Introduction
Allergic diseases represent a global health problem of increasing prevalence and severity. In some European countries, up to 50% of children demonstrate Immunoglobulin E (IgE) sensitization (‘atopy’) to inhalant or food allergens.Reference Bousquet, Anto and Auffray 1 Allergic diseases are complex multifactorial disorders, with both genetic and environmental components interacting during early life and probably in utero. Given the developed ‘plasticity’ of the prenatal and postnatal periods, environmental exposures, acting on genetically susceptible hosts, may have the greatest potential during these periods to influence the development of disease in early childhood.
In utero and early-life events appear to be important for the development of many chronic inflammatory-immune diseases including allergy and asthma.Reference Henderson and Warner 2 There are a number of environmental exposures during pregnancy and early life, which pose a potential risk for, or protect against, allergic disease development in the infant; these include inhalants such as allergens and molds.Reference Granum and Lovik 3 In addition, nutritional–metabolic, psychosocial and other lifestyle factors have been hypothesized to play key roles.Reference Bousquet, Anto and Auffray 1 Epidemiological and immunological studies suggest that dietary modification in fetal life, as well as fetal under- or over-nutrition, could affect the development of allergic diseases.Reference Chatzi, Torrent and Romieu 4 There are, however, conflicting data regarding the role of timing of introduction of solid foods, duration of breastfeeding and exposure to pets as potential factors, which could be targets for prevention or intervention in modulating the development or maintenance of allergic diseases.
In this study, we report the effects of environmental and nutritional factors during the prenatal and postnatal periods on atopic sensitization (atopy) and the development of allergic diseases (reported diagnosis of wheezing, eczema, allergic rhinitis or food allergy) in the prospective birth cohort Family Atherosclerosis Monitoring In earLY life (FAMILY) Study.Reference Morrison, Anand and Yusuf 5 , Reference Morrison, Atkinson and Yusuf 6
Methods
Study design
The FAMILY Study is a McMaster University/Population Health Research Institute-based longitudinal birth cohort established in 2005, and designed to examine early-life determinants in chronic disease development, with a primary emphasis on nutritional, metabolic and cardiovascular variables and their relationship with the development of childhood obesity and cardiovascular disease risk factors, as well as childhood allergy and asthma.Reference Morrison, Anand and Yusuf 5 , Reference Morrison, Atkinson and Yusuf 6
The study population included 901 babies, 857 mothers and 530 fathers (Table 1). Among them, data of 825 children, 783 mothers and 493 fathers were analyzed, based on availability of questionnaires and skin prick testing. In all, 857 women and 530 fathers were recruited after obtaining informed consent during pregnancy, drawing from three hospitals in Hamilton and Burlington, ON, Canada.
Table 1 Description of the study cohort
n=783 represents the number of pregnant mothers.
a As defined by self-reported eczema, rhinitis, asthma (wheezing) or a positive skin test.
Prenatal and postnatal determinants in the child and family for the development of obesity and cardiovascular disease, as well as select factors for allergic disease development, were measured as described in the following paragraph. Questionnaire data were collected from mothers during pregnancy (24–28 weeks of gestation) and at the 1-year visit.
Data were also collected on fathers (at the mother’s initial pregnancy visit) and on children at 1 year of age. These questionnaire data included information on pre-existing diseases, food dairies, nutritional supplements, medications, tobacco use, socio-demographic variables, maternal pet ownership and maternal exposure to farm animals. Data were also collected on duration of breastfeeding, infant food intake, medications taken by the infant during the 1st year of life and the mother during pregnancy, maternal smoking status prenatally and postnatally, vaccination history of the infant, and signs and symptoms of allergic disease in the infant (see Tables 2 and 3).
Table 2 Proportion of children with prenatal exposures by allergic outcomes at the 1st year
n is lower for atopy based on the available skin prick testing data.
Table 3 Proportion of children with postnatal exposures by allergic outcomes in the 1st year
The present study was approved by Research Ethics Boards of the participating hospitals (Hamilton Health Sciences, St Joseph’s Hospital – Hamilton, Joseph Brant Memorial Hospital, Burlington, ON, Canada).
Skin prick testing
Skin prick testing was performed by standardized methods on fathers during the prenatal visit and on mothers and babies at the 1-year visit. A positive control (histamine), a negative control (diluent), standard series of five food allergen extracts (wheat, egg, milk, peanut and nuts) and 14 inhalant allergen extracts [cat dander, dog dander, horse, feather, cockroach, house dust mite (Dermatophagoides farinae), house dust mite (Dermatophagoides pteronyssinus), Altenaria mold, Hormodendrum mold, Aspergillus mold, grass, tree, ragweed, weeds] were applied to the volar surface of the forearm for adults. In infants, multi-testers (Multi-Test II; Lincoln Diagnostics) were used – a positive control (histamine), a negative control (diluent), standard series of six food allergen extracts (wheat, egg, milk, soy, peanut and mixed tree nuts) and eight inhalant allergen extracts (cat dander, dog dander, feather, cockroach, house dust mite, grass, tree, ragweed). A skin prick test was deemed positive if the greatest diameter of the wheal was at least 2 mm greater than the greatest diameter of the saline control.Reference Skassa-Brociek, Manderscheid and Michel 7
Primary allergic outcomes for analysis
Primary outcomes at 1 year of age included atopy, food allergy and allergic disease. We defined atopy as having at least one positive skin test. For the purposes of these analyses, we defined ‘allergic disease’ as the development of asthma, wheezing, allergic rhinitis or eczema as reported by the mother. We defined food allergy in the child by 1 year of age as reported by the mother or if there was a positive skin test for food allergens in the child at 1 year.
Statistical analysis
Statistical analyses were conducted using SAS version 9.2 for Unix (SAS Institute Inc., Cary, NC, USA). Frequencies and proportions were calculated for all of the prenatal and postnatal exposures potentially associated with the outcomes of interest: allergic disease, atopy and food allergy. Multivariable logistic regression modeling was used to consider the effect of prenatal and early childhood environmental exposures, maternal nutritional exposures during pregnancy, postnatal breastfeeding and child nutritional exposures, and the use of medication in children on the development of allergic disease. For exposures that were highly correlated between the prenatal and postnatal periods, the prenatal measure was included (e.g. dog in home). Exposures with a univariate P-value of >0.10 were entered into the multivariable model. The final model was determined using backward selection methods, keeping variables with a P<0.05.
Results
There were 783 families with a 1-year follow-up visit. The follow-up cohort included 825 children (51% male and 90% singletons) with an average age of 13 months at follow-up and a mean gestational age of 39 weeks at birth. Of the 825, 818 had allergic disease measures captured: 576 had skin prick testing, whereas 793 completed self-reported allergy measures (Table 4).
Table 4 Allergic outcomes in children
n=825 represents the number of children (which included twin pregnancies accounting for the difference in numbers).
Rates of allergic disease and associated phenotypes noted in the FAMILY Study cohort were as follows: atopy 24.5%, food allergy 17.5%, cow’s milk allergy 4.8%, eczema 16.0% and wheezing 18.6% (Table 4). Rates of allergy were comparable within each ethnic group.
Multivariate analyses of prenatal and postnatal exposures
Prenatal and postnatal environmental and dietary exposures, which were associated with allergic disease with P<0.10 in the univariate analysis, were included in a multiple logistic regression (Table 5). Use of antibiotics in the child’s 1st year [odds ratio (OR): 2.04, 95% confidence interval (CI): 1.45–2.88] had an effect in increasing the odds of allergic disease, whereas use of acetaminophen in the 1st year (OR: 0.68, 95% CI: 0.51–0.92) and prenatal maternal exposure to dogs at home (OR: 0.60, 95% CI: 0.42–0.84) were ‘protective’ against allergic disease. Early introduction to soymilk was not included in these analyses, as its use is likely directly related to possible allergy. Postnatal exposure to pet dogs was not included as it was strongly correlated to prenatal dog exposures at home. Combined exposure to smoking during pregnancy and prenatal period and secondhand smoke exposure did not show significance, which can partly be because of very small numbers exposed to smoking in this cohort.
Table 5 Predictors of allergic disease (n=642)
CI, confidence interval.
a Multiple logistic regression model (n=642), including all exposures that were significant with a P<0.10 in univariate models.
b Multiple logistic regression model (n=787) keeping exposures significant with a P<0.05 using backwards selection methods.
Discussion
The natural history of the development of allergic diseases in childhood likely reflects a complex interplay between genetic factors and environmental exposures.Reference Prescott 8 Despite advances in our understanding of immunological responses associated with many chronic inflammatory diseases, there is still uncertainty surrounding the timing and the precise identity of the specific environmental factors that impact the development of allergic disease in early life. The FAMILY birth cohort provides an opportunity to study early-life, including in utero, exposures on the development of atopy and allergic disease in a general population rather than a high-risk cohort of atopy. We identified a number of exposures that were positively or negatively associated with allergy. Although some of the major findings of this study were consistent with findings from other birth cohorts (e.g. dog exposures), others such as negative associations with postnatal acetaminophen were unexpected.
Relative to the general Canadian population,Reference Soller 9 where self-reported food allergy is 7%, eczema is 17% and asthma in children is 13%, the children of the FAMILY Study cohort overall had high prevalence rates of allergic diseases and associated phenotypes: atopy 24.5%, food allergy 17.5%, cow’s milk allergy 4.8%, eczema 16.0% and wheezing 18.6% (Table 4). The higher prevalence of food allergy in part can be due to the inclusive definition used for food allergy in our study.
The protective association between early-life exposure to dogs and development of allergic disease has been studied extensively. We have likewise found a protective effect of exposure to dog during pregnancy and the 1st year of life on the development of allergic disease. Pelucchi et al. Reference Pelucchi, Galeone and Bach 10 reviewed data on the relationship between pet exposure and occurrence of atopic dermatitis (AD). The authors found that children with regular contact with pets had a 25% reduced risk of AD. Interestingly, this finding was specific to dog, as opposed to cat, exposure. The protective factor associated with dog exposure in the prenatal and postnatal periods is still relatively unknown, but may be related to the effects of contact with dog-associated microbial agents during the 1st few years of life. In support of the ‘hygiene hypothesis,’ regular contact with animals, and thus increased exposure to microbial products, including endotoxins, in pregnancy or during early life, has been linked to reduce atopic sensitization and allergic disease,Reference Pelucchi, Galeone and Bach 10 presumably through modulation of infant immune responsesReference Bufford, Reardon and Li 11 or alterations in the gut microbiome.Reference Azad, Konya and Guttman 12
We found a negative association between atopy and predominantly breastfeeding up to 17 weeks on multivariate analysis. There are several studies showing a protective effect of breastfeeding in the first 4–6 months of life for the development of asthma, eczema and cow’s milk allergy. Kramer et al. Reference Kramer, Aboud and Mironova 13 found a reduction in risk of Atopic Dermatitis (AD) with exclusive breastfeeding by age 3 months but did not find any clear reduction in asthma, allergic rhinitis, positive skin prick testing and food allergy.
Exposure to antibiotics during the 1st year of life was associated with the development of atopy and allergic disease. This effect has been seen in multiple other studies. Marie-Josée et al.Reference Marie-Josée, Évelyne and Jean-Luc 14 and Ong et al.Reference Ong, Umetsu and Mandl 15 found an increased risk of asthma in children with antibiotic use during the first 6 months of life. Tsakok et al.Reference Tsakok, McKeever and Yeo 16 showed increased risk of eczema with exposure to antibiotics in the 1st year of life. These findings are consistent with our multivariate analyses. Recent reports have confirmed previously documented associations of antibiotic use by newborns with altered infant microbiota profiles, including long-term reduction in microbiota diversity.Reference Russell, Gold and Hartmaan 17 Furthermore, antibiotic use during infancy is associated with increased risk for atopic disease later in childhood.Reference Kozyrskyi, Bahreinian and Azad 18 Shreiner et al. Reference Shreiner, Huffnagle and Noverr 19 explains this phenomenon as the ‘microflora hypothesis,’ in which the antibiotic-induced changes to microbiota resulted in increased susceptibility to T Helper Cells 2 (Th2) cytokine-specific responses. This has public health and policy implications, given that Persaud et al.Reference Persaud, Azad and Chari 20 as part of the Canadian Healthy Infant Longitudinal Development (CHILD) Study found that neonates in Canada are routinely exposed to antibiotics in the perinatal period. In a further analysis of the CHILD study, Arrieta et al.Reference Arrieta, Stiemsma and Dimitriu 21 showed the association between gut microbiota and development of asthma. Nonetheless, caution should be exercised in interpreting our findings, as the influence of early-life antibiotic use on development of allergic disease could at least partially be explained by early respiratory infections themselves, a potential cause of asthma, and thus the requirement for increased antimicrobial medications.Reference Mai, Kull and Wickman 22 We do not have data on respiratory infections in FAMILY, and therefore cannot exclude this possibility.
Secondhand tobacco smoke exposure during pregnancy, but not direct maternal smoking, was found to be significantly associated with the development of allergic disease (P=0.0180). The role of smoking, both direct and secondhand, in the development of allergic disease is well established.Reference Hancox, Subbarao and Sears 23 – Reference Jakkola, Kosheleva and Katsnelson 25 A recent study by Simons et al.Reference Simons, To and Moineddin 26 demonstrated that children whose mothers smoked or were exposed to secondhand smoke at home during pregnancy were more likely to develop asthma. These studies reinforce the concept that the development of allergic disease is not entirely dependent on maternally derived tobacco smoke, and secondary sources of smoke exposure are perhaps equally important. In our study, the lack of statistical significance for the association between maternal smoking and allergy may be explained by insufficient numbers of subjects to achieve significance, as only a small number of mothers smoked during pregnancy (7%) and only a small number of babies (4.7%) were exposed to secondhand smoke.
Exposure to acetaminophen was found to have an inverse association with allergic disease and atopy in our study. The relationship of acetaminophen use has been studied previously in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort,Reference Shaheen, Newson and Sherriff 27 with opposing results to ours. Studies have demonstrated increased risk of infant wheezing with acetaminophen use in earlyReference Shaheen, Newson and Henderson 28 and lateReference Shaheen, Newson and Sherriff 27 pregnancy, and increased incidence of asthma and atopy upon exposure to acetaminophen during the 1st year of life.Reference Wickens, Beasley and Town 29 , Reference Wang, Liu and Chen 30 In contrast, our study showed a negative association between atopy/allergy development and acetaminophen use in the 1st year. A possible explanation for this discrepancy is that these other studies evaluated allergic disease in children at ages 2–6Reference Wang, Liu and Chen 30 and 5–7 years,Reference Wickens, Beasley and Town 29 whereas our study looked at the same parameter in 1-year-old children. Our findings suggest a temporal relationship underpinning the association of acetaminophen use with atopy and allergic disease, with a changing gradient from infancy through school age. Of course it is difficult to separate the possible role of infections, which may cause fever and trigger the use of acetaminophen in the 1st year of life.Reference Almqvist, Wettermark and Hedlin 31
Although birth cohorts provide a unique opportunity to examine factors that may alter the development of atopy, there are limitations that must be considered. Clearly, these studies are not interventions; and observed associations do not necessarily imply causation. Reverse causation is a possibility that needs to be considered. For example, mothers may change an infant’s diet or remove pets from the home if signs of atopy in the infant are noted. Although our primary end point of atopy was objective, the number and specific types of allergens tested were arbitrary. We included food allergy and allergic disease as end points. These included skin test positivity but also relied on questionnaire data for the development of eczema, asthma, allergic rhinitis or food allergy; we did not independently (i.e. by physical examination) assess for these conditions. Finally, although the associations we have found may be important for the development of atopic disease in the 1st year of life, they may not continue to be important factors at subsequent ages. As such, it would be important to further investigate the development of allergic disease in FAMILY Study children to ages 3 and 5 years to determine whether these associations remain or are altered.
Conclusions
The FAMILY Study cohort demonstrated high rates of atopy and allergic disease at 1 year of age relative to the general Canadian population. Some previously reported exposures, such as exposure to dogs in the prenatal period and exposure to antibiotics were demonstrated in our study to be significantly associated with atopy and allergic disease. Novel factors such as postnatal infant use of acetaminophen were negatively associated with our targeted outcomes.
Further research is needed to understand the role of these associations in the development of atopy and allergic disease.
Acknowledgments
The authors are grateful to all the families who participated in the FAMILY Study and the whole FAMILY Study team, which includes interviewers, nurses, laboratory technicians, research scientists, statisticians, volunteers and managers. The authors also acknowledge the support of CIHR, the Population Health Research Institute and the Allergy, Genes and Environment Network.
Financial Support
This study was funded by the Canadian Institutes of Health Research, Heart and Stroke Foundation of Ontario and the Population Health Research Institute.
Conflicts of Interest
None.
Ethical Standards
The study was approved by the Research Ethics Boards at the participating hospitals (Hamilton Health Sciences, St Joseph’s Hospital, Hamilton and Joseph Brant Memorial Hospital, Burlington, ON).
