Variables in the multivariate model.

*<p>Gestational age (OR for a reduction of 1 week).</p>†<p>Birthweight by gestational age z-score (OR for a reduction of 1 standard deviation).</p>‡<p>Birthweight by gestational age z-score, values >2 (OR for an increase of 1 standard deviation).</p>•<p>Maternal level of education (OR for lowest vs. highest educational level).</p>□<p>Maternal age at birth (OR for an increase of 1 year).</p><p>For GDM the multivariate model showed an increased risk with low gestational age and a U-shaped distribution for maternal SDS, with the highest risk for mothers born with increasingly low birth for gestational age or increasingly high birthweight by gestational age. For pre-eclampsia the model showed an increased risk with low gestational age and a linear association with SDS. The model was corrected for maternal age and education, and the confounding factors were distributed as expected, with a significant effect of maternal educational level on GDM (higher educational level leads to lower risk of GDM) and a significant effect of maternal age on pre-eclampsia (low maternal age leads to higher risk of pre-eclampsia). No significant interactions between SDS and maternal age or educational level were found.