Contribution of Gestational Weight Gain on Maternal Glucose Metabolism in Women With GDM and Normal Glucose Tolerance

Fernanda L. Alvarado; Perrie O'Tierney-Ginn; Patrick Catalano


J Endo Soc. 2021;5(2) 

In This Article


Participant Characteristics

A total of 46 women met inclusion criteria for this analysis. No subject had diabetes prior to a planned pregnancy, based on the 75-g OGTT. All subjects had normal preconception thyroid, renal, and liver function and negative islet cell antibodies during each period of the study. Seventeen (37%) women developed GDM.

In the complete cohort, 98% of the subjects were Caucasian. Maternal characteristics of the NGT and GDM women pre-gravid are summarized in Table 1. Risk factors for GDM were significantly higher in the women who developed GDM. The second significant pre-pregnancy difference between the groups was decreased insulin sensitivity in women who developed GDM in late pregnancy (P = 0.01). There were no significant differences in insulin response and disposition index. Because of the original study design, there were no significant differences in maternal age, parity, weight, FM, %BF, and FFM.

Relative and Absolute Change During Pregnancy

As shown in Table 2, absolute and relative changes, defined as changes between values taken at 34 to 36 weeks and the pre-pregnancy period—in body composition, insulin sensitivity, insulin response, and disposition index—were similar between GDM and NGT women. Absolute change in FM was slightly, but not significantly (P = 0.06), higher in the NGT women.

Relationships between Δ in maternal glucose metabolism and Δ in body composition are shown in Table 3. Absolute change in insulin sensitivity was negatively correlated with Δ in body weight in the entire cohort (R2 = 0.20, P = 0.001); however, when analyzed by GDM status, a significant correlation was observed only in the NGT women (R2 = 0.24, P = 0.002).

To better examine the potential associations of adiposity and glucose metabolism, we performed the analysis using Δ FM and Δ %BF. The Δ in insulin sensitivity was negatively associated with Δ in FM in the combined cohort (R2 = 0.13, P = 0.008) but was not significant in either of the NGT or GDM groups. There was no correlation between maternal Δ in insulin sensitivity and Δ %BF.

We found no association between Δ in insulin response and Δ in body composition: Δ FM, Δ %BF and Δ FFM. The Δ disposition index was negatively correlated with Δ body weight in GDM (R2 = 0.26, P = 0.01), as shown in Table 3.

We performed a power analysis for the simple linear regression for both NGT and GDM subjects. For the NGT subjects, an n = 29 achieves 83% power to detect 32% difference on the R2. However, for the GDM subjects, an n = 17 achieves 37% power to detect 18% difference on the R2.

Last, as shown in Table 4, we performed a multilinear regression analysis to assess the effects of each maternal characteristic and their relationship to the Δ in insulin sensitivity during pregnancy. In the analysis, we adjusted for maternal age, gestational age at delivery, maternal pre-pregnancy weight, parity, Δ body weight, and group (NGT and GDM). The dependent variable was Δ insulin sensitivity. This model explains 33% of the variation of Δ in insulin sensitivity. The 2 variables that were significantly associated with Δ insulin sensitivity were gestational age at delivery (ΔR2 = 0.16, P = 0.03), and Δ body weight (ΔR2 = 0.09, P = 0.02). We ran a model including baseline insulin sensitivity; however, adjusting for this variable in the model increased collinearity, and decreased the contribution of Δ body weight (ΔR2 = 0.03, P = 0.68).

We performed a power analysis for the multiple linear regression and determined that an n = 46 achieves 86% power to detect 38% difference on the R2 between Δ in insulin sensitivity and Δ in body weight with an alpha level of 0.05.