# Table 2 Effects obtained from structural equation modeling

Effects on BMI z-scores (i. e. std) per GWG (kg/week)
Period Direct effect ± stderr Indirect effect ± stderr Total effect ± stderr (estimated via OLS regression)
t1 0.105 ± 0.060* 0.104 ± 0.043* 0.208 ± 0.042*
t2 0.255 ± 0.116* 0.241 ± 0.100* 0.498 ± 0.064*
t3 0.002 ± 0.098 0.199 ± 0.083* 0.198 ± 0.056*
Total gwg 0.608 ± 0.252*
Effects on waist circumference (cm) per GWG (kg/week)
Period Direct effect ± stderr Indirect effect ± stderr Total effect ± stderr (estimated via OLS regression)
t1 0.538 ± 0.254* 0.176 ± 0.181 0.714 ± 0.179*
t2 1.644 ± 0.489* 0.408 ± 0.421 2.055 ± 0.268*
t3 0.308 ± 0.413 0.337 ± 0.347 0.641 ± 0.237*
Total gwg 1.03 ± 1.062
1. The rows in this table represent the trimester-specific effects of GWG broken down to the direct part and the part mediated via total GWG. Significant (p < 0.05) are marked by *. The first column can be interpreted longitudinally over the whole pregnancy period: a mother gaining additional weight of 1 kg/week in t1 and t2 and compensating this gain in t3 arriving at no additional total GWG attains the direct effects of t1 and t2 (0.105 + 0.255 = 0.36) accounting for an increment in offspring’s BMI of 0.36 z-scores, but avoids the effect of total GWG. A mother continuing to gain additional weight of 1 kg/week in t3 would attain the direct effects of t1 – t3 plus the effect of total GWG (0.105 + 0.255 + 0 + 0.608 = 0.97). 