- Contrary to previous findings, having two alcoholic drinks per week raises the risk of miscarriage in early pregnancy.
- Drinking earlier in pregnancy carries greater risks.
- To play it safe, women might want to stop drinking when they start trying to get pregnant.
Just two cocktails, beers or glasses of wine a week can increase a woman's chances of early miscarriage by 66 percent, found a surprising new study that included more than 90,000 women.
Despite some evidence that light drinking during pregnancy is OK, the new findings offer the strongest evidence yet that playing it safe might be the best strategy for women who are gestating.
Still, many questions remain, as previous studies have offered mixed results and scientists still don't have a good mechanism to explain how such small doses of alcohol might affect fetal development.
"You should never give a recommendation based on a single study," said lead author Anne-Marie Nybo Anderson, an epidemiologist at the University of Copenhagen. "But if I was to give a recommendation to my daughter, I would say that if you plan a pregnancy and if you want to be careful and do everything you can in order to not harm your future baby, then I think you should stop drinking when you start trying to become pregnant, and then after the first four months you can be a little more relaxed."
There's no doubt that heavy or binge drinking during pregnancy is dangerous to a developing fetus. But research on the effects of low levels of alcohol consumption has offered contradictory results, with conclusions depending on the way studies are designed and even the country where studies are done.
Some studies have offered reason to be concerned. Others have suggested that a regular drink or two should be just fine. A 2010 study in the Journal of Epidemiology and Community Health, for example, found no signs of cognitive or emotional harm in young kids whose moms had consumed a couple drinks a week while carrying them. That study followed children up to age five and even found that the sons of light drinkers scored slightly higher than boys whose moms had stopped drinking altogether during pregnancy.
For the new study, Nybo Andersen and colleagues took advantage of a natural experiment in Denmark. For several years in the late 1990s, the country's national health board loosened its recommendations for pregnant women, suggesting that they generally avoid alcohol but if they decided to drink, they should not have more than a drink a day and they should not drink every single day.
For Danish women, the new guidelines had the effect of turning moderate drinking during pregnancy into standard and permissible behavior. No longer was having a glass of wine with a baby on board connected to depression or other negative health habits. It was just a normal thing to do, and many women indulged -- before the health board reversed its guidelines and became stricter again after a few years.
During that period of societal permissiveness, a national program enrolled more than 100,000 women during their pregnancies to take part in a Birth Cohort study. Among the many questions asked of them about their lifestyles, health behaviors and pregnancies during digital telephone interviews, women reported how much alcohol they had drunk before and how much they drank after getting pregnant. Mothers entered the study as early as six weeks into their pregnancies, and follow-up interviews recorded the eventual outcomes of their births.
Overall, about 55 percent of the Danish women continued to drink alcohol while pregnant during the study period, the researchers reported in the International Journal of Epidemiology. Just 2 percent averaged more than four drinks a week, and most drank far more moderately. But, the study found, it only took two drinks a week to raise the risk of miscarriage during the first 16 weeks by 66 percent.
Women who drank between half a drink and one and a half drinks each week showed a 19 percent increased risk in pregnancy loss, but only at the beginning of the second trimester. After the first four months, light drinking didn't seem to make a difference in rates of miscarriage or stillbirth.
An average of about 15 percent of pregnancies that get as far as week six end in miscarriage, Nybo Andersen said, and rates are higher than that for older moms -- making low to moderate amounts of alcohol a potentially significant risk. The new findings surprised even the experts.
"I was pretty sure that alcohol intake at such low amounts in pregnancy would not have dramatic effects," Andersen said. "But what we found was that there was quite a high risk for miscarriage with moderate alcohol intake, and there was a stronger risk the earlier in pregnancy we looked."
The study could not address whether one drink once in a while makes a difference. And, so far, science has no explanation for how, exactly, such a small amount of alcohol might harm a fetus.
With so many unknowns, it's possible for seemingly contradictory findings to be simultaneously true, said Fred Bookstein, an applied statistician who has studied fetal alcohol spectrum disorders at the University of Washington, Seattle, and the University of Vienna.
Fetuses might just fall into two groups, he offered: Some that are vulnerable to low-dose alcohol exposure in the early weeks and some that aren't. The first group would be the ones that would account for higher miscarriage rates in mothers who drank moderately during pregnancy. The second group would make it to birth undamaged despite the exposure, and those children would show no adverse effects into childhood.
It's hard to turn a study like this into universal recommendations, Bookstein added. Statistically, for example, it appears to be safe to drink up to one and a half drinks each week during the first four months and then responsibly after that.
But nuanced messages about drinking during pregnancy are often too easily interpreted as either a license to drink freely, or in this case, as permission to drink as a tool for ending unwanted pregnancies. Both interpretations would be inaccurate and dangerous.
"So, no, I wouldn't use this paper as the basis of any sort of public-health advisory," Bookstein said. "It's just too fraught."