* Winner of the Poster competition
A common event experienced by most women during pregnancy is the nausea and vomiting of pregnancy known as morning sickness or pregnancy sickness.1-5 Women with morning sickness are less likely to have a spontaneous abortion.1-3,5 Studies on spontaneously aborted fetuses show that many have congenital anomalies - frequently congenital heart defects.6
Congenital heart defects (CHD) are the most common serious birth defects affecting five to ten children per 1,000 live births, and factors associated with their occurrence could help us understand their pathogenesis or identify prevention potentials.7 Intrigued by the above-mentioned facts and by one of the findings in the Atlanta Birth Defects Case-Control Study (decreased risk for CHD in children of mothers who used antinausea medications [ANM] during pregnancy), we further explored this finding.8
The purpose of this study was to determine the association between maternal nausea during pregnancy (NP) and the child's risk for CHD, and to specifically address these questions: (1) Does having NP modify a woman's risk of having an infant with CHD? (2) If so, is the risk related to the severity of NP or to the use of antinausea medication? (3) Is the modified risk specific for particular CHDs?
We used data from the population-based Atlanta Birth Defects Case-Control Study, conducted 1982-1983.8 The study design and methods have been described in detail previously.9,10 Case infants (n = 998) had nonsyndromic CHDs and control infants (n = 3,029) had no congenital defects. Infants were born in the period of January 1, 1968 to December 31, 1980. Information on pregnancy history, maternal exposures, and other variables was obtained through a telephone interview. The questions on NP were:
(1) Did you have morning sickness? (2) In which month of pregnancy did you first have the nausea? (3) How many months did you have nausea? (4) How often did you have nausea? (5) Did you take any medicine for the nausea, for instance, Alka-Seltzer, Pepto-Bismol, or Bendectin? (6) What kind of medicine did you take? (7) In which week of pregnancy did you start taking it? (8) For how many weeks did you take it?
Mothers were divided into two main groups, based on reported presence or absence of NP. By dichotomizing three original variables describing NP - onset (within the first two months, the vulnerable period for heart development, vs. later), frequency (every day vs. less often), and duration (two months or more vs. less) - we constructed eight levels of NP "severity."
Level 1, the most severe NP, was early, frequent, and long-lasting, and Level 8, the least severe, was late, infrequent, and short-lasting NP (Table 1).
To determine whether maternal use of antinausea medication for NP altered a child's risk for CHD, we stratified the analyses by use or nonuse of ANM within the first eight weeks of pregnancy and by type of medication - Bendectin (doxylamine, dicyclomine [dropped from the formulation in 1976], vitamin B6 [pyridoxine]) or other. For the analyses by individual defects, we used a classification of CHDs similar to the one used in the Baltimore-Washington Infant Study (BWIS): hierarchical allocation of the principal lesion.11 For example, the group of ventricular septal defects (VSDs) does not include cases with complex CHDs.
The relative risk of having a child with a CHD was estimated by calculating the odds ratio (OR) and 95 percent confidence interval (CI), using the Mantel-Haenszel or Fisher exact tests. We used logistic regression to adjust for maternal age, race, education, parity, previous miscarriages, smoking, alcohol use, vitamin use during pregnancy, and child's year of birth. Because adjusted ORs did not differ significantly from the crude ORs, we reported the latter. A chi square test was used to evaluate linear trends in proportions.
Morning sickness was reported by 71.2% of case mothers and by 69.8% of control mothers, and its frequency decreased with age (test for trend, p < 0.05), (Figure 1). Among women less than 20 years old, NP was significantly less common in case mothers (p < 0.05). The relative risk for a CHD in the offspring was significantly lower in mothers who reported early, frequent, and long-lasting NP (Level 1): OR = 0.81, CI 0.67-0.99. This effect tended to disappear with less severe NP (Table 1) and the trend was statistically significant (chi square = 8.24, p = 0.004). Further analysis was limited only to early NP (Levels 1 to 4, combined) and mainly to Level 1 NP. Both Level 1 NP and early NP were associated with lower risks for CHDs among children of mothers who reported use of ANM (Table 2, Figure 2). For Bendectin, the ANM most often used, the risk was significantly lower for early NP and for Level 2, and close to significant for Level 1 NP.
To get an idea of the "net" effect of the ANM, we used as a referent a group of women who had NP but did not use ANM. For early NP both the group of "all medications" users and the subgroup of "Bendectin" users were at a lower risk for CHD in the offspring: OR = 0.80, CI 0.64-1.00 and OR = 0.70, CI 0.50 - 0.94, respectively (OR not presented in Table 2).
For individual heart defects and Level 1 NP the relative risk was decreased for a number of defects, including some complex CHDs, but significantly lower only for atrial septal defects (Table 3).
Our findings on NP and/or ANM are consistent with some other studies. NP was associated with a slightly decreased risk for CHDs in the BWIS as well, OR = 0.90, CI 0.82-1.00.12 In the meta-analytical study on Bendectin, the pooled relative risk for oral clefts and for CHDs was lower for children of mothers who used Bendectin.13 Like others, we found that the frequency of NP decreased with age.14 The finding that, among women less than 20 years old, NP was less common in case mothers, has not been reported previously.
In interpreting the study results and discussing the possible reasons for the associations found in this study, we should consider its potential strengths (a population-based study with a large number of case and control infants) and limitations (difficulties in quantifying NP severity, in separating effects of NP from effects of ANM, a potential recall bias - long interval, in some instances up to 12-13 years between child's birth and the mother's interview). A recall bias, however, was unlikely to contribute to the association in this study: if case mothers recalled nausea and medication use better than control mothers, then a recall bias would have acted against this association, not in its favour.
What is the biological plausibility for the associations found in our study? Are there common factors that contribute to both NP and normal heart development? Could the use of ANM, particularly Bendectin, have beneficial effects on heart development? Although we must be cautious in interpreting our findings, we would like to discuss several possibilities.
NP has been linked with high human chorionic gonadotropin (hCG) blood levels.15-17 Other peptides of early pregnancy (placental peptide 14, inhibin) parallel in time hCG secretion, while a growth factor, activin, is known to stimulate it.18 However, the relationship between these peptides and NP has not been studied yet. Animal studies on activin suggest a role in heart development.19,20 It is plausible that a certain level of these hormones/peptides may contribute to normal heart development.
The lower risk for CHDs associated with use of Bendectin may be due to one of its components - most probably to vitamin B6 (10 mg/tablet). Women who took the recommended dose (2 to 3 tablets) received about ten times the recommended daily allowance of pyridoxine.21 In support of this hypothesis are the following facts: (1) Pregnant women have higher needs for pyridoxine, whose "antagonists" - oestrogens - have higher levels in pregnancy.22,23 (2) Vitamin B6 prevents some chemically-induced birth defects (cleft palate) in animals.24 (3) Use of multivitamins in early pregnancy has been associated with lower risk for some CHDs, and it is possible that more than one of the multivitamin components (pyridoxine included) contributes to this.25-27
Thus we hypothesize that: NP during early gestation is a marker of normal production of pregnancy hormones and growth factors that may contribute to normal heart development. There may be common genetic or environmental factors that cause insufficient production of hormones/growth factors, lack of NP, and abnormal heart development. A possible reason for the lower risk for CHD among children of mothers who took ANM may be that they had more severe NP, which could not be measured by our scale. Alternatively, maternal use of high doses of vitamin B6 (contained in Bendectin) may have had a protective effect on the embryonic heart development. More details on this discussion can be found elsewhere.28
If the associations found in this study are confirmed by others, future investigation into the underlying mechanisms could help elucidate some pathogenetic aspects of CHDs and reveal opportunities for their primary prevention.