The association between abortion and nausea and vomiting of pregnancy

Paolo Mazzotta, MSc, Laura A. Magee, MD, FRCPC, MSc, Gideon Koren, MD, FACCT, FRCPC
The Motherisk Program, Division of Clinical Pharmacology and Toxicology, Department of Pediatrics and Research Institute, the Hospital for Sick Children, Toronto, Department of Pediatrics, Pharmacology, Pharmacy and Medicine, the University of Toronto, Toronto, Ontario, Canada

Supported by the Motherisk Research Fund and a grant from Duchesnay Inc., Laval, Quebec


For planning and expectant mothers, the Motherisk Program in Toronto provides counselling on the safety of chemical or physical exposures during pregnancy and lactation. One of the most common queries to Motherisk concerns the safety of antiemetic therapy. This is not surprising, given that up to 80% of women experience nausea and vomiting of pregnancy (NVP), and even if symptoms are not severe, they may interfere with family, social and occupational functioning.1,2

In the course of providing counselling services, Motherisk has become aware of women who reported that they terminated otherwise wanted pregnancies because of NVP. Although the published literature documents increased maternal morbidity (i.e., dehydration, weight loss, electrolyte imbalance, vitamin deficiency, thyroid and/or liver dysfunction), foetal morbidity (i.e., low birth weight), and a decreased incidence of spontaneous abortion associated with severe NVP,3 only three case reports (only one of which was in the medical literature) could be located that described an association between severe NVP (i.e., hyperemesis gravidarum [HG]) and therapeutic abortion.4,5

Study 1

In our first study, the objectives were two-fold: i) to examine whether or not NVP is associated with therapeutic abortion, by obtaining population-based statistics (unavailable in Ontario) from the Office of Population Censuses and Statistics (OPCS), London, England, and if such an association were demonstrated, ii) to generate potential hypotheses for such an association, based on the Motherisk experience.


Population-based statistics are most appropriate for determining the natural history of a disease, and such statistics were not available for Ontario. Therefore, relevant data from England and Wales were obtained from the OPCS in London, England. For the years 1979-1992, for which relevant data were available, the number of legal abortions for both "excessive vomiting in pregnancy" (ICD code 643) and all "maternal medical" indications was sought from the OPCS (Personal communication, Abortion Statistics, OPCS, London, England, United Kingdom, November, 1995). The incidence was expressed as the number of legal abortions for NVP per i) 100,000 legal abortions (total), and ii) 100,000 maternities (mothers delivered of live or stillborn infants at > 24 weeks' gestation), and iii) percent legal abortions for maternal indications.


Pregnancy termination for ICD code 643 in England and Wales
Table 1 shows that in England and Wales between 1979 and 1992, there were 25-59 cases per year of legal abortion for ICD code 643. This corresponds to a median annual incidence of termination for "excessive vomiting of pregnancy" of 25.7 (range: 15.6-46.6) per 100,000 legal abortions, 6.0 (3.7-9.5) per 100,000 maternities, and 97% (60-100%) of all terminations for maternal indications. Since 1979, the rate of termination for NVP has fallen relative to all legal abortions (p = 0.0001, R2 = 0.71), however, there was no significant change from 1984 onwards (p = 0.53, R2 = 0.06). No data were available with which to interpret this finding further.

Table 1 - Legal abortions for "excessive vomiting in pregnancy" (ICD code 643) to residents of England and Wales over the study period 1979-1992


Study 2

In a Motherisk study, an invitation to women to report their range of experience with morning sickness was advertised in newspapers, women's and family magazines, and the electronic media in Canada and the United States.


A structured questionnaire was used to gather information on various clinical aspects of the nausea and vomiting, on therapies used, and on the effect on the family and quality of life. Women who had had therapeutic abortions were matched by gravity to women who considered termination but eventually continued their pregnancy as well as to controls who were the next callers after the index cases of therapeutic abortion. The three groups were compared in several characteristics by non-parametric methods.


Voluntary reporting of women's experience with nausea and vomiting of pregnancy.
Over two months, 1,100 women were interviewed regarding their experience of morning sickness. Of those, 17 reported having terminated otherwise wanted pregnancies due to severe NVP. In none of these cases were there other causes for termination of pregnancy, e.g., socioeconomic or medical, identified, except for one case where hyperemesis worsened a medical problem.6 In no case was pregnancy terminated due to life-threatening complications of nausea and vomiting. An additional 42 women reported having considered termination of pregnancy as a way to eliminate their severe symptoms.

Women with HG who eventually terminated pregnancy had a mean of 9.3 ± 6.5 episodes of vomiting daily, lost on average 5.9 ± 3.8 kg of body weight, and lost 20 ± 21 days of work. Women with HG who considered termination but eventually continued with their pregnancies lost much more weight than either those who terminated pregnancy or control women. Their children were of significantly smaller birth weight than those of control subjects (2,942 ± 592 g vs. 3,737 ± 337 g, p = 0.0025) (Table 2).

Table 2 - Characteristics of women with severe morning sickness and outcome of pregnancies

Data are presented as means ± standard deviation (range)

* Women who intended to but did not terminate lost significantly more weight than women in the other two groups (p < 0.05)
^ Babies of women with morning sickness were significantly smaller than those of women in the control group (p = 0.0025)

Only four (23%) of the women who terminated pregnancy reported day-to-day support from their spouses or partners in carrying out home and family chores during the HG. This is compared to 60% of those who eventually did not terminate (p < 0.05). Of the ten who terminated and were offered antiemetics, dimenhydrinate (Gravol) was the most widely used (n = 9; 90%); despite being used as recommended on the package, it partially relieved symptoms in only two cases, as did prochlorperazine (Stemetil) in one case. Doxylamine succinate with vitamin B6 was used in two cases; in only one was a potentially effective dose prescribed. All women with severe manifestations of HG reported unique and quite homogeneous psychological symptoms (Table 3), including depression, feelings of isolation and hopelessness, and fear. Women said they often felt that the nausea and vomiting were not taken seriously enough by spouses and physicians alike.

Table 3 - Characteristic Behavioural Symptoms of Women Who Terminated or Considered Terminating Pregnancy



Population-based statistics from the OCPS in the U.K. demonstrated that NVP is associated with elective pregnancy termination. The incidence of legal abortion reported to be due to NVP was low, in the order of six per 100,000 maternities, but NVP was by far the most common indication for termination for maternal medical reasons.

What could explain the observed association of NVP with elective termination? Obviously, severe symptomatology is likely to be an underlying feature, but there are other potential contributing factors raised by our second study. It shows a grim picture of the medical management of this condition. Many women were not offered antiemetics because medical staff believed common antiemetics were teratogenic, or women would not take them due to their own fears of such effect. Additionally, our study suggests that women who terminate pregnancies differ from those who eventually continue pregnancies to term by having less support from their husbands or partners.

Motherisk is still hearing from women who report that they have been told that their symptoms are "psychological", and that they are "creating their own problems." This would suggest that there still exists the perception that NVP represents a conscious rejection of the pregnancy.7 It is also not unusual for Motherisk to be told that patients' doctors are unwilling to prescribe antiemetic medication, as well as to hear from patients that they are unwilling to take any medication, no matter how reassuring the reproductive risk profile may be. These comments would suggest that there is still misperception, by patients and health professionals, about the risks of antiemetic medication. Firstly, thalidomide would have been identified as a major teratogen at the pre-marketing stage of drug development if evaluated by today's testing standards. Secondly, Diclectin (currently marketed only in Canada) has been proven not to be a teratogen, and other antiemetics in common use, e.g., dimenhydrinate, are not known to be teratogenic.

There are obvious limitations to our data. The population-based statistics on the association between NVP and legal termination of pregnancy were not linked to medical records, therefore, neither the diagnosis of NVP nor other potential reasons for termination could be confirmed. However, it must be acknowledged that the vast majority of women did not list medical indications as their reasons for choosing abortion, and that NVP, which has long been regarded to have a strong psychological component,8 cannot be assumed to be a more socially acceptable indication for abortion than personal choice. Where the OPCS data are from England and Wales, it is noted that legal abortion is also widely available in Canada; the OPCS data should therefore be relevant. Finally, the Motherisk callers who reported that they terminated pregnancies due to NVP described their experience retrospectively, and their reports were not confirmed by reviewing medical records. However, their experiences were presented as a hypothesis-generating exercise and should be regarded in this light.

In conclusion, NVP is associated, albeit uncommonly, with legal abortion. Although termination is obviously an extreme example, the association between legal abortion and NVP should serve to highlight the morbidity which can result from this pregnancy-induced disorder.


The authors wish to thank the Office of Population Censuses and Statistics, London, England, for providing valuable and otherwise unavailable information for this manuscript.


  1. Gadsby R, Barnie-Adshead AM, Jagger C. A prospective study of nausea and vomiting during pregnancy. Br J Gen Pract 1993;43:245-8.
  2. O'Brien B, Naber S. Nausea and vomiting during pregnancy: effects on the quality of women's lives. Birth 1992;19:138-43.
  3. Hod M, Orvieti R, Kaplan B, Friedman S, Ovadia J. Hyperemesis gravidarum. A review. J Reprod Med 1994;39:605-12.
  4. Chatwani A, Schwartz R. A severe case of hyperemesis gravidarum. Am J Obstet Gynecol 1982;143:964-5.
  5. Pouliot S. "J'ai dû me faire avorter 4 fois". Dernière Heure. 1995, Octobre 21.
  6. Mazzotta P, Magee L, Koren G. Therapeutic abortions due to severe morning sickness. Unacceptable combination. Can Fam Physician 1997;43:1055-7.
  7. Farkas G, Farkas G Jr. The psychologic etiology of the hyperemesis gravidarum. In: Morris N, Karger BS, eds. Psychosomatic Medicine on Obstetrics and Gynecology. Basel: S. Karger,1972;175-7.
  8. de la Ronde SK. Nausea and vomiting in pregnancy. J SOGC 1994;16:2035-41.


Mazzotta P, Magee L, Koren G. Therapeutic abortions due to severe morning sickness: an unacceptable combination. Motherisk Newsletter 1996;6:1-3.

Back to home