Associations of Abnormal Maternal Glucose Regulation in Pregnancy with Offspring Adiposity, Insulin Resistance, and Adipokine Markers During Childhood and Adolescence

Abstract

Objective: To examine the associations of abnormal maternal glucose regulation in pregnancy with offspring adiposity, insulin resistance, adipokine, and inflammatory markers during childhood and adolescence.

Study design: Project Viva is a prospective prebirth cohort (n = 2128 live births) initiated from 1999 through 2002 in Eastern Massachusetts, US. During the second trimester of pregnancy, clinicians used 2-step oral glucose challenge testing to screen for gestational diabetes mellitus. In the offspring, we measured anthropometry, insulin resistance, adipokines, lipids, and inflammatory markers in mid-childhood (n = 1107), early adolescence (n = 1027), and mid-adolescence (n = 693). We used multivariable linear regression models and generalized estimating equations adjusted for child age and sex, and for maternal age, race/ethnicity, education, parity, and smoking during pregnancy; we further adjusted for prepregnancy body mass index (BMI).

Results: In mid-adolescence (17.1 [0.8] years of age), offspring of mothers with gestational diabetes mellitus (n = 27) had a higher BMI z-score (β; 95% Cl; 0.41 SD; 0.00, 0.82), sum of skinfolds (8.15 mm; 2.48, 13.82), homeostatic model assessment for insulin resistance (0.81 units; 0.13, 1.50), leptin z-score (0.40 SD; 0.01, 0.78), and leptin/adiponectin ratio z-score (0.51 SD; CI 0.09, 0.93) compared with offspring of mothers with normoglycemia (multivariable-adjusted models). The associations with BMI, homeostatic model assessment for insulin resistance, and adiponectin seemed stronger in mid-adolescence compared with earlier time points. The associations were attenuated toward the null after adjustment for maternal prepregnancy BMI.

Conclusion: Exposure to gestational diabetes mellitus is associated with higher adiposity, insulin resistance, and altered adipokines in mid-adolescence. Our findings suggest that the peripubertal period could be a key time for the emergence of prenatally programmed metabolic abnormalities.

Keywords: Project Viva; fetal programming; gestational diabetes; life course epidemiology.

Figure 1.

Flow diagram for study participants

Figure 2.. GDM exposure and offspring cardiometabolic outcomes at three time points (7y, 13y, and 18y).

For outcomes in which we observed an association in mid-adolescence, we tested potential interactions between GDM exposure and child age using GEE models. For outcomes with interaction p-values ≤ 0.20, we illustrated the predicted estimates and their 95% CI at three time points (7y, 13y, and 18y). We adjusted models for 1) child sex and age at outcome measurement; 2) model 1 + adjusted for maternal age, race and ethnicity, education, parity, smoking status during pregnancy; and 3) model 2 + adjusted for pre-pregnancy BMI. The grey shading (lighter to darker) represents the models adjusted for precision and confounding variables, and the shapes represent the three time points (7y, 13y, and 18y).