Eczema, asthma and allergy Jean Gelding”, Pauline M. Emmett, Imogen S. Rogers Unit of Paediatric and Perinatal Epidemiology, Institute of Child Health, University of Bristol, 24 Tyrtdall Avenw, Brisrnt BSS ITQ, UK Abstract The literature in relation to the development of atopic and allergic disorders has been reviewed, in order to assess the claim that prolonged and exclusive breast feeding protects against the development of such disorders. The data in the literature show little consistent evidence to identify any protective association between breast feeding and either eczema, wheezing/asthma or other types of atopy or allergic response. 0 1997 Elsevier Science Ireland Ltd. Keywords: Breast feeding; Eczema; Asthma; Wheezing; Allergy; Atopy 1. Introduction Received wisdom states that breast feeding protects a child from allergy. Advice to mothers, particularly if there is a family history of asthma, eczema or hay fever, has been that exclusive breast feeding is vital to protect the child from developing allergic or atopic responses. In this paper we examine the evidence for this point of view. There are three types of studies addressing this question-(i) studies of whole populations of children, regardless of family history; (ii) studies confined to children with a family history of asthma or eczema; (iii) studies confined to children with a specific disorder. Each published study based on reliable numbers will be discussed in some detail. 2. Population studies The first population study was reported in 1936 [I]: follow-up of 20 061 infants to nine months of age reported dramatic differences in prevalence *Corresponding author. 0 1997 037%3782/97/$17.00 PII SO378-3782(97)0005%3 Elsevier Science Ireland Ltd. All rights reserved of eczema with s122 J. Gold@ et al. I Early Human Development 49 Sup@ (1997) SIZl-S130 feeding habit. In comparison with totally breast fed infants, those partially breast fed had twice the prevalence, and those never breast fed, seven times the prevalence. These differences were highly significant. No subsequentpopulation study has been able to reproduce such dramatic findings however, and the findings may be consequent upon the artificial feed preparation used (boiled cows’ milk with cane sugar). The Christchurch Child Development Study is a prospective study in New Zealand following a cohort from birth. Important analyses have assessedcomponents of the diet during the first four months of the child’s life and its relationship with early childhood eczema [2]. The authors found that the rate of eczemaincreasedin almost direct proportion to the number of different types of solid food that the child had been given during the first four months, but that breast feeding itself had no significant relationship with eczema rates. The 1110 children were studied further and the analysis showed no association between the rates of asthma and either breast feeding or the introduction of weaning foods [3]. This was regardlessof whether or not there was a parental history of asthma. A study from Tuscan, Arizona carefully followed up 1246 healthy infants born between 1980 and 1984 [4]. Detailed assessmentsof the health of the children were made at defined intervals by paediatricians. Feeding patterns were identified at well-baby clinic visits and from questionnairescompleted by the parents. By the end of the first year of life, 33% of children had had a lower respiratory tract illness and 20% of children had had wheezing attacks. The study showed that among children breast fed for less than one month or not at all, 12.3% had started wheezing by four months of age; the prevalence for children breast fed for one to three months was 8.1% and for four + months was 5.2% (P < 0.005). Logistic regression in this study allowed for maternal education, maternal smoking, ethnic origin, child’s sex, whether either parent reported childhood respiratory problems themselves, whether others slept in the same room as the child and the type of day care. When all confounding variables were entered into the model, bottle feeding was still significantly associated with early wheezing illnesses [odds-ratio (OR) 1.71but with much reduced significance (PCO.05). A national sample of 5672 children in the United States concentrated on the analysis of factors relating to the development of asthmaor wheeze [5]. In this study, asthma was defined as diagnosed by the local doctor. Analysis of the data using logistic regression allowing for birthweight, maternal age, gestation, the child’s body mass index, triceps skinfold thickness, energy intake as well as socio-economic factors such as low income and living in the centre of a city, resulted in a non-significant association between breast feeding and asthma. Those bottle fed had an OR for asthmaof 1.1 (95% CI, 0.78-1.56) and for wheeze (i.e. wheezein the past twelve months apart from colds or ‘flu) an OR of 1.31 (95% CI, 1.01-1.70) P < 0.05. It should be noted here, however, that parental smoking was not considered,and may well have confounded the result. A study of allergic diseasesin children was mounted in Sweden on the basis of a questionnaire sent to 20 000 parents of school children aged seven, ten and fourteen J. Gold& et al. I Early Human Development 49 Suppl (1997) SIZI-5’1.30 s123 years of age [6]. The authors documented asthma prevalence as 2.4%, eczema at 7.8%, and total allergic diseasesincluding hay fever at 16.9%. There was considerable geographic variation throughout the country and, as expected, a parental history of allergy was associated with a two to nine times increased incidence in the offspring. Children, both of whose parents had a history of allergy, were shown to suffer a cumulative effect. The authors showed from the data available that although breast feeding appearedto postpone allergic disease,this was only true of children with both parents having a history of allergy (P < 0.01). It should be noted, however, that multivariate analyses were not undertaken in this study. A study of the effects of breast feeding was carried out in the Isle of Wight in the UK [7]. By one year of age, it was shown that those children one or both of whose parents had a history of atopic disease were more likely themselves to exhibit manifestations of allergy, than infants from non-allergic families. The authors reported that breast feeding apparently gave some protection against respiratory tract allergies, but eczemawas found as often in the initially breast fed as in the initially bottle fed. On follow-up of thesechildren to their fourth birthday, the authors showed that breast feeding was still not associatedwith eczema,whether or not one included those known to have been exclusively breast fed as opposed to those who also received supplementary milk. In fact more eczema was reported in the breast fed group. At four years, the initial apparent association with asthma found at one year had vanished and the breast fed group of children were as likely to have developed asthmaas the non-breastfed group. The authors explained the difference between the one and four year results by the inclusion of many casesof bronchiolitis in the first year. There have been several studies from the 1970 birth cohort which followed-up 14 000 infants born in one week throughout Great Britain. An analysis was made of the children who by the age of ten were reported to have wheezed after twelve months of age [S]. Seventeenpercent of children had had such a history. Logistic regression, taking account of maternal smoking, paternal smoking, paternal absence. sex of the child, family allergy symptoms, crowding in the household, dampnessof the bedroom, gas cooking and heating and social status, showed an OR for not being breast fed of 1.22 (95% CI, 0.97-1.53), which was not statistically significant. Analyses from this study concerning childhood eczema considered the development of this disorder by the age of five. Unadjusted associations showed a positive relationship between duration of breast feeding and onset of eczemain the first five years [9], but logistic regression analyses resulted in this association disappearing [lo]. The authors concluded “There is very little evidence to suggestan association between breast feeding and eczema as reported by the mother. and certainly no evidence of any protective effect of breast feeding. Furthermore the possibility of such a relationship pertaining solely to infants of mothers with a positive history of atopic disorders can be discounted:... the first order interactions between breast feeding and the two atopic history variables were both non-significant”. Population studies have also assessedrelationships with allergy using skin prick criteria. Such a study was undertaken of 930 children aged between nine and fifteen S124 J. Gelding et al. I Early Human Development 49 Suppl (1997) S121-S130 years in six Italian towns [ 111.They were skin prick tested and defined as atopic if at least one skin prick test caused a weal greater than 3 mm diameter. Their family history of allergy, personal allergy, medical history, allergic symptoms, exposure to allergens, infant feeding and parents’ occupations were recorded and logistic multiple regression analysescarried out. The significant findings concernedthe area in which they lived, housing conditions, exposure to environmental allergens, history of rhinitis, asthmaand/or eczemaand breast feeding. However, the relationship of atopy with breast feeding was not in the direction anticipated by the authors,-both unadjusted and adjusted data indicated that breast feeding was positively associated with atopy. A study in a totally different environment in eastern Nepal of children aged between five and fifteen also tested for allergy using the skin prick method [12]. The authors identified one in five children as being skin prick positive to at least one of the allergens, and one in five as having symptomsof allergic disease,but pointed out that there was no relationship between the symptomsand the skin prick positive tests. They found no relationship whatsoever between the skin prick test results and the length of time the child was breast fed. It should be pointed out that this was an area in which breast feeding was initiated in all except three of the 293 children involved, nevertheless,more than half these children had ceasedto be breast fed by the fourth month of age. A study from Denmark identified 39 children from 1749 that were being followed prospectively from birth as fulfilling criteria for the development of cows’ milk allergy [ 131.Seventeenof them had developed the symptoms during breast feeding, before the age of three months. By our calculations from their published data this implies that 1 in 100 children receiving breast milk apparently exclusively are likely to develop cows’ milk allergy. The authors found, however, that all 39 children had been given formula in the newborn nursery. Only 210 of the 1749 infants were not given such supplements-none developed cows’ milk allergy (but this was not statistically significant). A Canadian study followed 787 babies until the secondyear of life and undertook detailed challenge studies to identify cases of cows’ milk allergy. This occurred among 26 (7.0%) of 374 breast fed and 33 (8.0%) of 413 bottle fed children-a non-significant result. The authors concluded: “There is no evidence from this data that babies brought up on the breast for a few months are protected from cows’ milk allergy” [14]. In Finland [15] the effect of exclusive breast feeding through the first birch pollen seasonwas examined in 59 children and compared with 67 children on cows’ milk formula and 27 children who started breast feeding but were weaned onto formula during the first birch pollen season.The authors showed that exclusive breast feeding throughout the first birch pollen season did not prevent birch pollen allergy and indeed feeding cows’ milk-based formula during this seasonseemedto protect from subsequentdevelopment of such an allergy. Similar results were obtained in regard to allergy to grass pollen. The authors suggestthat immunological stresson the young infant provided by the introduction to cows’ milk at an early age may have had this unexpected benefit. 1. Gelding et al. ! Early Human Development 49 Suppl (1997) SIZI-S130 3. Studies of children fit25 with a family history A number of studies of children with a family history have been undertaken. One in London enrolled 73 infants with a parent who had a history of hay fever or asthma [ 161.The parents kept a diary of their feeding methods and smoking habits, and the date of the first cows’ milk feed was recorded. The children were clinically examined at birth, at three, six and twelve months, and annually thereafter. Almost half (36) of the children developed eczemaand 32 developed one or more episodesof wheezing, Fifty-five percent of the children had received exclusive breast milk during the first four weeks of life, but only 12% were totally breast fed for the first three months. There was no evidence from this study that breast feeding was protective-whether one measuredatopy by a positive skin test, the presenceof eczemaor a combination of the two. Indeed, the highest rate of the latter category (skin test positive and eczemapresent) was in children who were breast fed for longer than thirteen weeks (five (55%) out of nine), comparedwith 16 (25%) of 64 children who were breast fed for less than this. The strongest relationship with wheeze was parental smoking. Businco et al. [17] had followed 101 babies with a familial history of asthma,hay fever or eczemauntil the age of two. If the mother was successfully breast feeding by two weeks, she was encouraged to restrict her own intake of cows’ milk to 200 ml/day and to have no more than two eggs/week. If she needed/wanted to supplement or give up altogether, the baby was given a soy basedformula. Weaning foods were not given until six months and then restricted to those thought not to be allergenic until nine months. The comparison group had either never breastfed, or not done so for longer than two weeks. The babies were given cows’ milk formula, and no restriction was placed on weaning food. This study resulted in the statementthat breast feeding + soy supplements“exerts a prophylactic effect on the development of atopic disease”. This widely quoted study did no statistical tests-and indeed had they done so they would have had to report no significant effect [ 1S]. A study from Northwick Park Hospital [19] followed 92 children who had at least one atopic parent, to the age of five years. Skin prick tests and immunoglobulin E (IgE) antibody levels were measured over time. The authors stated that “Breast feeding was associatedwith an increase in the prevalence of positive results in skin tests,but was not associatedwith detectableIgE antibodies to food proteins... and was not associatedwith protection against the development of disease”. Kaufman and Frick [20] followed 94 infants of allergic mothers to the age of two: 38 were breast fed-two (6%) of these developed asthma, whereasten ( 18%) of the 56 bottle fed infants did so (p < 0.06). Although the authors did not report the prevalence of eczemain breast as opposed to bottle feeding, they noted that among those who had eczema,those who were bottle fed were significantly more likely to develop asthmathan those who were breast fed-but the numbers involved were not given-and consequently are impossible to interpret. In Canada, 37 children who had an older sibling with allergic disease and were exclusively breast fed for the first six weeks of life or longer, were compared by Chandra with 37 who had a similar older sibling but were fed cows’ milk formula [21]. There were statistically significantly lower rates of eczema V’< 0.001). S126 J. Golding et al. I Early Human Development 49 Suppl (1997) S121-S130 recurrent wheezing (P<O.Ol), high serum IgE (P <O.OOl), and other biochemical markers indicating a reduced risk of allergy in the breast fed as opposed to the formula fed children. A further study by Chandra and colleagues was a randomised controlled trial of three different formulae-a whey hydrolysate formula, a conventional cows’ milk formula and a soy based formula; these three groups were also compared with a group of mothers who breast fed exclusively for four months or longer. There were 72 children in each of the four groups. It is not clear to what extent the breast fed group was similar to that of the three randomised groups, but the authors concluded that “Exclusive breast feeding for more than four months is partially protective against the development of atopic disease among high risk infants. Among those not breast fed, feeding a milk/whey hydrolysate formula to infants at high risk because of history of atopy among first degree relatives reduces the incidence of atopic disease and this approach is significantly beneficial compared with breast feeding without maternal dietary restriction or feeding a soy based formula” [22]. The largest and most notable study was of 519 infants in South Wales whose mothers were participating in a randomised controlled trial of withholding cows’ milk, and substituting soya milk for those who were not breast fed; infants were followed up and examined for evidence of allergic disease at three, six and twelve months of age. Analysis of the factors in these children that predicted wheeze, treated the study as though it was an observational survey [23]. Logistic regression allowed for the number of siblings, the season of birth, the sex of the child, whether damp was present in the home, mother’s smoking habit and social class of the household. In the presence of these factors, the authors identified a statistically significant decrease in wheezing in children who had ever been breast fed (OR 0.53, 95% CI, 0.33-0.86, P < 0.01). There was no significant difference, however, in the prevalence of eczema between those who had been breast fed at all and those who had never been breast fed. Subsequent follow up to age seven of 453 of these children [24] showed that although the reduced rate of wheezing persisted in children who had received breast milk, this was only found in children who were not atopic. Burr and colleagues showed that wheezing with onset in the first two years was much more likely to have an infectious aetiology than wheezing with later onset. They concluded that breast feeding protects against wheezing episodes initiated by infection but not against asthma. 4. Studies confined to children with specific histories A study from Iraq [25] considered 100 children with wheezy bronchitis and showed that more of this group were breast fed when compared with children of similar ages who were attending a vaccination clinic. The authors suggested that the mechanism involved rapid growth of the breast fed infants in this community, since the immune properties of breast milk are likely to have given protection from severe infection and resultant malnutrition. In Iraq supplementing breast milk with cheap .I. Golding et al. I Early Human Development 49 Suppl (1997) S121-S130 S12? starch products such as biscuits is a common practice, resulting in overweight children who are commonly seen with wheezy bronchitis. Investigators in Israel followed prospectively the clinical course of diseasein 80 atopic wheezing babies from six months to five years [26]. They statedthat there was no correlation between changesin IgE levels and early feeding with breast or bottle milk, but that persistent wheezing in these children was associatedwith both bottle feeding and parental smoking. Their statistical analysis did not adequatelydistinguish whether one of these factors explained the other. An Australian study which had identified 371 seven year old children with wheezing and followed them up until 21 years of age, categorised the types of wheezing and asthma into four grades [27]. The authors found that those subjects with the most severewheezing at age 21 were more likely to have been breast fed in infancy and for longer periods than the group with no symptomsor those with mild or moderateasthma.The authors suggestthat “further investigation is neededbefore the hypothesis that breast feeding protects against the development of allergic phenomena can be accepted”. In one study 49 children with eczema who were exclusively breast fed and had never received anything except breast milk were compared with 34 similarly fed infants but without eczema[28]. Almost half the children with eczemawere shown to be hypersensitive to foods such as egg and cows’ milk (23 out of 49). Breast milk samples were collected from nineteen mothers after the ingestion of egg, but ovalbumin was detectedin similar concentrationswhether the mother had a child with eczema or not. A study from Canada[29] investigated nineteen children with IgE mediated allergy associatedwith positive skin prick tests to peanut, cows’ milk and/or egg. In all. seventeenof the children had been breast fed, ten exclusively for at least five months. Since reaction to the foods had occurred on first exposure to the food in almost all instances, it was proposedthat sensitisation had occurred antenatally or via the breast milk. A study of the mothers of sensitisedchildren showed that they did not consume more of these foods than the mothers of non-sensitised children and indeed, avoidance of the foods by some mothers did not ensure freedom from sensitisation. They concluded “Breast feeding by itself cannot be guaranteedto protect against the development of food allergy”. 5. Discussion Chandra [30] suggestedthat there are “at least five possible reasonsfor expecting breast fed infants to show a reduced occurrence of atopic disease.Firstly, breast fed infants are less exposed to ‘foreign’ dietary antigens. Secondly, human milk is postulated to contain factors that mature the intestinal mucosa,thereby allowing early ‘closure’ of macromolecular absorption. Thirdly, by reducing the incidence of infection and altering the gut microflora that can act as an adjuvant for ingested food proteins, the possibility of sensitisation may be decreased.Fourthly, human milk has anti-inflammatory properties that will also decreasemacromolecular uptake. Finally, S128 J. Gelding et al. I Early Human Development 49 Suppl (1997) S121-S130 the presence of various cytokines in human milk may play an important role in modulating the development of allergic diseases”. How clear is it though that there is a reduction in atopic disorders? Reviews of the literature on breast feeding and the development of allergy in the past tended to infer that breast feeding did indeed prevent atopic or allergic disorder, but probably only among children of affected parents [31-331. The data from general population studies certainly lack a clear pattern. Although the earliest study showed a protective effect for breast feeding in relation to development of eczema in the first nine months [ 11, other studies [2,7,10] showed no sign of any protective relationship. For asthma and/or wheezing, no protective effect was found in one study [3], it was evident in others [4,5,7,8] although in one of these the effect was found up to one year but not thereafter [7], and in another [8] statistical significance was not achieved. More general assessment of allergy using a combination of asthma, eczema, and/or hay fever [6] concluded that although there was an initial protective effect of breast feeding, this was only found where both parents had a history of allergy, and even in this group, the reduced prevalence was only a postponement of development of the disorder. Indeed skin prick testing showed, in one population, that allergic response was significantly more likely in children who had been breast fed [ll]. A distinction between ever and never breast fed was unable to be made in the Nepal study as almost all (99%) children had received some breast milk. Nevertheless, prolonged duration did not show a benefit in comparison with short duration [12]. Among the five studies of children with a familial history, one showed no relationship with eczema, wheezing or positive skin test [16], one showed increased IgE antibodies among those who had been breast fed [19], one showed no relationship with eczema but a significant reduction in wheezing in the first year of life [23]; only the two studies by Chandra and colleagues, however, claimed substantial benefits for breast feeding [21,22] but it is not clear how comparable the groups were in his studies. Some of the confusion in the studies may be due to confusion of diagnoses. This was amply shown by the seminal study of Burr and colleagues [23,24]. Although breast feeding was significantly associated with a reduced prevalence of wheezingthis seems to have been entirely due to a reduced prevalence of lower respiratory infection not of asthma. Another possible explanation for the striking difference between the protective findings in the study of the 1930s may lie in either the fact that prolonged unsupplemented breast feeding was common at the time, or that mothers had a more restricted diet and were less likely to expose their infants to allergens via the breast milk. Criticisms of studies that fail to find a protective effect have tended to suggest either the failure of a mother to continue to breast feed for long enough, or to have introduced non-milk supplements too early or have failed to avoid, while breast feeding, all those substances that might provoke an allergenic response in her child. Reality, however, is that mothers are unlikely to conform to such restrictions without J. Golding et al. I Early Human Development 49 Suppl (1997) Sl21-St30 s129 themselvesundergoing immense stressand anxiety. From the information available it is the authors’ opinion that breast milk itself is likely to have such major benefits in preventing infection that no further justification is needed for breast feeding. The stressto a mother of restricting her own and her child’s diet in the hope of preventing allergy (with consequentguilt if the child develops asthmaor eczema)is not justified on the evidence available. References [l] Gruelee CG, Sanford HN. The influence of breast and artificial feeding on infantile eczema. J Pediatr 1936;9:223-5. 121 Fergusson DM, Horwood LJ, Beautrais AL, Shannon ET, Taylor B. Eczema and infant diet. Clin Allergy 1981;11:325-31. [3] Fergusson DM, Horwood LJ, Shannon FT. Asthma and infant diet. Arch Dis Child 1983;58:48-51. [4] Wright AL, Holberg CJ, Martinez FD et al. Breast feeding and lower respiratory tract illness in the first year of life. Br Med J 1989;299:946-9. [5] Schwartz J, Gold D, Dockery DW, Weiss ST, Speizer FE. Predictors of asthma and persistent wheeze in a national sample of children in the United States: Association with social class, perinatal events and race. Am Rev Resp Dis 1990;142:555-62. [6] Aberg N, Engstrom I, Lindberg U. Allergic diseases in Swedish school children. Acta Paediatr Scnnd 1989;78:246-52. [7] Hide DW, Guyer BM. Clinical manifestations of allergy related to breast and cow’s milk feeding. Pediatrics 1985;76:973-5. [8] Neuspiel DR, Rush D, Butler NR, Golding J, Bijur PE, Kurzon M. Parental smoking and post-infancy wheezing in children: a prospective cohort study. Am J Public Health 1989;79: 168-7 I. 191 Golding J, Peters TJ. The epidemiology of childhood eczema: I. A population based study of associations. Paediatr Perinat Epidemiol 1987;1:67-79. [IO] Peters TJ, Golding J. The epidemiology of childhood eczema: II. Statistical analyses to identify independent early predictors. Paediatr Perinat Epidemiol 1987;1:80-94. [ 111 Astarita C, Hanis RI, de Fusco R, Franzese A, Biscardi D, Mazzacca FRM, Altucci P. An epidemiological study of atopy in children. Clin Allergy 1988;18:341-50. [ 121 Hope RA, Bangham CRM. The prevalence of immediate positive skin tests in Nepalese children. Clin Allergy 1981;11:263-71. [13] Host A, Husby S, Osterballe 0. A prospective study of cows‘ milk allergy in exclusively breast fed infants. Acta Paediatr Stand 1988;77:663-70. [ 141 Gerrard JW, MacKenzie JWA, Goluboff N, Garson JZ, Maningas CS. Cows’ milk allergy: prevalence and manifestations in an unselected series of newborns. Acta Paediatr Stand Suppl 1973;234:3-21. [ 151 Saarinen UM, Kajosaari M, Backman A. Birch pollen allergy in children: role of milk feeding during the firs1 birch season of life. Allergy 1982;37:345-50. [ 161 Cogswell JJ, Mitchell EB, Alexander J. Parental smoking, breast feeding and respiratory infection in development of allergic diseases. Arch Dis Child 1987;62:338-44. [17] Businco L, Marchetti F, Pellegrini G, Cantani A, Perlini R. Prevention of atopic disease in eat risk newborns’ by prolonged breast feeding. Ann Allergy 1983;51:296-9. [ 181 Kramer MS. Does breast feeding help protect against atopic disease? Biology, methodology and ;I golden jubilee of controversy. J Pediatr 1988;112:181-90. 1191 Rowntree S, Cogswell JJ, Platts-Mills TAE, Mitchell EB. Development of IgE and IgG antibodies to food and inhalant allergens in children at risk of allergic disease. Arch Dis Child 1985:60:727--X 1201 Kaufman HS, Frick OL. Prevention of asthma. Clin Allergy 1981;l l:549-53. [21] Chandra RK. Prospective studies of the effect of breast feeding on incidence of infection and allergy. Acta Paediatr Stand 1979;68:691-4. s130 J. Gold@ et al. I Early Human Development 49 Suppl (1997) S121-S130 [22] Chandra RK, Singh G, Shridhara B. Effect of feeding whey hydrolysate, soy and conventional cow milk formulas on incidence of atopic disease in high risk infants. Ann Allergy 1989;63:102-6. [23] Burr ML, Miskelly FG, Butland BK, Merrett TG, Vaughan-Williams E. Environmental factors and symptoms in infants at high risk of allergy. J Epidemiol Community Health 1989;43:125-32. [24] Burr ML, Limb ES, Maguire MJ, Amarah L, Eldridge BA, Layzell JCM, Merrett TG. Infant feeding, wheezing and allergy: a prospective study. Arch Dis Child 1993;68:724-8. [25] Sahib El-Radhi, A, Majeed M, Mansor N, Ibrahim M. High incidence of rickets in children with wheezy bronchitis in a developing country. J R Sot Med 1982;75:884-7. [26] Geller-Bernstein G, Kenett R, Weisglas L et al. Atopic babies with wheezy bronchitis: Follow-up study relating prognosis to sequential IgE values, type of early infant feeding, exposure to parental smoking and incidence of lower respiratory tract infections. Allergy 1987;42:85-91. [27] Martin AJ, Landau LI, Phelan PE. Natural history of allergy in asthmatic children followed to adult life. Med J Aust 1981;2:470-4. [28] Cant A, Marsden RA, Kilshaw PJ. Egg and cows’ milk hypersensitivity in exclusively breast fed infants with eczema and detection of egg protein in breast milk. Br Med J 1985;291:932-5. [29] Gerrard JW, Perelmutter L. IgE-mediated allergy to peanut, cows’ milk and egg in children with special reference to maternal diet. Ann Allergy 1986;56:351-4. [30] Chandra RK. Interactions between early nutrition and the immune system. In: Ciba Foundation Symposium 156--The childhood environment and adult disease. Chichester: Wiley, 1991. [31] Burr ML. Does infant feeding affect the risk of allergy?. Arch Dis Child 1983;58:561-5. [32] Host A. The influence of early allergen contact on the development of atopy in childhood. Allergologie 1989;12: 186-91. [33] Zeiger RS. Prevention of food allergy in infancy. Ann Allergy 1990;65:430-45.