Adriëtte Oostvogels

Results The children were evaluated at an average age of 5.6 years; they were predominantly of Dutch origin (79%) and 47% were girls. Of the mothers, 14.7% had overweight and 4.7% had obesity. Accelerated weight gain was present in 9% of the children and accelerated weight-for-length gain in 24%. Compared with children who were approached for the 5-year measurements but were not included (n=4478; due to non-response, no fasting blood sample, no data on postnatal growth or outcome variables; Figure 1), the current subgroup of children (n=1459) were more frequently of Dutch origin, had higher birth weight, higher gestational age, and older and higher educated mothers (Supplementary Table S1). Table 1 presents the mean maternal pBMI and postnatal weight and weight-for- length gain for different baseline characteristics. Higher maternal pBMI and higher postnatal growth were mostly observed among women with lower education, of non-Dutch descent, and in women with hypertensive disorders during pregnancy. Women with high pBMI had children with higher birth weight, while children born with lower birth weight had higher postnatal growth. Higher postnatal growth was also observed in children with shorter duration of breastfeeding. Children spending more than 2 hours per day behind a screen and with higher energy intake at age 5-6 years had experienced higher postnatal weight gain. The results of the linear regression analyses assessing the association between maternal pBMI, postnatal weight and weight-for-length gain and the metabolic components are presented in Table 2. In unadjusted models (model 1) we found independent associations between maternal pBMI and several metabolic components and the metabolic score of the offspring. In adjusted models for maternal and offspring’s characteristics (model 2), there was little change in the effect sizes. Adding postnatal weight gain to the model (model 3) or postnatal weight-for-length gain (model 4) did not change the association between maternal pBMI and the metabolic components/score, indicating no mediating role of postnatal weight or weight-for-length gain. When testing for interaction between maternal pBMI and postnatal weight gain in the association between maternal pBMI and the metabolic components/score (model 3), we found a significant interaction term in the association with SBP (p=0.021) and the metabolic score (p=0.047). The highest SBP and metabolic scores were found for children of mothers with overweight/ obesity who had accelerated postnatal weight gain compared to the other groups (Figure 2). When testing for interaction between maternal pBMI and postnatal weight-for-length gain in the association between maternal pBMI and the metabolic components (model 4), we found a significant interaction term in the association 182 Chapter 7