Background and aim: In Switzerland, gestational diabetes mellitus (GDM) affects about 11% of pregnant women and is associated with a 7-fold increased diabetes risk. Also, postpartum weight and fat mass (FM) represent essential risk factors for type 2 diabetes. However, because of the important changes that happen during and in the year after pregnancy, it is unknown whether body composition measures in pregnancy correlate or can predict FM and distribution in the postpartum period. The aim of the study was to determine in women with GDM if body composition measures during pregnancy can predict total body FM and visceral adipose tissue (VAT) at 1-year postpartum, assessed by the gold standard dual-X-ray absorptiometry (DXA). Associations between skinfold thickness (SF), bioelectrical impedance analysis (BIA) and DXA measures of body composition and their changes, both during pregnancy and in the postpartum period were also investigated.
Methods: This study is a substudy of the randomized controlled trial MySweetHeart within a multiethnic cohort of pregnant women with GDM recruited between 2016 and 2020. So far, 194 women with GDM were 1:1 randomized into intervention vs control group (treatment-as- usual). Weight, BMI, SF (biceps, triceps, subscapular and suprailiac) and BIA measures were compared at the baseline visit in all women, and in women from the control group at 6-8 weeks and 1-year postpartum. At 1-year postpartum, body composition by DXA was also performed and its prediction by the means of different measurement methods during pregnancy was assessed.
Results: Mean BMI before pregnancy was 25.7 ± 5.30 kg. Of the 97 participants in the control group, 90 completed the 6-8 weeks and 62 the 1-year postpartum visit so far (ongoing follow- up). SF measures (calculated as % fat mass and sum of 4 skinfold thicknesses in mm) correlated only weakly to other body composition methods during pregnancy (r≤0.44), and in the early postpartum (r≤0.52), but better in the late postpartum (r≤0.74). Correlations with weight were much stronger for BIA measures (r = 0.75-0.99 at all 3 time points, all p<0.001). FM assessed by BIA decreased by 5.40 ± 3.80 kg (~17%, p<0.001) between the third trimester of pregnancy and 1 year postpartum. Total FM and VAT at 1-year postpartum as assessed by DXA was not or poorly predicted by different SF measures in pregnancy, while it was moderately to strongly predicted by body weight (r = 0.75-0.91, p<0.001) and BIA measures (r = 0.77-0.94 p<0.001).
Conclusion: FM is around 17% higher in the third trimester of pregnancy compared to the late postpartum. Body composition measured by BIA correlates better than SF measures to weight during and after pregnancy. Because of their inaccuracy, SF measures during pregnancy and in the early postpartum should be avoided. Weight and BIA, but not SF, measures in pregnancy can well predict both FM and VAT measured by DXA at 1-year postpartum and could be used for risk stratification.