Background
Low birth weight (LBW) leads to a higher risk of perinatal mortality and morbidity, including impaired growth and cognitive development [
1]. More long-term complications include higher risks for high blood pressure [
2] and cardiovascular disease [
3], impaired glucose tolerance and type 2 diabetes [
4], early age at menarche [
5] and menopause [
6], and reduced bone mineral density [
7] and osteoporosis [
8]. LBW can relate to one or both of premature birth and fetal growth restriction, or being constitutionally small, and risk of LBW can be related to such factors as ethnic differences, multiple birth pregnancies, maternal age at birth, fetal environmental factors such as exposure to alcohol, smoking or illicit drugs, maternal nutrition during pregnancy, poor socioeconomic status [
9] and genetic defects [
10]. Another risk factor appears to be hyperemesis gravidarum [
11‐
13], a severe form of nausea and vomiting in pregnancy that can lead to maternal dehydration and weight loss. Treatment of severe nausea and vomiting in pregnancy with anti-emetics may even be associated with a reduction in the prevalence of LBW [
14,
15], although such findings are by no means universal [
16‐
18].
Whilst the association between hyperemesis gravidarum and higher risk of LBW is reasonably well established, what is not so clear is whether potentially less severe nausea and/or vomiting in pregnancy is also associated with the risk of delivering LBW babies. The only recent related evidence suggests that it may be associated with being small for gestational age (SGA) due to fetal growth restriction, one of the main reasons for a baby having a LBW [
19]. This study was therefore designed to test the hypothesis that nausea and vomiting in pregnancy, of insufficient severity to require treatment, is associated with the risk of delivering a LBW baby. To do this we used data collected for the Cambridge Baby Growth Study.
Discussion
Vomiting in pregnancy, not treated with anti-emetics, is associated with a higher risk of giving birth to LBW babies in this study. This is consistent with reported associations between LBW (or related phenotypes such as SGA) and hyperemesis gravidarum [
11‐
13,
25‐
28], although such associations are not universal findings [
29‐
32]. Nausea and vomiting with no reference to hyperemesis gravidarum has also been associated with a higher risk of LBW [
33] and decreased birth weight [
34] in some other studies, although no difference in risk was reported in others [
19,
35‐
37]. Although a systematic review [
38] reported a lower risk of LBW in association with nausea and vomiting in pregnancy the studies that it was based looked at anti-emetic use to categories study participants [
14,
15]. This is therefore very different to our own study where we specifically excluded women who took anti-emetics in case these drugs affected pathways involved in regulating LBW risk [
39]. Similarly the very large Norwegian Mother and Child Cohort Study [
40] found reduced population rates of LBW in association with nausea and vomiting in pregnancy, but did so without specifically excluding those women who took anti-emetics in pregnancy. Our study therefore presents results a slightly different population to those examined in other investigations, and one where we specifically investigated a potentially milder negative aspect of pregnancy than hyperemesis gravidarum.
This higher risk for LBW associated with vomiting, and nausea but just in the second trimester, appeared to be in the first two trimesters of pregnancy with an apparent lack of risk in the third trimester. Studies are ongoing to try and discover whether or not nausea and vomiting in pregnancy and resulting associations are genetically mediated [
41,
42]. With an overall odds ratio of 3.5 for LBW in our population, vomiting in early pregnancy may be a marker of risk for LBW that is useful for its prediction in conjunction with other risk factors. Where routine ultrasound is available small babies tend to be assessed as either SGA (possibly due to fetal growth restriction) and/or premature rather than LBW, but the link with vomiting may be useful in areas where such scans are not available and LBW is used to encompass them both. Nausea and vomiting in pregnancy are thought to be protective towards the embryo/fetus in terms of reducing exposure to food borne harmful substances such as infective microorganisms [
43], and they can lead to changes in the maternal dietary intake [
44]. There is evidence that this can lead to positive effects on the fetus such as decreased rates of miscarriage and congenital malformations [
45]. Whilst potentially being an advantage in a mild form, in excess it is possible that this vomiting might reduce nutrient delivery to the fetus leading to the greater risk of LBW [
46]. The fact that in our population as a whole there was no apparent decrease in mean birth weight despite the higher prevalence of LBW suggests that whilst there is a negative impact of vomiting on birth weight for some babies, in other babies a protective advantage may be evident [
43]. The Kernel density estimation plot for birth weight in our population (Fig.
1(a)) would appear to be consistent with this suggestion (as birth weight density around the mean appeared to be higher in those women affected by vomiting).
LBW often relates to the baby undergoing fetal growth restriction and/or being born prematurely. Given the increase in birth weight usually observed in male babies, in general there is also a slightly increased risk of LBW in female babies compared to males [
9]. In our population despite the association between vomiting and LBW, we did not find further significant independent associations between exposure to vomiting in pregnancy and the prevalence of SGA (a group that may have been enriched with babies who underwent fetal growth restriction) or premature births or the gestational age at birth. Looking at the Kernel density estimation plots for birth weight in our population below the cut off for LBW the line for women who experienced vomiting was clearly a little higher than those representing the other groups (Fig.
1(a)). The differences between the groups below the cut offs in the gestational ages at birth (to assess the densities related to prematurity) and the birth weights for gestational age (to assess the densities related to SGA) plots were smaller though (Fig.
1 (b) and (c)). Further studies are required to validate these findings in other cohorts, especially those that are better able to distinguish whether the LBW relates more to SGA (and therefore probably fetal growth restriction), prematurity or a mix of the two.
Whilst we could not ascribe the increased risk of LBW to fetal growth restriction or premature birth we did note a slight excess of pregnant women carrying female babies. This excess has previously been observed with hyperemesis gravidarum [
47]. We recently reported that serum GDF-15 concentrations around week 15 of pregnancy were higher in those women reporting vomiting in the second trimester of pregnancy in the Cambridge Baby Growth Study [
48], a hormone that may stimulate vomiting in pregnancy [
49]. Interestingly we have also observed an increased concentration of week 15 serum GDF-15 concentrations in women carrying females compared to those carrying males in our population (Petry et al., unpublished observation), which may explain the slight excess of pregnant women carrying females in the group that experienced vomiting.
The main strengths of this study are its prospective nature and the fact that its design enabled us to study a group of women with a potentially milder, albeit still unpleasant phenotype than has been tested in other published studies. We have therefore uniquely been able to observe that the risk of giving birth to a LBW baby is higher with vomiting. The study’s main limitation is that the nausea and vomiting and the taking of any anti-emetics to treat it were self-reported. The analysis could therefore have been affected by recall bias [
50], although given that the women were encouraged to fill in their questionnaires as their pregnancies progressed rather than retrospectively the effects of this may have been limited. Indeed effects of any recall bias were clearly not sufficient to prevent us from discovering our association between increased circulating GDF-15 concentrations and second trimester vomiting [
48]. Another limitation of our study may be a slight lack of statistical power to investigate potential effects of confounders
. However there was clearly sufficient power in our study to test associations between the maternal experience of nausea and vomiting in pregnancy and the risk of delivering LBW babies. The final main limitation may be exclusion of women who took anti-emetics during pregnancy. This is because the threshold for nausea and vomiting where women sought and were then prescribed anti-emetics would have varied from participant to participant, and therefore the exclusion could have been rather self-selecting. However the advantage of doing this is the lack of potential confounding in our study of these drugs affecting pathways involved in regulating fetal growth, a strength of the study.
Acknowledgements
We thank all the families that took part in the Cambridge Baby Growth Study, and we acknowledge the crucial role played by the research nurses especially Suzanne Smith, Ann-Marie Wardell and Karen Forbes, staff at the Addenbrooke’s Wellcome Trust Clinical Research Facility, and midwives at the Rosie Maternity Hospital.