Discovering the causes of atopy

Patterns of childhood infection and fetal growth may be implicated

The marked increase in the prevalence of childhood asthma, eczema, and hay fever in Britain over the past 30 years or more is largely unexplained. However, it is likely to be attributable to a rise in the prevalence of atopy. This is characterised by exaggerated Th2 cell responses to common allergens with production of raised concentrations of allergen specific IgE. Although we now understand more about the genetics of atopy and the role of Th1 and Th2 cells in the control of IgE, the environmental causes of atopy have eluded us. Of increasing interest is the potential roles that patterns of childhood infection and fetal growth and maturation might have in the inception of atopy.

The number of older siblings has been shown to be inversely related to the prevalence of adult hay fever and infant eczema.1 This observation led Strachan to propose in the BMJ in 1989 that atopy may have increased because of a fall in exposure to infections in early childhood through improved hygiene and reductions in family size and overcrowding in the home.1 Children are likely to experience more severe infections at an earlier age when the number of their older siblings is greater. Thus, it was suggested that infections in early childhood might protect against atopy and that successive cohorts of children have progressively lost this protection.

Several studies have since confirmed that family size, and birth order in particular, are associated with hay fever and atopy.2 A mechanism has been proposed by which early infection by viruses and bacteria, through the preferential induction of Th1-type cytokines, could prevent atopic sensitisation,3 although these effects may depend more on infective dose4 than on age at infection.

More direct evidence that childhood infection might prevent atopy comes from a recent historical cohort study in Guinea-Bissau, West Africa, which found that young adults who had experienced measles in childhood during a severe epidemic were significantly less likely to be atopic than those who had been vaccinated and not had measles.5 We do not know whether the findings for measles may apply to other respiratory viruses which are more difficult to study in population based studies. However, measles may be special in that it can cause severe damage to the thymus6 and has been associated with reductions in cell mediated immunity three years after infection.7

In this week's BMJ Matricardi and colleagues (p 999) describe how Italian military students who were seropositive for hepatitis A were less likely to be atopic and to have atopic disease than those who were seronegative.8 Adult seropositivity for hepatitis A is likely to be a marker of predominantly childhood infection. This suggests that hepatitis A infection, and perhaps other enteric infections in childhood, might prevent atopy. However, after hepatitis A status was controlled for, a substantial association between birth order and atopy remained, suggesting that the number of older siblings may tell us about the effects of infections other than hepatitis A.

There are several puzzles concerning the “infection hypothesis.”9 It is not clear, for example, why studies have not found consistent associations between family size and asthma,2 nor why preschool nursery attendance, which is known to promote cross infection and more severe infection, does not seem to be associated with a reduction in atopy.10 Further insights may be gained by more detailed studies in countries where there is greater variation in the burden of childhood infectious disease. Virologists and immunologists must collaborate with epidemiologists if we are to really understand the role of infections in the development of atopy.

The growing body of evidence linking patterns of fetal growth to adult disease11 has focused attention on the role of the prenatal environment in the aetiology of atopy and atopic disease. Babies who develop atopy in infancy have evidence of an altered T lymphocyte phenotype at birth.12 Also, Olesen and colleagues report the findings from two historical cohort studies in Denmark in this week's BMJ (p 1003).13 They observed in one study that babies who weighed 500 grams or more above average at birth were at increased risk of atopic dermatitis in childhood, compared with babies of average birth weight. In both studies they found that babies whose gestational age at birth was ≥41 weeks were also at increased risk compared with babies born at term.

These findings are in keeping with a previous study of adults, which also found that higher birth weight and postmaturity were associated with a raised concentration of serum total IgE, a marker of atopy.14 Interestingly, that study found that larger head circumference at birth was a more powerful predictor of raised adult IgE than was birth weight and explained the association between birth weight and IgE. It has been argued that a larger head circumference for a given birth weight indicates a disproportionate pattern of fetal growth arising as a consequence of undernutrition in late gestation in a fetus on a fast growth trajectory. This could lead to impaired growth of the trunk and thymus. It is speculated that disproportionate growth and postmaturity, which are associated with a reduction in thymic weight, may alter the balance of Th1 and Th2 cell populations in the thymus in favour of Th2 cells.14

In contrast to these intriguing findings, however, other studies have either reported no association of birth weight and gestational age with atopy10 or have found that lower birth weight15 and lower gestational age16 were associated with increased risk. Recently initiated prospective studies, which have recruited women early in pregnancy, are likely to unravel the complexities of these associations and greatly increase our understanding of the relations of prenatal nutrition and growth to atopy and atopic disease in children. This will hopefully bring us nearer to our ultimate goal of primary prevention.