Expecting Better

I Wrote That It’s OK to Drink While Pregnant. Everyone Freaked Out. Here’s Why I’m Right.


Photo by Kati Neudert/Shutterstock

When I was pregnant, I wondered, as many women do: Can I have a drink? It is well-known that drinking to excess during pregnancy is dangerous, and perhaps less well known but still true, that even one or two episodes of binge drinking can be harmful. But what about an occasional glass of wine with dinner?

Expert opinions on this differ. What to Expect When You’re Expecting says no alcohol. Panic-Free Pregnancy says an occasional drink is fine. A 2010 survey asked obstetricians, “How much alcohol can a pregnant woman consume without risk of adverse pregnancy outcomes?” Sixty percent of the OBs said none, but the other 40 percent said some alcohol was fine. The American Congress of Obstetricians and Gynecologists (ACOG) says no amount of alcohol has been shown to be safe, but the U.K. equivalent (the Royal College of Obstetricians and Gynecologists) says that while not drinking is the safest option, “Small amounts of alcohol during pregnancy have not been shown to be harmful.”

My obstetrician said a few drinks a week was fine. But as with everything else, amid this disagreement, I needed to go to the data myself.

I reviewed many, many studies, but I focused in on ones that compare women who drank lightly or occasionally during pregnancy to those who abstained. The best of these studies are ones that separate women into several groups—for example: no alcohol, a few drinks a week, one drink a day, more than one drink a day—and that limit the focus to women who say they never had a binge drinking episode. With these parameters, we can really hone in on the question of interest: What is the impact of having an occasional drink, assuming that you never overdo it? 

I summarize two studies in detail in my book: one looking at alcohol consumption by pregnant women and behavior problems for the resulting children up to age 14 and one looking at alcohol in pregnancy and test performance at age 14.  Both show no difference between the children of women who abstain and those who drink up to a drink a day. I summarize two others in less detail: one looking at IQ scores at age 8 and a more recent one looking at IQ scores at age 5. These also demonstrate no impact of light drinking on test scores. 

I argue that based on this data, many women may feel comfortable with an occasional glass of wine—even up to one a day—in later trimesters. (More caution in the first trimester—no more than two drinks a week—because of some evidence of miscarriage risk.)

Although this discussion takes up only a small share of the book, it has garnered the loudest reaction, much of it outrage. NOFAS, a fetal alcohol syndrome advocacy group, issued a press release even before the book came out saying I was harmful and irresponsible. Amazon reviews of the book—at least some of them by people who explicitly said they would never read it—attacked me and anyone who had a drink during pregnancy as an alcoholic. One commented on my daughter: “Emily Oster claims that her 2-year old daughter is perfectly healthy, yet the full impact of the alcohol exposure on her child will not be evident until the adolescent years.”

The president of ACOG has vehemently disagreed with me, saying in a radio interview about occasional drinking that alcohol in pregnancy is more dangerous than heroin or cocaine. Of course, there has been occasional public agreement from OBs (and much more private agreement).

Some of the arguments made in response to the book are tangential. Commenters wish that there was more in the book about the dangers of fetal alcohol syndrome, more discussion of the risks of binge drinking. I spend only a page on this, since it is not the question I believe most readers of the book are asking. 

Some of them are philosophical. People ask, “Why take the risk?” since there is no benefit to the baby. But this ignores the fact that we are always making choices that could carry some risk and have no benefit to the baby. Driving in a car carries some risk to your baby, and your fetus does not benefit from that vacation you took. Or they ask, “Is it so hard to give up drinking for nine months?” The answer is, of course, no, but because you might enjoy the occasional beer, it seems worth at least asking the question about the risks.

Then there is the criticism that I cherry-picked studies to fit the story. This certainly isn’t the case; the fact that the book doesn’t summarize all 23,000 studies in PubMed on alcohol in pregnancy reflects the desire to identify the most reliable and largest and present those. Still, it’s reasonable to ask whether there are studies that I missed that tell a different story.

One fact that has been cited to me a number of times, including by the ACOG president, is: “One in 7 children with fetal alcohol syndrome had a mother who drank one to eight drinks per week in the first trimester.” The implication is that even light drinking early on (which would be much closer to one than eight drinks) is dangerous. But this claim doesn’t come from a study; it comes from a statement made in a letter to the editor, and it’s therefore impossible to evaluate critically. One to eight drinks a week could mean eight drinks on one night, for instance, and that is known to be dangerous.

Another study that has been mentioned prominently relates prenatal alcohol exposure to behavior problems in young adulthood. Although some have suggested that this paper identifies impacts of having one drink per day, the analysis actually relates behavior problems to a measure of average daily intake—which includes people having more than that, sometimes a lot more. It’s true that some people evaluated in this study drink lightly, but others do not, and by lumping them together it is very difficult to draw conclusions about the light drinkers.

There is a much more technically complex study that I certainly would have included in the book if it had come out in time. It shows that light maternal drinking is associated with small IQ decreases for people with some particular genetic variants. Light maternal drinking is also associated with small IQ increases in people with some other genetic variants. This suggests that further studies may be useful in evaluating genetic risks, although it doesn’t provide a lot of guidance at this time.

The bottom line is that the criticism fails to identify studies that have the features we would want: a population that is never binge drinking and a data analysis that looks separately at women who drink lightly and those who drink more. In the book I discuss one study like this, which does argue there are impacts on behavior at one drink per day, but the study fails to adjust for differences across groups, like whether the father lives at home or if there was prenatal cocaine use, among other things.

Like alcohol, Tylenol, caffeine, and anti-nausea drugs like Zofran are substances that—in moderation—are thought to be safe during pregnancy. But they are also substances that in excessive doses could be dangerous. Some women decide that they will therefore avoid them altogether because they cannot be sure. And many women, seeing the evidence in the book on alcohol, will still choose to avoid it. 

But others will see the data, like the data on caffeine or Tylenol, and choose to have an occasional drink, as I did. The value of the data is not that it leads us all to the same choice, just that it introduces a concrete way to make that choice.