2.1 Early food introduction is only efficient to prevent allergy to this specific food

Clinical trials show that early introduction of food allergens only works for the specific foods that were introduced. In the Learning Early About Peanut Allergy (LEAP) study, patients who were early introduced peanut showed a reduced prevalence peanut allergy at 60 months compared with patients who avoided peanut (4.8% vs. 18.6%, respectively, p < .001)¹⁰ but similar rates of IgE sensitization to other foods than peanut.¹¹ In other words, early introduction of peanut may prevent only peanut allergy, but will not prevent the development of almond or hazelnut allergy. One consequence is the theoretical need to introduce all possible food allergens as early as possible in infants, which is not physically feasible. Even focusing only on the main allergens can lead to large amounts of different foods and be an additional source of stress for parents, who are already dealing with the transition from breastfeeding to solid food.¹²

Participants from the Enquiring About Tolerance (EAT) study, a RCT that evaluated the impact of early introduction of six allergens on FA, reported the difficulties they experienced with the study diet.¹³ Three main issues emerged, illustrating the practical difficulties met in the early introduction of solid foods: (1) infant refusal, (2) concerns about allergic reaction, and (3) practical problems (related to lifestyle and food preparations).¹³ These difficulties are reflected in the poor adherence to the study protocol in the early-introduction group, compared with the standard-introduction group (42.8% vs. 92.9% of the participants, respectively).¹⁴ As a consequence, in the intention-to-treat analysis, there was no difference in the frequency of FA to one or more of the six intervention foods between the standard-introduction group and the early-introduction group (p = .32).¹⁴ Conversely, in the per-protocol analysis, the prevalence of any FA was significantly lower in the early-introduction group than in the standard-introduction group (2.4% vs. 7.3%, p = .01).¹⁴

2.2 Early oral introduction of food allergens is not enough, regular intake is also required

Recent guidelines for the early introduction of solid foods had an impact on parental behavior, but recent data question their effectiveness in real life. As an illustration, a study showed that the incidence of peanut allergy in Australia was similar before and after the modification of guidelines regarding the time of peanut introduction in children (3.1% to 2.6%, respectively; difference: −0.5% [95% CI: −1.4% to 0.4%]; p = .26).¹⁵ This suggests that the age of introduction of food allergens may not be the only relevant factor affecting the effectiveness of the intervention. The same team showed that the frequency of peanut consumption greatly varied between children, from once to more than four times a week.¹⁶ This suggests that a regular intake is needed once a food allergen is introduced to maintain an immune tolerance. This concept has also been described in patients who spontaneously outgrew food allergies¹⁷ or after food immunotherapy, where regular consumption appeared needed to maintain the tolerance state.¹⁸ It was later shown in the context of primary prevention for cow's milk and peanut.^19-21 Despite the lack of RCT-based evidence, this concept is largely supported by clinical experience and the Canadian Society of Allergy and Clinical Immunology (CSACI) published a statement highlighting the importance of a regular ingestion of food allergens to prevent FAs.²² That being said, the minimal frequency of consumption to maintain oral tolerance, as well as the time window during which this exposition must be maintained, is unknown. Furthermore, if regular consumption appears feasible for certain foods such as cow's milk, others could be more problematic in young children, including nuts or shellfish.

3.1 Evidence that skin is a route of sensitization to food allergens

At the time of solid food introduction, a significant part of children is already sensitized to food allergens. That has been illustrated in clinical trials exploring the role of early introduction of food allergen in preventing FA, in which participants were tested by skin prick tests (SPT) and/or food-specific IgE (sIgE). For instance, among the 834 patients aged 3–11 months screened to participate in the LEAP study, 76 (9.1%) had a SPT to peanut ≥4 mm without ever having consumed orally peanuts.²³ Of the 163 atopic children aged 3–4 months and exclusively breastfed who were included in a study from Nishimura et al., 56.3% in the “food introduction” group and 61.4% in the “avoidance group” were sensitized to at least one food allergen, defined by sIgE > 0.1 kU/L.²⁴ Other studies reported similar findings,^{14, 25} including in non-atopic cohort.¹⁴ These data challenge the concept of a window of opportunity for the oral introduction of food allergen, or at least highlight the limit of the concept and the need for complementary strategies in preventing FAs.

The hypothesis of skin as a possible route of sensitization to food allergens has emerged many years ago. That was strongly supported by the fact that atopic dermatitis (AD), a chronic inflammatory skin disease, is a major risk factor for FA and is a major determinant of atopic trajectories.²⁶ As an illustration, a recent meta-analysis found that 49.8% (95% CI: 44.4–55.1) of children with AD had an IgE sensitization to at least one food allergen and that 31.4% (26.9–36.1) an IgE-mediated FA,²⁷ while a systematic review determined that having AD is a strong predictor of food sensitization at 3 months (odds ratio [OR]: 6.18, 95% CI: 2.94–12.98).²⁸

An ex vivo experiment also demonstrated that the stimulation of circulating memory T cells isolated from peanut-allergic children with peanut extract induced the predominant proliferation of peanut-specific CD4⁺ T cells expressing the skin-homing marker CLA (cutaneous lymphocyte antigen), whereas similar stimulation of memory T cells from non-allergic controls induced the predominant proliferation of memory T cells expressing the intestine-homing marker α4β7 integrin.²⁹ Another study demonstrated that peanut-sensitized children had a greater percentage of CLA-expressing effector T cells than non-sensitized children.³⁰ These results suggest that in sensitized children, the initial immune response to peanut occurred in the skin whereas in tolerant children, it occurred in the gastrointestinal tract. That supports the dual allergen exposure hypothesis, according to which oral exposure to food promotes immune tolerance, whereas exposure through other epithelium such as skin promotes allergies.

Sensitization through the skin requires the presence of food allergens in the environment. This presence is allowed for instance by consumption of food allergens by other people living in the same house, which has been found to be a risk factor of FA.³¹ One explanation for this is that house dust may carry food allergens in sufficient quantities to promote IgE sensitization.^32-34 The risk of sensitization is also correlated with the concentration of peanut protein found in the home environment.³⁵ In consequence, even in the absence of consumption by children, food allergens are found in bedding and play areas.³⁴ Skincare products are other possible sources of food allergens, especially those labeled “natural” or “ecological” that may contain food-derived proteins.³⁶ The use of such skincare products has been shown to be associated with an increased risk of FA,^37-39 and children with AD should avoid such type of products.

3.2 The skin barrier dysfunction promotes sensitization to food allergens

One important role of the skin is to serve as a barrier to limit the penetration of environmental allergen into the organism.⁴⁰ The barrier function of the skin involves chemical, physical, mechanical, immune, and microbial aspects that can be disturbed in diseases, such as in AD.⁴⁰

Neonates and infants have a more permeable skin barrier than adults.⁴¹ Transepidermal water loss (TEWL), an indirect measure of skin barrier function, has been shown to improve progressively in the first years of life, reaching adult values by 5 years of age.⁴¹ An increased TEWL at birth or during infancy, suggestive of skin barrier dysfunction, has been shown to be predictive of AD onset.^{42, 43} Newborns also have a higher skin pH than adults (>6.6 vs. <5, respectively).⁴⁴ Acidic skin pH, which is an important determinant of a competent skin barrier function by maintaining the commensal microbiota, regulating epidermal differentiation and lipidic barrier stabilization and other mechanisms, is obtained after several weeks.⁴⁵ The epidermis of both children and adults with AD has a higher pH compared with healthy individuals.⁴⁶

These observations support a place for an early emollient-based intervention to optimize the skin barrier function in newborns. The use of emollients was historically aimed at preventing the occurrence of AD flare-ups. Recently, it has been proposed that application of emollients on newborn skin rapidly after birth could help to prevent not only AD but also the atopic comorbidities. This approach was substantiated by animal models where the application of allergen on skin that had undergone barrier disruption through tape-stripping or detergent application led to type-2 immune responses and production of allergen-specific IgE, as well as anaphylactic reaction on subsequent oral exposure.^47-49

3.3 Emollients may prevent AD occurrence

Recent RCTs have evaluated the impact of early application of emollient to prevent the occurrence of AD.⁵⁰ These studies had heterogeneous methods, with several assessing the cumulative incidence of AD while on emollient treatment (meaning that participants were preventively treated for AD, without a wash-out period)⁵¹ and others the point prevalence of AD several months after emollient treatment (reflecting a real preventive effect of emollient application).⁵² Despite this limitation, there was a trend toward a preventive effect of emollient application from the first weeks of life on AD, and considering AD as an important risk factor of FA,⁵³ one can assume that preventing AD may subsequently prevent FA.

5.1 The nature of emollient is crucial

A retrospective study from Perkin et al.⁵⁸ reported results from a survey completed by participants of the EAT study and questioning their habits on emollient use during the first 3 months of their babies' life. Based on this questionnaire, this study suggested that the risk of FA increased dramatically with the frequency of emollient use.⁵⁸ In this study cohort of 3-month infants, the use of emollient during the first months of life was associated with the development of FA, in infants without eczema with an adjusted OR of 1.18 (95% CI: 1.07–1.30) for developing FA, and in infants with eczema with an adjusted OR of 1.20 (95% CI: 1.11–1.31).⁵⁸ However, emollients used by participants were various and not comparable. Many of them were also known to increase the penetration of exogenous chemical into the skin, such as olive oil and other plant oils,⁵⁹ which were the products the most frequently used as emollients in the study.⁵⁸

Moreover, several skincare products, whose some of which are considered as emollients, are also known to induce or aggravate skin barrier defects. In mouse models with normal skin, commercial products claimed to benefit from “dry and damaged skin” or “hypersensitive skin” increased TEWL and skin pH, and decreased the skin hydration after 4 days of twice-daily applications.⁶⁰ A clinical study performed in volunteers with a previous history of AD demonstrated that an emollient commonly used (including in the Perkin et al. study⁵⁸), the Aqueous cream BP, induced significant damage to the skin barrier as demonstrated by the TEWL increase and a concomitant alteration in the stratum corneum integrity.⁶¹ That could be explained by the fact that the Aqueous cream BP contains a surfactant, the sodium lauryl sulfate (SLS), to create an oil-in-water emulsion that is known to enhance the skin penetration of exogenous chemical and to be an irritant.^{62, 63}

Reassuringly, some emollients obviously improve barrier functions, but the ability to reduce TEWL can vary between products even when following comparable application protocols.^{64, 65} The emollient that has been shown to have the best effect on the skin barrier is a ceramide-based emollient⁶⁶ that had been shown to reduce TEWL more drastically than other emollients⁶⁴ and, consistently, was shown to have an important preventive effect on FA occurrence.⁶⁷

5.2 Emollients should be applied at least once a day, on the entire body, to prevent food allergies

Barrier defect affects the entire skin in atopic babies and children, including both lesional and nonlesional skin, as demonstrated in patients with AD.^{68, 69} This is why it is usually recommended to apply emollient on the entire body to prevent AD flare-ups and not only on specific sites. Moreover, studies exploring the effect of emollients on the skin barrier usually required at least one application per day^{65, 70} and a recent study even showed that an emollient applied twice daily may have a stronger impact on AD prevention than applied once a day.⁷¹ However, some protocols deviate from these practices without any evaluation of their equivalence in terms of skin barrier repair. For instance, the protocol used in the PreventADALL study was bathing infants for a maximum of 10 min with the addition of emulsified oil (paraffinum liquidum and trilaureth-4-phosphate) and application of an emollient only on the entire face, at least 4 days per week.⁵⁴ In addition, it was previously described that emulsification of lipids with water reduces their moisturizing effect.⁷⁰ In the study from Dissanayake et al., the protocol consisted in the application of an emollient “particularly on the cheeks and the peri-oral area,” not on the entire body (this was allowed, but not recommended).⁷² In contrast, in the Lowe et al. study as well as in the STOP-AD trial, the ceramide-based emollients were applied twice a day, all over the children's body.^{52, 67}

The discrepancies between the interventions used in the various trials (Table 1) may offer plausible explanations for the different conclusions. Future trials should ensure the use of emollient-based strategies with demonstrated ability to restore and maintain the barrier function and avoid recourse to skincare products that aggravate skin damage. This may include emollient reducing the skin pH⁴⁶ and delivering ceramide, free fatty acids, and cholesterol.⁷³ Unfortunately, few emollients currently offer these features and those that do are generally expensive, which is a major limitation in the context of prevention.