Gut Microbiota and the Use of Probiotics and Prebiotics

Data on microbial profiling and its relationship to disease are still not sufficiently detailed to consider specific microbiota modifications as a food allergy prevention strategy. However, some emerging data suggest that changes in microbiota could influence food allergies, offering further support for the microbial exposure hypothesis (West et al., 2015).

Marrs et al. included five studies in their systematic review that investigated characteristics of gut microbiota, two of which used food challenge outcomes and three that used food sensitization parameters (Marrs et al., 2013). The two manuscripts that ranked highest in quality and measured food allergy were from the same study of Spanish infants who were diagnosed with IgE-mediated cow milk allergy by milk challenge at a tertiary referral center. Differences in microbiota were identified but unfortunately none of the results was adjusted for diet. The Marrs review also included 11 RCTs in which microbial supplementation was the intervention as a potential prevention or treatment of food allergies or sensitization. Although the quality varied, the two highest quality studies that measured food allergy by OFC to assess whether microbial supplementation may be used to prevent or treat food allergies or sensitization found no benefit.

More recent data originate from the Canadian Synergy in Microbiota (SyMBIOTA) study, part of a larger Canadian research effort on the microbiota. This large 6-year longitudinal study is using metadata and samples from the Canadian Healthy Infant Longitudinal cohort to discern relationships between infant fecal microbiota and each of a group of factors, including antibiotic use, pets, and food sensitization (Kozyrskyj, 2015). Their data suggest that lower species richness in microbiota of infants (N=166, ages 3 and 12 months) might be a predictor of food (i.e., for egg, milk, and peanut) sensitization (SPT at age 12 months), even when adjusting for birth delivery mode, antibiotic use, or breastfeeding (Azad et al., 2015). Their research also revealed that sensitization occurred after the changes in microbiota diversity and richness, two commonly used indexes. Therefore, this ratio could potentially be used as a predictor of food sensitization, a potential surrogate for food allergies. Each quartile increase in richness at 3 months was associated with a 55 percent reduction in risk for food sensitization by 1 year (adjusted odds ratio [aOR] 0.45; 95% confidence interval [CI]: 0.23-0.87).

One meta-analysis of 10 RCTs (Kong et al., 2014) reported no significant difference in the incidence of food allergies comparing prenatal and postnatal probiotics supplementation with placebo or control. However, the food allergy assessments were not described in the meta-analysis. The World Allergy Organization (WAO) has recently conducted a systematic review on the relationship between supplementing the diet of pregnant or lactating women or infants with probiotics and allergy diseases. Six trials explored the relationship with food allergies but none of them made the direct comparison of probiotics versus no probiotics in pregnant women or in breastfeeding women for prevention of allergy in their children. None of the trials found differences in food allergy with probiotic supplementation (Cuello-Garcia et al., 2015). Two additional observational studies found during the committees' evidence-based search did not find an association between the addition of probiotics to infants' diets (Loo et al., 2014; West et al., 2013). The most recent work on the effect of prebiotics in food allergy, also conducted by the WAO (Cuello-Garcia et al., 2016), is a guideline that seems to be based on a systematic review. The methods of systematic review, however, were not fully reported and no other source or citation to the systematic review was found. The guideline is based on studies investigating the relationship between prebiotics consumption by women during pregnancy or lactation and by healthy infants for preventing various allergic symptoms, including food allergy. Only one intervention study assessed the risk of developing food allergy in infants consuming an infant formula containing oligosacharides (Ivakhnenko and Nyankovskyy, 2013). That study (N=240) found that infants who had been fed with breast milk or oligosaccharide-supplemented infant formula had significantly fewer allergic reactions to food products compared to the infants fed the standard formula (3.92 percent and 4.84 percent versus 16.98 percent, respectively; P<0.05).

The committee concludes that, at this time, only a few studies have been conducted on the relationship between changes in the microbiota and food sensitization and, therefore, the evidence supporting this relationship is limited. RCTs on probiotic and prebiotics supplementation are few and have methodological limitations. Therefore, the committee concludes that the evidence is limited and does not yet support a decrease in food allergy risk from the use of probiotics or prebiotics by pregnant and lactating women or by infants. Additional research would be needed before recommending the use of prebiotics or probiotics to prevent the onset of food allergies.

Route of Delivery

The composition of the gut microbiota is influenced by route of delivery. Vaginally-delivered infants harbor bacterial communities resembling their mother's vaginal microbiota. In contrast, infants delivered by cesarean section have bacterial communities similar to those found on the skin surface (Dominguez-Bello et al., 2010). In light of the fact that the gut microbiome plays a central role in the development of immune regulation and oral tolerance, it is not surprising that investigators have examined the question of whether caesarean delivery increased the risk of food allergy.

In their systematic review, Marrs et al. identified 13 publications. Of these, five identified food allergy through OFCs. All 13 publications, except for the study of lowest quality, reported an increased risk of developing food allergy or food sensitization in children delivered by cesarean section (Marrs et al., 2013). Six of these associations were significant. However, only two included clinical food allergy diagnoses. Of the studies included for review, these two studies yielded the highest quality data. The studies used 2,803 consecutive mother-infant pairs from a Norwegian birth cohort surveyed at 12, 18, and 24 months. When children were challenged with food orally using open or double-blind protocols, cesarean section was associated with a significantly higher risk for cow milk allergy. This occurred only in the subgroup of children with atopic mothers, however (aOR: 9.6 [95% CI: 1.8-52.4]) (Eggesbo et al., 2005). They also observed a nonsignificant 60 percent increase in egg allergy risk up to age 2 years (Eggesbo et al., 2003).

The Marrs review also included a prospective nested case-control study of 16,237 infants in Finland, ages 0 to 2 years (Metsala et al., 2010). Infants whose parents had received a reimbursement for the cost of specialized formula based on diagnosis of cow milk allergy were recruited, and the allergy was certified by a pediatrician using clinical exam, symptoms, elimination diet, SPT, and elevated sIgE or open challenge test (Metsala et al., 2010). Controls were randomly selected infants who were matched for age, sex, and delivery hospital. A significant relationship between cesarean delivery and cow milk allergy was observed (aOR: 1.18; 95% CI: 1.10-1.27).

Lodge et al. conducted a more recent review of systematic reviews and found two systematic reviews that included six original studies (Lodge et al., 2013). An association between cesarean section delivery and increase in food allergy is seen in only the three smallest studies. Two of these studies used specific IgE to food allergens as the outcome measurement. No conclusion was reached by the authors due to methodological flaws (i.e., small size studies or inaccurate food allergy measurement).

Since the Marrs' systematic review, six prospective cohort studies investigating associations between cesarean delivery and allergy risk have been published. They include studies conducted in Australia (Peters et al., 2014), France (Pele et al., 2013), the United Kingdom (Grimshaw et al., 2014), the United States (Luccioli et al., 2014; McGowan et al., 2015), and a five-country study (Depner et al., 2013) totaling 25,688 cases and controls. Overall, these studies found no significant associations between cesarean delivery and a variety of food allergies. The age of the children in the studies ranged from 0 to 5 years, and most included physician-diagnosed food allergy. Minimum criteria for diagnosis were sIgE to food allergen or a positive SPT. However, Luccioli et al. used physician diagnosis based on parental report (Luccioli et al., 2014). The largest study was the Australian HealthNuts Study (Peters et al., 2014), which recruited 5,276 infants at immunization clinics. These infants (2,848 of the total recruited) were investigated for open challenge-proven egg, peanut, and sesame allergy. However, no significant association was demonstrated with mode of delivery (Peters et al., 2014). Two retrospective case-control studies from Finland (N=3,181) (Pyrhonen et al., 2013) and the United States (N=291) (Dowhower Karpa et al., 2012) also did not show an association between cesarean delivery and food allergy.

The variation in association between mode of delivery and risk of food allergy may be partly explained by the fact that some studies have been unable to distinguish between whether cesarean delivery had been done on an elective or emergency basis (e.g., Koplin et al., 2012a; Peters et al., 2014). Emergency cesarean delivery is generally associated with rupture of membranes. As a result, the baby has some exposure to vaginal commensal bacteria during labor. However, the exposure is not usually to the same extent as vaginal delivery. However, because the proportion of emergency cesarean deliveries is usually relatively small compared to elective cesarean deliveries, we would still expect to see some association between mode of delivery and food allergy. This would be true even in those studies that could not differentiate emergency from elective cesarean deliveries, particularly in the larger and better powered studies. It also should be noted that the association between cesarean delivery and allergic risk could be misinterpreted due to the potential for reverse causation similar to breastfeeding.

Only a few observational studies have been conducted on the relationship between food sensitization or food allergy and cesarean delivery. The studies have methodological limitations. Therefore, the committee concludes that, at this time, evidence to support an increased risk for food sensitization or food allergy due to giving birth by cesarean delivery is limited. Strong evidence is unlikely to be forthcoming because of the ethical inability to randomize a population to deliver a baby by cesarean section. However, additional prospective research studies are needed.

Antibiotic Use

Antibiotics are known to cause short-term and, in some cases, lasting alterations in the microbiota (Faa et al., 2013). Infants can be exposed to antibiotics pre-, peri-, or postnatally as individual exposures or multiple exposures across this time, when the microbiome is not well established and is more susceptible to perturbations. The Marrs et al. systematic review reported no relationship between antenatal or postnatal antibiotic exposure and increased risk of food allergy (Marrs et al., 2013).

Since 2012, two prospective cohort studies of food allergic children have been published that were not included in the Marrs systematic review (Marrs et al., 2013). Studies in Finland (Metsala et al., 2013) and the United Kingdom (Grimshaw et al., 2014) and one retrospective case control study from the United States (Dowhower-Karpa et al., 2012) investigated associations between antibiotic exposure and food allergy risk. In those infants whose mother used antibiotics before or during pregnancy, respectively, the Finnish prospective, nested case-control study (N=16,237) reported a statistically significant 26 percent (aOR: 1.26; 95% CI: 1.20-1.33) and 21 percent (aOR: 1.21; 95% CI: 1.14-1.28) increased risk for cow milk allergy (determined by OFC) (Metsala et al., 2013). An even greater risk of cow milk allergy (aOR: 1.71; 95% CI: 1.59-1.84) was reported in infants who were treated with antibiotics between birth and 1 month of age (Metsala et al., 2013).

However, two other studies described below showed no statistically significant association. Cases (N=41) and controls (N=82) in the UK study were drawn from the Prevalence of Infant Food Allergy (PIFA) study (Grimshaw et al., 2014). Children in this study were part of the larger EuroPrevall birth cohort. Food allergy was diagnosed using SPT, physical exam, clinical history, sIgE, and DBPCOFC. Maternal antibiotic use during or after pregnancy or during breastfeeding was not associated with increased risk of food allergy in the infant. However, administration of the antibiotic to the infant was not assessed (Grimshaw et al., 2014). In a retrospective case (N=99) control (N=192) design, Dowhower Karpa et al. found no association between peripartum or neonatal antibiotic exposure and food allergy, diagnosed by positive sIgE or SPT (Dowhower Karpa et al., 2012).

Thus, taking together the results of the Marrs systematic review (Marrs et al., 2013) and the three studies published since, only one study (Metsala et al., 2013) has reported a link between antibiotic use and food allergy. The strengths of that study is the large sample size (more than 16,000 children) and the prospective design. However, additional studies are needed to conclusively demonstrate a link between antibiotic use in early life and food allergy risk.

Only a few studies have explored the relationship between food allergies and antibiotic use. The committee concludes that evidence from observational studies suggesting a link between antibiotic use in early life and food allergies is limited. Additional studies with information on the type and dose of antibiotic, the timing of exposure along the perinatal continuum, and whether the infant is repeatedly exposed are needed to conclusively demonstrate a link with food allergies.

Animal Exposure

As noted above, the premise of the “Hygiene” and “Old Friends” hypotheses is based on the concept that the lack of early childhood exposure to infectious agents, symbiotic microorganisms, and/or parasites increases susceptibility to allergic diseases and asthma by suppressing the natural development of the immune system (Strachan, 1989).

The Marrs review reported on four studies investigating associations between farm and animal exposure and food allergy (Marrs et al., 2013). In their review, only the HealthNuts Study supported the microbial hypothesis. The study reported data on risk of pets and siblings for the development of challenge-proven egg allergy (Koplin et al., 2012a). It also assessed the role of these factors on any food allergy using latent class analysis, a sophisticated analytical epidemiological method (Peters et al., 2015). Marrs et al. also reported findings from the European Protection against Allergy Study in Rural Environments (PASTURE), which described a cohort of families living in proximity to farm animals in rural settings (Marrs et al., 2013). This study showed significantly less food sensitization in the cord blood of mothers who consumed raw cow milk (versus boiled milk) in the perinatal period. However, the authors applied a lower cutoff for sIgE concentration than is conventionally used (>0.2 versus 0.35 IU/ml), which may have overestimated the incidence of food sensitization (Ege et al., 2008).

Since 2012, several prospective cohort studies have investigated whether exposure to farm animals (Depner et al., 2013; Pele et al., 2013) or pets (Goldberg al., 2013; Grimshaw et al., 2014; Martin et al., 2015; Peters et al., 2015; Stelmach et al., 2014) influenced the risk of food allergy or food sensitization. Depner et al. performed an additional analysis of data from 686 children in the rural European PASTURE cohort (Depner et al., 2013). Again using sIgE as their diagnostic criterion for food sensitization, they explored the more traditionally used sIgE cutoff of 0.35 IU/ml compared to 0.2 IU/ml in their previous study by Ege et al. (2008). They found that allergen-specific IgE levels rarely exceeded 0.35 IU/mL (<3% of all children) at age 1 year and the 95th percentiles at 1 year were consistently less than 0.7 IU/mL (RAST class 2) for any IgE. The only exception was cat (1.3 IU/mL) (Depner et al., 2013). They also found that early life exposure to farm animals, such as sheep, goats, and rabbits, did not confer protection against food allergen sensitization. However, exposure to farming increased (P=0.0015) the risk of food allergen sensitization (aOR: 2.11; 95% CI: 1.33-3.34). A total of 793 (378 farm and 415 nonfarm) children were included in the analyses. Pele et al. also reported no effect of farm animal contact on food allergy incidence in more than 1,400 children participating in the PELAGIE mother–child cohort. However, mold or dampness in the home increased (P≤0.001) the incidence of food allergy (23.9% versus 8.8%, yes versus no) in this cohort, as measured by parent report (Pele et al., 2013).

All other prospective cohort studies published since 2012 investigated exposure to pets. Two studies with a total of 350 children reported no association between pets in the home (Israel) and food sensitization (measured by specific IgE to cow milk) (Goldberg et al., 2013) nor an association of pet ownership (United Kingdom) with food allergy risk (measured by DBPCOFC or convincing history of anaphylaxis) (Grimshaw et al., 2014). In contrast, Stelmach et al. reported an increased risk of food allergy based on diagnosis by a doctor following international guidelines (aOR: 1.48; 95% CI: 1.02-2.16) associated with pets in the home during pregnancy in a cohort of 501 children from the Polish Mother and Child Cohort Study (REPRO_PL cohort) (Stelmach et al., 2014).

Two studies from the HealthNuts cohort, a prospective, population-based cohort of 5,276 infants age 12 months in Melbourne, Australia, investigated whether direct exposure to pets (Koplin et al., 2012a; Peters et al., 2015) or the co-incidence of eczema (Martin et al., 2015) moderated the effect of pets on food allergy risk. Koplin et al. examined the relationship between environmental and demographic factors and egg allergy, the most common food allergy in infants and young children (Koplin et al., 2012a). Using SPT to egg white and oral food challenge at 12 months revealed that children with a pet dog at home (dog ownership ascertained by questionnaire) were less likely to develop egg allergy than those without a pet dog at home (aOR: 0.72; 95% CI: 0.52-0.99). Peters et al. observed that, compared to not having a dog in the home, having a dog significantly reduced the risk of multiple food allergies (including peanut) by 60 percent (aOR: 0.4; 95% CI: 0.21-0.73), whereas having a dog that was kept outside only (versus no dog) provided no protection. In this latter scenario, a significantly increased risk in egg allergy was actually observed (aOR: 1.56; 95% CI: 1.1-2.21) (Peters et al., 2015). Within the same cohort, Martin et al. compared the effect of dog or cat exposure on infants with (N=2,795) or without (N=1,903) eczema (Martin et al., 2015). Having a dog reduced the risk of food allergy in infants with eczema (aOR: 0.7; 95% CI: 0.5-0.9), but not in infants without eczema. A similar effect on food allergies was observed for infants with (aOR: 0.6; 95% CI: 0.4-0.9) or without eczema in homes with cats (Martin et al., 2015).

Results from studies exploring the relationship between animal exposures and food allergies are inconsistent. The few observational studies related to living on a farm found that exposure to farm animals offers no protection against food allergies. Also, from observational studies, the committee concludes that evidence is limited regarding the potential for a close interaction with a pet being more protective against a food allergy than pet ownership in general or having a pet who is restricted to outside the home. Further studies should be conducted on the nature of the association between exposure to farm animals or pet ownership and food allergies.

Exposure to Antigen Through Maternal Diet During Pregnancy or Lactation

Maternal diet during pregnancy and lactation has been of great interest in understanding the etiology of food allergies in offspring. The fetal programming hypothesis supports the idea that the maternal diet has long-term influence on children's health (Barker, 1990; Langley-Evans, 1997). Its application to food allergies would suggest that consuming specific allergenic foods during this critical period might be associated with the development of the immune system in utero that may later manifest itself as food allergies over the life course, given specific childhood exposures. Results from two prospective cohort studies (Bunyavanich et al., 2014; Frazier et al., 2014) (total N=9,482 mother–child pairs) show that a higher consumption of allergenic foods before or during pregnancy (e.g., peanut), as measured by a food frequency questionnaire, was associated with a reduced risk of having a child with food allergies. This finding supports the fetal programming hypothesis. The HealthNuts Study also assessed the role of allergen avoidance in pregnancy and lactation and the risk of challenge-proven egg allergy and found no association (Koplin et al., 2010). Another recent prospective cohort study (Pele et al., 2013) reported an association between maternal pre-pregnancy consumption of shellfish and food allergy (1.62; 95% CI: 1.11-2.37). However, this study assessed food allergy by parental report. Randomized studies on this subject have involved the elimination of certain allergenic foods as opposed to increasing their consumption among primarily high-risk families. Kramer and Kakuma conducted a high-quality systematic review that included three RCTs of foods avoided during pregnancy and/or lactation and the outcomes of egg and milk sensitization (but not food allergy itself) among women at high risk of having an atopic offspring (Kramer and Kakuma, 2012). In two of the RCTs (Falth-Mangnusson and Kjellman, 1987; Lilja et al., 1988) (total N=334), women either avoided or decreased their intake of cow milk and eggs beginning in the third trimester of pregnancy and this was associated with a nonsignificant reduction of egg sensitization in their infants at 6 months, but not at 18 months. Sensitization for cow milk allergy was not reduced at either time point. The remaining RCT (Appelt et al., 2004) (total N=497) had women totally avoid peanuts, nuts, and fish as well as decrease their intake of cow milk and eggs beginning in the third trimester through 1 year postpartum. This study found no significant associations with milk or peanut sensitization in offspring at age 1, 2, or 7 years. However, for egg sensitization, an increased risk was seen at age 2 years only (1.91; 95% CI: 1.03-3.5). This trial is published in abstract form only, with no details on the randomization being available.

Another recent systematic review by de Silva et al. found seven high-quality studies on maternal diets and also concluded that “overall, the evidence is not strong enough to recommend changing the diet or supplements of pregnant or breastfeeding women” to prevent food allergies in infants at normal or high risk of food allergies (de Silva et al., 2014).

The committee concludes that, to date, study findings provide limited evidence to support or discourage eliminating allergenic foods from the diet of pregnant or lactating women at high risk of having a child with allergies. Because the evidence about the benefits of consuming or eliminating allergenic foods during pregnancy and lactation is not clear, additional RCTs are warranted before providing advice in this regard. Studies exploring the effect on the development of food allergies in children of intake of allergens by the mother are in progress.

Breastfeeding

Breastfeeding is an important early life factor that determines an individual's gut microbiota and likely indirectly modulates immune responses. In addition, breastfeeding transfers bioactive compounds from the mother to the child that can also influence immune responses. However, the evidence assessing any potential link between breastfeeding and food allergies risk is not clear. Systematic analysis of observational studies on the protective effect of breastfeeding have shown conflicting results, and many of the studies included were conducted decades ago when food allergy was uncommon and methods of assessment were limited (Grimshaw et al., 2009). Most systematic reviews have failed to find a specific beneficial effect of breastfeeding on food allergy or food sensitization (de Silva et al., 2014; Kramer and Kakuma, 2012). Moreover, two cohort studies reviewed in de Silva et al. (2014) suggested that exclusive breastfeeding for 8 weeks did not reduce the risk of cow milk allergy (measured by parents report followed by SPT and oral food challenge) (Saarinen et al., 1999) and breastfeeding for 5 months or more may increase the likelihood of sensitization to egg in infants at high risk of atopy, although food allergy was not assessed (Wetzig et al., 2000). Importantly, the apparent negative effects of extensive breastfeeding may relate to the delayed introduction of first complementary foods rather than the effects of breast milk per se (see the section “Dual Allergen Exposure Hypotheses” in this chapter). Alternatively, these recent findings of increased risk of breastfeeding may simply be a misinterpretation of the data related to the reverse causation phenomenon (see the section “Methodological Limitations” in this chapter). One study found that the effects of breastfeeding on food sensitization can be modified by genetic variants relevant to allergic diseases (Hong et al., 2011).

Lodge et al. undertook a systematic review to assess the role of breastfeeding in food allergy (Lodge et al., 2015). The review included nine cohort and four cross-sectional studies. The numbers of participants ranged from 163 to 21,766 (cohort studies) and from 1,278 to 13,110 (cross-sectional studies). No association with food allergy was found for more versus less⁹ breastfeeding in the pooled estimate (6 cohort and 6 cross-sectional), although study heterogeneity was high. Various sub-analyses failed to find any protective association of breastfeeding for food allergy. The primary issue concerning the quality of these studies was the poor accuracy of outcome assessment. Only two studies used OFCs, the recognized gold standard for food allergy diagnosis; most studies relied on parental report of symptoms or on physician diagnosis.

The committee's review of the evidence found eight studies (seven cohort and one cross-sectional) that explored breastfeeding as a food allergy risk determinant. Although Ivakhnenko and Nyankovskyy showed that infants (N=240) who were breastfed had significant risk of developing an allergy to cow milk protein and had gastrointestinal symptoms of food allergy by age 18 months compared with those who were fed standard infant formula, the risk of bias of this trial was high due to unclear definitions and diagnoses of food allergy outcomes and high dropout rates (Ivakhnenko and Nyankovskyy, 2013). Two studies performed only unadjusted analyses so the results (mostly no significant associations) are likely to be confounded (Grimshaw et al., 2014; McGowan et al., 2015). The other four cohort studies showed associations between longer duration of breastfeeding (any or exclusive) and a lower risk of developing cow milk sensitization (Liao et al., 2014; N=258), food allergy (Stelmach et al., 2014; N=501; aOR: 0.88; 95% CI: 0.82-0.95), or multiple food allergy (predominantly egg) (Peters et al., 2015; N=5276; aOR: 1.17; 95% CI: 1.09-1.24) after adjusting for potential confounders. The single cross-sectional study did not find a significant association between exclusive breastfeeding (poorly defined) and food allergies among children (N=386) ages 0 to 18 years with atopic dermatitis. Luccioli et al. collected data from prospective cohort of children (N=1,363) who participated in the Infant Feeding Practices Study (IFPS) II and also found no significant relationship between breastfeeding for various periods and food allergies (Luccioli et al., 2014). Only some studies used OFC as an outcome measure (Grimshaw et al., 2014; Peters et al., 2015). The single cross-sectional study did not find a significant association between exclusive breastfeeding (poorly defined) and food allergies among children, ages 0 to 18 years, with atopic dermatitis (Mailhol et al., 2014).

As mentioned above, investigation of the role of breastfeeding in allergic disease is particularly prone to confounder bias because families who are at high risk of allergy are more likely to breastfeed, as recommended by some guidelines. In addition, the composition of human milk changes from colostrum to late lactation and throughout the day, and differs from mother to mother (Ballard and Morrow, 2013) and could therefore affect health outcomes of the child. Compounding the difficulties in this area is the inability to randomize to a nonbreastfeeding arm, as this would be unethical given the many well-established benefits of breastfeeding, such as protection against some chronic diseases, obesity, and infections.

The committee concludes that due to inconsistencies in results from prospective studies, the evidence that breastfeeding is protective against food allergies is limited. Strong evidence is unlikely to be forthcoming because of the ethical inability to randomize a population to breastfeeding alternatives. However, additional well-designed prospective research studies in infants at low and high risk for food allergy are needed.

Types of Infant Formula

Significant interest has been expressed in the use of modified infant formulas—especially partially hydrolyzed formulas (PHF), which include longer cow milk peptides, and extensively hydrolyzed cow's milk formulas (EHF), which include di- and tri-peptides derived from cow milk protein—as a way to avoid allergen exposure and prevent early childhood allergic disease. As a result of demand from families with a history of allergy seeking readily available primary prevention interventions, industry has responded with the development of a variety of “allergy prevention” formulae, and expert bodies have provided recommendations regarding their use for preventing allergies. Some infant feeding guidelines have recommended that hydrolyzed formula can be considered as primary prevention therapy for some allergic diseases. In the United States, a policy statement from the American Academy of Pediatrics indicated that in studies of infants at high risk of atopy, modest evidence supports the delay or prevention of onset atopic dermatitis by the use of hydrolyzed, and particularly extensively hydrolyzed, formulas (Greer et al., 2008). In Australia, the Australasian Society of Clinical Immunology and Allergy Guidelines: Infant Feeding and Allergy Prevention no longer recommends hydrolyzed formulas as primary prevention therapy for allergic diseases. The guidelines now state, “Based on a recently published review of studies (Boyle et al., 2016), no consistent convincing evidence supports a protective role for partially hydrolyzed formulas (usually labelled ‘HA' or Hypoallergenic) or extensively hydrolyzed formulas for the prevention of eczema, food allergy, asthma or allergic rhinitis in infants or children” (ASCIA, 2016b).

A Cochrane review supports the use of hydrolyzed formula to prevent allergy in high-risk infants who are unable to be completely breastfed but not for those infants who can breastfed (Osborn and Sinn, 2006, 2009). Critics of this Cochrane review have pointed out that it suffers from small-study publication bias (i.e., scarcity of small negative studies) (Lowe et al., 2013) and thus the beneficial effect of PHF was likely overestimated. Due to the methodological concerns and inconsistency of the findings of the studies included in the review, the authors themselves recommend that further larger trials be conducted. Subsequently, new evidence from a large intervention trial of 620 high-risk infants (the Melbourne Atopic Cohort Study) has emerged. Findings from this trial challenge the effectiveness of PHF (Lowe et al., 2011).

The German Infant Nutritional Intervention (GINI) study was a trial aimed at exploring the effect of hydrolyzed formulas (compared to cow milk formula) in preventing allergic diseases in infants at high risk of atopy. Infants (N=2,252) were randomly assigned at birth to receive partially or extensively hydrolyzed whey formula, extensively hydrolyzed casein formula, or cow milk formula as milk substitute for the first 4 months when breastfeeding was insufficient. In a follow up until the children were age 6 years, hydrolyzed infant formulas prevented eczema and allergic manifestation (atopic dermatitis, food allergy, allergic urticaria, asthma, and hay fever/allergic rhinitis) (von Berg et. al., 2008). However, subsequent results showed little evidence of an ongoing preventive effect between the ages of 7 and 10 years (von Berg et al., 2013a). These more recent findings have not yet been incorporated into the Cochrane review. Likewise, the European Academy of Allergy & Clinical Immunology (EAACI) systematic review included both the Cochrane review and the Melbourne Atopic Cohort Study as well as the preliminary GINI results but did not include the latest results from the GINI study. Therefore, their conclusion supported the protective effect for PHF. Interestingly, the most recent findings from the GINI study suggest that casein-predominant EHF might be expected to have a greater biological effect than PHF because the formula is more extensively modified (von Berg et al., 2013a,b). However, most infant feeding guideline recommendations are based on the reality that PHF is both cheaper and more palatable than EHF and therefore should be considered instead of EHF. Additionally, in some countries EHF is only available with a prescription, which significantly increases costs to the health care system.

Most recently, Boyle et al. conducted a systematic review and meta-analysis of studies to determine whether feeding infants with hydrolyzed formulas reduces their risk of allergic disease (Boyle et al., 2016). Their search yielded 37 intervention trials of more than 19,000 participants, although few studies included in the meta-analysis were published in the past 10 years. For the majority of studies, infants were considered to be at high risk of allergy because a first degree relative had a history of allergic disease. Overall, the pooled data showed no significant reduction in risk of any food allergy in infants ages 0 to 4 years when they were fed EHF or PHF compared to standard cow milk formula. On concluding the review, the authors found that previous studies suffered from unclear or high risk of bias. The review also showed evidence of conflict of interest and had inadequate methods of randomization and treatment allocation (selection bias). The authors recommended that international infant guidelines should be revised to remove the recommendation that hydrolyzed formula protects against allergic disease. In addition, a review of systematic reviews also stated that evidence is insufficient to conclude that the use of hydrolyzed formulas may reduce food allergy or sensitization when compared with standard formula in children with high atopy risk, and no evidence supports hydrolyzed formulas over breast milk for prevention of food sensitization or food allergy (Lodge et al., 2013).

The committee concludes that the studies on the effects of PHF or EHF for preventing food allergies have methodological flaws and their findings are inconsistent. Therefore, evidence on the effect of PHF or EHF for the prevention of food allergies is limited. If this area were to be investigated, high-quality RCT studies on the effects of PHF and EHF to determine whether hydrolyzed infant formulas influence the onset of food allergies would be needed before the use of these formulas could be recommended for prevention.

Adequate Early Life Skin Barrier Function

It is important to note that mutations leading to filaggrin loss-of-function appear to be equally common among individuals with asymptomatic food sensitization and those with true food allergy (Tan et al., 2012a), suggesting that filaggrin confers a risk for food sensitization—the first step to food allergy—but not for food allergy itself. Previous studies reporting an association with food allergy were not designed to untangle any differential effect between sensitized tolerant and sensitized allergic individuals (Brown et al., 2011). Recent data from the Isle of Wight birth cohort used path analysis to demonstrate that the effect of filaggrin loss-of-function mutations on food allergy at age 10 years occurred indirectly through an effect on eczema and food sensitization in early childhood (Venkataraman et al., 2014). Together, these findings suggest that skin barrier function plays a role in sensitization status but not in food allergy or tolerance.

Two recent RCTs have investigated the application of daily moisturizer from birth in an attempt to reduce infantile eczema. Although the studies are small in size, the results support the idea that the integrity of the skin barrier is related to preventing food allergy. One RCT in the United States and the United Kingdom (N=124) examined the effects of an intervention that consisted of the use of an emollient at least once per day on neonates at risk of atopic dermatitis (Simpson et al., 2014). Atopic dermatitis was measured at 6 months. The results demonstrated a significant protective effect against atopic dermatitis (relative risk [RR]: 0.50; 95% CI: 0.28-0.9; P=0.017). The second trial examined the effect of using a moisturizer from the first week of life on eczema as a primary outcome and egg sensitization (but not allergy) as a secondary outcome in a group of 118 neonates at high risk of atopic dermatitis (Horimukai et al., 2014). At 32 weeks postnatal age, application of moisturizer to neonates was effective at preventing atopic dermatitis after 32 weeks, but unfortunately the trial showed no evidence of a reduction in sensitization to egg white in this relatively small study of 118 infants. However, a higher proportion of infants with atopic dermatitis showed egg sensitization compared with infants without atopic dermatitis.

The committee concludes that limited but consistent evidence on mutations on the filaggrin gene and on preventing eczema at early age suggests that impairment of skin barrier function plays a role in sensitization status as the first step on the path to food allergy.

Timing of Introduction of Solids and Infant Feeding

The dual antigen exposure hypothesis states that the second factor in the two steps to food allergy is the delay in oral allergen exposure. Until recently, delayed introduction of solids and particularly allergenic solids into the infant's diet was a strategy adopted in many countries with the aim of reducing or preventing food allergies. Although exclusive breastfeeding for the first 6 months of life has been universally recommended in all countries to promote its health benefits (WHO, 2016), as described above, no evidence indicates that exclusive breastfeeding prevents the development of food allergies.

In 2008 and 2009, specific dietary advice to avoid peanuts in the United Kingdom and the United States, respectively, was rescinded (Greer et al., 2008; NHS, 2015a) based largely on the premise that evidence was insufficient to promote avoidance as a strategy to prevent food allergies. More recent advice does not state whether infants should actively receive allergenic foods, and if so at what age. Indeed, a recent nationwide UK dietary survey showed that only 8 percent of children younger than age 1 year had consumed any foods containing peanut (McAndrew et al., 2012).

The EAACI systematic review includes three cohort studies that found that the concept of delaying solid foods or cow milk consumption until 4 months of age does not appear to confer any benefit in terms of food allergies (de Silva et al., 2014). Most recently, evidence has been accumulating about the benefits of introducing allergens early. This section will focus on the most recent RCTs that evaluate the benefits of introducing allergens early in life.

In 2008, Du Toit et al. found that the level of peanut allergy in Jewish children in the United Kingdom was 10-fold higher than that of Jewish children in Israel and that median consumption of peanut protein was 0 g per month in the United Kingdom versus 7.1 g per month in Israel¹⁰ (Du Toit et al., 2008). Based on these results, Du Toit et al. conducted a large RCT to formally assess whether early introduction of peanut prevented the development of peanut allergy at age 5 years (Du Toit et al., 2015). The LEAP (Learning Early about Peanut Allergy) study randomized 640 highly atopic children with severe eczema and/or egg allergy to either consumption or avoidance of peanut at ages 4 to 11 months and the intervention continued until the children were age 5 years. The results showed that early consumption of peanut reduced the prevalence of peanut allergy (diagnosed by DBPCOFC) at age 5 years by more than 80 percent. The reduction was effective in children who were either SPT negative or SPT positive to peanut (wheals of 1, 2, 3, or 4 mm). As 17 percent of the LEAP cohort had peanut-specific IgE ≥0.35 at entry into the study and 27 percent had detectable IgE (≥0.1 kU_A/L), prevention of peanut allergy was occurring for the majority of children after IgE sensitization had occurred; this represents secondary prevention (Du Toit et al., 2013). However, for infants in the group who were SPT negative at enrollment and who had no detectable IgE, early consumption of peanut also reduced the prevalence of peanut allergy (6 percent and 1 percent in the peanut avoidance group and in the peanut consuming group, respectively). This primary prevention strategy also was effective in a secondary analysis in children of different races.

As reviewed in Chapter 4, at the moment, we do not have definitive biomarkers to define tolerance. It is of interest that during the LEAP study, an early rise in peanut-specific IgG4 and peanut-specific IgG4/IgE ratio occurred in the peanut-consuming group (Du Toit et al., 2015). A high peanut-specific IgG4/IgE ratio was associated with protection against peanut allergy. Although the peanut-specific IgG4/IgE ratio decreased in the original peanut-consuming group during the period of peanut avoidance in the follow-up LEAP-On Study (Du Toit et al., 2016), it remained significantly higher than in the original peanut avoidance group. Interestingly in LEAP, peanut-specific IgE was not significantly different between the original peanut consuming and peanut avoidance groups throughout the study. However, peanut-specific IgE to Ara h 2 started to decline in the original peanut consuming group after 2.5 years of consumption, and continued to decline despite 1 year of peanut avoidance in that group between ages 5 and 6 years (Du Toit et al., 2016). This suggests potentially that high production of allergen-specific IgG4 may be important in the initiation of tolerance and that inhibition of IgE synthesis may be important in long-lived tolerance.

In order to determine whether early introduction of peanut was effective at preventing peanut allergy in the absence of ongoing peanut consumption, the LEAP-On Study was designed (Du Toit et al., 2016). Children (N=566) from the original LEAP cohort, irrespective of whether they were in the original peanut consumer or avoidant group, were asked to completely avoid peanut consumption for 1 year and then their peanut allergy status was determined by OFC, SPT, and specific IgE. Despite high adherence to this protocol of avoidance, the protective effects of early consumption remained and the original peanut consuming group had a 74 percent reduction in peanut allergy at age 6 years compared to the original peanut avoidant group. Another follow-up study, the LEAP Adlib Study, is currently being designed. In this trial the original LEAP participants will continue to be followed up for a 4-year period of ad libitum consumption of peanut to determine whether the effects of early introduction remain protective.

In regard to other foods, observational studies suggest that delayed introduction of egg (Koplin et al., 2010), cow milk (Katz et al., 2010), and wheat (Poole et al., 2006) are associated with an increased risk of those respective food allergies. Various trials are in progress to confirm or refute these observations (see Table 5-1). Early evidence from Koplin and Allen (2013, p. 830) “suggests that if a window of opportunity for promoting tolerance exists, it may be different for each food” (Koplin and Allen, 2013, p. 830). However, further investigation is required. In the large HealthNuts study, where egg allergy was determined by challenge (among other food allergens), it was found that early introduction (age 4 to 6 months) of hen egg in the infant's diet protected against the development of egg allergy, but introduction after 6 months of age was associated with significantly increased risk of developing egg allergy and even more so if introduced after age 9 months (Koplin et al., 2010). The results from the LEAP study (Du Toit et al., 2015) also are supported by data from the Solids Timing for Allergy Research (STAR) trial, which randomized 86 infants with eczema to egg avoidance or early regular egg consumption from age 4 months. The study found a lower, nonsignificant prevalence of egg allergy by 12 months in the intervention group (33% versus 51%; P=0.11) (Palmer et al., 2013). In a large birth cohort study conducted in Israel, IgE-mediated cow milk allergy did not occur in infants (N=13,019) who had received cow milk–based formula regularly in the first 2 weeks of life. In contrast, children who had formula milk introduced at age 3 to 4 months had the highest rate of cow milk allergy (Katz et al., 2010).

The EAT (Enquiring about Tolerance) intervention trial, which has recently been published, also examined the effects of early introduction of common allergenic foods. Unfortunately, compliance with intervention in this trial was low and the intention to treat analysis did not reveal a protective effect from early introduction of solids. In contrast, the per protocol analysis did suggest that early introduction of other common allergenic foods into the diet of infants may protect against the development of food allergies in general (Perkin et al., 2016). In the EAT study, exclusively breastfed infants (N=1,303) were recruited in the general population and randomly assigned at age 3 months to either introduction of six allergenic foods (cooked egg, peanut, cow milk, sesame, white fish, and wheat) (Early Introduction Group) or to the current recommended practice of exclusive breastfeeding until approximately 6 months of age (Standard Introduction Group). The primary outcome was determined to be food allergy between 1 and 3 years of age determined in nearly all participants by DBPCOFC. The study showed a modest and nonsignificant 20 percent overall decrease in the rate of food allergies in the Early Introduction Group (5.6 percent compared to 7.1 percent in the Standard Introduction Group). However, in a per protocol analysis, the prevalence of any food allergy was significantly lower in the Early Introduction Group compared to the Standard Introduction Group (2.4 percent versus 7.3 percent; P=0.01) representing a 66 percent reduction in the prevalence of overall food allergy. The effects were most apparent for peanut allergy in the per protocol analysis (0 percent in the Early Introduction Group versus 2.5 percent in the Standard Introduction Group; P=0.003) and for egg allergy (1.4 percent versus 5.5 percent; P=0.009). These changes also were accompanied by decreases in SPT to the foods in the Early Introduction Group. A dose–response analysis revealed that 2 g of peanut protein or egg white protein per week appeared to be most protective against these food allergies. Interestingly 2 g of peanut protein per week is the dose that was observed in the Du Toit et al. study in Israel where children appeared to be protected against peanut allergy (Du Toit et al., 2008).

The EAT study shows that early introduction of foods was safe, as the intervention group did not experience an increased number of reactions compared to the controls. However, it is difficult to make any certain conclusions from the EAT study about the efficacy of early introduction of foods, given that efficacy was seen only in the per protocol group. Although careful analysis did not show any evidence of bias that could account for these results, it is not possible to completely exclude unmeasured bias.

A number of factors appeared to be associated with nonadherence to early introduction of foods relating to atopic predisposition. These include ethnicity, family life, readiness to eat solid foods, and parental perception of possible food allergic reactions (IgE- or non-IgE-mediated). The EAT study therefore suggests that if early introduction of allergenic foods from 3 months of age is to be adopted as a prevention strategy, numerous potential obstacles must be overcome with respect to implementation of adherence. Importantly, early introduction of allergens in the LEAP study or the EAT study did not reduce duration of breastfeeding (Feeney et al., 2016). It is noteworthy, however, that the participants in the EAT study are from the general population rather than a high-risk population and therefore any effect size may be less pronounced compared to the LEAP study. Furthermore, the intervention was more complex because it involved six foods, not one.

Diet diversity Two studies have examined the role of diversity of early life food exposures, which may be one factor that coincides temporally with the rise in food allergy, in the development of food sensitization and food allergy. A prospective birth cohort study of 856 children found that increased diversity of complementary foods introduced in the first year of life was associated with a reduced risk of food allergy (Roduit et al., 2014). In another prospective longitudinal study of 123 participants, the authors found that dietary patterns in the first year of life consisting of more fresh fruit and vegetables and home-prepared meals were associated with less challenge-proven food allergy by the age of 2 years (Grimshaw et al., 2014).

The committee concludes that results of the LEAP trial provide strong evidence that early introduction of peanut (between 4 and 11 months) is protective against peanut allergy in infants who are at high risk (as defined by early onset eczema or coexistent egg allergy). Limited evidence from observational studies also suggests that delaying the introduction of egg, cow milk, and wheat to decrease risk of those food allergies has no benefits. Results from one RCT show a not significant decrease in food allergy if allergenic foods (i.e., cooked egg, peanut, cow milk, sesame, white fish, and wheat) are introduced starting at 3 months of age. More studies are necessary to assess whether early introduction of other allergenic foods, in addition to peanut, affect food allergy.

Vitamin D

Vitamin D has become increasingly recognized as an important regulator of immune response (Adams and Hewison, 2008). 1,25(OH)₂D can be converted from 25(OH)D locally based on widespread expression of vitamin D activating enzyme CP27B in a broad spectrum of cells involved in immune response, such as macrophages, B cells, and T cells. This active form of vitamin D exerts its function through interaction with the vitamin D receptor (VDR), which is also present in the above immune cells. Vitamin D has been demonstrated to inhibit the differentiation of B lymphocytes to plasma cells and suppress immunoglobulin production (Chen et al., 2007). However, the effects of vitamin D on T lymphocytes are more complicated. Vitamin D has been shown to inhibit T cell proliferation and production of Th1 cytokines, which induces a shift in the balance between Th1 and Th2-type cytokines toward Th2 dominance (Cantorna et al., 2004; Iho et al., 1985; Reichel et al., 1987). In contrast, in CD4+ and CD8+ T-cells from human cord blood, vitamin D inhibits IL-12-generated interferon (IFN)-γ production and IL-4 production, as well as IL-4-induced expression of IL-13.

It has been hypothesized that in the presence of vitamin D, T regulatory cells function normally to suppress inappropriate Th1 and Th2 responses to environmental exposures leading to disease (Litonjua and Weiss, 2007). Research suggests that vitamin D deficiency might impair epithelial barrier integrity, which would in turn result in increased and inappropriate mucosal exposure to food antigens and also a pro-sensitization immune imbalance that compromises immunological tolerance (Roider et al., 2013).

Two opposing hypotheses have been proposed regarding the connection between vitamin D and allergic disease in general. In 1999, Wjst postulated that excess vitamin D might be associated with an increased risk of allergic disease based on its effects on the shift in the T-cell phenotype from a balance on Th1/Th2 to aTh2 dominance, and parallel patterns of increased oral vitamin D supplementation with a “Western lifestyle” (Wjst, 2008; Wjst and Dold, 1999). In contrast, Litonjua and Weiss raised an opposite hypothesis, suggesting that vitamin D might protect against asthma and allergies (Litonjua and Weiss, 2007, 2008). They believed that the immune effects of vitamin D are probably found on dendritic cells and Treg cells, and that these effects may differ depending on the stage of human development.

Two lines of ecological enquiry support the more recent hypothesis that low vitamin D may increase the risk of food allergy. First, countries further from the Equator (and thus receiving lower ambient ultraviolet radiation) have recorded more pediatric admissions to the hospital for food allergy–related events, and more prescriptions of hypoallergenic formulas for the treatment of cow milk allergy and adrenaline auto injectors for the treatment of anaphylaxis in children, compared to countries closer to the Equator (Camargo et al., 2007; Mullins et al., 2009, 2010; Rudders et al., 2010). These findings appear to be independent of longitude, socioeconomic status, or physician density. Second, children receiving care at a large medical center in Boston for food-related acute allergic reactions were more likely to be born in autumn/winter than in spring/summer (Vassallo et al., 2010). Similar relationships of food allergy to birth seasonality have been reported in the Southern hemisphere (Mullins et al., 2011). Furthermore, children residing in Australia's southern states have twice the odds (95% CI: 1.2-5.0) of peanut allergy at age 4 to 5 years and three times (95% CI: 1.0-9.0) the odds of egg allergy than children in the northern states (Osborne et al., 2012). A recent study from Australia described that infants with vitamin D insufficiency were three times more likely to have egg allergy than those who had adequate stores of the vitamin, with the odds increasing to 10-fold among those with two or more food allergies. Furthermore, among food-sensitized infants, those with vitamin D insufficiency were six times more likely to be food allergic than tolerant (Allen et al., 2013). These effects were observed among infants with Australian-born parents but not those with parents born outside Australia. Genetic polymorphisms contribute to variation in vitamin D binding protein levels, explaining almost 80 percent of variation in levels (Koplin et al., 2016). Binding protein levels in turn alter the biological availability of serum vitamin D, with lower levels increasing the availability of serum vitamin D (25OHD₃). It was recently described that polymorphisms resulting in lower VDR levels appeared to compensate for adverse effects of low serum vitamin D on food allergy risk (Koplin et al., 2016), presumably by increasing the ability to use available vitamin D. These findings suggest that references ranges for optimal levels of serum vitamin D may need to take into account differences in VDR level.

A few studies have been published on the effect of maternal vitamin D status during pregnancy and the development of food allergy in offspring. A follow-up study from an RCT (N=164) reported that Vitamin D supplementation of the mothers during lactation may increase the risk of later food allergy up to 2 years of age (unadjusted analysis), although the authors reported high loss in subjects in the follow-up (Norizoe et al., 2014). However, results from cross-sectional studies (Allen et al., 2013) suggest that vitamin D sufficiency in infants age 1 year may be an important protective factor for food allergy at that age. Another cross-sectional study that followed a German birth cohort for 10 years reported that specific IgE for food allergens (OR: 1.07; 95% CI: 1.02-1.11) at age 10, as well as lifetime prevalence were significantly related to the vitamin D status (Wawro et al., 2014). Conversely, a study in Korea (N=226) showed that vitamin D deficiency increased the risk of sensitization to food allergens (Baek et al., 2014). In a longitudinal study (N=231), Jones et al. showed that maternal intake of supplemental vitamin D was significantly correlated with cord blood 25(OH)D₃ concentration (Jones et al., 2012). However, the associations between cord blood 25(OH)D₃ concentration and allergen sensitization, IgE-mediated food allergy, or eczema severity were not significant. Another prospective birth cohort study (N=378) in Germany reported that maternal and cord blood 25(OH)D₃ was positively associated with children's risk for food allergy within the first 2 years of life (Weisse et al., 2013).

Liu et al. reported that the combination of persistently low vitamin D status at birth and in early childhood (ages 1 to 3 years) increased the risk of food sensitization (defined as specific IgE ≥0.35 kUA/L to any common food allergen, that is, egg white, milk, peanut, walnut, soy, shrimp, cod fish, and wheat) (aOR: 2.03; 95% CI: 1.02-4.04); the risk was particularly higher among children carrying the C allele of rs2243250 (aOR: 3.23; 95% CI: 1.37-7.60) (N=460) (Liu et al., 2013).

Multiple genes are known to be involved in 25(OH)D₃ metabolism and regulatory pathways: genes encoding the molecules to convert 25(OH) D₃ into its bioactive form 1,25(OH)2D (i.e., CYP27B1) and then a water-soluble metabolite (i.e., calcitroic acid; CYP24A1), as well as the receptor complex of vitamin D (i.e., VDR, RXRA, RXRB) and vitamin D binding protein (i.e., GC). Liu et al. evaluated children in the Boston Birth Cohort (N=649) and did not find an association between vitamin D levels in cord blood and sensitization to food allergens in early childhood (Liu et al., 2011). However, when examined with candidate gene single nucleotide polymorphisms, a significant interaction was identified for an IL-4 gene polymorphism and three other genes, indicating a risk for sensitization. In an Australian study, Koplin et al. investigated whether polymorphisms in a VDR-binding protein gene (low, the GT/TT genotype; high, the GG genotype) could modify the relationship between serum vitamin D and food allergy (Koplin et al., 2016). The study (N=5,276) found that low serum 25(OH)D₃ levels (≤50 nM/L) at age 1 year had a modest association with food allergy, particularly among infants with the GG genotype (aOR: 6.0; 95% CI: 0.9-38.9) but the CI was wide. There was no association with food allergy in children with those with low serum 25(OH)D₃ levels and GT/TT genotypes (aOR: 0.7; 95% CI: 0.2-2.0; P interaction=0.014).

The committee concludes that the quantity of evidence on the role of vitamin D in the development of food allergy during critical developmental windows (in utero, infancy, and early childhood) is limited. Further research is needed to confirm or refute this relationship.

Lipids/Omega-3 Fatty Acids

Dietary fat consumption has been hypothesized to influence atopy development by modulation of IgE production (Black and Sharpe, 1997). Among the different dietary fats, the ones that have been studied most extensively are the omega-3 fatty acids. Omega-3 fatty acids are known to have anti-inflammatory and immune modulator properties (Wall et al., 2010). Current evidence suggests that the intake of omega-3 fatty acids has decreased from ancestral times, whereas the consumption of omega-6 has probably increased. Consequently, the dietary ratios of omega-6 to omega-3 fatty acids have changed over time from approximately 1:1 to almost 17:1 in certain industrialized societies (Simopoulos, 2002). The parallel increases in this ratio and in the prevalence of allergic disease, as well as information from experimental models, have elicited the hypothesis that dietary omega-3 fatty acids in early life may influence immune system development and immune cell function (Calder, 2013; Shek et al., 2012).

This hypothesis has been tested using a variety of experimental models, and the results of individual studies have been the focus of several reviews and meta-analyses that reveal the uncertainties that currently afflict this area of knowledge. Contributing to the current controversies are (1) the different experimental designs (observational versus RCTs), (2) the times of intervention and follow up, (3) the usually small size of the populations studied, (4) the different approaches to supplying the omega-3 fatty acids and the doses used, (5) the different periods investigated (fetal life, infancy, childhood), (6) the different outcomes examined, and (7) the potential confounder introduced by the wide-ranging presence of pro-allergenic pollutants and contaminants in fish, the major source of dietary omega-3.

The systematic review of Klemens et al., which reviewed the literature from 1950-2010, is considered to be of medium quality (Klemens et al., 2011). The review included three RCTs (Dunstan et al., 2003; Furuhjelm et al., 2009; Lauritzen et al., 2005; total N=264) of omega-3 fatty acids supplementation compared to olive or soy oil during pregnancy and/or lactation in a high-risk population for outcomes of food allergy, as defined by SPT and clinical diagnosis. When supplementation started during pregnancy egg sensitization decreased at 12 months of age (OR: 0.33; 95% CI: 0.16-0.70). Receiving the supplementation during pregnancy and/or lactation and food allergy at age 12 months were not significantly associated.

A recent Cochrane review, which included manuscripts published until August 2014, assessed the effect of omega-3 supplementation in pregnant and/or breastfeeding women on allergy outcomes (food allergy, atopic dermatitis, allergic rhinitis, and asthma/wheeze) in their children (Gunaratne et al., 2015). Overall, the results showed little reduction of allergic disease in the children resulting from maternal omega-3 supplementation during pregnancy and/or breastfeeding. Five trials reported food allergy outcomes (Dunstan et al., 2003; Furuhjelm et al., 2009; Lauritzen et al., 2005; Makrides et al., 2009, 2010). There was only one study where omega-3 supplementation reduced the incidence of IgE-mediated food allergies in children up to 12 months of age (Furuhjelm et al., 2009) (N=117; RR: 0.13; 95% CI: 0.02-0.95). Similarly, another recent review identified three RCTs (Dunstan et al., 2003; Furuhjelm et al., 2009; Palmer et al., 2012) and two follow-up studies (Furuhjelm et al., 2011; Palmer et al., 2013) with pregnant women whose infants were at high risk of atopy. After adjusting for potential confounders or after long-term follow-up only one study showed an association between maternal omega-3 fatty acid supplementation and lower risk of food sensitization (Newberry et al., 2016).

The committee concludes that the current evidence does not support a link between increased maternal omega-3 intake and a protective effect on childhood food allergy.

Folate

Emerging interest in the role of folate in immune development and allergic disease has been driven by the recent understanding that folate, a dietary methyl donor, can affect immune function and alter gene expression through epigenetic mechanisms (Brown et al., 2014). Concerns have been raised about whether folic acid supplementation during pregnancy and/or early childhood is a potential risk factor for the development of atopic diseases in children. As animal models have demonstrated, maternal supplementation with dietary methyl donors during pregnancy induces hypermethylation of key regulatory genes in lung tissue, resulting in subsequent allergic airway disease in offspring (Hollingsworth et al., 2008). Exposure to folate in utero can affect DNA methylation during fetal development in humans (Amarasekera et al., 2014), which can influence transcriptional activity. For example, hypermethylation can silence the expression of genes. During polarization of naive T helper cells to Th2 cells, methylation of the promoter region of the IFN-γ gene blocks transcription factor binding and thus expression of the IFN-γ gene (Jones and Chen, 2006). Consequently, increased folic acid intake could influence the expression of genes that may be involved in T-cell differentiation during gestation. In turn, this may influence the allergic predisposition in the neonate.

To date, most human studies on this topic have focused on asthma, with very limited number of studies specific to food allergy or food sensitization. An Australian study (N=484) assessed maternal folic acid intake and serum folate levels during the third trimester, and cord blood folate status at birth (N=285), and allergic outcomes at age 12 months, including IgE-mediated food allergy, eczema, and asthma, in offspring (Dunstan et al., 2012, p. 51). In their study, food allergy was defined as “a history of immediate symptoms following contact and/or ingestion and a positive SPT to the implicated food.”

However, maternal serum folate status and allergic outcomes were not associated (Dunstan et al., 2012). In a study of 2,834 Dutch children, maternal folic acid supplement intake across the whole pregnancy, and intracellular folate status (measured in the third trimester of pregnancy in 837 [29.5%] participants) was not significantly associated with specific IgE against hen egg, cow milk, peanut, and aeroallergens at age 2 years or eczema until age 6 to 7 years (Magdelijns et al., 2011).

A recent study that measured serum folate (at ages 2, 4, 6, and 8 years) in 138 U.S. children found that increased serum folate levels at or before age 6 years were significantly associated with increased incidence of sensitization to both food and aeroallergens, but not with serum total IgE, asthma, or wheezing at ages 6 or 9 years (Okupa et al., 2013). In the National Health and Nutrition Examination Survey (NHANES) (which covers ages 2 to 85 years), a cross-sectional study, serum folate levels were inversely associated with atopy, wheeze, and elevated total IgE levels (Matsui and Matsui, 2009).

Of note, the inconsistent results of previous studies are likely due to many reasons, including differences in sample size, participants' ages, clinical characteristics, allergic outcomes, methods used for measurement of folate status, and statistical methods used in the analysis.

The committee concludes that evidence to assess the causal association between folate and the development or prevention of food allergy is lacking. Further research to study this potential association is needed.

Other Nutrients

A prospective cohort study assessed the relationship between maternal dietary antioxidant intake (B carotene, vitamins C and E, copper, and zinc) during pregnancy and food allergy of the child at age 12 months among families at high risk (West et al., 2012). This study of 300 mother-infant dyads found a protective effect of vitamin C intake on food allergy, with higher intakes that were limited to one quartile of vitamin C intake. For copper, intake in the highest quartile also showed a protective effect. However, as previously noted, observational studies suffer from inherent methodological flaws. Thus, proper RCTs are required to determine the causal effect of the maternal diet on the etiology of food allergies in offspring.

The committee concludes that evidence to assess the causal association between other nutrients and the development or prevention of food allergy is lacking.

Do the Obesity and Diabetes Epidemics Have a Role in the Rise of Food Allergy?

The parallel increase in the prevalence of obesity and type 2 diabetes and allergic diseases raises the question of whether these conditions may be linked. Obesity is known to induce systemic inflammation, which might adversely influence the immature immune system and atopic outcomes. Increased adipose tissue also could lead to reduced adiponectin levels, which in turn down-regulates the secretion of IL-10 and decreases regulatory T cells (Hersoug and Linneberg, 2007). Although the precise mechanism underlying the link between obesity and allergic disease including food allergies remains to be elucidated, the hypothesis is biologically plausible.

Very limited data are available on the association between having overweight or obesity and food allergy. Observational studies have shown that obesity is associated with a higher risk of atopy (elevated specific IgE to allergen) (Ouyang et al., 2009; Visness et al., 2009; Xu et al., 2000). For example, data from the 2005-2006 NHANES demonstrated that children with overweight or obesity had a higher geometric mean of total IgE levels and were at a higher risk of atopy than children with normal weight. This association was driven largely by allergic sensitization to food allergens, and systemic inflammation (measured as serum c-reactive protein) in children with obesity may play a role in the development of allergy (Visness et al., 2009). In contrast, ample studies show the association between overweight and obesity and asthma in both children and adults (Baumann and Lorentz, 2013; Granell et al., 2014).

The role of maternal overweight and obesity and diabetes on the developing fetus and the subsequent risk of allergic diseases has not been well studied but deserve attention. In the prospective Boston Birth cohort, Kumar et al. reported that in term births, gestational diabetes was significantly associated with allergen sensitization in the child, and such association was also driven by food sensitization (Kumar et al., 2009). In contrast, others reported no associations between obesity measures and atopy (Jarvis et al., 2002; Ma et al., 2010), or inverse associations (Van Gysel et al., 2009).

Other Unsubstantiated Hypothesis for the Rise in Food Allergy

Media interest in food allergies has become significant and sustained as food allergies have become more common. As such, public conjectures about potential causes for the rise are widespread. In particular, awareness about unfortunate cases of food-induced anaphylaxis is high. Added to that is the increased awareness by various community or commercial organizations (such as schools, restaurants, airlines, and sporting clubs) of their need to be careful about how they provide foods for food allergic individuals. As a result, communities are greatly interested in why the prevalence of food allergy appears to be rising.

One of the most widely held theories, among the many that abound, as to why food allergy is on the rise holds that it is due to the increasing consumption of processed foods and food additives. Unfortunately, to date no significant research has been conducted on this issue. Websites and blogs tout the dangers of processed foods and food additives, and evidence from clinical observation suggests that some parents believe that food additives aggravate a range of clinical symptoms and signs, from difficult behavior and autism to gastrointestinal reactions. Clinically, the best way to understand whether a food is aggravating symptoms is to eliminate that food and later challenge with it—provided the risk of anaphylaxis has been excluded. However, the role of additives and preservatives in the development of food allergy in the first place has never been examined at the ecological or epidemiological level. In addition to understanding whether preservatives or additives have a direct toxicological effect on the developing immune system, it would be valuable to assess whether these substances actually influence the composition of the gastrointestinal microbiome.

Concerns over genetically modified crops (Nordlee et al., 1996) has resulted in consideration of the role that such foods may play in aggravating food allergy and in a requirement to assess the potential allergenicity of genetically modified crops (CAC, 2009; FAO/WHO, 2001). Although an online tool recently has been developed to help assess the role a novel protein may play in cross-reactivity (Goodman et al., 2016) based on criteria from the Codex Alimentarius Commission (CAC, 2009), current methodologies are considered inadequate to predict de novo allergenicity. Little or no research exists on whether the increased use of genetically modified crops could be linked to the rise in food allergy.

Numerous lay books and review articles argue that the increased consumption of fast food in the Westernized diet may have a significant impact on immunity (Myles, 2014). Although emerging indirect evidence suggests that fresh fruit and vegetables and food diversity might be important for an optimal and healthy start to life, to date little work has been done on their role specifically in preventing food allergy. Some of the first emerging evidence of diet diversity and its impact on food allergy development has been generated by the EuroPrevall study (Grimshaw et al., 2014). In a nested case-control within-cohort study of 41 infants using gold standard food challenge outcomes and 82 age-matched controls, the authors found that an infant diet with high levels of fruits, vegetables, and home-prepared foods is associated with less food allergy by 2 years of age. As an observational study, these results are subject to confounding but they generate a hypothesis worth testing in systematic trials.

The committee concludes that speculation abounds regarding why food allergy is on the rise. Although some ideas are based in appropriate theoretical frameworks, the absence of RCTs prevents firm conclusions to be drawn on their validity.

The Microbial Hypothesis

• Determine, using well-designed prospective studies, the role of mode of birth delivery (vaginal, emergency versus elective cesarean section) and early life microbiome composition on the development of food allergy.
• Assess, through well-designed prospective studies, potential links between food allergy and antibiotic exposure in children (studies should include information on the type, dose, and frequency of antibiotic exposure).
• Determine whether pet ownership is related to food allergy by using well-designed prospective studies.
• Assess, with RCTs, the potential benefits of prebiotics and probiotics to prevent the onset of food allergy.

Allergen Avoidance and Exposure

• Elucidate the relationship, if any, between breastfeeding and the onset of food allergy (may also influence through microbiome modulation) with well-designed prospective studies and take into account the potential effect of differences in breast milk composition.
• Determine, with RCTs, whether consuming or eliminating or avoiding specific allergenic foods during pregnancy and lactation has any benefits.
• Conduct RCTs, similar to the Learning Early About Peanut study, to determine whether early introduction of peanut products has benefit in individuals other than high-risk infants, who were studied in the original trial.
• Examine early introduction of allergenic foods in addition to peanut to determine whether this approach is beneficial in preventing the development of food allergy.

Nutrition Immunomodulation Hypothesis

• Assess, with RCTs, the potential role of specific nutrients, such as vitamin D, folate, or fatty acids, in preventing food allergy.