THE TERM HARELIP, USED TO DESCRIBE FACIAL CLEFTS, arose from a belief that the defect originated from the mother being startled by a hare during her pregnancy. Theconcept that congenital structural abnormalities of the fetus developed from events during pregnancy is called maternal impression and it continued to be seriously discussedwell intothe19thcentury. However, it isnowknown that the vast majority of congenital structural fetal malformationsareaconsequenceofaneuploidyorabnormalembryonic development in the first 8 weeks postconception. A series of studies suggest that other complications of pregnancy, such as intrauterinegrowthrestriction,pretermbirth,andstillbirth, alsohave theirorigins, at least inpart, inveryearlypregnancy. These observations raise the possibility that women who are at increased risk of adverse outcome in late pregnancy may be identifiable in the first trimester of pregnancy, with the potential for trials of screening and early intervention. The study in this issue of JAMA by Mook-Kanamori et al presents important new information about first-trimester determination of growth restriction. The authors measured the difference between the expected size of the fetus in the first trimester, based on menstrual history, and the actual size of the fetus, assessed by ultrasound. They demonstrated that when the fetus was smaller than expected in the first trimester, there was an increased risk of preterm birth, delivery of a small-for-gestational age (SGA) infant, and delivery of a low-birth-weight infant. The associations were independent of other maternal characteristics and the risk of these outcomes was increased 2 to 3 fold among fetuses in the lowest quintile of first-trimester growth. One of the problems in assessing fetal growth is to differentiate between infants who are growth restricted (grew to less than their genetically determined potential) from infants who are constitutionally small. The study by MookKanamori et al suggests that the associations between firsttrimester growth and the risk of delivering an SGA infant are due to growth restriction because smaller than expected size of the fetus in the first trimester was unrelated to maternal weight and height. Moreover, smaller size in the first trimester was associated with more rapid growth in infancy, suggesting ex utero catch-up growth to compensate for true growth restriction in utero. The findings of Mook-Kanamori et al confirm a previous study thatdemonstratedan increased riskof lowbirthweight, delivery of an SGA infant, and preterm birth among spontaneousconceptionsinwhichthefetuswassmallerthanexpected in the first trimester. The relative weakness of both studies is their reliance on menstrual data to establish the expected size of the fetus inearlypregnancy.Theseanalysesassumethat fertilization took place on day 14 of the menstrual cycle and the first day of the last menstrual period is taken as day 1. It is possible that fertilizationat laterdays in themenstrual cyclecould be associated with adverse outcomes and this would also be associated with smaller than expected size of the fetus in the first trimester. However, secondary analysis of a large-scale, multicenter cohort study of Down syndrome screening conducted in the United States analyzed the associations between early pregnancy growth measurements and the risk of delivering an SGA infant among approximately 1000 women who conceived after use of assisted reproductive technology, and for whom the date of conception was known. This study confirmeda linearrelationshipbetweensmaller thanexpectedsize in the first trimester and both lower birth weight and the proportionofSGAinfants.Hence, theassociation is still observed when the timing of conception is known. The analysis reported by Mook-Kanamori et al also described the maternal characteristics associated with reduced growth in the first trimester. After taking into account multiple comparisons, multivariate analyses demonstrated a significant positive relationship between crown to rump length and maternal age and a significant negative relationship between crown to rump length and hematocrit. The association between poor first-trimester growth and a high hematocrit level is of particular interest. Hematocrit normally declines in early pregnancy and higher hematocrit in early pregnancy is associated with increased perinatal mortality. The association could reflect a direct influence of high hematocrit levels on pregnancy (eg, through alteration in uterine blood flow). Alternatively, it could indicate that the absence of the normal decrease in hematocrit is a marker of some other determi-