Animals and allergy

Allergic disease is less common in countries where people live a more traditional lifestyle. Even within developing countries, there is a higher and increasing prevalence in urban communities, but asthma and hay fever still tend to be rare in rural areas [ 1–3]. A number of studies, including our own in Ethiopia, have sought to discover what environmental exposures are responsible for this difference, and have looked at a wide range of factors [ 4]. In the search for specific exposures to explain this pattern of disease, no one, until now, has focused on one of the more natural distinctions between town and country dwellers; reliance on farming for a livelihood. In this issue, three independent studies, carried out in Europe but in select locations where farming is still a relatively flourishing industry, report a reduced risk of hay fever or associated symptoms in children or young adults who lived their early years on a farm [ 5–7]. The underlying theme of these and a related publication [ 8] is that the specific factor responsible is contact with the bacterial milieu of the farmyard. In all four studies the difference in risk associated with growing up on a farm was substantial, such that hay fever prevalence was reduced by between 37 and 72% in farmers' children in the most recent publications ( Fig. 1a). Such a large effect is relatively unlikely to be entirely explained by diagnostic bias or by differential underreporting by farmers of symptoms in their children, and moreover one of the studies confirmed the association by clinical examination in a small subset [ 6]. It is also unlikely to arise from residual confounding since the effect is larger than most of the recognized risk factors for the condition. On the other hand, hay fever was but one of a range of allergic outcomes assessed in all four studies, with three out of four including both lifetime diagnosis and current symptoms, and the effect of farming exposure was not consistent in relation to these. In the largest two studies, each comprising over 10 000 individuals, a statistically significant negative association was found with diagnosed hay fever but not current symptoms in 5–7-year-old German children [ 5], and with the combination of allergic rhinitis or conjunctivitis in 18–24-year-old university students [ 7]. In the smallest study, of 1620 Swiss school children, the only symptom that was significantly less common in farmers' children after controlling for potential confounders was that described as sneezing attacks during the pollen season [ 8]. Although these studies, and the fourth in this series, by Riedler et al. of 2283 Austrian 8–10 year olds [ 7], individually collected information on farming in conjunction with data on a wide range of early life factors and not all were established to investigate the effects of farming as a primary hypothesis, the consistency of their findings provide persuasive evidence that the relationship is valid. These studies therefore raise some intriguing questions. . Adjusted odds ratio (95% CI) for the effect of farming on (a) hay fever and (b) asthma, ever Firstly, is there sufficient justification to conclude, as all four studies do, that the protective effect is general to the development of allergy, rather than specific to the expression of hay fever, or even just sensitivity to grass pollen? Certainly there seems to be little evidence from these studies that eczema or associated symptoms are less common in farmers' children. The evidence with respect to asthma is relatively inconsistent in that in the majority of studies asthma or asthma-like symptoms tended to be less common in farmers' children but the magnitude of effect was somewhat smaller than that for hay fever ( Fig. 1b) and the reductions in risk that remained statistically significant after adjusting for potential confounders occurred in relation to wheezing symptoms but not diagnosed asthma [ 5, 6]. To some extent, this weaker association may arise from the fact that childhood asthma probably comprises a major component of nonallergic as well as allergic disease [ 9], but it seems reasonable to conclude that whatever the underlying cause of the farming phenomenon, physiologically it relates more closely to hay fever than to asthma. At first sight, allergen skin test measurements provide further evidence for a generally lower risk of atopy in farmers' children [ 7]. However, the only sensitivities which were markedly and significantly reduced were those to grass and birch pollen, i.e. outdoor allergens. Furthermore, the questions used to ascertain hay fever were directly [ 8], or indirectly [ 10], more likely to measure sensitivity to pollen rather than other forms of allergic rhinitis. A lower prevalence of pollen than other allergies has also been reported in farmers themselves [ 11, 12]. The conservative conclusion from these collective findings would be that exposure to a farming environment is associated with a reduced risk of pollen sensitization and consequent reduced risk of hay fever, and a corresponding but lesser reduction in asthma in which other sensitizing agents predominate [ 13]. However, allergen sensitivity was also measured in rather small numbers of children in the Swiss study, and when individual ‘indoor’ and ‘outdoor’ allergens were aggregated, both were significantly reduced. At this stage, the evidence is inconclusive, but in trying to explain the reduction in hay fever and pollen sensitivity those mechanisms which should also invoke reductions in sensitivity to other allergens should not be discounted. According to this interpretation then, children who might be presumed to be most exposed to pollen allergen are least sensitive to it. Whether in fact pollen exposure is markedly different for those living on, as opposed to those living in the region of a European farm is actually far from clear. However, though low power precluded statistical significance, we found a very similar disparity in our comparisons of urban and rural Ethiopia. Sensitivity to mixed threshings was less common in the farming communities of the rural area than in the town (OR 0.65), whilst sensitivity to house dust mite was actually increased in the rural areas [ 4]. The problem with the farming scenario emerging from studies in the developed world is the direction of the cause and effect. It is conceivable that some atopic individuals from farming families might choose to opt out of farming and accordingly a family history of allergic disease did appear to be reduced in farmers' children. With growing evidence that, for some aeroallergens at least, there might be an inherited component even in the disposition to specific sensitization [ 14], it is possible that this form of ‘healthy worker effect’ is even more selective, such that those remaining in the family farming tradition have lower rates specifically of hay fever and pollen allergy than those who turn to an alternative occupation. It seems fairly unlikely to provide the complete explanation, however, for reasons that are well described by Riedler et al. [ 7]. Hay fever is now at levels of about 20%, but has more or less doubled over recent decades [ 15]. To explain a 50% reduction in risk in farmers' children would require that all affected individuals from the previous generation left farming (in which case 10% of the remaining 90% would develop hay fever and the relative risk in farmers relative to non-farmers would be 11 to 20%, i.e. 0.52). Of course, selective departure from farming may have been taking place and aggregated over generations and it remains difficult to dismiss this possibility. Another non-causal explanation which has not been considered by these studies is that farmers' wives might be more likely to give birth at certain times of the year, to avoid busier farming seasons for example, and that the well documented month of birth effects on specific sensitizations contribute to a lower risk of pollen sensitization [ 16, 17]. Yet neither of these alternatives seem likely to explain the differences within a subsistence society like Ethiopia. Moreover, one might speculate on a causal mechanism whereby high exposure might reduce the risk of sensitization, and more generally the farming environment might protect against hay fever. The former does not fit easily with evidence that sensitization to house dust mite, for example, is dose related to exposure [ 18], but this perhaps reflects our past tendency to oversimplify the processes involved in the development of sensitization. More recent thinking has been directed by recognition of the fact that what distinguishes the allergen responder status of atopic individuals from that of nonatopic individuals is the nature of the TH1 and TH2 cell subsets that dominate their respective allergen-specific T-memory cells [ 19]. This picture was initially built upon the murine model, which has also provided clues to the early immune processes which may result in development of TH1- or TH2-type immunity. If the murine model is valid, the natural response to an allergen is the one that we see in nonatopic individuals who, after an unstable period in early life during which both TH1 and TH2 responses co-exist, develop a stable allergen-specific TH1 (nonatopic) response [ 20]. This ‘immune tolerance’ or ‘immune deviation’ almost always occurs in response to food allergens, to which, after a short transient stage, almost all people eventually develop immune tolerance [ 21]. However, according to this hypothesis, in some individuals there is a malfunction with this process for one or more aeroallergens and instead of developing immune tolerance, the TH2 phenotype dominates and they develop allergic sensitization. If true in man, it appears that what we need to explain is the better success rate of immune tolerance to these specific allergens in farmers children. It is suggested that one of the factors which regulates the TH1–TH2 resolution is the dose of the sensitizing allergen that is received. Indeed, a possible explanation for the difference in the overall efficiency of tolerance induction to dietary relative to inhaled allergens relates to the greater intensity of antigen exposure in the gastrointestinal tract relative to the lung [ 22]. Evidence from animal studies suggests that, whatever the genetic predisposition, tolerance even to inhaled allergen can be achieved by stimulation with a sufficiently high allergen dose [ 23]. However, whether tolerance to aeroallergens can be attained in this manner in humans, and at what dose and timing is as yet unknown [ 22]. Even if it were possible that the exposure of farmers' children to pollen is sufficiently high that type 2 T cell immunity is averted and tolerance developed, the logical extension to this theory would be that exposure to house dust mite and cat should be dramatically increased to prevent sensitization. Such a suggestion certainly cuts across current conventional thinking on the topic. Moreover, current opinion holds that the predominant drive to maturation of the immune system is confrontation with the microbial environment, particularly gastrointestinal tract pathogens or commensals [ 24]. Some gastrointestinal flora send TH1 signals which may hasten the establishment of the TH1/TH2 balance and skew immunity towards tolerance rather than sensitization. This immunological model of how infectious agents or bacteria might deviate the immune system away from the development of atopy was in fact pre-empted by the ‘hygiene hypothesis’ which made a very similar argument but motivated by the epidemiological observation that younger siblings have less allergy than older ones [ 25]. Yet it did not seem likely that further epidemiological studies in the developed world would provide the most useful means of identifying the specific exposures involved because differences across the population, like the family size effect itself [ 15], have waned over time. The strong protective effect of farming is therefore an important finding because it demonstrates that there are still groups of individuals close at hand who seem to be at much lower risk of allergic disease and are therefore a valuable resource for further study. The clear difference between farmers' and non-farmers' offspring is exposure to livestock. Two of the available studies suggest that this is the one characteristic which does explain some of the difference in risk between the two groups, with some evidence of a dose–response relationship with the level of animal contact [ 5]. After a great deal of speculation that exposure to animals, especially furry ones, was a risk factor for allergy and partly responsible for the increase in asthma in western civilizations [ 26, 27], a diametrically opposite protective effect of sharing an environment with an animal has been found in relation to dogs [ 28], and likewise pigs in the developing world [ 29]. Livestock are a potential source of all sorts of protozoans and bacteria including giardia, listeria, E. coli, salmonella, mycobacterium TB, toxoplasmosis, among others [ 30]; most of these are also contractible from some household pets. The association with farming and in particular with contact with farm animals is therefore consistent with, and complements, the current evidence on the role of infection in allergy. There is currently experimental work underway looking at the effect of contracting specific infections upon the developing immune system; the farming effect may provide an indicator of those which are likely to be influential. The authors of these various studies [ 5–7] between them have considered many of the noninfective explanations for the farming association. What is clear is that even though children of farming communities in the developed world share a similar genetic background, and much the same access to education, health care and general information as their peers in non-farming families, there are a number of potentially important differences in the environment in which they live, apart from animal contacts. On the basis of the home lifestyle information collected we know that their exposure to indoor pollution is quite different, with much higher levels of smoke from wood or coal fires but less cigarette smoke. There are also marked differences in diet, with greater consumption of saturated fats including whole milk, butter and eggs and a lower intake of margarine and snack foods. Some of these factors are among those proposed to modify airway inflammation which directly, or indirectly via immune deviation [ 31], may be involved in altering the risk of allergic disease. In fact none of these factors appeared to explain much if any of the farming effect in the published studies, but we cannot know whether the relevant component exposures were assessed. For example, it may be that the development of allergic disease is more strongly associated with maternal diet in pregnancy than the child's personal diet [ 32]. Alternatively, if ingredients such as unpasteurized milk are among those consumed more often by farmers and their families, it is possible that their diet, as well as their living environment, may contribute to an increased level of pathogen exposure in these children. The ultimate aim of all of this research is to establish what factor or factors are responsible for the increase in allergic disease over recent years, and to identify the environmental exposure which might be modifiable to prevent these conditions. In many countries of the developed world, where the prevalence of allergic disease has risen over recent decades, changes in the proportion of the population involved in farming are unlikely to be responsible. It is also clear that farming per se is not always linked to a reduced risk of allergy, and exposure of farmers to storage mites [ 33], pesticides [ 34], herbicides [ 35] and disinfectants [ 36] have all been implicated either directly or indirectly as potential causes of asthma. Nevertheless, the apparent protective effect of growing up on a farm may well provide clues which help to focus on the likely causal factors of allergic disease and to dismiss others. Much of the previous work in the field has assumed that levels of allergy are too high in the western world to look for such factors here, and has looked for answers in the developing world, where allergic diseases are just emerging. This remains important, but the farming studies provide us with a model which is relatively close at hand, and for which the differences in exposure are comparatively limited.