GDF15 levels during and after pregnancy
In NW, serum GDF15 increased 40-fold in T1 to 200-fold in T3 compared with nonpregnant values (i.e. after pregnancy), as shown in Fig 1A. OB showed a similar trend although the increases were slightly lower in T2 and T3 than in NW. After the initial postpartum decrease, GDF15 did not significantly change between 6, 12, and 18 months in either group. GDF15 was generally higher in OB than in NW after pregnancy, but the difference was significant only at 18 months.
In line with the differences in GDF15 levels between the two groups during pregnancy, GDF15 levels correlated negatively with body weight and BMI at T2 (R= –0.385, p = 0.003 and R= –0.336, p = 0.009, respectively) and T3 (R= –0.344, p = 0.003 and R= –0.289,p = 0.013, respectively). In analyses of body composition, GDF15 correlated negatively with FFM (but not FM) in NW at both T1 (R= –0.433, p = 0.007) and T3 (R= –0.336, p = 0.042). However, after adjustment for maternal BMI, only the negative correlation between GDF15 and FFM at T3 remained significant (Fig. 1B). When analyzing the BMI groups separately, GDF15 correlated negatively with FFM (but not FM) in NW at both T1 and T3 (R=-0.433, P=0.007 at T1 and R=-0.336,P =0.042 at T3). In the OB group, GDF15 during pregnancy did not correlate significantly with weight, BMI, or body composition, although the correlation between GDF15 and FFM at T2 narrowly missed significance (p = 0.051). After pregnancy, GDF15 correlated positively with FM at 18 months, but not after adjustment for BMI (Fig. 1C).
GDF15 levels were not consistently associated with changes in weight or body composition, apart from a negative association between early GDF15 change (T1 to T2) and gestational FM gain (T1 to T3) in NW (R= –0.382,p = 0.037).
In analysis by fetal sex, GDF15 at T1 was significantly higher in women carrying a female fetus than in those carrying a male fetus (Fig. 1 D). The difference remained significant after adjustment for maternal BMI. When analyzing the relationship between GDF15 and nausea, there was a trend for lower GDF15 at T1 in women not reporting nausea compared to those that did (12046±4298 ng/L vs 13893±5081 ng/L, respectively,p =0.12). Dividing GDF15 at T1 into quarters, the women with the lowest quarter of GDF15 reported lower incidence of nausea compared to the rest (33% vs 64%, respectively, p=0.04). This difference remained significant after BMI adjustment (p =0.04), but was non-significant after adjustment for fetal sex (p =0.16).