Can we predict atopic disease using perinatal risk factors?

Impressive experimental evidence is accumulating that beside genetic factors, initial allergen exposure of the immunologically naive immune system during fetal and early postnatal life induces T-cell priming and predisposes to the development of allergy or tolerance [ 1–3]. Epidemiological studies are necessary to evaluate the role of factors present at birth in the development of atopic diseases and to back up the evidence available from experiments. Such studies provide patterns of associations and produce hypotheses which fit to biological mechanisms, contributing to the proof of causal relationships [ 4, 5]. In this context, large studies are especially valuable because inferences from observations are more reliable, data can be stratified to evaluate dose–response relationships, and if historical data is added the temporality of events can be ascertained. Consistent results from large epidemiological studies are especially valuable to support a good hypothesis [ 4, 5]. The study by Bråbäck and Hedberg linked nationwide and population-based data for some 150 000 male Swedish conscripts [ 6] to the data of the Medical National Birth Register of the years 1973, 1974, and 1975. Two outcome measures were used, current asthma and allergic rhinitis in the past 12 months. These atopic phenotypes develop through the interaction of several genetic and environmental causes [ 7]. Bråbäck and Hedberg explored the influence of factors recognizable at birth [ 6]. Associations of such indicators need a judicious interpretation. The number of older siblings was inversely related to the prevalence of allergic rhinitis, but not of asthma without rhinitis. Other epidemiological studies point in the same direction: the prevalence of hay fever [ 8–11], atopic sensitization [ 10, 12] and eczema [ 8, 11, 13, 14] decreased with sibship size. These observations fit to the hypothesis that a larger number of siblings poses a greater risk of infection in early life. Early infection could promote a permanent allergen specific immune response of the TH1 type and protect from TH2 reactions in later life [ 15]. Bråbäck and Hedberg observed a relationship between gestational age and the prevalence of allergic rhinitis but not of asthma without rhinitis [ 6]: the more mature the newborn the higher the risk for later allergic rhinitis. This is in line with several other studies: the sensitization rate was also lower in preadolescent girls in Munich who had been born prematurely and had required ventilatory support [ 16]. A high gestational age was associated with an increased risk of atopic dermatitis in children in a large cross-sectional study in Denmark [ 11]. The biological basis for these observations might be the shift of the maternal immune response from TH1 towards TH2 during pregnancy, which might also progressively influence the immune response of the developing fetus [ 3]. In contrast, prematurity was found to be a risk factor for wheezing during infancy and asthma in preschool children and adolescent girls [ 16–19]. Diminished lung function after birth is a predisposing factor for wheezing illness in infants [ 20]. Prematurity, in utero smoke exposure, and gender may influence fetal lung development and lead to diminished respiratory function after birth [ 21, 22]. Smaller lungs and/or smaller airways may also be associated with young maternal age, which was shown to be a risk factor for infant wheezing, childhood asthma and also for asthma in the Swedish conscripts already mentioned [ 6, 23, 24]. These observations illustrate that a single indicator such as prematurity may be related to opposing causal factors, which can reduce and increase the risk of atopic disease, respectively. Genetic factors were not taken into account in the study on Swedish conscripts [ 6]. These powerful risk factors could have improved the fit of the models. In view of the large sample size, the odds ratios obtained for the perinatal factors are rather small, which suggests, that either the perinatal indicators were only weakly associated with the real causes or the influence of the contributing component causes is only small compared with other risk factors which were not analysed or have not yet been recognized. Risk factors are not predictors of atopic disease, especially when there is only a minor elevation of risk. It is difficult to draw conclusions from statistical models which do not include the strongest variables. In the German multicentre atopy study MAS-90 [ 25], as in other studies parental history of atopic diseases was a strong risk factor for atopic outcome in the offspring. Yet most infants who developed an atopic phenotype and were sensitized in the first 2 years of life did not have atopic parents ( Fig. 1). This demonstrates that neither genetic nor environmental factors can reliably predict that one person and not another will be atopic. In conclusion, the impressive study by Bråbäck and Hedberg [ 6] shows that — genetics apart — some indicators recognizable at birth may be of aetiological relevance for subsequent allergic rhinitis and asthma. The effects are small and have little predictive value in the individual case. Do they act in persons with or without a family history? What are the causal relationships behind these and other indicators? The puzzle remains. . Population based prevalence of atopic disease in German infants during their first 2 years of life (manifestation plus sensitization) according to the history of atopic disease (life-time prevalence) in their parents (source of data: MAS-90 [ 25]).