In early pregnancy, nausea with or without vomiting is very common. 1 From the clinical point of view, nausea with or without vomiting should be distinguished from hyperemesis gravidarum that is characterized by serious and persistent vomiting that interferes with fluid intake and nutrition. Several studies have investigated the patterns of nausea with or without vomiting, and these have recently been reviewed. 2 The condition has a cumulative incidence between 50% and 80% among pregnant women. 3,4 Nausea with or without vomiting may occur as early as 2 weeks after fertilization 2 and in 50% of instances recedes by gestational weeks 14 to 15, although a small proportion of women may still experience nausea with or without vomiting until the 22nd week of gestation. Nausea with or without vomiting is generally considered a phenomenon that accompanies normal pregnancies, but the underlying physiology is poorly understood. 5 Several authors have suggested that high levels of human chorionic gonadotropin (hCG) may contribute to nausea with or without vomiting, 2 whereas others have considered estrogens, 1 maternal prostaglandin E2, 6 thyroid hormones, 7 other hormones, 1,2 and genetic, cultural, and lifestyle factors, including smoking and caffeine intake. 1,2,8,9
We have prospectively evaluated the role of four maternal hormone levels–estradiol (E2), estriol (E3), progesterone, and prolactin (PRL)–as well as sex hormone–binding globulin at the 16th and 27th weeks of pregnancy, in relation to the occurrence of nausea with or without vomiting by the 27th week of pregnancy, in a group of pregnant white women in Boston.
MATERIALS AND METHODS
Between March 1994 and October 1995, a total of 402 eligible women were identified at the Beth Israel Hospital in Boston in the context of an international study exploring perinatal events and conditions that could explain the large difference in breast cancer incidence between white and Asian women. 10 According to the protocol of the international study, the American women studied in the present investigation, in addition to being white, had to be younger than 40 years and have a parity of no more than two. They were not eligible if they had taken any hormonal medication during the index pregnancy, if they had a prior diagnosis of diabetes mellitus or thyroid disease, or if the fetus had a known major anomaly. A health professional met all pregnant women coming for their first routine prenatal visit to the collaborating maternity clinic, ascertained their eligibility to participate, explained to them the objectives of the study, and obtained informed consent. 10 The procedures followed were in accordance with the ethical standards for human experimentation established by the institutional review boards of the Harvard School of Public Health and Beth Israel Hospital. Of the 402 eligible women, 77 refused to participate, nine were subsequently excluded because of a spontaneous or induced abortion, two were excluded because of twin birth, ten were lost to follow-up after the initial meeting, and 42 had missing values in one or more of the parameters used in the analysis. Thus, 262 pregnant women were eventually included in the analysis.
Details on the questionnaire administered, medical record reviewing, blood collection, and sample storage have been given in an earlier publication. 10 The occurrence of nausea with or without vomiting was ascertained through in-person interviews conducted around the 27th week of pregnancy, when the period at risk for nausea with or without vomiting would have been completed and the cumulative incidence of the phenomenon could be ascertained. Women were asked if they had experienced either nausea without vomiting or nausea with vomiting at any time during their pregnancy. We defined nausea with or without vomiting as the presence of either.
Hormone levels were measured in maternal blood around the 16th and 27th completed weeks of gestation. Hormone measurements were done for all women in the study, so that these measurements were unrelated to possible nausea with or without vomiting reporting and the possibility of differential misclassification bias was minimized. In the present analysis, E2, E3, sex hormone–binding globulin, progesterone, and PRL were evaluated.
Estradiol-17β in 25 μL of serum, diluted 1:20, was measured with a time-resolved competitive solid-phase fluoroimmunoassay (AutoDELFIA E2 kit; Wallac Oy, Turku, Finland), based on competition between europium-labeled E2 and sample E2 for polyclonal anti-E2 antibodies (derived from rabbits). A second antibody, directed against rabbit immunoglobulin (Ig) G, is coated onto the walls of microtiter plates and binds the IgG E2 complex. Cross-reactivities for estrone and E3 at the 50% inhibition level were 0.75% and 0.40%, respectively.
Unconjugated E3 in 50 μL of serum was measured with a similar time-resolved competitive solid-phase fluoroimmunoassay method (AutoDELFIA unconjugated E3 kit). Cross-reactivities at 50% inhibition were 37% for estriol-3-sulphate, 37% for estriol-3-glucuronide, and less than 0.1% for E2.
Sex hormone–binding globulin in 25 μL of serum, diluted 1:100, was measured with time-resolved noncompetitive solid-phase sandwich fluoroimmunoassay (AutoDELFIA sex hormone–binding globulin kit). A recombinant human sex hormone–binding globulin preparation was used as reference.
Progesterone in 25 μL of serum, diluted 1:8, was measured with a time-resolved competitive solid-phase fluoroimmunoassay (AutoDELFIA progesterone kit).
Prolactin in 25 μL of serum was measured with time-resolved noncompetitive solid-phase sandwich fluoroimmunoassay, in which two monoclonal antibodies (derived from mice) are directed against two separate antigenic determinants of the PRL molecule (AutoDELFIA PRL kit).
In the statistical analysis, women who had reported nausea with or without vomiting were compared with pregnant women who had reported neither, with respect to age, parity, body mass index (BMI), smoking and coffee drinking during pregnancy, and gender of the offspring. The analysis was done both by simple crosstabulation and by modeling the data through multiple logistic regression. Logistic regression allowed the estimation of odds ratios (ORs) for nausea with or without vomiting for a specified increment (set at one standard deviation [SD]) of each of the studied hormones at the corresponding sampling, with simultaneous adjustment for the exact gestational age at blood sampling and several maternal factors, and mutually among the studied hormones. The statistical significance level was set at a P value (two tailed) of .05. All analyses were conducted using SAS Software 8.0 (SAS Institute Inc., Cary, NC).
Two hundred nine of the 262 women (79.8%) reported that they had experienced nausea with or without vomiting by the 27th week of their pregnancies. Of these 209, 79 had experienced vomiting as well as nausea, whereas the remaining 130 had only experienced nausea. Table 1 shows the distribution of women in these three categories (no nausea, nausea without vomiting, vomiting), by maternal characteristics. In the same table, multiple logistic regression–derived ORs for possible determinants of reported nausea with or without vomiting other than hormone measurements are given, together with the corresponding 95% confidence intervals (CIs). None of the studied factors was significantly related to nausea with or without vomiting, although the cumulative incidence of the condition might be higher among biparae than among primiparae (OR 1.75; 95% CI 0.88, 3.47).
In Table 2, mean and median values of the serum levels of the studied hormones are presented for the 16th and 27th gestational week samplings, both for women with nausea with or without vomiting and for women with neither nausea nor vomiting symptoms. This table serves descriptive purposes, because mutual confounding is not accounted for.
Table 3 shows logistic regression–derived ORs (and 95% CIs) for nausea with or without vomiting for increments of each of the studied hormones equal to one SD at the corresponding sampling (16th or 27th gestational week). For each hormone and sampling, three ORs are given: a crude one, one after adjustment for the maternal variables indicated in Table 1, and a third adjusting also mutually among the studied hormones. There is no evidence that E3, sex hormone–binding globulin, or progesterone, at either the 16th or the 27th week of gestation, is significantly or substantially related to reported nausea with or without vomiting. Moreover, there is no evidence for substantial confounding among the measured hormones affecting the results concerning these three compounds.
With respect to PRL, there is again no evidence of powerful confounding, as indicated by the relative stability of the OR estimates before and after adjustment for the maternal and hormonal variables. It is also evident that there is an inverse, substantial, and statistically significant association of PRL with reported nausea with or without vomiting at both the first and the second samplings (P < .01 in both instances). The fully adjusted OR for the first sampling is 0.58 (95% CI 0.41, 0.84), and for the second sampling it is 0.59 (95% CI 0.42, 0.82). Thus, per SD increase in PRL levels at either time point, women were 40% less likely to experience nausea with or without vomiting. When models were separately fitted for vomiting and for nausea without vomiting, in comparison with the absence of both symptoms, the regression coefficients were similar.
Finally, with respect to E2, the crude estimates of OR in both samplings were close to the null value, but became substantially elevated after adjustment for maternal variables and the other measured hormones. Adjusting for potential confounders, a one-SD increase in E2 increased the odds of nausea with or without vomiting 1.61 times (95% CI 0.98, 2.66; P ∼ .06) at the 16th gestational week and 1.48 times (95% CI 0.97, 2.27; P ∼ .07) at the 27th week of gestation (Table 3).
Using prospectively collected blood samples we found strong evidence that maternal PRL blood levels are significantly inversely associated with reported nausea with or without vomiting and weaker evidence that maternal E2 blood levels may be positively associated with this condition. In contrast, we found no evidence that E3, progesterone, or sex hormone–binding globulin plays a major role in the occurrence of nausea with or without vomiting. We also found no evidence that maternal age, prepregnancy BMI, smoking or coffee drinking in pregnancy, and gender of the offspring are important determinants of reported nausea with or without vomiting. With respect to parity, a positive association could be noted, but it was not statistically significant. The intensive investigation during the last 2 decades of perinatal events and conditions in relation to offspring health and disease 11 requires a better comprehension of the biology that underlies various pregnancy manifestations. Our investigation represents a contribution to that effect, as well as to the understanding of pregnancy physiology.
The strengths of the present investigation are its prospective nature, its implementation on a well-defined cohort, and control for exact gestational age at blood sampling. Misclassification with respect to exposure and outcome variables is unavoidable, but these errors are generally nondifferential and thus unlikely to account for the observed associations. If very minor nausea with or without vomiting were not reported, then the results would apply only to the nontrivial nausea with or without vomiting that is actually reported. There is no evidence, however, that very minor nausea with or without vomiting is discounted by women, who are generally very attentive to the various manifestations of their pregnancy, or that it has a different etiology. A weakness of this study, and most other studies on this topic, is that not all hormones and compounds of suspected etiological importance could be evaluated. Moreover, measurements at around the 27th week of pregnancy follow rather than precede nausea with or without vomiting, and their relevance hinges on the assumption that hormones during pregnancy tend to track, at a higher or lower level, throughout most of the duration of the pregnancy. This assumption is difficult to document because it requires serial measurements, but underlies the interpretability of most point measurements of hormones that are undertaken during a pregnancy.
Of the hormones that have not been measured in this study, hCG is the one that has received particular attention in earlier studies, mostly because hCG levels vary concomitantly with the appearance, peak, and recession of nausea with or without vomiting. 2 The empiric evidence for hCG as a causal agent for nausea with or without vomiting, however, has been assessed as limited. 1,2,6,7,12 A powerful argument against hCG playing a major role in nausea with or without vomiting is that nausea is not common in women with choriocarcinoma, which is characterized by high levels of hCG. 1,6
The inverse association of PRL with reported nausea with or without vomiting is strong and robust in our study and was evident in both samplings, making it unlikely that it was a chance phenomenon. Earlier studies, however, have been less conclusive. 13–15 Thus, the findings should be replicated with hormone measurements conducted at a time more proximal to the occurrence of nausea with or without vomiting before an attempt is made to identify plausible physiologic mechanisms. In contrast, the suggestive positive association between E2 and nausea with or without vomiting in our study is compatible with evidence that exogenous estrogens, in oral contraceptives, may occasionally cause nausea. 16,17 Our data do not support earlier reports of a possible inverse association between progesterone and nausea with or without vomiting, and we found no evidence in the earlier literature to contradict our findings that neither E3 nor sex hormone–binding globulin plays a role in nausea with or without vomiting. 1,18
With respect to maternal demographic and lifestyle variables, our findings agree with some earlier reports. 3,19 Our results do not indicate that tobacco smoking reduces the risk of nausea with or without vomiting, or that coffee intake increases the risk of this condition. 2,8,20 However, the findings of our study are not incompatible with these earlier reports because the respective CIs in our study are large enough to accommodate a modest inverse association for tobacco smoking and a weak positive association for coffee drinking.
The results of this prospective investigation point to lower levels of PRL and, perhaps, higher levels of E2 as contributing to or correlating with the occurrence of nausea with or without vomiting in pregnancy. We have found no evidence that E3, progesterone, or sex hormone–binding globulin is associated with this condition.
1. Jarnfelt-Samsioe A. Nausea and vomiting in pregnancy: A review. Obstet Gynecol Surv 1987;42:422–7.
2. Furneaux EC, Langley-Evans AJ, Langley-Evans SC. Nausea and vomiting of pregnancy: Endocrine basis and contribution to pregnancy outcome. Obstet Gynecol Surv 2001;56:775–82.
3. Jarnfelt-Samsioe A, Samsioe G, Velinder GM. Nausea and vomiting in pregnancy–a contribution to its epidemiology. Gynecol Obstet Invest 1983;16:221–9.
4. Huxley RR. Nausea and vomiting in early pregnancy: Its role in placental development. Obstet Gynecol 2000;95:779–82.
5. Dally A. Thalidomide: Was the tragedy preventable? Lancet 1998;351(9110):1197–9.
6. Gadsby R, Barnie-Adshead A, Grammatoppoulos D, Gadsby P. Nausea and vomiting in pregnancy: An association between symptoms and maternal prostaglandin E2. Gynecol Obstet Invest 2000;50:149–52.
7. Panesar NS, Li CY, Rogers MS. Are thyroid hormones or hCG responsible for hyperemesis gravidarum? A matched paired study in pregnant Chinese women. Obstet Gynecol Scand 2001;80:519–24.
8. Weigel RM, Weigel MM. Nausea and vomiting of early pregnancy and pregnancy outcome. A meta-analytical review. Br J Obstet Gynaecol 1989;96:1312–8.
9. Cook DG, Peacock JL, Feyerabend C, Carey IM, Jarvis MJ, Anderson HR, et al. Relation of caffeine intake and blood caffeine concentrations during pregnancy to fetal growth: Prospective population based study. BMJ 1996; 313:1358–62.
10. Lipworth L, Hsieh CC, Wide L, Ekbom A, Yu SZ, Yu GP, et al. Maternal pregnancy hormone levels in an area with a high incidence (Boston, USA) and in an area with a low incidence (Shanghai, China) of breast cancer. Br J Cancer 1999;79:7–12.
11. Signorello LB, Trichopoulos D. Perinatal determinants of adult cardiovascular disease and cancer. Scand J Soc Med 1998;26:161–5.
12. Soules MR, Hughes CL Jr, Garcia JA, Livengood CH, Prystowsky MR, Alexander E 3rd. Nausea and vomiting of pregnancy: Role of human chorionic gonadotropin and 17-hydroxyprogesterone. Obstet Gynecol 1980;55:696–700.
13. Ylikorkala O, Kauppila A, Haapalahti J. Follicle stimulating hormone, thyrotrophin, human growth hormone and prolactin in hyperemesis gravidarum. Br J Obstet Gynaecol 1976;83:529–33.
14. Ylikorkala O, Kivinen S, Reinila M. Maternal serum prolactin and its response to TRH in normal and complicated early pregnancy. Clin Endocrinol (Oxf) 1979;10:523–9.
15. Jarnfelt-Samsioe A, Bremme K, Eneroth P. Non-steroid hormones and tissue polypeptide antigen (TPA) in emetic and non-emetic pregnancy. Acta Obstet Gynecol Scand 1986;65:745–51.
16. Darney PD. OC practice guidelines: Minimizing side effects. Int J Fertil Womens Med 1997;Suppl 1:158–69.
17. Rosenberg MJ, Meyers A, Roy V. Efficacy, cycle control, and side effects of low- and lower-dose oral contraceptives:A randomized trial of 20 micrograms and 35 micrograms estrogen preparations. Contraception 1999;60:321–9.
18. Jarnfelt-Samsioe A, Bremme K, Eneroth P. Steroid hormones in emetic and non-emetic pregnancy. Eur J Obstet Gynecol Reprod Biol 1986;21:87–99.
19. Jarnfelt-Samsioe A, Eriksson B, Waldenstrom J, Samsioe G. Some new aspects on emesis gravidarum. Relations to clinical data, serum electrolytes, total protein and creatinine. Gynecol Obstet Invest 1985;19:174–86.
20. Zhou Q, O'Brien B, Relyea J. Severity of nausea and vomiting during pregnancy: What does it predict? Birth 1999;26:108–14.