Parental preconception BMI and infant birthweight
Figure 1 presents model estimates for parental BMI associations with infant birthweight. There was weak evidence for an interaction between parental BMIs for median birth weight (β=-0.98; 95%CI=[-1.90, -0.05], p=0.04) with infants having increasing birth weight with increasing parental BMI, when one parent has normal weight (Figures 1A and 1C). When either parent was overweight or obese, although median infant birthweight are higher (maternal β=15.9; 95%CI=[1.63, 30.1], p=0.03 and paternal β=7.33; 95%CI=[0.297, 14.4] p=0.04), they were not associated with increasing BMI of the other parent (Figure 1B and Figure 1D).
Maternal overweight and obesity was associated with increased risk of extreme LGA (95th percentile: β=-25.1; 95%CI=[5.07, 45.1], p=0.01), while there was no effect of increasing paternal BMI (90th percentile: p=0.50; 95thpercentile: p=0.43). In both the median, 90th and 95th percentile models, there was no evidence of a non-linear association between maternal BMI and birthweight (median: p=0.61; 90th percentile: p=0.32; 95th percentile: p=0.70). However, in the 5th and 10th percentile regression there was a strong non-linear association between maternal BMI and infant birthweight (p=0.002 and p=0.03). Such that infant birthweight increased with maternal BMI, approximately 27.5kg/m2or lower, but for maternal BMIs in the obese range (>30kg/m2), the 5thpercentile of infant birthweight plateaued, indicating a greater divergence from the median baby weights and greater risks of SGA. There was no effect of paternal BMI and risk of SGA infants (5th percentile: p=0.60; 10thpercentile: p=0.52).