PREVENTION OF FOOD ALLERGY IN INFANTS: TIMING OF FOOD INTRODUCTION

I will start with several "facts" that provide some of the underpinning for the discussion.

1. FOOD ALLERGY OCCURS IN 4 - 6% OF CHILDREN Sampson 2004.  Host found that cow milk protein allergy (CMPA) is the commonest food allergy in early childhood with an incidence of 2 to 3% in the first year of life. Host 2002.
   
   
2. SEVEN FOODS CREATE THE MAJORITY OF REACTIONS (cow milk, eggs, peanut, nuts, fish, shellfish, seeds).
   
   
3. ALLERGY TO COW MILK, EGG AND SOY FREQUENTLY REMIT. Thong 2004  In a study by Host et al. allergy to cow milk protein had resolved in 77% of children by age 2 Host 2002 with an overall remission rate of approximately 85 to 90% Host 2002.
   
   
4. ALLERGY TO PEANUT, NUTS, FISH PERSISTS. A recent study in Montreal found that over 1% of children are allergic to peanut. Kagan 2003. The prevalence of peanut sensitivity in industrialized countries is increasing Sicherer 2003. Sensitivity to peanut is outgrown in about 20% of instances Skolnick 2001. That is far less than the number outgrowing milk allergy but still better than the original suggestion that peanut allergy is never outgrown.
   
   
5. ALLERGIC MARCH: IN INFANCY CHILDREN BECOME ALLERGIC TO FOODS AND AFTER AGE 3 TO AEROALLERGENS AND MUCH LESS COMMONLY TO FOODS. The "allergic march" was first described by Foucard 1973, confirmed by Zeiger 1995 and Zimmerman 1988.
   
   

SUMMARY OF THE 2004 CRITICAL REVIEW OF THE LITERATURE Muraro 2004

• At present prospective interventional studies in high-risk infants show evidence of the effect of dietary allergy-preventive measures as regards food allergy, especially cow’s milk protein allergy, and eczema
• Feeding exclusively human milk for at least 4 months is associated with a lower cumulative incidence of cow milk allergy until 18 months.
• No convincing evidence for a preventive effect of maternal diet during pregnancy or lactation.
• A documented hypoallergenic formula (extensively hydrolysed formula) combined with avoidance of solid foods for 4-6 months reduces the cumulative incidence of cow milk allergy and eczema.
• Partially hydrolysed formula may have an effect, although it seems to be less than that of extensively hydrolysed formula at present.
• No evidence for a preventive effect of a diet after the age of 4-6 months, although this needs additional investigation.
• This preventive effect has only been demonstrated in high-risk infants.

RECOMMENDATIONS OF THE CRITICAL REVIEW PANEL:

1. Breastfeeding is highly recommended for all infants irrespective of atopic heredity. Although the number of high quality observational and interventional studies is limited, the following evidence-based recommendations should be followed:
2. A dietary regimen is effective in the prevention of allergic diseases in high-risk patients
3. The most effective dietary regimen is exclusively breastfeeding for at least 4-6 months or, in case of lack of breast milk, formulas with documented reduced hypoallergenicity for at least 4 months combined with avoidance of solid food and cow’s milk for the same period.
4. Based on this analysis no conclusive evidence for a protective effect of a maternal exclusion diet during pregnancy or lactation could be shown.
5. In addition, no controlled studies have yet determined the role of peanut avoidance in the prevention of atopic diseases (i.e. avoiding peanut until after age 3).
6. Prospective prevention studies currently available indicate that soy formulas are as allergenic as conventional cow’s milk formulas and are not effective for prevention of allergy.

There are a number of points that I have problems with in the "critical review".

1. They simply dismiss the question of when to introduce peanut or nuts by saying there is no data. However that begs the question. Many of the studies cited refer to cow milk or egg allergy and they are the most trivial food allergies since the literature supports the conclusion that most infants outgrow milk and egg sensitivity (and usually by 18 months). Similarly the dermatologists would probably argue that most children outgrow eczema by age 2 or 3. The real question is asthma and there the data is fuzzy. It is confused because preschool asthma, as the critical analyses states, is primarily non-atopic. In fact I think the evidence is strongest that breast-feeding prevents non-atopic preschool asthma (as shown in both the Tucson and Dunedin cohorts).
2. They dismiss the Sears contrarian study (Dunedin Cohort) which described increased allergy and asthma in breast-fed infants Sears 2002  by citing the weaknesses and never referring to it again. However that same population database has given us a lot of good data on the relationship of allergy to asthma and Sears has followed the population from childhood to adulthood. Sears has claimed that the preventative effect of breastfeeding on allergy and asthma is seen in short term studies but not long term studies ( N Engl J Med. 2004 Jan 15;350(3):304).
3. They dismiss the Martinez dataset (Tucson Cohort) where breastfeeding reduced nonatopic asthma but not atopic Wright 1995 Wright 2001. That dataset has given important information on the outcome of nonatopic preschool asthma vs atopic asthma. In the 2003 JACI review of the data coming out of that cohort Taussig 2003, it was suggested that infants in daycare or with more sibs have more wheezing in early childhood but less late onset asthma. They found that breast-fed infants had less early viral-induced wheeze but the children with maternal asthma and elevated IgE actually had more asthma if breast-fed. They postulated that the "hygiene hypothesis" might account for this perverse effect that is, infants suffering more infections might develop less allergy. This is accounted for in the critical review by "reverse causation" selection artefact that is mothers with atopy might breast-feed their infants longer with the hope of preventing allergy but this selects for a more allergy-prone population being breast-fed longer.
4. The critical review concluded that the recommendations apply to "high risk" populations but they suggest that the data from three studies that show increased risk of asthma in breastfed could be due to "reverse causation" i.e. highest risk families choosing to breast feed the longest. (papers by Wilson, Wright and Bergmann). But that is the "high risk" group where we are most commonly trying to provide recommendations.
5. Table 5, cites 9 studies of high risk infants with level 1b evidence proving that strict breast milk reduces atopy, food allergy and eczema (in levels of evidence, level 1b is strong evidence but the studies are small numbers of subjects). But 3 of the 9 studies are by the same author, R.K. Chandra and probably the same study population.
6. The three meta-analyses on this subject by Gdalevich, Mimouni et al. show an effect on asthma Gdalevich 2001  and eczema Gdalevich 2001  but not allergic rhinitis (actually a slight effect). So the latter study is dismissed as comprising too few numbers of studies for analysis Mimouni 2002.

One of the publications they clearly drew from was by Zeiger et al. who reported the effect of maternal and infant avoidance of allergenic foods on food allergy in a prenatally randomized, controlled trial of infants of atopic parents. The protocol of the prophylactic-treated group (N = 103) included: (1) maternal avoidance of cow's milk, egg, and peanut during the third trimester of pregnancy and lactation and (2) infant use of casein hydrolysate (Nutramigen) for supplementation or weaning, and avoidance of solid foods for 6 months; cow's milk, corn, soy, citrus, and wheat, for 12 months; and egg, peanut and fish for 24 months. In the control group (N = 185), mothers had unrestricted diets. The cumulative prevalence of atopy was lower at 12 months in the prophylactic-treated (16.2%) compared to the control (27.1%) group (p = 0.039), resulting from reduced food-associated atopic dermatitis, urticaria and/or gastrointestinal disease by 12 months (5.1% versus 16.4%; p = 0.007), and any positive food skin test by 24 months (16.5% versus 29.4%; p = 0.019), caused primarily by fewer positive milk skin tests (1% versus 12.4%; p = 0.001). The prevalences of allergic rhinitis, asthma and inhalant skin tests were unaffected. The authors concluded that reduced exposure of infants to allergenic foods appeared to reduce food sensitization and allergy primarily cow milk allergy during the first year of life Zeiger 1989. However when the children were followed to age 7, there were no differences in the prevalence of allergy, asthma, atopic dermatitis and food allergy Zeiger 1995. This suggested that since most children outgrow milk sensitivity anyway, it did not prevent sensitivity to peanut and other major food allergens.

SUMMARY OF RECOMMENDATIONS AMERICAN ACADEMY OF PEDIATRICS AAP2000

COMMENTS: The recommendations from the American Academy of Pediatrics are similar to the consensus document published recently in that they advise breast-feeding without other food proteins for 6 months and they encourage breast-feeding for a year. They recommend restrictions in the maternal diet while breast-feeding including peanut and tree-nuts (almonds, cashews etc.) with consideration given to restricting cow milk proteins, egg and fish. The 2004 consensus document 2004 does not recommend restrictions of maternal diet during breast-feeding since there is little supporting evidence.

Maternal dietary restriction during pregnancy and breast-feeding: There are certainly studies that demonstrate detectable food protein in the breast milk of lactating women including: cow milk protein  Axelsson 1986 Host 1988 Sorva 1994,  egg Cant 1985 and peanut  Vadas 2001. Since infants who are exclusively breast fed have been found to develop allergy to these foods it has been inferred that transfer of foreign food protein into breast milk might be responsible for this sensitization. From that inference, recommendations were made to restrict foods in the diet of lactating women. However when restricted diets during late pregnancy and breast-feeding were examined in controlled studies, there was found to be no difference in infant allergy in breast-fed infants whose mothers restricted their diet and those who did not. LATE PREGNANCY:  Falth-Magnusson 1987  Falth-Magnusson 1992  Lilja 1989 LATE PREGNANCY AND LACTATION: Herrmann 1996  Lilja 1991  AND LACTATION: Hattevig 1990  Sigurs 1992. Although the last two studies by the same group found some difference in the degree of sensitization to milk and egg between the restricted and non-restricted groups with infants in the restricted group having lower specific IgE to milk and egg than infants in the non-restricted group. So they became sensitized equally but perhaps were not boosted to the same degree. Thus there is no evidence that restricting the diet in late pregnancy and lactation changes the infant sensitization rate. Peanut has not been studied.

CONCLUSIONS:

• Can allergy be prevented by breast-feeding and late introduction of solid foods in infancy?
  The data suggest that strict breast-feeding with no supplementation or infant feeding with extensively hydrolysed formulas might be able to reduce the prevalence of atopy in high risk infants (i.e. infants born in families with parental or sibling allergy). However there are contrarian studies that suggest that breast-feeding might increase the risk of atopic asthma while reducing the risk of viral-induced, non-atopic preschool wheezing. Nevertheless, strict breast-feeding for 4 - 6 months should be recommended.
• Should pregnant and breast-feeding mothers restrict their own diet to prevent sensitization of the infant?
  There is clearly evidence that food protein crosses into the breast milk of lactating women suggesting the potential for sensitizing their offspring. However studies examining the result of restricting the diet of breast feeding mothers did not find a reduction in the level of atopy. These studies did not focus specifically on peanut and nuts. Since many highly allergic infants have IgE antibody to peanut in their serum even prior to having eaten peanut Zimmerman 1989 and this sensitization might have occurred during breast-feeding, I will continue to recommend that nursing mothers in high risk families should restrict peanut, nut and shellfish in their diet.
• At what age should the most allergenic foods be introduced to infants?
  There have been no studies directed specifically to peanut and nut introduction in high risk children. So it is not clear whether waiting "until the food sensitization window closes" actually prevents sensitization in some children who would have otherwise become sensitized to peanut by an earlier introduction. Certainly some of those high risk children have become sensitized before peanut was knowingly introduced into their diet - see the Zimmerman study cited above. However there is a general observation in the literature that the younger the infant at the time of introduction of allergenic foods, the more likely the sensitization. I will continue to recommend late introduction of peanut, tree nuts and shellfish in children at risk for atopy, i.e. after age three, with the hope that the "Allergic March" makes them less susceptible to sensitization at that age. If studies are published that specifically examine the question with peanut, nuts and shellfish then I will modify the recommendations to conform to the evidence. It could be argued that fish and seeds are also highly allergenic foods but I will not recommend that those foods be withheld until after age three because fish is a very healthy food and seed allergy is not that common. Milk and egg can probably be introduced to at risk infants within the second 6 months of life. Hopefully there will be further studies on the high risk population that specifically examine the question of introduction of the more allergenic foods into the diet of high risk infants.