Katja Radon Unit for Occupational and Environmental Epidemiology and Net Teaching Institute for Occupational and Environmental Medicine Ziemssenstr. 1 80336 Munich Germany
Background: It has been argued that the inverse association between exposure to farm animals and nasal allergies observed in children and adults might be because of self-selection.
Aims: We aimed to assess the health-based selection out of farming in adults.
Material and methods: A cross-sectional study was carried out in a rural region. Overall, 4053 inhabitants (63%) aged 18–44 years responded to a questionnaire on respiratory diseases, life-time exposure to farming environments and potential confounders. For 2678 of these, specific immunoglobulin E to common allergens was available. The outcome was: (i) sensitization and symptoms of nasal allergies (symptomatic sensitization); (ii) sensitization without symptoms of nasal allergies (asymptomatic sensitization).
Results: Farm animal contact in childhood was associated with a decreased risk of symptomatic and asymptomatic sensitization. Continued exposure to farm animals in adulthood further decreased the odds ratio of symptomatic (odds ratio 0.2; 95% confidence interval 0.1, 0.4) but not asymptomatic sensitization (0.7; 0.4, 1.1). Starting farm animal contact in adulthood even increased the odds ratio of asymptomatic sensitization (2.4; 1.1, 5.2).
Conclusions: The preventive effect of childhood contact to farm animals against sensitization continues into adulthood. However, in adulthood self-selection based on symptoms and underreporting of symptoms might also play a role.
A large number of studies reported that exposure to farm animals in childhood might protect against respiratory allergies in children (1–4). Furthermore, several studies indicated that this protective effect of childhood exposure to farm animals is retained in adulthood (5–8). Prior et al. (9) even reported that exposure to farming environments in adulthood results in a loss of sensitization. In contrast, in a cross-sectional study of primary school children, only first year of life exposure to animal confinement buildings was associated with lower prevalence of atopic diseases (1). Methodological issues might explain part of the different findings in these studies. While, in the study by Prior et al. (9), selection bias because of losses to follow-up might play a role Riedler et al. (1) could only dichotomize the exposure to farm animal contact during the first year of life or later. Therefore, the lower prevalence of respiratory allergies among adults might either be explained by a protective effect in childhood lasting until adulthood, by self-selection of adults with allergic sensitization out of farming, by a protective effect of farm animal contact in adulthood itself (10) or by a combination of these.
Therefore, we aimed to assess the health-based selection out of farming in a population-based survey in an agricultural area.
Materials and methods
Using random numbers, a sample of 7080 adults (69% of the population) aged 18–44 years living in a rural area in North Germany (Lower Saxony) were invited to answer a mail-in questionnaire and to take part in a clinical examination including blood sampling within a cross-sectional study. The local population registry provided the addresses. Up to two postal reminders have been sent. Subjects not responding within 6 weeks after the first mailing were contacted by phone. Of the 4720 subjects who returned the questionnaire (66.7%), 4053 were born in the western part of Germany and thus eligible for clinical examination. Overall, serum samples of 2678 subjects were available (62.9%).
The study protocol was approved by the Ethical Committees of the Ludwig-Maximilians-University Munich and the Chamber of Physicians of Lower Saxony.
For exposure to farming environments, questions from the questionnaire of the Allergies and Endotoxin-study questionnaire were used (2). Questions on respiratory diseases and on potentially confounding factors for these conditions were taken from the questionnaire of the European Community Respiratory Health Survey (11). Nasal allergies were defined as a positive response to the question ‘Do you have any nasal allergies, e.g. hay fever?’
Specific immunoglobulin E (IgE) against a panel of inhalant allergens (timothy grass, rye, mugwort, birch, Dermatophagoides pteronyssinus, Cladosporium herbarum, cat and dog) was measured in all serum samples (SX-1, Pharmacia CAP system; Phamacia, Freiburg, Germany). A specific IgE concentration of ≥ 0.7 kU/l (corresponding to radioallergosorbent test class 2 or higher) was regarded as positive. In addition, all blood samples were tested for specific IgE against farm-related allergens (chicken, turkey, pig, cattle and Aspergillus fumigatus). Overall, only seven participants were sensitized against farm-related allergens only (0.3%). Therefore, only results for the ubiquitous allergens were considered in the analyses.
The outcome was sensitization against inhalant allergens. In order to assess the potential self-selection based on symptoms, we additionally stratified the outcome:
1Subjects with sensitization to common allergens and self-reported symptoms of nasal allergies (‘symptomatic sensitization’).
2Subjects with sensitization to common allergens not reporting symptoms of nasal allergies (‘asymptomatic sensitization’).
Subjects without symptoms and sensitization served as control.
The following exposure categories were defined:
1No regular farm animal contact ever.
2No regular farm animal contact during childhood (age < 18 years) but living or working on an animal farm during adulthood.
3Regular farm animal contact at any time during childhood [living on an animal farm or regular (at least once a week) contact to stables with farm animals] not continuing through adulthood.
4Regular farm animal contact during childhood and living or working on an animal farm during adulthood.
We used cross-tabulation to visualize bivariate distributions of categorical predictors and outcome. Age, gender, active smoking, number of siblings, having siblings or parents with a history of allergy (allergic rhinitis, asthma or atopic dermatitis) and higher level of education were considered as potential confounders in the final multiple logistic regression models.
Statistical analyses were carried out using SAS 9.1 statistical package.
The mean age (SD) of the 2678 participants was 34.0 years (7.2) when compared with 32.8 years (7.7) in the eligible population. The proportion of women among the participants was slightly higher than in the source population (51.1%vs 47.8%).
Twenty percent of the study population were sensitized against common allergens (n = 535). Of these, 281 (52.5%) reported nasal allergies. As in previous studies, lower number of siblings, atopic parents or siblings and younger age were associated with allergic sensitization. These results were confirmed in the multivariate models (Table 1).
Table 1. Prevalence and adjusted odds ratios [OR; 95% confidence intervals (CI)] for sensitization against common allergens
N = 2678*
Sensitization against common allergens
OR† (95% CI)
*Because of missing data in some of the predictors, the numbers do not necessarily add up to 2678.
†Mutually adjusted for all other variables listed in the table.
‡Living on a farm, regular (at least once a week) contact to farm animals in stables or working on an animal farm.
§Symptoms of allergic rhinitis, asthma or atopic dermatitis.
¶At least 12 years of schooling.
Only as adult
1.46 (0.77, 2.79)
Only as child
0.65 (0.50, 0.85)
As child and adult
0.44 (0.31, 0.64)
1.37 (1.07, 1.76)
1.83 (1.43, 2.34)
0.64 (0.50, 0.81)
1.20 (0.87, 1.65)
1.09 (0.83, 1.43)
Number of siblings
0.72 (0.56, 0.92)
Higher level of education¶
1.01 (0.77, 1.33)
2.69 (1.73, 4.18)
2.31 (1.51, 3.54)
1.97 (1.33, 2.91)
1.39 (0.95, 2.06)
Overall, 1373 subjects (56.8%) had contact with farm animals either in childhood or in adulthood. Of these, 75 (3.1%) started farm animal contact in adulthood, 877 (36.3%) reported regular farm animal contact only in childhood and the remaining 421 (17.4%) continued farm animal contact from childhood through adulthood.
Contact to farm animals only in adulthood
The prevalence of sensitization against common allergens in subjects who reported contact to farm animals only in adulthood was similar to the prevalence among subjects without contact to farm animals. This result was confirmed in the multivariate models (odds ratio 1.5; 95% confidence interval 0.8, 2.8; Table 1). Stratifying the outcome based on symptoms of nasal allergies (Fig. 1), the risk of asymptomatic sensitization was even increased in subjects with recent contact to farm animals (2.4; 1.1, 5.2). No association between adult contact to farm animals and symptomatic sensitization was seen (1.0; 0.4, 2.6).
Contact to farm animals only in childhood
Subjects who gave up contact to farm animals in adulthood were less likely to be sensitized against common allergens than participants who never had regular contact to farm animals (0.7; 0.5, 0.9). Stratifying the outcome for symptoms of nasal allergies did not change the association. However, for asymptomatic sensitization, the difference was not statistically significant in the multivariate model (0.8; 0.5, 1.1).
Contact to farm animals in childhood and adulthood
Subjects with continuous contact to farm animals had the lowest prevalence of allergic sensitization (0.4; 0.3, 0.6). The reduction in risk of sensitization was significantly larger for symptomatic subjects (0.2; 0.1, 0.4) when compared with asymptomatic subjects (0.7; 0.4, 1.1).
Our study suggests that
1Contact to farm animals in childhood protects from sensitization to common allergens in adulthood.
2Continuing contact to farm animals in adulthood does not further contribute to the reduction in risk. The additional decrease in the prevalence of sensitization in this group may at least partly be because of self-selection out of farming or underreporting of symptoms. In addition, unknown confounders might play a role.
3Starting exposure to farm animals in adulthood might even increase the risk of sensitization.
One major limitation was the cross-sectional character of the study. Therefore, we do not know the prevalence of symptoms and sensitization at the time subjects started or gave up exposure to farming environments. Studies in occupational epidemiology have shown that there is no substantial loss of sensitization after cessation of exposure while symptoms might improve (12).
It has been indicated that, especially, early contact to farming environments might decrease the risk of sensitization (1). We, therefore, repeated our analyses restricting childhood exposure to the first 3 years of life. This did not change the risk estimates (data not shown).
We used a valid and reliable questionnaire instrument for respiratory symptoms and potential confounders that has been extensively tested within the European Community Respiratory Health Survey (11). Subjects involved in farming might tend to underreport their symptoms. This might help to understand why the decrease in sensitization was most pronounced in symptomatic subjects with continuous exposure to farm animals. Likewise, this might explain why the increased risk of sensitization was confined to asymptomatic subjects who started contact to farm animals later in life.
Despite reasonable response rate, selection bias cannot be completely ruled out. Attempts to contact nonrespondents indicated that many of them were born in Eastern Europe. Because of potential differences in childhood exposures, these subjects were excluded from the clinical part of the study. Therefore, this should not bias our results. However, subjects who did not participate in the clinical part of the study but answered the questionnaire were significantly less likely to report symptoms of allergic rhinitis and less likely to have farm contact. Therefore, our results might underestimate the true association.
With respect to our outcome definition, we also tested whether the cut-off value used for specific IgE might influence our results. However, using a specific IgE concentration of 0.35 kU/l as cut-off for sensitization revealed similar risk estimates (data not shown). We did not include atopic dermatitis or asthma in our definition of symptomatic subjects. With respect to atopic dermatitis the question turned out to be rather unspecific. Regarding asthma, only 13 (2%) of sensitized subjects without nasal allergies reported asthma. Therefore, the results did not change including asthma in our definition of symptomatic subjects (data not shown).
It is well accepted today that compartmentalization of the immune system occurs mainly before the age of 5 years (13). Our findings are supported by animal studies using lipopolysaccharides (LPS) as surrogate exposure for contact to animal houses that have shown that only exposure to LPS before allergic sensitization reduced the risk of allergic sensitization. This reduction in risk might even be because of prenatal exposure (14).
Overall, only childhood contact to farm animals might decrease the risk of sensitization to common allergies.
We are grateful to Alexandra König, Martina Dutschke, Bernhard Schwertner and Matthias Netsch for their help in the field work. Jessica Kellberger, Evelyn Braun, Stefanie Goy, Ana Bogdan, Freydis Vogel, Michaela Paul and Stefanie Sprung are gratefully acknowledged for data entry and cleaning. We thank the participants for their cooperation. Parts of this paper have been used for the PhD thesis of Anja Schulze.
The study was supported by the grant from the Lower Saxony Ministry of Women, Labor and Social Affairs; European Union.