis, & AR	Study closed to recruitment but ongoing No statistically significant difference in AD: present in 478 (8%) of 596 infants in control group, 64 (11%) of 575 in skin intervention group, 58 (9%) of 642 in food intervention group, and 31 (5%) of 583 in combined intervention group. Risk difference of 3·1% (95% CI – 0·3 to 6·5) for skin intervention versus control and 10% (−2·1 to 4·1) for food intervention versus control Reports of Itching, dry skin, and urticaria were similar in all groups

POEM: Patient Oriented Eczema Measure; EASI: Eczema Area and Severity Index; IDQoL: Infants’ Dermatitis Quality of Life; RCT: Randomized controlled trial

Table 2.

Randomized Trials Assessing at Least One Ceramide-containing Emollient for Atopic Dermatitis

Prevention type	RCT Author & Study Name	Participant Number & Study Arms	Population	Emollient	Duration, Frequency, & Location	Outcome measures	Results
1ry	Yonezawa K, et al. Effects of moisturizing skincare on skin barrier function and the prevention of skin problems in 3-month-old infants: A randomized controlled trial157	227 Control group: 113 Intervention group: 114	1 week to 3 months of age	For bathing, Pigeon Baby Soap [Pigeon, Tokyo, Japan] or Kewpie Baby Body Soap, Cow Brand Soap [Kyoshinsh a, Tokyo, Japan] As moisturizer in intervention group: Pigeon Baby Milk Lotion [Pigeon] or Atopita Milky Lotion© [Tampei Pharmace utical, Tokyo, Japan]	12 weeks Application 1–2x daily with bathing every 2 days in intervention group Vs Daily bathing without moisturizer in control group	1 ry: TEWL, SCH, skin pH and sebum secretion at 3 months of age 2ndry: Incidence of skin problems (e.g. Diaper dermatitis) at 3 months of age	Statically significant lower TEWL (mean ± SD, 14.69 ± 7.38 vs 17.08 ± 8.26 g/m2 per h, P = 0.033) and higher face SCH (60.38 ± 13.66 vs 53.52 ± 14.55, P = 0.001) and higher body SCH (58.89 ± 12.96 vs 53.02 ± 10.08, P < 0.001) in intervention vs. control group Diaper dermatitis at 1–3 months of age significantly lower in intervention vs. control group (6.3% vs 15.9%, P = 0.022; RRR= 0.399 [95% confidence interval [CI], 0.174–0.919]). Skin problems at 1–3 months of age significantly lower in intervention vs. control group (42.1% vs 55.2%, P = 0.064; relative risk ratio, 0.762 [95% CI, 0.569–1.021]). Risk of dry facial skin was reduced with intervention by 48.4% (relative risk ratio, 0.516; 95% CI, 0.333–0.799]) Risk of dry body skin was reduced with intervention by 39.7% (relative risk ratio, 0.603; 95% CI, 0.399–0.911)
1ry	Eric Simpson A Community-based Assessment of Skin Care, Allergies, and Eczema (CASCADE) NCT03409367	1250 2 arms	Up to 2 months of age Born at > 25 weeks of gestation	Choice of over the counter lipid-rich emollient (Vaseline/Vanicream/Cerave Healing Ointment/Cerave cream/Cetaphil cream)	24 months Once daily full body application	1ry: Cumulative incidence of AD at 24 months of age 2ndry: AD diagnosed by Children’s Eczema Questionnaire (CEQ) at 12 and 24 months Any prescription topical medication or over-the-counter hydrocortisone usage Parental report at 3, 6, 9, 12, 15, 18 and 24 months of provider-diagnosed AD and sleep loss of infant Parental report of immediate food allergy symptoms or a provider diagnosis of food allergy that was confirmed by prick testing or IgE blood test at 12 and 24 months Asthma risk using a modification of the asthma predictive index (API) at 12 and 24 months Global Health Status using one question from the PROMIS PGH-7 instrument at 12 and 24 months Atopic dermatitis severity as measured by parental report of eczema, age of onset Time to onset of AD as measured by provider-recorded date of first diagnosis retrieved from review of health record AD symptom severity as reported by POEM Atopic dermatitis severity by Parent-reported global severity of eczema assessment IDQoL instrument	Study ongoing- no results published to date.
1ry	Dissanayake et al. Skin care and synbiotics for prevention of AD or FA in newborn infants: A 2 × 2 factorial, randomized, non-treatment controlled trial160	605 pregnant women recruited - 549 infants enrolled Randomized into 4 treatment arms: Skin care and synbiotics. (n=137) Synbiotics only (n=137) Skin care only (n=138) No intervention (n=137)	Infants	Locobase ® REPAIR cream (Daiichi Sankyo, Japan)	6 months 2–3x daily	1ry: Development of AD at 1 year of age 2ndry: Prevalence of FA, as reported in the questionnaires at 1 year Sensitization to any food or inhalant allergen at 9 months of age EASI score at 9 months of age TARC score at 9 months of age	Cumulative incidence of AD (itchy skin condition lasting 2 or more months) at 12 months of age: 30.9% in group 1, 32.1% in group 2, 38.6% in group 3, and 25.6% in group 4 – No statistically significant difference Incidence of AD AS DEFINED BY UKWPC was 20.4% in group 1, 14.7% in group 2, 21.0% in group 3, and 18.8% in group 4 – No statistically significant difference No statistically significant difference in AD incidence after adjusting for emollient application rate (80% of the parents/guardians in the emollient groups applied emollient at least 2x/day) No statistically significant difference in AD rate when comparing group 1+3 (received emollients) to 2+4 (no emollients) TARC levels not significantly different between all infants of all groups TARC levels of infants who developed AD by 9 months of age also not significantly different between all groups EASI scores showed mild disease in most babies who developed AD by 9 months of age (only 5/49 babies with AD had moderate disease) Highest EASI score was 13.6. Median scores did not differ significantly among groups Prevalence of FA at 1 year of age, sensitization to any food or inhalant allergen at 9 months of age and allergen-specific IgE showed no difference in AD incidence between groups
1ry	Lowe A, Su J, Tang M, et al. PEBBLES study protocol: RCT to prevent AD, FA and sensitization in infants with a family history of allergic disease using a skin barrier improvement strategy172 Phase III ACTRN12617001380381 NCT03667651	760 2 arms (380 in each)	Up to 3 weeks of age 1st degree relative has a self-reported hx of AA, AD, hay fever/AR or FA	EpiCeram®	6 months 6 grams 2x daily	1ry: Presence of AD in first 12 months of life assessed using UKWP criteria and/or visible AD at time of examination 2ndry: FA, based on skin prick tests, hx of reactions & OFC (if wheel ≥ 1 mm) at 12 months Adverse events (AEs) to EpiCeram® Skin barrier function as assessed by TEWL at 6 weeks & 12 months Food sensitization (positive skin prick test) at 12 months of age	Study ongoing- no results published to date.

OFC: oral food challenge; RCT: Randomized controlled trial; TARC: Thymus and activation-regulated chemokine

Table 3.

Randomized Trials Assessing Combination of Emollients and Topical Steroids for Atopic Dermatitis

Prevention type	RCT Author & Study Name	Participant Number & Study Arms	Population	Emollient	Duration, Frequency, & Location	Outcome measures	Results
2ndry	Yamamoto - Hanada et al. Early aggressive interventi on for infantile AD to prevent developm ent of FA: a multicenter, investigator-blinded, randomized, parallel group controlled trial (PACI Study)-protocol for a RCT168 UMIN000028043	650 Control group (standard guideline-based treatment): 325 Intervention group (early aggressive topical steroids & Heparinoid cream application): 325	7–13 weeks old Diagnosis of AD (UKWP criteria) with rash development within previous 28 days	Alclometas one dipropionate (Almeta®) to face, Betametha sone Valerate (Rinderon®-V) and Mometasone Furoate (Fulumeta®) to body Heparinoid cream (Hirudoid® Soft ointment) as emollient	28 weeks Whole body (except scalp) treatment Variable application depending on week in study & flares	1ry: OFC-proven IgE-mediated hen’s egg allergy at 28 weeks of age 2ndry: OFC test scores at 28 weeks Total IgE antibody titer & Egg white, ovomucoid, milk, soy, wheat, & peanut-specific IgE & IgG4 antibody titers in serum at 28 weeks EASI scores at 2, 4, & 8 weeks after study entry & at the age of 28 weeks POEM scores weekly through study Percentage of disease-free days Dose of rescue medication (applications) used IDQoL and DFI questionnaires at 2, 4, & 8 weeks after study entry & at 28 weeks Cumulative incidence of IgE-mediated FA assessed by a doctor during study Cumulative incidence of wheezing assessed by a doctor during study (parental reports evaluated by doctor)	Study ongoing- no results published to date.

OFC: oral food challenge; POEM: Patient Oriented Eczema Measure; EASI: Eczema Area and Severity Index; IDQoL: Infants’ Dermatitis Quality of Life; DFI: Dermatitis Family Impact, RCT: Randomized controlled trial

Figure 6:

Effect of twice-daily EpiCeram® vs standard skin care for the first 6 months on skin prick test reactivity to food allergens at 12 months. Results were limited to intervention participants who were treated at least 5 times per week on average. Per protocol analyses revealed a significant reduction in food sensitization at 12 months in the treatment group (0%, 0 of 21) compared to the control group (19%, 7 of 36). Reproduced with permission from Lowe AJ, Leung DYM, Tang MLK, Su JC, Allen KJ. Ann Allergy Asthma Immunol. 2018 Feb;120(2):145–151.

Although the experimental and mechanistic studies, to date, build a strong case towards targeting the skin to prevent AD to thereby FA, two large preventative emollient RCTs have been disappointingly negative (BEEP and PreventADALL); there are several reasons why they might not show a significant effect, as outlined below:

1. The age at which emollients were first applied: If emollients are applied after significant sensitization has progressed, it may difficult to prevent FA. In PEBBLES (Table 2), reduction in sensitization was only found in infants where application commenced before 3 weeks. Miyaji et al.⁴⁴ (Supplementary Table A) showed that the time period until the of proactive topical steroid treatment in AD was a risk factor for FA at 24 months of age.

2. The rate of adherence with the study intervention: The PEBBLES study (Table 2) only showed a reduction in sensitization rates if emollients were applied twice daily for ≥5 days a week. In contrast, in the PreventADALL¹⁵⁸ and BEEP¹⁵⁹ study, adherence was defined as emollient application once a day, at least 3–4 times a week (less than the 5 days in PEBBLES study).

3. Emollient properties: The properties of emollients on skin hydration, skin pH, restoration of the skin barrier and immune function are all important properties to consider in the choice of emollient for the prevention of FA¹⁶² (Supplementary Table C). Emollients can have high pH levels, and high skin pH which can damage skin proteins and lipids, leading to dryness, irritation, itching and increased TEWL.¹⁶³ BEEP and PreventADALL RCTs (Table 1) both used petrolatum-based emollients which have been shown to have lower benefits on TEWL compared to ceramide-containing trilipid creams used in PEBBLES (Table 2).^{164, 165}

4. There may be a disconnect between AD and FA development depending on what needs to be reduced (e.g. dryness or inflammation or both). This disconnect was shown in other treatments for the prevention of AD that did not result in a reduction of IgE (with probiotics).¹⁶⁶

5. AD is not just a barrier defect, but also an inflammatory process. Targeting inflammation early and aggressively with topical steroids has been shown to be associated with a significant reduction in food sensitization and allergy in infants with moderate to severe AD.⁴⁴ Thus treating only the barrier defect in AD may not be enough to protect against the development of FA.

6. The role of environmental food exposure in these studies has not been assessed. It is not known if parents washed their hands prior to application of cream and whether increasing the contact of hands onto the skin of the child may have inadvertently increased topical food exposure.

SUMMARY

There is an increasing body of evidence to support the role of a disrupted, inflamed skin barrier being the root cause for the development of food sensitization and allergy.¹⁶⁷ Interest in restoring the skin barrier to prevent AD and FA has gained ground in recent years. Oral tolerance induction has offered a completely new strategy for the prevention of FA; results are compelling for PA²⁴ and moderate for egg allergy.²⁵ If this strategy were to work for other food allergens, it would require the introduction of multiple food allergens very early on in life. Targeting the skin to prevent AD or reduce the duration and severity of AD could prevent the development of FA, by either preventing epicutaneous sensitization, or by increasing the window of opportunity to induce oral tolerance by early allergenic food introduction. However, results from two large RCTs using preventative emollient therapy have had initial negative results for the prevention of AD (BEEP¹⁵⁹ and PreventADALL¹⁵⁸) and FA (BEEP)¹⁵⁹ and other RCTs are ongoing (CASCADE- A Community-based Assessment of Skin Care, Allergies, and Eczema, and PEBBLES) (Tables 1 and 2). There are also secondary prevention studies (Table 3) using both emollients and proactive topical steroid treatments ongoing or planned (PACI study: Prevention of Allergy via Cutaneous Intervention Study and SEAL-Stopping Atopic Dermatitis and Allergy Study).¹⁶⁸ Regardless as to whether these studies are positive or negative in their findings, they will help inform this important area of research.

BOX 1: Major milestone discoveries.

1. 1911: Oral tolerance induction is first described in an animal model.¹⁶⁹

2. 1994: The first report of epicutaneous sensitization leading to IgE mediated FA is followed by large body of evidence in experimental models.⁷¹

3. 1996: The Dual Allergen Exposure Hypothesis is first proposed in 1996.¹⁷⁰

4. 2003: The ALSPAC birth cohort, shows an association between peanut oil containing creams applied to the skin of children with AD and the development of PA.⁴⁰

5. 2006: FLG loss-of-function mutations are confirmed to be major predisposing factor for AD,⁹⁹ thus arguing that AD is a disease resulting from a defective skin barrier.

6. 2008: FLG loss-of-function were shown to be associated with challenge proven IgE-mediated PA, even after adjusting for AD.¹⁰⁷

7. 2009: High household peanut consumption was associated with the development of PA.¹²⁹

8. 2014–2015: High environmental exposure to peanut (measured in dust) was shown to increases the risk of peanut sensitization and likely allergy in children with mutations in the FLG gene or with AD.¹²⁹

9. 2015: LEAP randomised controlled trial demonstrated that oral peanut ingestion from 4–11 months of age prevented PA in high-risk infants.^{23, 24}

10. 2013: HealthNuts study found that by 12 months of age, 3.1% of children had challenge proven FA and 9% were egg allergic.¹ There is therefore a narrow window of opportunity to induce tolerance.

11. 2014: Japanese,¹⁵⁶US and UK¹⁷¹ preventative emollient therapy pilot studies showed that early emollient therapy in high risk children reduced the onset of AD by 33–50%.

12. 2015–2016: TEWL at 2 days was shown to predict not only AD,¹⁵³ but also FA at 2 years,¹⁵⁴ providing an early marker of disrupted skin barrier predictive of later disease.

13. 2017: PEBBLES pilot study demonstrated a trend towards a reduction in AD and food sensitization in infants treated with emollient therapy for 6 months.¹⁶¹

14. 2019: Early (≤4 months) versus later (>4months) proactive use of topical steroids in infants with AD was shown to be associated with a 45% reduction in FA at 2 years (retrospective study).⁴⁴

15. 2019: A novel non-invasive technique using skin tape stripping identified skin barrier dysfunction in AD with FA subjects.¹³⁷

16. 2020: Two large preventative emollient RCTs using petrolatum-based creams/bath oils show no significant effect on the prevention of AD (BEEP¹⁵⁹ and PreventADALL¹⁵⁸) or FA (BEEP)¹⁵⁹.

BOX 2: Future research perspectives.

1. To determine at what age interventions to prevent the development of food sensitization and allergy should commence, and the duration of these interventions.

2. The degree of adherence to therapies targeting the skin required to achieve a reduction in food sensitization and allergy.

3. The importance of different emollient properties (skin hydration, skin pH, restoration of the skin barrier and immune function) for preventing food sensitization and allergy.

4. The role of topical anti-inflammatory agents as part of the skin care regime to prevent food sensitization and allergy.

5. The role of reducing environmental food allergen exposure on reduction of sensitization and allergy.

6. Reduction in S. aureus colonisation in at risk children on the prevention of FA.

ABBREVIATIONS:

AA

Allergic asthma

AD

Atopic dermatitis

AEs

Adverse Events

API

Asthma Predictive Index

AR

Allergic Rhinitis

BEEP

Barrier Enhancement for Eczema Prevention

CER

Ceramides

CEQ

Children’s Eczema Questionnaire

CLA+

Cutaneous Lymphocyte Antigen

DC

Dendritic cell

DFI

Dermatitis Family Impact

EASI

Eczema Area and Severity Index

EOS CER

fatty acid ultra long chained ceramides

FA

Food allergy

FLG

Filaggrin gene

TSLP

Thymic stromal lymphopoietin

IDQoL

Infants’ Dermatitis Quality of Life

IL

Interleukin

ILC2

Type 2 innate lymphoid cells

IP

Intraperitoneal

LC

Langerhans cells

LN

Lymph node

OFC

Oral food challenge

OVA

Ovalbumin

PA

Peanut allergy

PEBBLES

Prevention of Eczema By a Barrier Lipid Equilibrium Strategy

PGH

Pediatric Global Health

PreventADALL

Preventing Atopic Dermatitis and ALLergies in Children

POEM

Patient Oriented Eczema Measure

PROMIS

Patient Reported Outcome Measurement Information System

RCT

Randomized controlled trial

S. aureus

Staphylococcal aureus

S

Sphingosine

SC

Stratum corneum

SCORAD

SCORing Atopic Dermatitis

SDI

Shannon Diversity Index

SEB

Staphylococcus enterotoxin B

sIgE

Allergen-specific IgE antibodies

SPT

Skin prick testing

TARC

Thymus and activation-regulated chemokine

TEWL

Transepidermal water loss

Th

T helper cells