DISCUSSION
In this study, we examined the effect of GDM on children allergic diseases. Based on eight studies (two case-control studies and six cohort studies), this meta-analysis found that compared to the normal, the children born to GDM women were at a 17% higher risk of allergic diseases (OR:1.17, 95%CI:1.10-1.25). Based on the types of allergic diseases, we conducted a subgroup analysis to reduce heterogeneity (25),providing that the effect estimates of each allergic disease: 1.13 (OR:1.13,95%CI:1.01-1.27) in asthma group, 1.12 (OR:1.12,95%CI:1.08-1.16) in wheezing group, and 1.41 (OR:1.41,95%CI:1.31-1.53) in atopic dermatitis group. GDM may increase the risk of allergen sensitization, but the relationship was not statistically significant (OR:1.45,95%CI:0.40-5.23). We also explored the source of heterogeneity in the study design. No heterogeneity was found in the case-control study, but higher heterogeneity in the cohort study. For sensitivity analysis, we excluded studies in clearly-defined GDM patients and not clearly defined GDM patients respectively, and found that children born to GDM women regardless of whether GDM was clearly defined all confronted an increased risk of developing allergic diseases (OR:1.10,95%CI:1.04-1.17;OR:1.19,95%CI:1.10-1.28)
In previous studies, Rusconi et al. (19) found an association between maternal diabetes during pregnancy and early-onset wheezing in children (OR:1.72, 95%CI:0.99–3.00). Kumar et al. (22) found that GDM increased the risk of atopic dermatitis (OR:7.2, 95%CI:1.5-34.5) and early allergen sensitization in childhood (OR:5.7, 95%CI:1.2-28.0). In a Swedish record linkage study, Aspberg et al. (26) reported that maternal diabetes was associated with a 19% higher risk of having a child who would be prescribed with anti-asthmatic medication (OR: 1.19, 95%CI:1.12–1.28).
In the study of childhood asthma, Paula’s findings (7) were inconsistent with ours (OR:1.13, 95%CI: 1.01-1.27), probably because the children of their studies were all born as very preterm (<32 weeks)(OR: 0.96, 95%CI: 0.58-1.69) and late preterm (34+0–36+6weeks) (OR: 0.78, 95%CI: 0.63-1.11). The inconsistency may also be related to sample size, and uncontrolled confounding factors, such as antenatal steroids and early-life feeding, or exclusivity of breastfeeding. In the study of children wheezing, the results of Rusconi et al. (19)(OR: 0.94, 95% CI:0.47-1.86) and Liu et al. (21) (OR: 0.85, 95%CI:0.67-1.09) were inconsistent with ours (OR:1.12, 95% CI:1.08-1.16). Only children with late-onset wheezing were studied in both studies. Liu et al. did not explain the results of late-onset wheezing unassociated with GDM. Rusconi believed that the misclassification of children in the late-onset wheezing group could lead to the inconsistent results. In the study of atopic dermatitis in children, the results of Kumar et al. (22) (OR: 0.74, 95%CI:0.11-4.91) and Paula et al. (7)(OR: 0.62, 95%CI:0.23-1.69) were inconsistent with ours (OR: 1.41, 95%CI:1.31-1.53). In both studies, only children born as premature (<37 weeks and moderate preterm 32+0–33+6 weeks) were included. The reasons that preterm is not related to GDM have been described above. In the studies about children’s allergen sensitization, the three sets of results reported by Kumar et al (OR: 0.37, 95%CI:0.07-1.88) (OR: 0.26, 95%CI:0.04-1.64) (OR: 0.38, 95%CI:0.07-2.19)(22) were inconsistent with ours (OR: 1.45, 95%CI:0.40-5.23). The children studied by Kumar et al. were all born as preterm.
Currently, no mechanisms are revealed to interpret the association between GDM and childhood allergic diseases. We explored the possible mechanisms through this meta-analysis. In GDM, fetal exposure and limited response to high cord blood glucose (27) may cause fetal hyperinsulinemia, hypoxia (28), and ultimate lung immaturation. In addition, the maternal autoantibody may cross the placenta to directly act on the fetal lung development (29). An epidemiological study of human adults in the Netherlands revealed a link between GDM and atopic diseases in early childhood, and the mediatory role of insulin resistance mediates between obesity and allergen sensitization (30). Mothers with GDM have higher levels of TNF-α, leptin and visfatin, and lower levels of adiponectin. Adiponectin can reduce allergic inflammation in mouse models. Therefore, GDM-decreased adiponectin may have a negative impact on the development of infant immunology (31-34). Apart from this, maternal diabetes predisposes mothers to premature births who carry a greater risk of neonatal alveolar surfactant deficiency. In addition, high blood sugar level may significantly inhibit fetal alveolar surfactant synthesis and secretion during late pregnancy (35, 36). Further researches are needed to determine the mechanisms between GDM and children allergic diseases.