Objectives: To assess the association between maternal and fetal characteristics and discrepancy between last normal menstrual period and early (<20 weeks) ultrasound-based gestational age and the association between discrepancies and pregnancy outcomes.
Design: Hospital-based cohort study.
Setting: Montreal, Canada.
Sample: A total of 46,514 women with both menstrual- and early ultrasound-based gestational age estimates.
Main outcome measures: Positive (last normal menstrual period > early ultrasound, i.e. menstrual-based gestational age is higher than early ultrasound-based gestational age, so that the expected date of delivery is earlier with the menstrual-based gestational age) discrepancies > or =+7 days, mean birthweight, low birthweight, stillbirth and in-hospital neonatal death.
Results: Multiparous mothers and those with diabetes, small stature or high pre-pregnancy body mass index were more likely to have positive discrepancies. The proportion of women with discrepancies > or =+7 days was significantly higher among chromosomally malformed and female fetuses. The mean birthweight declined with increasingly positive differences. The risk of low birthweight was significantly higher for positive differences. Associations with fetal growth measures were more plausible with early ultrasound estimates.
Conclusions: Although most discrepancies between last normal menstrual period- and early ultrasound-based gestational age are attributable to errors in menstrual dating, our results suggest that some positive differences reflect early growth restriction.