Exposure to Dogs and Cats in the First Year of Life and Risk of Allergic Sensitization at 6 to 7 Years of Age FREE

The increasing prevalence of asthma in the United States and other developed countries over the last few decades has been a cause for concern.^1- 2 While many factors appear to be involved in the development of childhood asthma, allergic sensitization to common allergens has consistently been shown to be related to the risk of developing asthma and to the risk of asthma persisting from childhood into adulthood.^3- 6 Many studies have attempted to elucidate relationships between environmental exposures, especially during infancy, and the risk of allergic sensitization in later life.^7- 8 These studies are based on the theory that an individual's genetic predisposition to allergic disease is activated or enhanced by early allergen exposure.^4,7,9 The outcome of interactions between genetic influences and allergen exposures may be influenced by other environmental exposures, such as passive exposure to environmental tobacco smoke.^7,9 If these relationships were better understood it might become possible to reduce the prevalence of allergic sensitization and perhaps asthma in children.

Exposure to dogs and cats during infancy has been thought to increase the risk of subsequent allergy to these animals.^8,10- 12 This assumption is primarily based on a few retrospective studies reporting an increased likelihood of allergic sensitization following exposure during infancy.^10- 12 Some studies, however, have suggested that exposure to dogs or cats during infancy is associated with reduced risk of allergic disease.^13- 18 Others have shown that children growing up on farms, especially farms with animals, were less likely to be allergic than were children growing up in urban environments.^19- 20

This analysis is part of the Childhood Allergy Study, a prospective birth cohort study designed to simultaneously evaluate multiple relationships between early environmental exposures and subsequent allergic sensitization and asthma.^21- 30 Among the variables considered were parental allergy histories, parental smoking, IgE levels in cord blood, month of birth, concentrations of dust mite and cat allergen in the child's bedroom at age 2 years, and pet exposure. In this analysis, we specifically examined exposure to dogs or cats in the first year of life and a child's risk of later allergic sensitization to common allergens after adjusting for potential confounding associations. We also examined relationships between early dog and cat exposure and allergen-specific serum IgE concentrations, lung function, methacholine airway responsiveness, and asthma.

The selection of children for the Childhood Allergy Study has previously been described.²¹ Briefly, all pregnant women living in an area of northern, suburban Detroit, defined by contiguous ZIP codes, and belonging to a health maintenance organization, were eligible to participate if their infants were born between April 15, 1987, and August 31, 1989. Only infants born at term (36 or more weeks' gestational age) with valid measurements of cord serum IgE concentration were entered into the study. The study was approved by the institutional human rights committee, and written informed consent was obtained when the mothers were enrolled, at the time of the first home visit, and prior to the clinical evaluations.

Study nurses interviewed mothers prior to delivery to obtain information concerning each parent's level of education; presence of allergies in general and of hay fever and asthma specifically; and parental smoking habits. The number of siblings was also noted along with other data about the home. Cord serum IgE concentrations were measured for all infants as previously described.²¹

We contacted parents by telephone when infants were aged 1 year to obtain information on prespecified variables of interest, including the presence and number of pets in the home during the first year. The number of dogs and cats reported at this time was used for this analysis. When children were aged 2 years, nurses visited each child's home to obtain information about the home environment and to collect dust samples from the child's bedroom, as well as urine samples from the child for measurement of urinary cotinine as a biomarker of passive cigarette smoke exposure. The dust samples were analyzed for concentrations of mite (Der p 1 and Der f 1) and cat (Fel d 1) allergens using a monoclonal antibody-based enzyme-linked immunosorbent assay as previously described.²⁸ We have documented the validity of parental smoking histories in this cohort with reference to children's urinary cotinine concentrations.²⁶ Questionnaire-based parental smoking histories from the first year of the child's life were used for these analyses because there were fewer missing values than for urinary cotinine concentrations. Follow-up telephone interviews also were conducted when the children were aged 3, 5, and 6 years, and a second home visit was conducted when the children were aged 4 years.

Clinical evaluations for allergic sensitization and asthma were performed when the children were aged 6 to 7 years. In addition to general medical histories and physical examinations, these evaluations included skin prick testing with commercial extracts of dust mites (Dermatophagoides farinae, D pteronyssinus), dog, cat, short ragweed (Ambrosia artemisiifolia), and blue grass (Poa pratensis), along with saline and histamine controls (all extracts and controls, Pharmaceutical Division, Bayer Inc, Spokane, Wash). Skin prick test results were considered positive if the product of perpendicular wheal diameters was 4 mm or more associated with a flare of at least 10 mm, and if there was no response to the negative control. Atopy was defined as a positive skin prick test result with any of the 6 allergens tested. Blood samples obtained during the evaluation were assayed for total serum IgE concentration and concentrations of allergen-specific IgE antibodies using a commercial assay (AlaSTAT, Diagnostic Products Corp, Los Angeles, Calif). Allergen-specific IgE testing included the same 6 allergens used for skin prick testing in addition to Alternaria species. Total and allergen-specific serum IgE levels were expressed in international units per milliliter (1 IU/mL corresponds to 2.4 ng/mL). Allergen-specific IgE levels of 0.35 IU/mL or higher were considered to be a positive test result in accordance with the manufacturer's recommendation. Seroatopy was defined as any positive test result for an allergen-specific IgE concentration. Numbers of children with seroatopy may have been slightly higher than those with atopy because one additional allergen, Alternaria, was also used to define seroatopy. Because a study published after the start of this study suggested that cockroach sensitization may be associated with asthma,³¹ a random sample of 100 sera were assayed for cockroach-specific IgE and only 2 sera were positive.²⁴ Given the low prevalence of detectable cockroach sensitization, no further testing for cockroach-specific IgE was performed.

At the time of skin prick testing, children were defined as having current asthma if a parent reported that they had been diagnosed by a physician as having asthma and that they had asthma symptoms or used asthma medications in the preceding 12 months. Pulmonary function tests were performed as previously described, and the results are presented as the percentage of predicted using standard equations.²⁹ Methacholine airway responsiveness was determined as previously described.²⁹ After baseline spirometry and no response to a control saline challenge, 5 doses of methacholine (0.025-25 mg/mL) were administered through a dosimeter (Pulmonary Data Services, Louisville, Colo). Methacholine airway responsiveness was defined as a fall in forced expiratory volume in 1 second (FEV₁) of 20% or more from the postsaline challenge value at a concentration of administered methacholine of 10 mg/mL or less.²⁹

The power of this study was originally based on the ability to detect a small to medium effect (0.2) for a χ² test as defined by Cohen.³² With this assumption, and assuming an α level of .05, the power to detect significant associations between outcomes in 3 exposure groups is greater than 90% with an overall sample size of 470. Given the same assumptions, it also is possible to stratify the data for a variable with approximately equal prevalences (ie, sex) and still have power greater than 70% with 235 in each group. If the prevalence of a variable is low in the cohort, such as current asthma, the power is much lower.

The collected data were first examined for potential imbalances between those children lost from the study and those who were retained. χ² Tests were used to compare the relative percentages.³³

Pet exposure in the first year of life was defined as an ordinal variable with 3 categories: no dog or cat exposure, exposure to 1 dog or cat, and exposure to 2 or more dogs or cats. The highest strata was truncated at 2 or more because of the small sample size above this level. Information about pet exposure at age 6 to 7 years was used to create the same 3 categories of pet exposure as was used for the first year of life. Binary variables of interest (eg, atopy [yes or no], specific skin prick test positivity) were analyzed according to pet exposure category using a χ² test for 2 × 3 contingency tables.³³ We did not have a preconceived hypothesis concerning a relationship between exposure to varying number of dogs and cats and the risk of allergic sensitization. Therefore, we tested the general hypothesis that outcomes differed across categories of dog and cat exposure rather than testing for trends with increasing exposure.

The relationships between pet exposure categories and continuous variables, such as percent predicted forced vital capacity (FVC) or total serum IgE, were evaluated with a 1-way analysis of variance technique.³⁴ Each continuous variable was transformed to natural logarithmic equivalents to reduce positive skewing prior to analysis. If the range of the variable to be logarithmically transformed included zero, 1 was added to the variable prior to transformation. When logarithmic data transformation did not result in a near normal distribution, such as for dust mite concentrations, a Kruskal-Wallis test was used. The number of children in each analysis varied slightly because of missing data.

Atopy and seroatopy were each used as a dependent binary variable in a linear logistic regression assessing 2 indicator variables for pet exposure: exposure to 1 dog or cat or exposure to 2 or more dogs or cats.³⁵ These models were fitted without other variables and with other potentially confounding variables, including cord serum IgE concentration, child's sex, having older siblings, parental smoking, mother or father with a history of asthma, and total bedroom dust mite allergen levels at child age 2 years. The logistic model is appropriate for modeling binary dependent variables. It makes minimal assumptions about the distributional properties of the independent variables and the exponentiation of the coefficient allows for estimation of the odds ratio (OR). We chose to include the number of dogs and cats as 2 binary indicator variables avoiding assumptions concerning the direction of any associations. Using the Hosmer-Lemeshow test, we found no evidence to doubt the validity of the models.³⁵

We analyzed the entire data set and data sets defined by sex. In all analyses, an α = .05 criteria was used to determine statistical significance. There was no attempt to impute data; all analyses were performed on all available data. SAS v8.0 (SAS Institute Inc, Cary, NC) was used for all analyses.³⁶

A total of 1194 pregnant women were potentially eligible for entry into this study, and consent for participation was obtained from 953 women. Infants of 106 of these women were not enrolled in the study because a cord blood sample was not obtained, leaving 847 eligible newborns. Six of the cord blood samples were thought to be contaminated by maternal blood,^21,23 and an additional 6 children were found to be ineligible when each child's data were examined and verified prior to the 6- to 7-year evaluations, yielding 835 eligible children enrolled at birth. Of the 835 children initially enrolled, 235 had been lost to follow-up by age 6 years, and 126 of those contacted at age 6 years declined participation in the clinical evaluation. Thus, 474 (57%) of the 835 eligible children initially enrolled completed the clinical evaluation for allergic sensitization and asthma at an average age of 6.7 (SD, 0.17) years. Characteristics of children who were evaluated at 6 to 7 years did not differ significantly from those of children who did not undergo clinical evaluation at age 6 to 7 years, including whether the parent had a history of asthma or hay fever or whether there were dogs or cats in the household (Table 1). Also, interactions between each variable, any exposure to dogs and cats in the home in the first year of life, and whether the child participated in the clinical evaluation were not statistically significant (Table 1). When the relationship between maternal and paternal histories of asthma, allergies, and hay fever and presence of 2 or more dogs or cats in the household was evaluated, no significant associations were found.

The parents of the children in this study were relatively well educated and almost all (804 [96.3%]) described themselves as white, non-Hispanic. Characteristics of the children completing the study are presented in Table 2. Boys and girls were approximately equally represented. The presence of a dog or cat in the home did not differ significantly between parents with a history of asthma, allergies, or hay fever and those who did not report these conditions.

To investigate the relationships between dog and cat exposure and allergic sensitization, we initially compared the 184 children with any dog exposure in their first year of life to the 220 children without either dog or cat exposure. Children exposed to a dog were less likely to have a positive skin test result to dog allergen (3.3% vs 8.6%, P = .03) and detectable dog-specific IgE (3.7% vs 8.7%, P = .06) at follow-up. Any exposure to a dog was also associated with lower total serum IgE levels (geometric mean, 23.8 IU/mL vs 33.1 IU/mL for no dog or cat exposure; P = .04). The inverse association between dog exposure and allergic sensitization was further examined in relationship to number of dogs (Table 3). The reference group remained the 220 children with neither dog nor cat exposure in the first year of life. An apparent dose-response effect for atopy and seroatopy was found across the 3 exposure categories. Atopy was present in 33.6% of the children without dog or cat exposure, in 29.7% with exposure to 1 dog, and in only 8.3% with exposure to 2 or more dogs (P = .009). The prevalence of seroatopy was 38.5% with no pet exposure, 36.7% with exposure to 1 dog, and 12.9% with exposure to 2 or more dogs (P = .02). The same analyses were performed with cat exposure during the first year. Using the same reference group of 220 unexposed children, the patterns toward less prevalent allergic sensitization with exposure to cats were similar to those observed with dog exposure, but none of the associations reached statistical significance. For example, the prevalence of atopy declined from 33.6% to 31.4% to 23.1% with exposure to no dogs or cats, 1 cat, or 2 or more cats, respectively (P = .54 for comparison across categories), while seroatopy declined from 38.5% to 34.5% to 26.1% (P = .45).

Based on finding similar relationships between dog and cat exposures in the first year of life and allergic sensitization at age 6 to 7 years, relationships were further analyzed by simultaneously considering combined dog and cat exposure. Combining dogs and cats increased the number of children in each category, allowing further exploration of the relationships through stratification of the data by sex. Atopy and seroatopy were each present in about one third of children. When all children were considered, the prevalence of skin prick test positivity to dog allergen, outdoor and indoor allergens, atopy, and seroatopy were significantly different across the 3 exposure categories and generally decreased with increasing pet exposure (Table 4). The pattern of decreasing skin prick test positivity to cat allergen with increasing exposure was similar but the relationship(s) failed to reach statistical significance.

When boys and girls were considered separately, different patterns emerged from the data (Table 4). Exposure to a single dog or cat in the first year of life was associated with an increased prevalence of atopy and seroatopy in girls while both outcomes declined in boys exposed to a single dog or cat. Lower prevalences of skin prick test positivity to dog, cat, and indoor and outdoor allergens, and of methacholine airway responsiveness were consistently found in association with exposure to a single dog or cat with boys but not with girls. Measurements of lung function were also related to dog and cat exposure for boys but not for girls. The prevalence of methacholine airway responsiveness in boys was 25.5% when there had been no dog or cat exposure, 20.3% with exposure to 1 dog or cat, and 5.1% with exposure to 2 or more dogs or cats (P = .03). In girls, the prevalence of methacholine responsiveness was unchanged across pet exposure categories. Similarly, the mean percent predicted FVC and FEV₁ increased significantly across pet exposure categories among boys but not among girls, and were highest with exposure to 2 or more dogs or cats. Thirty-three (7%) of 473 children had current asthma. The prevalence of current asthma was lower in boys who had been exposed to 2 or more dogs or cats in infancy compared with no exposure (5.1% vs 11.8%, respectively), but the difference across exposure categories was not statistically significant (P = .43) and no difference was seen for girls.

For all children, exposure to 2 or more dogs and cats in the first year of life was associated with a lower total serum IgE at age 6 to 7 years, but the analysis across exposure categories was not statistically significant (P = .09) (Table 5). In analyses stratified by sex, exposure to more dogs or cats was associated with significantly decreased geometric mean IgE among boys (P = .02) and among children with a parental history of asthma (P = .03), but not among girls (P = .42) or among children without a parental history of asthma (P = .31).

When mean total dust mite concentrations in the child's bedroom at age 2 years were compared for homes with no pet, 1 dog or cat, and 2 or more dogs or cats, the median (5th percentile, 95th percentile) dust mite concentrations were not significantly different (2.0 [0.5, 37.1] µg/g dust; 1.8 [0.5, 16.6] µg/g dust; 1.1 [0.5, 43.8] µg/g dust; P = .27).

Logistic regression analysis was used to adjust for the effects of possible confounding variables (cord serum IgE concentration, levels of house dust mite allergen in the child's bedroom at age 2 years, child's sex, an older sibling, passive exposure to parental tobacco smoke, and parental history of asthma) on relationships between dog and cat exposure and risks of atopy and seroatopy. After adjusting for all of these variables, exposure to 2 or more dogs or cats was still associated with significantly lower risks of atopy (OR, 0.31; 95% confidence interval [CI], 0.14-0.72) and seroatopy (OR, 0.43; 95% CI, 0.19-0.96) in all children (Table 6, model 1).

When the analysis was further adjusted for current exposure to dogs or cats using the same exposure categories as in the first year of life (no exposure; exposure to 1 dog or cat; exposure to 2 or more dogs or cats at age 6-7 years), the risk of atopy and seroatopy associated with exposure in the first year of life remained significantly decreased (atopy: OR, 0.23; 95% CI, 0.09-0.60; seroatopy: OR, 0.33; 95% CI, 0.13-0.83) (Table 6, model 2). When the variable for dog or cat exposure at age 6 to 7 years replaced the variable for exposure in the first year of life, no statistically significant associations were found. For example, with all children included in the analysis, risks of atopy and seroatopy among children with 2 or more dogs or cats at age 6 to 7 years were not significantly different than those among children with no current pet exposure (for atopy: OR, 0.79; 95% CI, 0.44-1.85; P = .79; for seroatopy: OR, 0.81; 95% CI, 0.43-1.94; P = .81).

In this prospective study we found that exposure to 2 or more dogs or cats in the first year of life was associated with a lower prevalence of allergic sensitization at age 6 to 7 years regardless of exposure to dogs and cats at age 6 years. This inverse relationship was consistent whether skin prick tests for 6 common aeroallergens or tests for 7 allergen-specific IgE concentrations were considered as primary outcomes. The inverse relationship was present for both indoor (dust mites, dog, and cat) and outdoor (ragweed, grass, and Alternaria) allergens. The relationships remained significant after adjusting for a number of variables that may be risk factors for allergic sensitization or that could have been associated with pet ownership, including cord serum IgE concentration, house dust mite exposure, older siblings, parental smoking, and parental history of asthma.^4,37- 42

Other studies have also reported lower prevalences of allergic sensitization or symptoms related to allergic diseases in association with early exposure to dogs and cats,^{13,15- 16,43- 44} but a systematic review of the literature concerning this question concluded that previous exposure to dogs and cats increased the risk of asthma and wheezing in children older than 6 years.⁴⁵ The conclusions of this systematic review differ from the results of 2 large prospective birth cohort studies.^43- 44 Nafstad et al⁴³ found that after using logistic regression to adjust for potential confounders, being exposed to pets in early life reduced the risk of asthma (OR, 0.7; 95% CI, 0.5-1.1) and allergic rhinitis (OR, 0.6; 95% CI, 0.4-1.0) in a birth cohort of 2531 children followed to age 4 years. In a birth cohort of 1246 children in Arizona followed up to age 13 years, Remes et al⁴⁴ reported that children who had 1 or more dogs in the home at birth were significantly less likely to develop frequent wheeze than children without early dog exposure, but neither early exposure to dogs or to cats was associated with skin prick test positivity or total serum IgE concentrations. Remes et al did not find a difference between children exposed to 1 dog compared with those exposed to 2 or more dogs. They also found that the inverse relationship between dog exposure and frequent wheeze was predominantly among children without a parental history of asthma. Reasons for the differences in allergic sensitization outcomes between our study and the study by Remes et al are not clear, but may include differences in climate where the birth cohorts were located and differences in keeping pets inside the home.

In a recent cross-sectional study in children by Platts-Mills et al,¹⁸ and also in a subsequent study in adults,⁴⁶ a bell-shaped dose-response relationship between cat allergen exposure and cat-specific sensitization was observed. Decreased levels of cat-specific sensitization were associated with both the lowest and the highest cat allergen exposure groups. Platts-Mills and colleagues also found that cat-specific IgG antibody levels increased with increasing cat exposure, and were highest in the highest cat exposure group. They suggested that high levels of cat allergen exposure induced a modified T helper cell type 2 (T_H2) response with production of cat allergen–specific IgG and IgG4 antibodies without allergic sensitization. This interesting hypothesis is not entirely consistent with the data presented in our study because we found that allergen-specific IgE antibodies to dust mites, ragweed, and grass (allergens unrelated to dog and cat) were also less prevalent in children exposed to dogs and cats in the first year of life.

Other researchers have suggested that the protective effect of dogs and cats is not related to allergen exposure but rather to increased exposure to bacterial endotoxin associated with household pets.^15,47 Endotoxin exposure is hypothesized to shift the developing immune system away from a T_H2-type pattern of response, which favors development of allergic sensitization, toward a T_H1-type response. Studies in animals have shown that concomitant exposure to endotoxin and allergen will prevent allergic sensitization normally induced by the allergen.⁴⁸ Recent studies have shown that endotoxin levels in homes are inversely related to T_H2-type cytokine production by lymphocytes of children residing in the homes and that the presence of household dogs is related to higher levels of indoor endotoxin.^47,49 Our data are consistent with the hypothesis that exposure to 2 or more dogs or cats, and therefore exposure to higher levels of endotoxin, is associated with a T_H1 pattern of immune response and less allergic sensitization.

There were several associations with dog and cat exposure that were evident for boys but not for girls, including lower total serum IgE concentrations, lower prevalence of methacholine airway responsiveness, and better lung function. These differences in associations between boys and girls are puzzling, but others have also observed differences between boys and girls in factors related to asthma.^50- 52 Consistent with the pattern of the results of total serum IgE concentrations, methacholine airway responsiveness, and FEV₁ in boys, the prevalence of asthma was also lower in those boys exposed to 2 or more dogs or cats compared with those who were unexposed (5.1% vs 11.8%). This difference was not statistically significant across pet exposure categories; however, only 39 boys were exposed to 2 or more dogs or cats in the first year of life. Assuming prevalences that we found in our cohort, a study designed to detect a statistically significant difference in the prevalence of asthma among boys exposed to 2 or more dogs or cats would have required a final cohort of at least 1327 children followed up to age 6 to 7 years.

An important strength of this study is the prospective design using a population-based cohort of children followed yearly from birth. The prevalences of allergic sensitization, methacholine airway responsiveness, and asthma found in our cohort were similar to those reported by others studying children of similar ages.^53- 55 The mean values for total serum IgE were also similar to those reported by others.^55- 56 Animal exposure was ascertained when the child was 1 year old, not years later.^10,14- 15 Assessing animal exposure prior to assessing outcomes reduces concern of misclassification of exposure and recall bias. Information on other factors potentially related to risk of allergic sensitization, most importantly family history, was collected, allowing adjustment for the potential confounding effects of these other variables.⁵⁷ Another strength is the multiple objectively measured outcomes.^10,14 The association between pet exposure and less allergic sensitization was found with both in vivo (skin prick test reactivity) and in vitro (allergen-specific IgE levels) tests. The persons performing the skin prick tests, allergen-specific IgE tests, spirometry, and methacholine challenges were unaware of study hypotheses at the time the tests were performed, making systematic measurement bias unlikely.

Bronchial hyperresponsiveness is frequently stated to be a major component of asthma that can be objectively measured.^58- 60 Our findings of reduced methacholine airway responsiveness in boys with exposure to 2 or more dogs or cats suggest that exposure to dogs and cats may be associated with a reduced risk of asthma, at least in boys. While we did not find a statistically significant association for current asthma in this study, the prevalence of asthma in boys exposed to 2 or more dogs or cats was 57% lower than in unexposed boys, a difference that would likely be significant in a larger population.

There are limitations to our study. As with most prospective studies, some children did not complete the entire study. However, we found no important differences between children examined at age 6 to 7 years and those who were not examined. In addition, we could not detect differences in the relationship between dog and cat ownership and parental history of asthma, allergy, or hay fever among those examined and not examined. A second caveat is the limited racial, socio-economic, and geographic diversity of our study population, suggesting that our conclusions can only be applied to similar populations of white children. Since our follow-up was limited to an average age of 6.7 years, we do not know if the associations we found will persist as the children grow older, but others have found that the association between dog and cat exposure and a lower risk of allergy-related symptoms persisted to age 12 to 13 years.^15,44 Sample size is another limitation of the study. A larger sample would have allowed more reliable estimates and detailed examinations of the differences between boys and girls and between children with and without parental histories of asthma. A final caveat is that we did not consider exposure to dogs and cats outside the child's home.

In this prospective study designed to examine multiple potential risk factors for allergic sensitization, we found that exposure to 2 or more dogs or cats in the first year of life was associated with a significantly lower probability of subsequent allergic sensitization to common aeroallergens. Exposure to 2 or more dogs or cats was also associated with significantly lower serum IgE concentration, less methacholine airway responsiveness, and better lung function in boys but not in girls. The association between pet exposure and decreased prevalence of allergic sensitization remained unchanged after adjustment for potentially confounding variables. These findings suggest that exposure to more than 1 dog or cat in the first year of life may reduce a child's risk of allergic disease.