Food allergy is a recognized public health concern, for which preventative strategies are required. Although an intervention that adequately protects against the development of food allergy has still to be identified, limited benefits have been shown for the prevention of related allergic conditions such as eczema, and to a lesser extent asthma and rhinitis; these benefits are usually limited to at-risk populations. Prevention strategies need to be tested using randomized controlled study designs that account for the numerous methodological challenges, safety concerns, and necessary ethical limitations.
The prevalence of IgE-mediated food allergy has increased in the last 2 decades, with approximately 3–6% of children in the developed world being affected ; however, the increase in food allergy is best described for peanut allergy (PA). Although genetic factors are important in the development of food allergy, the increase in food allergy has occurred in a short period and is therefore unlikely to be the result of germline genetic changes alone. It seems plausible that 1 or more environmental exposure(s), or lack of exposure(s), may through epigenetic changes result in the interruption of the default immunologic state of tolerance to food. Strategies aimed at the prevention of food allergy are therefore required: primary prevention strategies seek to prevent the onset of IgE sensitization; secondary prevention seeks to interrupt the development of food allergy in IgE-sensitized children; and tertiary prevention seeks to reduce the expression of end-organ allergic disease in children with established food allergy.
This article highlights important conclusions of the many reviews in this field and adds to current evidence.
Methodological challenges
The fact that no single intervention, or combination of interventions, is able to repeatedly show a strong protective effect against food allergy reflects either on the interventions themselves or the study methods used to measure them. This section examines the methodological concerns that complicate the interpretation of these studies (summarized in Table 1 ).
| Issue | Problem | Recommended Approach |
|---|---|---|
| Study design | Most studies in this field are observational studies | Randomized double-blind, placebo-controlled (RDBPC) trials reduce unmeasured and unknown sources of bias |
| Reverse causality | Early signs of suspected allergic disease (eg, eczema) may influence feeding patterns | Although challenging (and not always possible), trials should adopt RDBPC methodologies |
| Randomization | There are necessary ethical restraints that surround nutritional intervention studies in infancy | Breastfeeding should always be encouraged. Studies that wish to assess the effect of complementary feeding can then randomize breastfed children, from 4 months onwards and when exclusive breastfeeding no longer satisfies the infant |
| Blinding of dietary interventions | It may prove difficult to blind the dietary intervention/s | Some dietary interventions cannot be blinded |
| Determination of food allergy | Few studies make use of oral food challenges (OFCs) for the diagnosis of food allergy. The diagnosis of food allergy may therefore be inadequate both at study entry and exit. Too many studies rely on allergy testing (SPT and/or Sp-IgE) for the determination of food allergy | Aim to perform OFCs in all participants. For children who do not undergo OFCs, a priori diagnostic algorithms (which make use of the combination of history, examination, skin prick testing, and Sp-IgE) are required to reach a best possible diagnosis |
| Surrogate markers | IgE sensitization, eczema, asthma and rhinitis are often used as surrogate markers for food allergy | As above |
| Natural history of food allergy | Tolerance is anticipated for many, but not all, childhood food allergies | Account for natural remission rate of food allergy before assessing for a study effect |
| Nomenclature | There is insufficient consensus with respect to the terminology used for common allergic conditions, particularly in early childhood. Definitions of generic terms such as allergy or atopy are open to variability | Consensus with respect to the allergy nomenclature will facilitate research in this field |
| Determination of diet | The determination of food consumptions is usually by retrospective food frequency questionnaires (FFQs), which are prone to many forms of bias | Use should be made of prospective food diaries, which have been validated for context, language, and consistency |
| Dietary variables and measurement thereof | Few dietary analyses consider all variables; these include age of introduction, quantity ingested (individual and cumulative quantity), frequency of exposure, variability of allergens, and allergen processing and concomitant breastfeeding at time of commencing complementary feeds | Well-designed validated tools are required to accurately record all dietary variables |
| Definitions: weaning | Use of the term weaning is not consistent and indicates the introduction of solid foods only | WHO recommends that the term weaning should be replaced by the term complementary feeding. which incorporates any nutrient-containing food or liquid (other than breast milk) given to young children |
| Differing patterns of complementary feeding | Endless permutations make study effects difficult to compare (eg, many exclusively breastfed infants receive early top-up infant feeds with cow’s milk formula) | Studies need to try to capture the following dietary variables: dose, timing, recurrence of exposure, variability of allergens, and allergen quality such as degree of processing |
| Outcome classifications | Many studies refer generically to allergy: allergic disease may encompass 1 or more of asthma, eczema, and hay fever. Definitions for each of these conditions are also open to great variability | A priori clinically validated definitions required |
| High-risk markers | Many studies are aimed at high-risk atopic populations; such populations are difficult to define | Studies should include entire study populations (ie, both low and high risk). At-risk populations should be defined a priori. Better at-risk markers are required |
| Separation of specific effects when interventions are combined | Multiple interventions are often studied at different time points. For example, probiotic administration may be administered to mother (during pregnancy and/or breastfeeding) and/or newborn infant. This strategy makes it difficult to determine the specific effect of each intervention at each time point | Preliminary proof-of-concept studies need to separate the effects of each intervention |
| Introduction of complementary feeds is associated with multiple variables | The early introduction of solid foods has been associated with cultural and socioeconomic factors as well as specific factors such as maternal age, formula feeding, and maternal smoking | Regression analysis should control for as many relevant confounders as possible, especially in observational studies. This strategy highlights the need for randomized controlled studies |
| Monitoring adherence | Monitoring of adherence to dietary interventions is difficult | Better tools for monitoring dietary adherence are required |
A major limitation of many food allergy prevention studies lies in the study design; linked to this are the necessary ethical limitations that apply to allergy prevention, particularly when applied to the in-utero environment and in early childhood. For example, strict ethical restrictions surround randomization of infants to anything but breast milk.
A second major limitation of studies in this field is the phenotypic description of food allergy. Few studies make use of food challenge diagnostic procedures; although tolerance is adequately determined by open food challenge, the gold standard for the determination of food allergy is the double-blind, placebo-controlled food challenge; these challenges are laborious and sometimes difficult to perform, particularly in young children. In addition, entry-level oral challenges cannot be performed in those children assigned to the avoidance arm of intervention studies, with the consequence that the true allergic phenotype of such infants at time of enrolment or exit from studies remains uncertain. This limitation is particularly problematic for the diagnosis of cow’s milk allergy, which is a common, and frequently studied, childhood allergy.
The most frequently used surrogate marker for the determination of food allergy is the determination of IgE sensitization by skin prick test (SPT) results and/or serum specific IgE (sp-IgE) measurement. Although IgE-mediated food allergy requires a state of sensitization, most sensitized children are not food allergic. Many variables are also associated with the determination of sp-IgE and SPT, which further limit the use of this surrogate marker.
Eczema is also a commonly used surrogate, which itself has many limitations. Although eczema is strongly associated with food allergy, the 2 are not synonymous, as evidenced by studies that show an improvement in eczema but not in food allergy. Without proper study randomization it is often difficult to establish if studies that show protection against the development of eczema are showing the treatment thereof (ie, tertiary prevention, as opposed to primary prevention). Many studies do not adequately assess eczema severity; there is therefore a risk that beneficial effects are restricted to mild eczema, which runs a transient course, when compared with moderate to severe eczema. Similar limitations hold true for the use of asthma and rhinoconjunctivitis as food allergy surrogates.
Additional limitations of the surrogate markers used arise because of the inconsistencies in nomenclature used and difficulties in accurately diagnosing these conditions. For example, wheezing in infancy does not always mean that the child has asthma. Moreover, lung function is difficult to measure in infants.
There are many other limitations to studies in this field. For example, nutritional interventions are prone to both selection bias and reverse causality. Such bias may arise when atopic families (if aware of public health recommendations) are motivated to alter maternal and/or infant dietary practices. The effects of reverse causality are highlighted in various studies and for different allergic outcomes. For example, in the Avon Longitudinal Study of Parents and Children (ALSPAC) study, a history of an allergic reaction to peanut is associated with prolonged breastfeeding. However, when adjusted for infantile eczema by regression analysis, there was no effect of breastfeeding on the development of PA. This phenomenon is also true for asthma; Fussman and colleagues, in a large (n = 696) international study, reported that reverse causation negates the finding that the consumption of cow’s milk is associated with asthma, as shown by the finding that the presence of asthma in the months before assessment led to a reduction in further exposure to cow’s milk. Likewise, Lowe and colleagues, in a prospective birth cohort of 620 infants, reported that early symptoms of eczema prolong the duration of exclusive breastfeeding.
Childhood food allergies are dynamic, with the general trend being for resolution of many, but not all, food allergies during the first decade of life (a time course sometimes referred to as the food allergic march). Other atopic conditions such as childhood eczema and asthma also share unique natural histories. Study planning should therefore take these natural histories into account when planning the introduction of interventions as well as when measuring potential outcomes.
The inclusion of a placebo in nutritional studies is not always practical, or safe. For example, in the LEAP (Learning Early about Allergy to Peanut) peanut intervention study a placebo snack is not used for safety reasons. If it were possible to create an equivalent peanut placebo, children consuming the placebo snack might be at risk of consuming peanut-containing foods (to which they might unknowingly be allergic). In addition to representing a major safety concern this action would nullify any immunologic consequence that may have resulted from the avoidance of peanut.
Study interventions should be safe for both mother and child. Safety concerns have nonetheless arisen in select studies. For example, dietary interventions have been noted to compromise fetal and maternal well-being, and probiotics have been shown to increase rates of sensitization and allergic outcomes in separate studies.
Many interventions are introduced to both mother and child. This strategy complicates the understanding of specific study effects because it is unclear whether the immunologic effects were achieved prenatally or postnatally or whether effects should be attributed to a single or multiple factors.
The determination of dietary intake is usually performed by food frequency questionnaires (FFQs); FFQs are known to be subject to substantial forms of bias. Studies do not always make use of validated FFQs, a complex but important undertaking. FFQs do not always assess all relevant dietary variables, such as age of introduction, recurrence of exposure, quantity (single and cumulative) of exposure, variability of allergens eaten, and allergen processing. In addition, it is often difficult to disguise those questions that relate to the specific food/s of interest. Prospective food diaries are cumbersome because they demand detailed information and effort on the part of parents. It is particularly difficult to measure food allergen exposure that occurs via routes other than the oral route (eg, through an abraded skin barrier). For example, the nursing mother who ingests peanut butter is also likely to transfer this allergen to the infant through kiss and touch contact. In addition, it is often the nursing mother who determines consumption patterns within the household, which further increases (or decreases) the opportunity for environmental food allergen exposure to foods that the mother likes (or dislikes). A different problem arises if the intervention is one of avoidance, because the elimination of 1 or more foods from the diet is likely to affect the diet. Such changes may be anticipated and therefore measured, or unknown and missed.
Many studies in this field are aimed at at-risk children; however, at-risk populations are difficult to define. For example, approximately 10% of children without an allergic first-degree relative develop allergic disease, compared with 20% to 30% with single allergic heredity (parent or sibling) and 40% to 50% with double allergic heredity. In addition, the definition of the term atopy is inconsistent.
Observational study designs (ie, most studies in this field) are particularly vulnerable to bias from both unmeasured, and unknown, sources; study hypotheses would ideally be assessed using only randomized placebo-controlled, double-blind studies.
Onset of sensitization and food allergy
It remains unclear when prevention strategies should be implemented. Prerequisites for the development of food allergy (particularly in genetically susceptible individuals) are believed to include allergen exposure, uptake, recognition, and processing. It is therefore important to determine whether sensitization occurs in-utero or afterwards. That in-utero sensitization to foods is possible is suggested by the early detection of sp-IgE and clinical presentation of IgE-mediated food allergies within the first few days of life (but usually within the first years of life). This finding is also true for non–IgE-mediated food-induced immunologic reactions such as colitis induced by cow’s milk protein, eosinophilic esophogitis, and colitis induced by food protein.
The fetal immune system is sufficiently mature and able to respond to both food allergens and aeroallergens. It has also been shown that food allergens and aeroallergens can pass transplacentally. The analysis of sp-IgE to foods in cord blood has been restricted to occasional studies. Two large birth cohort studies were unable to show measurable food sp-IgE in cord blood, even in those children who subsequently developed clinical or immunologic food sensitization.
Summary
There is no firm evidence to support the hypothesis that sensitization and allergy to foods commences in-utero.
Onset of sensitization and food allergy
It remains unclear when prevention strategies should be implemented. Prerequisites for the development of food allergy (particularly in genetically susceptible individuals) are believed to include allergen exposure, uptake, recognition, and processing. It is therefore important to determine whether sensitization occurs in-utero or afterwards. That in-utero sensitization to foods is possible is suggested by the early detection of sp-IgE and clinical presentation of IgE-mediated food allergies within the first few days of life (but usually within the first years of life). This finding is also true for non–IgE-mediated food-induced immunologic reactions such as colitis induced by cow’s milk protein, eosinophilic esophogitis, and colitis induced by food protein.
The fetal immune system is sufficiently mature and able to respond to both food allergens and aeroallergens. It has also been shown that food allergens and aeroallergens can pass transplacentally. The analysis of sp-IgE to foods in cord blood has been restricted to occasional studies. Two large birth cohort studies were unable to show measurable food sp-IgE in cord blood, even in those children who subsequently developed clinical or immunologic food sensitization.
Summary
There is no firm evidence to support the hypothesis that sensitization and allergy to foods commences in-utero.
Maternal diet (during pregnancy and/or breastfeeding) and the prevention of food allergy
This section examines the effect of maternal diet during pregnancy and/or lactation on the development of food allergy.
There are studies that assess the effects with respect to allergy prevention through maternal dietary avoidance of 1 or more common food allergens during pregnancy and/or lactation. Kramer and Kakuma in a Cochrane review assessed the evidence for allergy prevention through prescribing an antigen avoidance diet during lactation. Three trials were included involving 209 women; findings suggest a strong protective effect of maternal antigen avoidance on the incidence of atopic eczema during the child’s first 12 to 18 months of life. The investigators note the methodological shortcomings in all 3 trials and argue for caution in applying these results; in particular, the high incidence of atopic eczema in the control groups is of concern and may be explained by non-blinding or de-blinding of the examining physicians. One trial reported that a restricted diet (egg and cow’s milk) during pregnancy was associated with a small but statistically significant lower mean gestational weight gain and there were trends toward increased preterm delivery and lower birth weight. Given the uncertainty of these findings, and potential safety concerns, no allergy organizations recommend the avoidance of either egg or milk during pregnancy. The American Academy of Pediatrics and the UK Government Department of Health had previously suggested that at-risk families may wish to avoid peanut; these recommendations have been withdrawn, like in many organizations that offer similar, nonrestrictive, dietary recommendations for the prevention of allergy. Two large prospective birth cohort studies, one based on the Isle of Wight, UK and the ALSPAC study, showed no effect of maternal peanut consumption in pregnancy or lactation on the development of immunologic or clinical reaction to peanuts on follow-up at 4 to 6 years of age. Likewise Fox and colleagues, in a questionnaire-based study, found no effect of maternal peanut consumption during pregnancy or lactation. More recently, 2 studies have suggested that peanut exposure during pregnancy and lactation may be associated with higher rates of PA; however, these studies made use of less rigorous methodologies.
There are studies that examine the role of dietary patterns on other allergic disease outcome. For example, the protective effects against childhood wheezing associated with a Mediterranean diet are largely a result of dietary effects in pregnancy. In the ALSPAC UK birth cohort, health-conscious dietary patterns were positively associated with eczema, total IgE, forced expiratory volume after 1 second, and forced expiratory flow, and negatively associated with early wheezing and asthma. The processed dietary pattern was positively associated with early wheezing and negatively associated with atopy and forced vital capacity; however, when controlling for confounders, these effects were substantially attenuated and became nonsignificant, suggesting that in this cohort dietary patterns in pregnancy did not predict asthma and related outcomes in the offspring.
Breast milk contains low concentrations of dietary proteins, which are present in maternal serum such as gliadin, peanut, β-lactoglobulin, and ovalbumin. β-Lactoglobulin is found in the breast milk of as many as 95% of mothers consuming cow’s milk during lactation. Whether at-risk infants are protected by the many beneficial immunologic properties of breast milk or put at risk by this low-dose allergen exposure is an ongoing debate. No studies have modified maternal diet during lactation only. There are studies that modify the maternal diet during both pregnancy and lactation. Neither the study by Hattevig and colleagues nor the study by Herrmann and colleagues reported a protective role against infant food allergy through maternal dietary avoidance of cow’s milk, egg, and fish during either pregnancy or both pregnancy and lactation. However, the study by Herrmann and colleagues did note effects for eczema.
Summary
There is little evidence to suggest that manipulation of the maternal diet during pregnancy and/or breastfeeding has any protective effect on the development of food allergy; preventative effects are noted for eczema. Such strategies have been shown to potentially compromise the nutritional well-being of both mother and child.
Complementary infant feeding and the prevention of food allergy
This section examines the effect of complementary feeding on the development of food allergy. The World Health Organization (WHO) now recommends that the term weaning be replaced by the term complementary feeding, which incorporates any nutrient-containing food or liquid other than breast milk.
Exclusive Breastfeeding Versus Mixed Feeding/Complementary Feeding
Although there is universal consensus that breast milk remains unchallenged as the milk of choice for all infants, there is conflicting advice with respect to the age at which complementary feeding should occur. Whereas most allergy/gastrointestinal opinion leaders suggest that complementary feeding may occur (if the infant is ready) from 4 months of age onwards WHO recommends that complementary feeding should occur only after 6 months of age.
Breast milk provides a rich and favorable source of important immune-regulating substances such as immunoglobulins, lactoferrin, lysozymes, oligosaccharides, long-chain fatty acids, cytokines, nucleotides, hormones, antioxidants, and maternal immune cells. The importance of these factors was shown by Verhasselt and colleagues, who investigated whether the exposure of lactating mice to an airborne allergen affects asthma development in progeny; they found that airborne antigens were efficiently transferred from the mother to the neonate through milk and that tolerance induction did not require the transfer of immunoglobulins. Breastfeeding-induced tolerance relied on the presence of transforming growth factor β (TGF-β) during lactation, was mediated by regulatory CD4+ T lymphocytes, and depended on TGF-β signaling in T cells. Breast milk may affect other processes capable of directly influencing allergic disease expression such as food antigen absorption and processing. It is frustrating therefore that breastfeeding rates remain below WHO targets.
Studies that support a protective effect of breastfeeding over cow’s milk formulae date back to the 1930s, when Grulee and Sanford, in a large (n ≈ 20,000) observational study, reported a protective effect with respect to the development of eczema in the first 12 to 48 months of life. A systematic review of 12 prospective studies (8183 infants) found that exclusive breastfeeding in the first months of life is associated with reduced rates of subsequent asthma (odds ratio [OR] 0.70; 95% confidence interval [CI] 0.60–0.81). Many, but not all, of the observational studies that followed supported these early findings. Nonetheless, there is a consensus among reviews in the field that breastfeeding offers at least some protection against the development of allergy. For example, Muraro and colleagues suggest an overall protective effect (for at-risk children) of exclusive breastfeeding during the first 3 months of life on atopic eczema, asthma, but not childhood allergic rhinitis. Although the protective effects are most consistent for eczema, the underlying immunologic mechanisms to explain this phenomenon remain unclear. In support of those previous studies that report a protective effect of breastfeeding over asthma is the observational study from Australia by Oddy and colleagues. These investigators followed a large cohort (n = 2602) of children (enrolled before birth) and reported a protective effect for asthma (strict diagnostic criteria used) and atopy (positive reaction to common aeroallergens) at 6 years of age. After adjustment for sex, prematurity, and maternal smoking during pregnancy, they found that exclusive breastfeeding for less than 4 months was associated with an increased risk for current asthma (ie, there is a substantial reduction in risk of childhood allergy [asthma and atopy] at 6 years of age if infants are exclusively breastfed for at least the first 4 months of life); it may also be that the benefits of breastfeeding persist into the teenage years. Saarinen and Kajosaari, in an observational community-based study, reported early (less eczema and food allergy at 1–3 years of age) and late (respiratory allergy at 17 years of age) effects in infants who were exclusively breastfed for 6 months compared with those who were breastfed for 3 months or less.
No studies report a clear benefit of breastfeeding over cow’s milk formula on the development of food allergy; this is true even for premature infants (with an increased gut permeability and an immature gut immune system). Lucas and colleagues, in a large (n = 777) randomized interventional study of premature infants, compared the effect of human milk, standard preterm formula, and nutrient-enriched preterm formula. At 18 months after term there was no difference in the incidence of allergic reactions between dietary groups. However, in the subgroup of infants with a family history of atopy, those infants who received preterm formula rather than human milk had a significantly greater risk of developing 1 or more atopic conditions (notably eczema) by 18 months. Furthermore, De Jong and colleagues in a large (n = 1693) randomized intervention study (the BOKAAL study) found that early (first 3 days of life) high-dose exposure to cow’s milk (as frequently occurs in nurseries) was not associated with an increase in allergic disease or symptoms. In addition, no increase between the groups was found in sensitization or allergy to cow’s milk (up to 5 years of age).
There are observational studies that controversially report an increased risk for the development of allergic disorders in breastfed infants. Sears and colleagues in a large (n = 1037 children) observational study followed children until 21 years of age. These investigators found that breastfeeding (for at least 4 weeks) does not protect against childhood atopy and asthma. Significantly more breastfed children were atopic to common aeroallergens at age 13 years than non-breastfed children. Breastfeeding also increased the likelihood of current asthma at age 9 years and at age 21 years. Findings were similar when breastfeeding was considered over longer periods (8–12 weeks). However, there are many criticisms of this study; these include that data for breastfeeding were retrospectively gathered when infants were age 3 years; definitions for atopic heredity included parental atopy only, hence adjustments were not made for eczema or sibling atopy. The number of exclusively breastfed children in this study was low, with nonsignificant findings when considering the effects of exclusive breastfeeding only. Likewise, Bergmann and colleagues, in the large Multicentre Allergy Study observational birth cohort (n = 1314 infants born in 1990), found that each month of breastfeeding increased the risk of developing atopic eczema in the first 7 years by approximately 3%. It was noted that breastfeeding persisted for longer if at least 1 parent had eczema, the mother was older, did not smoke in pregnancy, and the family had a high social status. Whereas many studies report the greatest benefit for the prevention of allergies lies with at-risk groups, this study suggests that breastfeeding does not prevent eczema in children with a genetic risk. However, multiple regression analysis cannot exclude reverse causation (ie, mothers may extend exclusive breastfeeding and duration of breastfeeding once infants develop eczema). Isolauri and colleagues observed that atopic eczema in exclusively breastfed infants improved when breastfeeding was stopped, which suggests that the protective effect may be caused by the reduction in allergic disease expression (tertiary prevention) rather than primary prevention. However, this finding cannot easily be assessed because this study did not include controls to confirm that this effect was caused by the cessation of breastfeeding Hence, these studies suggest that prolonged breastfeeding could maintain eczema, at least in some infants, and that cessation of breastfeeding acts to treat the eczema as opposed to prevent it.
As detailed earlier, there are many potential mechanisms by which breastfeeding may reduce allergic outcomes. Alternatively, breastfeeding may serve as a surrogate for other important factors, the most obvious of which is the absence of complementary feeding (see definition in earlier discussion). Prospective birth cohorts that show a preventive effect of breastfeeding on allergy report prolonged breastfeeding (>4–6 months) and late introduction of solids (>4–6 months). Most studies in this field consider weaning to be the introduction of solid foods only. However, the biophysical properties of allergens are complex and there is no reason to believe that the allergenic potential of liquid feeds is different from that of solid or semisolid feeds. For example, both cow’s milk and hen’s egg allergy (EA) are common childhood allergies despite being ingested as liquid and solid, respectively. It is therefore arbitrary to restrict the usage of the term weaning to solids. An infant who is breastfed while receiving cow’s milk formula supplementation is no more or less weaned than a breastfed infant who is fed rice cereal mixed with expressed breast milk.
Kramer and Kakuma performed a systematic review of the available evidence concerning the effects of exclusive breastfeeding for 6 months versus exclusive breastfeeding for 3 to 4 months followed by mixed breastfeeding (complementary liquid or solid foods with continued breastfeeding) to 6 months, on eczema, asthma, and other atopic outcomes. This extensive review covers 20 independent, observational studies. These investigators were unable to establish evidence for a significant reduction in the risk of atopic eczema, asthma, or other atopic outcomes amongst those infants who were exclusively breastfed for 6 months compared with those exclusively breastfed for only 3 to 4 months followed by mixed feeding.
Summary
Evidence suggests that exclusive breastfeeding offers some protection against eczema and asthma, but not food allergy, when compared with cow’s milk protein formulas. Observational studies that report an increase in allergic disease because of breastfeeding are prone to reverse causality. Studies in this area suffer from true randomization because of necessary ethical limitations.
Cow’s Milk Hydrolysates and Other Milk Formulas
The Cochrane review by Osborn and Sinn found no evidence to support feeding with a hydrolyzed formula for the prevention of allergy compared with exclusive breastfeeding. For high-risk infants who are unable to breastfeed there seems to be a consensus among reviews that the use of hydrolyzed milk formula offers at least some protection against allergic disease, and in particular eczema. Although these findings are reflected in recent summary papers, in view of methodological concerns and inconsistency of findings, Osborn and Sinn recommend for future research in this field that large, well-designed trials should be used, comparing formulas containing partially hydrolyzed whey or extensively hydrolyzed casein with cow’s milk formulas. One such study is the German Infant Nutritional Interventional (GINI) Study. This is a large (n = 2252) randomized multicenter study in which Von Berg and colleagues allocated at-risk infants to 1 of 4 milks (cow’s milk formula, partially or extensively hydrolyzed whey formula, or extensively hydrolyzed casein formula). Overall, a significant reduction in the incidence of atopic dermatitis was achieved in this study using the hydrolyzed formula; the preventive effect of partially hydrolyzed whey formula and extensively hydrolyzed casein formula on allergic manifestations (defined as physician’s diagnosis of atopic dermatitis, food allergy/intolerance, and allergic urticaria; asthma and rhinitis were added at the later time points) and atopic dermatitis found in the 2 previous analyses at age 1 year and until 3 years have been confirmed at age 6 years in both the intention-to-treat and the per-protocol analyses. The results found with extensively hydrolyzed whey formula in this 6-year analysis are interesting; until 3 years, this formula showed only a small preventive effect that never reached significance, whereas at age 6 years it showed a preventive late-onset effect on atopic dermatitis and allergic anifestations in the per protocol population similar to that of the other 2 hydrolysates and a considerable, although not significant, effect on rhinitis. The GINI study is ongoing.
Although the clinical benefits reported by the GINI study are convincing, it remains unclear whether dietary modification has prevented allergic disease. The challenge in interpreting these findings lies in the relationship between eczema and cow’s milk allergy, because the preventative effect of select formula may simply show tertiary prevention (ie, eczema treatment, as opposed to the prevention of eczema). The GINI study was not able to clearly define the end point of food allergy by double-blind, placebo-controlled food challenge (because many parents declined). Thus, the reduction in eczema could be caused by a true preventative effect of infant dietary modification or alternatively it could reflect the beneficial effect of removing cow’s milk protein from the diet of infants with eczema and milk allergy. It is difficult to separate these effects when the study intervention (modified formula) is introduced at, or before, the expected age of onset of eczema. In addition, the GINI study does not include the findings of the breastfed infants (n = 945) in the analysis (because this group differed significantly from the formula-fed group).
Soy Formula and Other Mammalian Milks and the Prevention of Allergy
Soy formulas have long been used as cow’s milk formula alternatives. Osborn and Sinn, in a recent Cochrane review, concluded that, on current evidence, the use of soy formulae could not be recommended for the prevention of allergy or food intolerance in infants at high risk. No study reported an increase in soy allergy. There is also no evidence to support the use of other mammalian milks for the prevention of food allergy.
Summary
There is evidence that for high-risk infants who are unable to exclusively breastfeed the use of hydrolyzed infant formula may protect against the development of allergic disease, particularly eczema. It may be that protective effects better reflect tertiary prevention (ie, reduction of disease expression in infants with preexisting milk allergy) rather than true primary prevention. Overall, with 1 notable exception, the studies suggest a priority of an extensively hydrolyzed formula over a partially hydrolyzed formula.
Complementary Feeding (with Solid Foods)
Before 2001, WHO recommended that infants be exclusively breastfed for between 4 and 6 months, with the introduction of complementary foods thereafter. In 2001, after expert consultation and the systematic review by Kramer and Kakuma, the recommended period of exclusive breastfeeding was extended to the first 6 months of life. Advice in this regard is conflicting, with other allergy organizations recommending exclusive breastfeeding for between 4 and 6 months. Since 1975, there has been a significant trend in developed countries for the later introduction of solid foods. For example, in the United Kingdom, the proportion of infants given solids by 8 weeks of age decreased from 49% in 1975 to 24% in 1980 and 1985 and 19% in 1990. This decrease has coincided with a 3-fold increase in allergy in children. Reasons for differences in complementary feeding are complex and early weaning has been associated with cultural and socioeconomic factors, as well as specific factors such as maternal age, formula feeding, and maternal smoking. All of these factors need to be controlled for in study analyses.
A review of the evidence of the relationship between the early (defined as <4 months of age) introduction of solid foods to infants and the development of allergic disease was recently performed by Tarini and colleagues. Thirteen studies met their criteria for review, of which only one was controlled. Studies were not limited to at-risk study populations. These investigators conclude that there is insufficient evidence to suggest that, on its own, the early introduction of solids to infants is associated with an increased risk of asthma, food allergy, allergic rhinitis, or animal allergies. Tarini and colleagues note the consistent association between the persistence of eczema and the introduction of solid foods before age 4 months that is supported by long-term follow-up studies and the dose-dependent nature of the association.
Reverse causality has been proposed as an explanation for many of the study findings in this field. There are 2 studies that may be less severely affected by the effects of reverse causality. The first study by Fergusson and colleagues was performed in New Zealand in 1977 (ie, before the publication of WHO weaning recommendations); feeding practices would not therefore have been influenced by these recommendations. These investigators report a 2.9-times greater risk of chronic or recurrent eczema amongst children fed 4 or more solids before 4 months of age compared with those not fed solids before 4 months of age. This difference was still apparent at 10 years of age. However, individually, exposure to cow’s milk, egg, cereals, vegetables, meat products, or fruit did not increase the risk for the development of atopic dermatitis. Zutavern and colleagues studied a large population-based, multicenter cohort; they controlled for the effects of reverse causality and assessed the effect of early life diet on allergy outcomes. No measurements were made for sensitization to foods or food allergy. There was no evidence for a protective effect of late introduction of solids on the development of preschool wheezing, transient wheezing, atopy, or eczema. On the contrary, there was a statistically significant increased risk of eczema in relation to late introduction of egg and milk. The late introduction of egg was also associated with a nonsignificant increased risk of preschool wheezing.
A common finding in these studies is that the introduction of 4 or more foods during the first 4 months of life may increase allergic outcomes but not increase risk for the development of food allergy. Whereas the influence of complementary feeding with solid foods during the fifth or sixth month of life is less clear, complementary feeding after the age of 6 months seems not to influence allergic outcomes. Why 4 or more, as opposed to 1, 2, or 3 foods, would confer an increased risk is unclear, but may simply reflect a greater opportunity for the food-allergic infant to encounter an allergen to which they are allergic. A later section covers in greater detail studies that relate to the early introduction of single common food allergens such as egg and peanut.
Summary
Observational studies suggest that complementary feeding with 4 or more solid foods in the first 4 months of life may increase the risk of developing allergic disease (but not food allergy). There is no evidence that delaying the introduction of solid foods beyond 4 months of age is protective, and some evidence suggests that the delayed introduction of solids may promote allergies.
Combined maternal and infant dietary measures and the prevention of food allergy
It seems intuitive that of all interventions aimed at the prevention of food allergy the combined approach should offer the greatest hope because it covers the many routes of allergen exposure at times of immune vulnerability and pliability. The past 2 decades have witnessed a trend to allergen avoidance, with early calls to support this ; however, more recently many have questioned the wisdom of strict allergen modification in the diet of pregnant and breastfeeding women and the newborn diet.
Kramer and Kakuma, in a Cochrane review, assess the evidence for the prevention of allergic disease through maternal dietary antigen avoidance during pregnancy or lactation, or both. Their analysis finds that the prescription of an antigen avoidance diet to a high-risk woman during pregnancy is unlikely to substantially reduce her child’s risk of atopic disease. These investigators warn against the potential dangerous nutritional consequences of overzealous and unsupervised dietary restrictions. However, they do acknowledge that the prescription of an antigen avoidance diet (to high-risk women) during lactation may reduce the child’s risk of developing atopic eczema, but better trials are needed.
Of the existing studies in this field, 2 pioneering randomized studies adopt a multi-intervention approach (in both mother and child) with subsequent long-term blinded follow-up. Clinical assessments were robust in both these studies. Zeiger and Heller report a significant reduction in milk sensitization and eczema before age 2 years but no differences for food allergy, atopic dermatitis, allergic rhinitis, asthma, any atopic disease, lung function, food or aeroallergen sensitization, or serum IgE level, at 7 years of age. The early dietary differences were almost entirely caused by cow’s milk allergy, so it is not surprising that the early differences between the 2 groups were not sustained, and disappeared by 2 years of age. No difference in skin prick testing or sp-IgE was shown for the other food allergens tested, including peanut, which was the most common skin- test–positive food allergen at 7 years of age, indicating that the beneficial effect of the dietary interventions was mainly in reducing allergy to cow’s milk.
Arshad and colleagues also reported that allergic disease (asthma, atopy, rhinitis, eczema) could be reduced, at least for the first 8 years of life, by combined food and house dust mite allergen avoidance in infancy. These investigators recruited infants at high risk of atopy and randomized them to either a prophylactic (n = 58) or control (n = 62) group. Infants in the prophylactic group were either breastfed (with the mother on a low-allergen diet) or given an extensively hydrolyzed formula. The control group followed standard UK Department of Health advice. Repeated measurement analysis, adjusted for all relevant confounding variables, confirmed a preventive effect on asthma, atopic dermatitis, rhinitis, and atopy. The protective effects were primarily observed in the subgroup of children with persistent disease (symptoms at all visits) and in those with evidence of allergic sensitization. Study powering did not allow for the assessment of food allergy at 8 years of age, but earlier transient effects show some protection, at least for cow’s milk allergy. The specific preventative contribution of the dietary and aeroallergen interventions is unclear.
Summary
There are randomized trials that adopt a multi-intervention approach (ie, dietary modification of both maternal diet during pregnancy and lactation and infant diet) that report a reduction in allergic disease. Although findings in 1 study were transient and no longer observed at 7 years of age, in a second study the effects in allergy reduction were still observed at 8 years of age. The effects with respect to a reduction in food allergy seem to predominantly apply to cow’s milk allergy. Caution is required before the recommendation of such interventions because of the potential for nutritional compromise in both mother and child.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
