We know that drinking alcohol during pregnancy can cause birth defects; the government-mandated warnings on alcoholic beverage labels constantly remind us of that fact. But toxicologists remind us that the poison is in the dose: what is the dose of alcohol that causes birth defects? Heavy drinking can cause fetal alcohol syndrome, but there is no evidence that light to moderate drinking can cause it. Alcohol has been implicated in a number of other adverse effects on pregnancy and on the fetus. We simply don’t know if there is a threshold dose below which alcohol intake is safe, so the default position of most medical authorities has been to advise total abstinence during pregnancy. This is not a truly evidence-based recommendation, but rather an invocation of the precautionary principle. Those advising complete abstinence have been accused of paternalism and bias by wine-lovers and other critics, for instance here and here.

The literature on alcohol and pregnancy is extensive and confusing. It addresses many different endpoints, looking at effects on children and on the pregnancy itself. The studies are inconsistent in how they define “moderate” or “light” drinking, and they rely on self-reports that may not be accurate.

It would be impossible to read and accurately summarize such a large body of literature (over 21,000 hits on PubMed!), but here are a few examples that illustrate the scope, diversity, and conflicting results of these studies:

This study found that total abstinence from alcohol during pregnancy was correlated with an increased risk of stillbirth. Alcohol is known to inhibit uterine contractions and has been used IV in hospitals to stop premature labor.

This systematic review found an association of maternal alcohol intake with acute myelogenous leukemia in children.

This study showed a correlation with placental abruption but not with placenta previa.

This study found no association between maternal drinking and autism in general, but it found that binge drinkers (5 or more drinks on a single occasion) were less likely to have autistic children. They reasonably commented that this association was probably non-causal.

This study found a blunted response to pain in alcohol-exposed neonates.

This study found that “children born to mothers who drank up to 1–2 drinks per week or per occasion during pregnancy were not at increased risk of clinically relevant behavioural difficulties or cognitive deficits compared with children of abstinent mothers.”

A study by Willford, Leech, and Day found that prenatal alcohol exposure equivalent to 3-6 drinks a week correlated with lower IQ scores in African-American children but not in whites. As one commentator points out, the same study showed that maternal consumption of cocaine correlated with increased overall IQ scores in white children: this “suggests that perhaps the standards for confounding factors and statistical significance might have been too low.”

I found several websites that mentioned a study of 33,300 California women but I haven’t succeeded in tracking down the study itself. It reportedly showed that even though 47% drank moderately during their pregnancies, none of their babies met the criteria for Fetal Alcohol Syndrome. The authors of this study reportedly concluded “that alcohol at moderate levels is not a significant cause of malformation in our society and that the position that moderate consumption is dangerous, is completely unjustified.”

When studies have conflicting results, we resort to meta-analyses and systematic reviews to try to make sense of the data.

A meta-analysis by Testa, Quigley and Eiden found a detrimental effect on infant mental development at age 12–13 months, but no effect in younger or older infants. And it found a statistically significant improvement in mental development in children of light drinkers at 18-26 months. It’s hard to interpret what this might mean, if it is not a statistical fluke.

A systematic review by Henderson, Gray and Brocklehurst looked for possible effects of light to moderate drinking on outcomes including miscarriage, stillbirth, intrauterine growth restriction, prematurity, birth weight, small for gestational age at birth and birth defects including fetal alcohol syndrome. They

… found no convincing evidence of adverse effects of prenatal alcohol exposure at low–moderate levels of exposure. However, weaknesses in the evidence preclude the conclusion that drinking at these levels during pregnancy is safe.

There are no studies showing harmful effects from 1–3 drinks a week.

Conclusion: The scientific evidence has not identified a threshold below which alcohol consumption during pregnancy is definitely safe, but neither has it shown any convincing evidence of harm at low levels of intake, and it has not ruled out the possibility that low levels might provide a small benefit.

In the absence of better data, we are left to cope with uncertainty. Individuals will have to make decisions based on their philosophy of risk tolerance and their own personal non-scientific reasons. For me, the unconfirmed possibility of a low level of risk was not enough to outweigh the enjoyment of an occasional glass of wine during my pregnancies.