Updated

Women who take the epilepsy drug valproate during pregnancy are three times more likely to have a child with an autism spectrum disorder, suggests new research based on close to 700,000 babies born in Denmark.

Previous studies have found more birth defects and lower intelligence among children of mothers who took valproate, but the new work represents the "strongest evidence to date" of a link between the drug and autism, according to an editorial published with the study on Tuesday.

The results don't prove the generic drug, also sold as valproic acid, causes autism. But researchers were able to account for a number of underlying factors - such as the age and health of the mothers and the babies' fathers - that make the study more convincing, Christopher Stodgell said.

"This finding isn't necessarily a brand new finding, but it's an important finding in that (researchers studied) really a much larger population, and they also looked at some other underlying drivers," said Stodgell, who studies the origins of autism at the University of Rochester Medical Center but wasn't involved in the new research.

Women "need to be very diligent about what the effects are if they're taking valproic acid," he said.

About one in 88 children has an autism spectrum disorder, or ASD, according to the U.S. Centers for Disease Control and Prevention. Those conditions range from autism itself to less disabling ones such as Asperger's syndrome.

For the new study, researchers tracked 656,000 kids born in Denmark between 1996 and 2006. Using a large prescription drug database, they found that just under 6,600 of the mothers of those children had epilepsy and 508 women took valproate while pregnant.

By 2010, 4.4 percent of the kids whose mothers had taken valproate during pregnancy were diagnosed with any ASD, including 2.5 percent with autism.

In contrast, 1.5 percent of all babies in the study had an ASD and 0.5 percent had autism, the study team reported in the Journal of the American Medical Association.

Mothers' underlying epilepsy didn't fully explain the link, according to Jakob Christensen from Aarhus University and his colleagues. In addition, autism rates were higher among children of women who used valproate during pregnancy than those who had previously used the drug but stopped before conceiving.

CONSIDER RISKS, BENEFITS

"Valproate is an effective drug, but it appears that it is being prescribed for women of childbearing potential at a rate that does not fully consider the ratio of benefits to risks," wrote Dr. Kimford Meador and David Loring from Emory University in Atlanta, in a linked editorial.

Valproate could affect maturation of a fetus's brain, Christensen suggested, including the signal-sending neurotransmitters.

Women who may become pregnant "certainly should discuss with their doctor if there are alternative treatments that would be reasonable," he said.

For those with certain syndromes or generalized epilepsy, there aren't necessarily other good options. Stopping valproate in that case isn't a good idea, Christensen said.

"It's also a risk if you have seizures, both for the mother and the unborn child. (Stopping medication) is not a thing that you take lightly," he said.

"Even those that are exposed to this drug, there's still a good chance - more than a 95-percent chance - that their child will never develop signs of autism."

The study didn't take into account whether women drank during pregnancy, or if they took folic acid - which has been tied to a lower risk of some birth defects.

Christensen said there are steps pregnant women on valproate can take to lower any risks to their baby, such as using the lowest possible dose and dividing it up during the day.