Pre-eclampsia is associated with a low estrogen level 1 and hyperemesis with an increased estradiol level. 2 Both pre-eclampsia 1 and hyperemesis 2 are associated with an increased human chorionic gonadotropin (hCG) level. Findings on hormonal levels in hyperemesis, however, have been inconsistent. 2,3
Should hormonal imbalances be present at conception, abnormal sex ratios at birth could be expected. 4 Female gender and twinning are associated with higher hCG levels, at least late in pregnancy, 5,6 potentially triggering pre-eclampsia and hyperemesis. Both conditions are more frequent in multiple pregnancies 7–9 and are associated with altered sex ratios at birth, although in opposite directions. 4,7,9–12 In this study, we provide further evidence on these associations by examining the joint effect of twinning and fetal gender in relation to pre-eclampsia and hyperemesis.
From the Danish National Board of Health we obtained birth records and discharge diagnoses for all births for which women were hospitalized during pregnancy with hyperemesis (9401 newborns) and/or eclampsia/pre-eclampsia (30303 newborns) from 1980 to 1996 in Denmark. This information was linked to the Danish National Birth Registry and Statistic Denmark’s Fertility Database 13 to obtain information on all the births of the identified mothers during the study period and on other covariates. As a source for the reference population we used a 10% random sample of 60,810 mothers (103,394 infants) from the general population.
The Fertility Database was updated to 1994 and we used data from this source to assess maternal citizenship and parity. We thus mainly limited this study to births occurring between 1980 and 1994.
We only included Danish citizens, as information was more complete for women who had been in the country for some time. While births from foreign mothers comprised 8% of the total random sample, they represented 17% and 5% of the babies whose mother was hospitalized with hyperemesis and pre-eclampsia, respectively. We further excluded multiple deliveries except twins, and all mothers who had both illnesses under study between 1980 and 1996, regardless of whether they were registered to the same or to a different birth. Since the Birth Registry recorded all live births but only stillbirths occurring after the 27th gestational week, we included only still- and live births with a gestation of at least 28 weeks. Gestational age was missing in 2.5% of the pregnancies. We kept in the data analysis births of infants with unknown gestational age and reanalyzed the data after excluding them. We excluded records with the most recent diagnosis coded as doubtful (patient was ‘under observation’).
We thus analyzed 24,065 births with pre-eclampsia and 6,084 with hyperemesis, using 76,804 births from non-affected mothers as a reference.
The total incidence of hospitalization between 1980 and 1994 among Danish pregnant women (provided that the pregnancy lasted at least 28 weeks) was 3.0% for pre-eclampsia and 0.8% for hyperemesis.
We examined sex ratio following a history of hyperemesis or pre-eclampsia by using pairs of first and second consecutive singleton births. Since we were able to ascertain maternal parity and citizenship from the older sibling, in this analysis the first child in the pair was born in the period 1980–1994, while the second child may have been born up to 1996.
The male-to-female sex ratio was 1.04 (95% CI = 1.02–1.05) in the reference population, 0.87 (95% CI = 0.82–0.91) and 1.10 (95% CI = 1.07–1.12) in births among mothers with hyperemesis and with pre-eclampsia, respectively. Twins were more frequent in pregnancies with either condition (Table 1).
In Table 2 we estimated the odds ratios of hyperemesis and pre-eclampsia according to gender and whether the mother delivered a singleton or twins. The analysis excluding women with missing gestational age resulted in virtually unchanged effect estimates.
Non-cohabiting women and women with a lower education had a higher risk of hospitalization for both hyperemesis and pre-eclampsia, but adjustment for these factors had only a minor effect on the estimates (data not shown). The proportion of female babies in births among women with hyperemesis was 54.2% with the first birth, (3196) 53.7% with the second, (2344) and 51.0% with the third. (596) When comparing twins with at least one female in the pair with male twins we saw a slight increase in the risk of maternal hyperemesis (OR = 1.38, 95% CI = 0.93–2.05).
While women with hyperemesis had more twins and female babies in their affected pregnancies, in their non-affected ones they had fewer twins (OR = 0.66, 95% CI = 0.47–0.92) and marginally less females (OR = 0.95, 95% CI = 0.89–1.01), compared with the pregnancies of reference births.
We observed the highest probability of a female baby (OR = 1.31, 95% CI = 1.23–1.41) when the first hospitalization for hyperemesis occurred during the first trimester of gestation (56% of all cases). Births with the first hospitalization occurring later had the same sex ratio as in the births in the reference population, although the odds ratio of female gender was 1.15 (95% CI = 0.96–1.37) when hyperemesis was first recorded at the time of delivery (in 8.5% of the cases).
Pre-eclampsia was weakly associated with male gender, while the association with twinning was strong (Table 2). The non-affected pregnancies of women with pre-eclampsia were characterized by a lower proportion of twins (OR: 0.83, 95% CI = 0.70–0.98), as seen in women with hyperemesis, but the proportion of boys was the same in the affected births as in the non-affected ones (52.3%).
When examining offspring sibling pairs to study the effect of history of hyperemesis on the gender of the second child, we found 24,616 women with neither hyperemesis nor pre-eclampsia with the first two births and 3,463 pairs whose mother had had hyperemesis with one or both births. We used the same reference population and 12,935 pairs of siblings whose mother had had pre-eclampsia in one or both of the first two births. Table 3 shows that hyperemesis in the second pregnancy but not in the first influenced the gender of the second baby. We saw the lowest sex ratio (0.80; 95% CI = 0.72–0.88) when hyperemesis had been diagnosed only with the second birth.
Pre-eclampsia with the first birth was, on the other hand, weakly associated with a higher proportion of males in second birth.
Hyperemesis was associated with twinning and with being pregnant with a female fetus, the latter only if the first hospitalization occurred during the first trimester of pregnancy, as previously reported. 10,12 The presence of at least a female in the twin pair was also associated with hyperemesis, which is consistent with higher hCG levels. 6 Births with eclampsia/pre-eclampsia were characterized by a slightly increased sex ratio, as well as a higher incidence of twinning independent of fetal gender.
Mothers with pre-eclampsia appeared to have slightly more male babies in their pregnancies where pre-eclampsia was not recorded. This result could be due to sub-clinical disease in these other pregnancies or it may indicate that a common factor is associated with both pre-eclampsia and fetal gender.
The threshold for hospitalization is most likely independent of fetal gender, which need not be the case for multiple pregnancies. The later onset of pre-eclampsia may explain the association between twinning and pre-eclampsia, although our estimates were similar to those reported by others. 8,9
We found that gender of the newborn and twinning were associated with two pregnancy-related diseases associated with intrauterine hormonal level. We cannot assess the direction of causation, although the persistent slightly skewed sex ratio towards males in pre-eclamptic women may point towards a peri-conceptional mechanism.
1. Innes KE, Byers TE. Preeclampsia and breast cancer risk. Epidemiology 1999; 10: 722–732.
2. Depue R, Bernstein L, Ross RK, Judd HL, Henderson BE. Hyperemesis gravidarum in relation to estradiol levels, pregnancy outcome, and other maternal factors: a seroepidemiologic study. Am J Obstet Gynecol 1987; 156: 1137–1141.
3. Jarnfelt-Samsioe A. Nausea and vomiting in pregnancy: a review. Obstet Gynecol Surv 1987; 41: 422–427.
4. James WH. Sex ratios of offspring and the causes of placental pathology. Hum Reprod 1995; 10: 1403–1406.
5. Steier JA, Myking OL, Bergsjo PB. Correlation between fetal sex and human chorionic gonadotropin in peripheral maternal blood and amniotic fluid in second and third trimester normal pregnancies. Acta Obstet Gynecol Scand 1999; 78: 367–371.
6. Steier JA, Myking OL, Ulstein M. Human chorionic gonadotropin in cord blood and peripheral maternal blood in singleton and twin pregnancies at delivery. Acta Obstet Gynecol Scand 1989; 68: 689–692.
7. Källen B. Hyperemesis during pregnancy and delivery outcome: a registry study. Eur J Obstet Gynecol Reprod Biol 1987; 26: 291–302.
8. Conrood DV, Hickok DE, Zhu K, Easterling TR, Daling JR. Risk factors for pre-eclampsia in twin pregnancies: a population-based cohort study. Obstet Gynecol 1995; 85: 645–650.
9. Conde-Agudelo A, Belizán JM. Risk factors for pre-eclampsia in a large cohort of Latin American and Caribbean women. Br J Obstet Gynaecol 2000; 107: 75–83.
10. Askling J, Erlandsson G, Kaijser M, Akre O, Ekbom A. Sickness in pregnancy and sex of child. Lancet 1999; 354: 2053.
11. Zhang J, Cai W. Severe vomiting during pregnancy: antenatal correlates and foetal outcomes. Epidemiology 1991; 2: 454–457.
12. Del Mar Melero-Montes M, Jick H. Hyperemesis gravidarum and the sex of the offspring. Epidemiology 2001; 12: 123–124.
13. Knudsen LB. The Danish Fertility Database. Dan Med Bull 1998; 45: 221–225.