Hyperemesis gravidarum during pregnancy and delivery outcome: a registry study

Bengt Källén, PhDEmbriol, MD
Head, Thornblad Institute, Head, National Registry of Congenital Malformations, Lund, Sweden

Data on approximately 100,000 annual births have been recorded in the Swedish Medical Birth Registry since 1973. Swedish medical record forms are standardized and information on births is collected from antenatal clinics, from hospital maternity wards and from the paediatric examination of newborns. From 1973 to 1981, data from the record forms were summarized by a secretary and the summary computerized in the Medical Birth Registry. Starting in 1982, the original records were copied directly into the Registry computer.

During the period 1973-1981, all diagnoses given to women during pregnancy were recorded as International Classification of Diseases (ICD) codes, including hyperemesis gravidarum; from 1982 onwards however, only severe conditions were noted. With such a database of information it was possible to study pregnancy outcome in 3,068 women in the years 1973 to 1981 who had received the clinical diagnosis of hyperemesis.1 The frequency of the diagnosis varied widely among delivery hospitals, from more than 1 per 100 to virtually none, with the average rate being 3.4 per 1,000. This may be related to the poorly defined diagnosis of hyperemesis. Likewise, there was variability in malformation registration, therefore these two variables were standardized for delivery unit in the analysis. As some other variables vary with the year of delivery, e.g., birth weight and maternal age, these were standardized for year of delivery.

In this hyperemesis study population there was a significant, although not strong effect of teenage pregnancies (relative risk 1.3) and at parity 1 (relative risk 1.07). As these variables co-vary, the effect of each one on the other was calculated after standardization.

The increased risk of emesis in cases of twin pregnancies is already known. It has been proposed by some authors that this effect may be explained by the higher hormone concentration, especially of human chorionic gonadotropin (hCG). In our population there was 2.2 times the risk, a markedly increased incidence, for twin pregnancies (irrespective of zygosity). There was also a higher rate of pregnancies with female infants, where only 47% of the infants were males, significantly fewer than the expected population rate of 51% (p < 0.001). Again, a hormonal aetiology has been ascribed, that is, an oestrogen effect causing emesis.

Among singleton infants there were more low birth weight infants, between 2,000 g and 2,500 g (0.01 < p < 0.05), and a slightly shorter gestational duration, notably 36-38 weeks (p < 0.001 for gestation less than 38 weeks: 381 found, 283.9 expected). These outcomes were both expected.

There was no effect on perinatal survival despite the lower birth weight and the higher twinning rate in the study population. There was a slight increase in congenital malformations (5.8% having a malformation diagnosis, where 5% are expected, p < 0.05). The congenital malformations for which the rate of occurrence was higher than expected were congenital hip dysplasia (60 versus 47, p = 0.05) and Down's syndrome (0.01 < p < 0.05). There were nine Down's infants born, against 3.2 expected (after stratification for year of birth, maternal age, and parity). However, in order to be able to determine the significance of this finding, a much larger population number would be required. Some commentary may be possible. It is known that roughly 70% of Down's foetuses undergo spontaneous abortion. Some authors have reported a lower rate of miscarriage in women with nausea and vomiting of pregnancy. It may be possible that under these circumstances the Down's foetuses are therefore being kept. The excess of congenital hip dysplasia cases as noted in this study had also been found in an earlier (1960s) prospective study on pregnancy outcome and antiemetic drug use.2

This early study in Malmö, Sweden, of antiemetic drug use in pregnancy grouped women into four categories: those having a miscarriage, those having an inducedabortion, those giving birth to a dead infant, and those giving birth to a live infant.2 The antiemetics taken were primarily promethazine, prochlorperazine, diphenhydramine and meclizine.

It was found that complaints of morning sickness were significantly higher in patients giving birth to a living, normal child (3,645/4,954, 74%) than those who aborted (212/423, 50%). Within the first group 591 (12%) used antiemetics, and in the abortion group 22 (5%) used emetics, i.e., a significantly greater antiemetic use in the patients producing a normal, healthy baby (p < 0.02). It was also found that the mean age of women having morning sickness was significantly lower than that of women not reporting it (p < 0.001), but there was no significant difference in the use, or lack of it, of antiemetics by those women.

Additional findings in this study included:

There is recent (1995-1997) data from the Registry on women who reported the use of antinauseant drugs during the first trimester. We identified 5,266 such women during the two-year period using either promethazine or meclizine. The majority of these women had morning sickness, but not hyperemesis. The women in this cohort were primarily young and of low parity. Similar to the earlier study of women with hyperemesis1, the findings showed a higher than normal frequency of infants of the female gender (sex ratio 0.92, versus 1.06 as the normal ratio) and of twins. There was no higher rate of low birth weight infants or of shorter gestation. No increased rate of occurrence of congenital malformations was noticed: there were seven infants with Down's syndrome, against 5.1 as the expected number. No increased incidence of congenital hip dysplasia was seen. The difference in this last finding may be due to more specific or improved diagnostic criteria over the intervening years.

Other points of interest are: the incidence of hyperemesis and the use of medication decreased with maternal age; the incidence of hyperemesis decreased with parity, however, use of antinauseants increased with parity; the risk ratio for multiple births was higher for hyperemesis than for medication use; there were opposite effects on birth weight and gestation length, with hyperemesis increasing risk for low birth weight and short gestation and drug use increasing gestation length and decreasing the risk of low birth weight.

In summary, even though a tentative association has been seen between NVP and two infant conditions (hip dysplasia and Down's syndrome), neither seems to be related to the use of antiemetics. Furthermore, no teratogenic property of antiemetic drugs has been demonstrated.


  1. Kallen B. Hyperemesis during pregnancy and delivery outcome: a registry study. Eur J Obstet Gynecol Reprod Biol 1987;26:291-302.
  2. Kullander S, Kallen B. A prospective study of drugs and pregnancy. II. Anti-emetic drugs. Acta Obstet Gynecol Scand 1976;55:105-11.

Back to home