Taking into account maternal depression as a factor, that translates to no association for use of antidepressants in the first trimester but an 87 per cent increased risk in the second or third trimester, Anick Berard, a researcher at the University of Montreal, and co-authors wrote.
The researchers said several mechanisms may be at work that explain the phenomenon, and they have to do with how serotonin affects brain development. Scientists believe the capacity of the brain to synthesise serotonin develops atypically in children with autism.
Because antidepressants are "likely to remain widely used" in the future, "a better understanding of the long-term neurodevelopmental effects of [antidepressants] on children when used during gestation is a public health priority," Berard wrote.
Doctors have long had mixed feelings about prescribing antidepressants to pregnant women. Previous studies have shown an increased risk of spontaneous abortion, prematurity and other physical issues and the US Food and Drug Administration has warned that Paxil may increase the risk of birth defects.
But many doctors have tended to weigh the harms of maternal depression - which may lead to poor nutrition, drinking, smoking, avoidance of medical care among other issues - as potentially more directly damaging to a fetus than the drugs.
The new study may shift the conversation.
"We need to think even more carefully about the prenatal environment. And not just about birth defects you can see at the time of delivery but about longer-term consequences," said Susan Hyman, a former chairperson of the American Academy of Pediatrics committee on autism, who is not affiliated with the study.
Surveys have shown that the use of SSRIs by pregnant women is rising in the United States - from less than 6 per cent in 1999 to 13 per cent in 2003 - and Hyman emphasised that the overwhelming majority of them do not have children with autism spectrum disorders.
"We would not want people to feel guilty," said Hyman, a professor of neurodevelopmental pediatrics at the University of Rochester Medical Centre. "That is not productive."
Scientists believe SSRIs balance the level of serotonin, which boosts mood. But the drugs also cross the placental barrier and the effect of such drugs on the brain development of a fetus at its earliest and more vulnerable state is still being studied.
The JAMA Pedatrics research adds to the growing number of studies about antidepressant and developmental delays, autism or Attention Deficit Hyperactivity Disorder (ADHD), but the picture is far from clear as the results have been mixed.
From a study design standpoint, the new study is stronger than some of the previous work because it is prospective, taking information about the pregnancies before they knew the outcome. But it has a number of limitations. First, the study used prescription data to determine which women were on antidepressants, but they may or may not have actually taken them. Perhaps more importantly, the data didn't have information about lifestyle, and it's possible that myriad factors, such as whether they smoke or their body mass index, may have influenced the findings.
Last April, Johns Hopkins University researchers reported in the journal Pediatrics that boys with autism were almost three times as likely to have been exposed to SSRIs than their typically developing counterparts. The team, which looked at nearly 1000 mother-child pairs, found that the effect appeared to be stronger in those exposed during the first trimester and in boys than in girls. A similar link was seen between a mother's SSRI use and developmental delays, but the sample was smaller and the researchers said those results should be interpreted with caution.
In another study, Harvard scientists analysing electronic medical records reported that children who were exposed to antidepressants in the womb were at 80 per cent increased risk of ADHD.
However, that same study found that children were not more likely to have autism spectrum disorders when accounting for a mother's medical history. That is, mothers with a history of psychiatric illness were more likely to have a child with autism regardless of whether or not they were on medication. Several past studies have suggested that a family history of depression may increase an individual's risk for autism.
One such study, which looked at nearly 670,000 children born in Denmark from 1996 to 2006 and used prescription information from the country's drug registry, found no association between antidepressant use and autism when researchers controlled for things like mood disorders in the mothers.
So if you're on antidepressants and pregnant or thinking about getting pregnant, what should you do?
Hyman said that it should still remain an individual decision, but that prospective mothers might want to have a serious discussion with their doctor about the possibility of other types of therapies for depression and anxiety, such as counselling.
Bryan H. King, a doctor at Seattle Children's Hospital, had similar advice, writing in an editorial accompanying the study that "it makes no more sense to suggest that [antidepressants] should always be avoided than to say that they should never be stopped".
King added that given the fact that the prevalence of autism is growing - up to 1 in 45 children in the United States - "the search for explanations, particularly among environmental factors, is critically important."
