What’s new in atopic eczema? An analysis of systematic reviews published in 2019. Part 1: Risk factors and prevention

This review is part of an annual evidence update on atopic eczema (AE), providing a summary of key findings from 18 systematic reviews published in 2019 on AE risk factors and prevention. Parental atopy, particularly AE, is a risk factor for offspring AE, and this risk is augmented both by the number of parental atopic diseases present and the number of affected parents. Low‐quality evidence suggests that autumn or winter birth increases childhood AE risk compared with birth in spring. There is some evidence to support filaggrin gene–environment interactions; however, this is limited by small underpowered studies. There is no evidence to suggest that polymorphisms in the –1082, –592 and –819 loci of the interleukin‐10 gene increase susceptibility to AE. There is no robust evidence to support a relationship between childhood AE development and either yoghurt consumption in the first year of life, gut microbiota variants, prenatal or infantile paracetamol exposure, maternal antibiotic exposure or air pollution. Three systematic reviews investigated the effect of probiotics given during pregnancy or infancy; although low‐quality evidence suggests benefits of combined probiotics, these studies were limited by significant heterogeneity. No relationship between the age at which complementary food and beverages are introduced and the risk of developing AE in infancy was identified. Consistent evidence showed no relationship between human milk feeding and infant AE development, aside from limited evidence suggesting a protective role in those with atopic heredity. This summary of recent evidence related to AE risk factors and prevention highlights the complex aetiology of AE.


Introduction
The aim of this first part of a two-part annual evidence update is to present key findings from systematic reviews (SRs) published or indexed in 2019 on the risk factors for, and the prevention of, atopic eczema (AE). We identified publications using our standard search strategy for publishing this series of evidence updates. 1 SR characteristics (Table S1) and meta-analyses (Table S2) were extracted in duplicate using standardized forms. The quality of each SR was appraised using the AMSTAR2 tool (Table S1). 2

Risk factors
Parental atopy A meta-analysis of 119 observational studies (241 651 participants) demonstrated a significant association between parental history of atopy and presence of AE in the offspring (OR = 1.81; 95% CI 1.65-1.99); in subgroup analyses of all the atopic diseases, parental history of AE showed the strongest association (OR = 3.30; 95% CI 2.46-4.42). 3 Relative effect sizes were similar for maternal and paternal atopy. The risk of offspring AE development increased in association with the number of parental atopic diseases present and the number of affected parents. This SR scored highly using AMSTAR2, indicating robust methodological quality of the review.

Season of birth
Meta-analyses of 9 quantitative studies (726 378 participants) demonstrated a weakly positive association between the development of childhood AE and birth in autumn and winter compared with spring. 4 However, no clear evidence of an association in the qualitative analysis (676 867 participants) was found. Potential concerns about this study include recall bias and AE misclassification because AE was diagnosed from questionnaires. Furthermore, there may be a risk of confounding, as only northern hemisphere countries were included.

Gene-environment interaction
One SR assessed gene-environment interaction (GEI) in the filaggrin (FLG) loss-of-function mutation, 5 which represents the strongest genetic association with AE. 6 There was limited evidence for GEI between FLG genotype and breastfeeding duration, having older siblings, phthalate exposure in household dust, urine phthalate metabolite levels, water hardness and early-life exposure to cats. All GEIs increased the risk of AE development, apart from prolonged breastfeeding, which decreased the risk. The authors highlighted that small participant numbers and underpowered studies were the reason for lack of replication, that reverse causality was possible owing to the timepoint of AE diagnosis and that adjusting for confounders was inconsistent, therefore cautioning the interpretation of these results.

Interleukin-10 gene polymorphisms
Two SRs and meta-analyses investigated the association between genetic polymorphisms within the interleukin (IL)-10 gene (IL10) and susceptibility to AE. 7,8 Both SRs investigated the same IL-10 gene loci, and 13 overlapping case-control studies were identified.
Both SRs had small samples sizes and neither study defined the AE outcome investigated.
Zhao et al. 8 concluded that overall, no significant association was detected in three single nucleotide polymorphisms (SNPs) of IL10 (-1082 G/A, -592 A/ C, -819 G/A), apart from in the recessive model for -819. Subgroup meta-analyses revealed ethnicityspecific effects. Similarly, Qi et al. 7 found no strong evidence to support a relationship between the three IL10 SNPs (I-1082 A/G, -592 A/C and -819 T/C) and susceptibility to AE.

MicroRNA
One SR investigated whether a link exists between micro (mi)RNA (noncoding RNA that controls gene expression), autism spectrum disorders and AE. 9 Using seven case-control studies (321 participants), the authors concluded that the miRNAs -146 and -155 are dysregulated in both conditions. However, statistical analyses were not conducted and most of the AMSTAR2 quality criteria were unmet.

Air pollution
One SR of 57 observational studies (2 693 223 participants) assessing air pollution and the risk of AE development concluded that small-scale exposure, such as truck traffic emissions, increases AE prevalence, whereas large-scale exposure to larger particles has no effect. 10 However the quality of evidence was low.

Maternal antibiotic exposure
One SR of seven observational studies investigated the effect of maternal antibiotic exposure on infant AE. 11 Meta-analyses found that maternal antibiotic exposure was significantly associated with AE by 1 year of age (1490 participants) but was not significant after 1 year of age (88 601 participants) or when antibiotic exposure occurred during the third trimester (62 971 participants). The small number of studies, with a low proportion (10-30%) of participants exposed to antibiotics, potentially affects the robustness of these results. heterogeneity between included studies that did not adjust for confounders was identified.

Prenatal and postnatal prevention Probiotics
Three SRs assessed the effect of probiotics on the development of AE, with overlap identified in the included studies. [13][14][15] The first, a small SR of 1805 participants. reported that probiotic supplementation in the pregnant mother was associated with a reduced risk of infant AE development. 13 The other two SRs assessed probiotic supplementation in both mother and infant, concluding that probiotic supplementation in the pregnant mother, breastfeeding mother or infant demonstrated risk reduction in childhood AE. 14,15 On subgroup analysis this relationship was insignificant with infantile supplementation alone. All three SRs conducted meta-analyses for different strains and concluded that mixed-strain probiotics were superior to single-strain probiotics, 13-15 but the results were limited by study heterogeneity.

Gut microbiota
One SR assessed 44 observational and interventional studies (7059 participants) and concluded that the role of the gut microbiome in the onset and severity of pre-existing AE remains unclear. 16 Few of the included studies adjusted for confounders, and multiple critical domains of the AMSTAR2 checklist were unmet.

Yogurt and fermented milk
One SR investigated the effects of yogurt and fermented milk products on childhood AE, 17 evaluating two prospective cohort studies (2591 participants). The authors concluded that yogurt consumption in the first year of life is associated with a lower risk of developing AE. Methodological weaknesses related to the literature search strategy and the authors' conflicts of interest were identified.

Complementary feeding
One SR undertaken by the United States Department of Agriculture (USDA) as part of the 'Pregnancy and Birth to 24 months' project included 20 observational studies (46 988 participants). 18 The study found no relationship between the age at which complementary foods and beverages were introduced and the risk of developing childhood AE. 18 Furthermore, no relationship was identified between the development of AE and the age of introduction of peanuts, tree nuts, sesame seeds, egg, cow-milk products, fruit, vegetables or meat. Across 15 studies (30 146 participants), there was limited evidence to suggest that introducing fish within the first year of life may reduce the risk of AE. This review scored highly using the AMSTAR2 tool.

Human milk
One SR from the same USDA 'Pregnancy and Birth to 24 months' project found, using mostly observational evidence from Europe (30 052 participants), inconclusive evidence on 'never' vs. 'ever' human milk feeding and 'shorter' vs. 'longer' durations of any human milk feeding with development of AE. 19 In childhood (10 185 participants), most associations between the duration of any human milk feeding and AE development were nonsignificant. A second SR assessed 27 prospective cohort studies (177 445 participants) from Europe, Asia and Australia, and found no association between AE and total breastfeeding. 20 Subgroup meta-analyses demonstrated some evidence for a protective role of 'total' and 'exclusive' breastfeeding in a cohort with atopic heredity. However, these findings should be interpreted with caution because heterogeneity was evident.
A third SR assessed 15 studies (2609 participants) investigating the effect of transforming growth factor (TGF)-b in human milk, and found no consistent association with the development of AE. 21 The methodological quality of this review was considered to be thorough and robust.

Conclusion
AE is a complex, multifactorial condition, and deconvoluting the risk factors and strategies for prevention is challenging. There is clearly a strong genetic component, but exact mechanisms and the role of environmental factors in AE aetiology requires further elucidation with well-conducted, appropriately powered studies.

Learning points
• Parental atopy, in particular AE, is a risk factor for offspring AE, which is further augmented by the number of parental atopic diseases present and the number of affected parents. • There is limited evidence in support of FLG GEI, including an effect of breastfeeding duration, having older siblings, phthalate exposure in household dust and urine phthalate metabolite levels, water hardness and early-life exposure to cats. • Studies on probiotic supplementation in pregnancy and infancy have provided limited evidence of benefit in reducing the risk of AE development; however, these reviews were limited by study heterogeneity. • There is no robust evidence to support a relationship between the development of childhood AE and yoghurt consumption in the first year of life, or with gut microbiota variants, prenatal or infantile exposure to paracetamol, or maternal exposure to antibiotics. • Moderate evidence suggests no relationship between the age at which complementary food and beverages are introduced and the risk of developing AE. • There is consistent evidence to support no relationship between infant AE and either human milk feeding or human milk TGF-b. 19

CPD questions
Learning objective To gain up-to-date knowledge on the latest evidence regarding risk factors and prevention of atopic eczema.

Question 1
In which scenario is the offspring most at risk of developing atopic eczema (AE)?