NEW YORK — Children born to mothers who took the anti-seizure drug valproate were five times more likely to be born with autism than those whose mothers didn't take the medication, a Danish study found.
The epilepsy drug was also tied to a three-fold increase of autism spectrum disorder, which includes Asperger syndrome and other developmental disorders, according to research published Tuesday in the Journal of the American Medical Association.
The findings show conclusively that valproate should be avoided and other treatments used instead to control seizures in women of childbearing age to reduce risk of autism in their unborn children, said Kimford Meador, a neurologist who wrote an accompanying editorial. If not, only the lowest dose of the drug should be given. Previous research has linked valproate's use during pregnancy to heart defects, spina bifida, cleft palates and cognitive problems including lower intelligence scores.
"This is an important risk factor and one that can be avoided or at least the risk reduced in women who don't need to take this and can take another drug," said Meador, a professor of neurology at Emory University in Atlanta, in an April 22 telephone interview. "This is the strongest evidence to date that there is a link between fetal exposure and childhood autism or autism spectrum disorder."
AbbVie Inc., based in North Chicago, Ill., sells valproate under the brand name Depakote. The drug is approved by the Food and Drug Administration to treat seizures, prevent migraine headaches and treat manic episodes associated with bipolar disorder. It is also used by doctors off-label for other conditions, particularly psychiatric disorders, according to the FDA website.
The agency in 2011 warned women of childbearing age that valproate was associated with lower cognitive scores in children whose mothers took the drug during pregnancy. It had previously warned that the medicine was linked to birth defects.
Christopher Stodgell, an associate professor in the Department of Obstetrics and Gynecology at the University of Rochester Medical Center in New York, who wasn't an author on Tuesday's paper but is studying how valproate causes autism, said the drug may turn on genes in the body too much, not at all or at the wrong time that could ultimately change the way the nervous system develops. He said though it's too soon to know for sure.
The effects do seem to occur early on in pregnancy, within the first few weeks, sometimes before a woman even realizes she's pregnant, Stodgell said.
"The risks outweigh the benefits but the caveat is the drug is important to some women," he said in an April 22 telephone interview. "It's important that these women understand what those risks are."
The study was one of the largest to show valproate's link to autism, he said.
Researchers from Aarhus University Hospital looked at children born in Denmark from 1996 to 2006 and identified those exposed to valproate during pregnancy. The analysis included 655,615 children.
During the course of the study, 5,437 children were diagnosed with autism spectrum disorder, including 2,067 who were diagnosed with childhood autism. Of those, 508 were exposed to valproate.
In 14 years of follow up, use of valproate during pregnancy was associated with an absolute risk of 4.42 percent for autism spectrum disorder and 2.5 percent for childhood autism, compared with a total risk of 1.53 percent for autism spectrum disorder and 0.48 percent for childhood autism, the study showed.
When the researchers limited it to just children of women with epilepsy, 432 children were exposed to valproate. The absolute risk for developing autism spectrum disorder was 4.15 percent and the risk for childhood autism was 2.95 percent. That compares to 2.44 percent risk for autism spectrum disorder and 1.02 percent risk for childhood autism for the 6,152 children not exposed to valproate, according to the paper.
Meador said more studies are needed to better understand how seizure drugs and other treatments affect an unborn child.
"There's still a great deal of valproate being used," Meador said. "The amount being used in women of childbearing age seems to be excessive given the risk benefit ratio. There's alternative drugs that have lower risks than valproate."