Linking antibiotic treatment in early life with childhood allergic disease

Childhood allergic diseases are common and lead to substantial morbidity and healthcare use. They include atopic eczema, allergic rhinitis, allergic sensitisation and asthma. There has been overall a steady increase in the prevalence of allergic disease over the last few decades, with some regional differences, which highlights the exacerbating role of environmental exposures in the development of these conditions.1 Climate change also influences the incidence and severity of allergic diseases, with immediate consequences for both patients and caregivers.2 We urgently need to understand the factors that affect the onset, severity, prevention and treatment of allergic diseases. In this issue of Acta Paediatrica, Puisto et al.3 report the findings of a nested case-control design that used data from ongoing Finnish probiotic intervention trials to investigate early life risk factors for developing childhood allergic diseases. They studied 433 children with a genetic predisposition to allergic disease, who were born to mothers with atopic eczema, hay fever, asthma or food allergies. By 2 years of age, 231 cases had met the clinical definition of atopic eczema, had a positive skin prick test for allergic sensitisation or had been prescribed inhaled corticosteroids as a proxy for obstructive respiratory symptoms. The 202 controls had not developed these conditions. Associations with early life environmental and lifestyle exposures were evaluated by models with and without potential confounders. Firstly, there is tremendous scientific value in conducting intervention studies, mainly for the primary trial outcomes but also for secondary analyses. Some very successful research programs have continued far beyond their original trial objectives. These programs include the ongoing American CAMP childhood asthma treatment trial, the Danish COPSAC nutritional intervention during pregnancy trial and the Dutch PIAMA childhood dust mite allergy intervention trial. Reassuringly, Puisto et al.3 found that giving probiotics to pregnant women decreased the risk of atopic eczema in offspring up to 2 years of age, similarly to previous reports from the trials,3 but not allergic sensitisation or inhaled corticosteroids. The new results showed that low maternal prepregnancy body mass index was associated with an increased risk of atopic eczema and allergic sensitisation. Also, intrapartum antibiotic use was similarly associated with an increased risk of atopic eczema, allergic sensitisation and inhaled corticosteroid use. In contrast, using antibiotics during the first 6 months of life was associated with a decreased risk of atopic eczema and allergic sensitisation. Exposure to pets prenatally and in early childhood was also linked with a decreased risk of allergic sensitisation.3 Obesity is a growing public health concern for children and adults, including pregnant women and women planning pregnancies. There are well-known positive links between childhood overweight or obesity and asthma and impaired lung function.4 Possible explanations include shared genetics, low-grade inflammation and the mechanistic impact on the lungs and airways. The opposite pattern was observed in the Puisto et al. study,3 as a low maternal body mass index was linked to an increased atopic eczema risk. This agreed with a meta-analysis that found that maternal underweight was associated with a higher risk of childhood atopic eczema.5 Previous studies suggest links between maternal nutritional problems during pregnancy and impaired (restrictive) lung function in offspring up to adulthood.6 The observations that maternal nutritional and metabolic status affect allergic disease in offspring deserve further research. Several studies have recognised links between antibiotic exposure and atopic eczema in early life. It is widely accepted that atopic eczema is a multifactorial disease and that genetic predisposition, altered barrier and/or immune functions and both prenatal and postnatal environmental exposures play a role. The risk of atopic eczema linked to prenatal antibiotic exposure is less established than the risk linked to early life antibiotic exposure.7 However, a meta-analysis concluded that exposure to antibiotics during pregnancy was associated with an increased risk of atopic eczema in infancy.8 Puisto et al. observed an increased risk of atopic eczema when there was in utero exposure to antibiotics and a decreased risk for both atopic eczema and allergic sensitisation if the child was exposed to antibiotic treatment in the first 6 months of life.3 These findings contrast with those of a large meta-analysis from 2018 that reported an increased risk of atopic eczema after antibiotic treatment in the first 2 years of life, but no association between antibiotic use and allergic sensitisation, defined as positive skin prick test.9 This difference may be due to the timing of antibiotic exposure, as the infant's microbiome develops quickly during their first 6 months, providing a possible susceptibility window. Puisto et al's finding that antibiotics during the first 6 months were associated with a decreased risk of atopic eczema and allergic sensitisation of approximately 50% is intriguing but should be interpreted with caution. We do not know what antibiotics were prescribed, the exact timing, or the indications for prescribing.3 Associations with antibiotic use may also suffer from reverse causation or confounding by indication, as immunologically incompetent children may need antibiotics early in life and have a concomitant vulnerability to later develop allergic disease.9 Puisto et al. also noted the risk of reverse causation, as the follow-up period overlapped with the 6-month period of antibiotic exposure.3 Reverse causation would typically lead to positive associations, namely antibiotics increasing the risk of the outcome. That was not the case here. Were the authors observing a beneficial effect of early antibiotics treatment of, for example, a bacterial skin infection that would otherwise have triggered atopic eczema development or persistence? Could the fact that only high-risk children were studied have influenced the results? We also need to note that the study only had a rather small number of patients in each exposure strata, including 17 atopic eczema cases exposed to antibiotics.3 A link, and in some studies, a dose-response relationship, between maternal prenatal and infant antibiotic use and childhood asthma has previously been established and reviewed.10 Indeed, Puisto et al. reported an association between intrapartum antibiotic treatment and inhaled corticosteroids, but no association with infant antibiotic use.3 The observed links between the increased risk of allergic conditions following antibiotic treatment may be explained by a perturbation in the child's gut microbiota and, in turn, altered immune responses. Maternal factors, such as disruption of the microbiota, for example due to antibiotic use and/or susceptibility to infections, could be passed on to offspring and may also explain these links.10 Future studies should particularly address, which antibiotics, and what doses, are linked to allergic disease and during what time frame, to identify susceptibility windows for possible interventions. Large randomised controlled trials are not generally feasible in this setting, and data triangulation is, therefore, important. As such, supportive data from controlled human exposure studies and animal models are also important. Another promising approach is to re-analyse data from other randomised controlled trials on antibiotic use, for example for infection, and allergy outcomes. This requires appropriate study designs and statistical models but may overcome the confounding and bias always seen in observational studies. Also, studying twins or siblings would control for confounding familial factors,7 which Puisto et al. partly managed by selecting genetically predisposed children.3 The Puisto et al. study adds very interesting observations to the debate on early life exposures and later allergic disease, but it is clear that further research is needed to elucidate causal pathways and involved mechanisms. EM is supported by grants from the Swedish Research Council, the Swedish Heart-Lung Foundation and Region Stockholm (ALF). The authors have no conflicts of interest to declare. Maura Kere Erik Melén