Association between alcohol consumption during pregnancy and risks of congenital heart defects in offspring: meta-analysis of epidemiological observational studies

Congenital heart defects (CHDs) are the most common group of congenital malformations, affecting nearly 1 % of live births throughout the world [1]. CHDs account for approximately one-third of all congenital anomalies and are the leading noninfectious cause of death in the first year of life [2]. Although massive breakthroughs have been achieved in cardiovascular diagnostics and cardiothoracic surgery over the past century, leading to the increased survival of newborns with CHDs, the etiology of most congenital heart defects is still unknown.

A number of chromosomal anomalies, certain maternal illnesses, and prenatal exposures to specific therapeutic drugs are recognized risk factors. It is difficult to establish the role of a single factor because in many cases, the cause of a defect is believed to be multifactorial, including environmental teratogens with genetic and chromosomal conditions [3]. Maternal alcohol consumption is associated with a variety of harmful effects to the fetus, as demonstrated by the range of impairments present in fetal alcohol syndrome [4]. Various clinical signs have been described, which led to the classification of different degrees of embryopathy, ranging from patients with minor symptoms, the so-called “alcohol effects”, to the most severely affected individuals [5]. Up to one-third of affected individuals have congenital cardiac disease [6].

A study by Jones et al. [7] was one of the first to report the association between maternal alcohol consumption and CHDs. However, the evidence since then has been mixed, with some studies showing positive associations and others providing null results. CHDs include distinct subtypes (e.g., conotruncal defects, left ventricular outflow track defects, septal defects), and there is potential for etiologic heterogeneity. Binge drinking is defined as consuming ≥5 alcoholic drinks at one sitting [8].

Because the matter of whether there is an association between maternal alcohol consumption and/or binge drinking and congenital heart defects remains uncertain, we conducted this meta-analysis for more sufficient evidence on this issue.

A computerized literature search was conducted in MEDLINE, EMBASE, and the Cochrane Library from their inceptions to December 1, 2014, by two independent investigators (Wen and Yu). We searched relevant studies using the following medical subject heading terms and/or text words: “congenital heart defect”, “heart abnormalities”, “CHD”, and “heart malformation” in combination with “alcohol”, “drinking”, “maternal alcohol consumption”, “maternal drinking”, “periconceptional drinking”, and “binge drinking”. In addition, we performed a broader search on environmental teratogens and CHDs and checked the reference lists of retrieved articles and relevant review articles to identify additional relevant studies.

To our knowledge, this is the first meta-analysis to report on the association between maternal alcohol consumption and CHDs. Although the studies included in our analysis varied in terms of case definition, control selection, and exposure assessment, the associations were largely consistent in the sub-analyses. Our findings indicated that maternal alcohol consumption during pregnancy might have no association with the risk of CHDs. Additionally, statistically significant heterogeneity was detected (Q = 46.52, P < 0.001, I ²  = 88.8 %), although no publication bias was indicated from Egger’s test (P = 0.156) while Begg’s test (P = 0.037) indicated a publication bias (Fig. 4). Because of the existence of heterogeneity, the power of Begg’s test lowed. Therefore we conducted the subgroup analyses and meta-regression analyses (Table 2). In the sub-analyses of design types, geographical region, period of pregnancy and adjusted for smoking, the results were almost consistent with the pooled RR. Meta-regression analysis showed that the different study design might result in the heterogeneity. Moreover, when findings were stratified by geographic reign, null results were found among American and European populations, although a rate of 69 % was found among Australian populations. However, the study on Australian populations was conducted in only one paper by one author [31], and there were 277 cases from 1983–2007 in his report. He observed positive associations between maternal alcohol consumption during pregnancy and the CHD subtypes he analyzed: Ventricular septal defect (VSD) (RR = 1.36, 95 % CI = 1.14–1.63) and Atrial septal defect (ASD) (RR = 1.77, 95 % CI = 1.27–2.46); conotruncal defects (RR = 1.50, 95 % CI = 0.98–2.30). O’Leary attempted to demonstrate that maternal drinking was associated with different types of CHDs, but the number of ASD cases he found was 60, only 0.3 % of the exposed group, and the rate of conotruncal defects was 0.1 % as well. A smaller sample size could be why this result differed from others. However, there is another possibility that is worth our attention. The differences between the American, European and Australian natures of alcohol consumption could also have caused these effects. As we know, different wine has different components. For example, beer is made from wheat and Humulus lupulus, whisky is mainly made from Hordeum vulgare, brandy is made from fruits such as grapes, and so on. Criqul MH found the so-called “French Paradox”, that the French have a relatively low incidence of coronary heart disease while having a diet that is relatively rich in saturated fats [37]. He indicated that wine made from grapes could be the primary explanation for the phenomenon. However, we found very few statistics regarding the ingredients of different wine, and from our point of view, this question deserves to be studied in the future. Among those statistics that found that maternal alcohol consumption had a positive association with CHD risk, Martinez-Frias reported that consumption of more than 116 grams per day could increase CHD risk [29]. We must here highlight that this study had only 67 cases and 20 controls, which is quite small compared with the other sample sizes, and this authors report also had no adjustment variables; this dose of alcohol consumption could lead to a number of diseases that could have an impact on CHD risk. The Williams study had the same problem as well. Based on the adjustment results, the potentially important confounding factor of smoking was excluded in the analysis of alcohol consumption. However, there was no association between maternal alcohol consumption and CHDs following adjustment for smoking, indicating that smoking has little influence on CHDs. However, Lee et al. [38] confirmed the association between maternal smoking and increased risk of CHDs by meta-analysis. The influence of maternal smoking on CHD risk needs further study.

The result of our meta-analysis is in reality rather confusing. As we all know, alcohol consumption has a number of effects on our health. However, our research shows that maternal alcohol consumption during pregnancy may have no association with the risk of CHDs. Ethyl alcohol as the main component of wine has been suggested to play a positive role in heart disease. We speculate that a small amount of alcohol may have little influence in increasing the risk of CHDs. However, these statistics do not intend to say that maternal drinking is safe; the fact is that even low levels of prenatal alcohol exposure, such as in a single dose, can produce the birth defect termed fetal alcohol syndrome (FAS), and as many as 54 % of live-born children with FAS present with some form of cardiac anomalies [39], e.g., septal defects and pulmonary stenosis [20], that can lead to developmental challenges, ongoing medical care and death. Additionally, findings have shown that alcohol consumption during pregnancy may affect the Wnt/β-catenin signaling that allows for normal gene activation and cardiogenesis [40].

Our study has a number of strengths, including the large sample (n = 19,160) used to estimate the effect of maternal alcohol consumption on CHDs as a group. Additionally, because the suspected heterogeneous etiologies can affect CHD risk during different times in pregnancy, we divided pregnancy into stages; still, no association was found in any of the stages. However, our study must also be considered in light of certain limitations. First, our analysis is limited to studies published in English. Evidence of publication bias was found, and there was heterogeneity in the component studies, which could have been attributable to study design, study populations, analytic strategies or other unknown factors. Second, we derived most of our data from case-control studies, which may be more prone to information bias than cohort studies. Third, the consumption quantity in each study varied, including times/week, drinks/week, drinks/occasion, etc. The highest and lowest intakes varied across studies, and the highest intake in one study could have been similar to the median or lowest intake in another, which could have biased the overall results. Additionally, based on the different methods that were used to assess and report alcohol consumption across studies, we established the dose-response relationship between alcohol consumption and CHDs. It showed that the risk of CHDs did not influence with the increase of alcohol consumption, even binge drinking. Fourth, because of our inability to fully adjust for various confounders, associations between alcohol consumption and CHDs could be attributed to other factors, such as maternal smoking, ethnicity, age, and education. Fifth, our analysis included a wide range of exposure periods, but heart anomalies develop during weeks 2–7 of gestation [41]. We suspect that our including studies that assessed exposure beyond the critical period may have biased our result toward the null. That is, the summary results may be a misestimate of the relative risk of CHDs associated with alcohol consumption.

In conclusion, our analysis indicates that maternal alcohol consumption has no association with CHD risk. However, the findings from our study need to be confirmed in future research in well-designed cohort or intervention studies. In addition, the underlying mechanisms call for further elucidation.

Although its effects are modest, drinking is relatively common among women of reproductive age and could have important public health consequences. Young women continue to drink although the adverse effects of drinking on reproductive health are known, and some women continue to drink even after they learn that they are pregnant. Decreased maternal alcohol consumption during pregnancy would result in improved reproductive outcomes and could contribute to reduced infant mortality and morbidity.