A new systematic review finds that exposure to the antiseizure drugs valproate and topiramate during pregnancy increases the risk of autism (ASD), intellectual disability (ID), and attention deficit hyperactivity disorder (ADHD) in offspring. Valproate was associated with up to four times the risk of ASD, five times the risk of ID, and 1.5 times the risk of ADHD; topiramate was linked to a doubled risk of ASD, four times the risk of ID, and a doubled risk of ADHD.
The drugs were also linked to impaired adaptive functioning and social skills more generally.
The researchers draw attention to the fact that these harms, along with the risk of congenital abnormalities, are well-known in the neurology field but that psychiatry appears to ignore them. The drugs—commonly prescribed for depression and bipolar disorder—are often used at higher doses in psychiatry, bringing more risk.
“That ASM use is linked to adverse neurodevelopmental outcomes has been appreciated in the field of epilepsy for many years, yet has been largely overlooked by psychiatry despite the widespread usage of ASMs in the treatment of disorders such as depression and bipolar and at higher dosages than is typically used in epilepsy.,” the researchers write.
Other antiseizure drugs (phenobarbital, levetiracetam, carbamazepine) were also linked to neurodevelopmental problems, but methodological issues in the original studies hindered the ability to draw conclusions. Many popular antiseizure drugs, such as gabapentin and pregabalin, did not have enough data to analyze. The one exception to the finding in this analysis was lamotrigine, which did not appear linked to neurodevelopmental outcomes.
Most studies also found that polypharmacy—taking multiple antiseizure drugs—was associated with worse outcomes in all neurodevelopmental domains.
The research was led by Eliza Honybun and overseen by senior authors Piero Perucca and Genevieve Rayner at the University of Melbourne. It was published in Neurology, a top neurology journal.
The study adds to a large body of research on the impact of psychiatric drugs in pregnancy. In particular, many studies have linked in-utero antidepressant use to poor outcomes for babies, including neonatal withdrawal syndrome. However, little research has focused on antiseizure drugs, though they are commonly prescribed, particularly for bipolar disorder.
The current study was a systematic review of 43 studies that assessed neurodevelopmental outcomes in children after their mothers used antiseizure drugs during pregnancy. The outcomes included measures of social development, adaptive functioning, emotional functioning, and official psychiatric diagnoses.
Of the included studies, 10 were population-based, 19 were prospective cohort studies, nine were retrospective cohort studies, and five were cross-sectional observational studies.
The review found that the quality of evidence was generally low. Some studies lacked a comparison group (those not exposed to antiseizure drugs) or did not control for confounding factors like maternal age, smoking, other diagnoses, etc. Some studies lacked a blinded assessment, lost many subjects to follow-up, or were simply too small to analyze. The specific outcomes were highly variable, with studies including different measures and measuring different types of outcomes. There was a high variation in the duration of follow-up and when the outcomes were assessed.
The researchers broke down their results by drug:
Valproate and topiramate: Drugs with evidence for a large increased risk
Valproate: 40 of the 43 studies included valproate. Exposure to the drug in utero was linked to worse adaptive and social functioning, as well as up to four times the risk of ASD, five times the risk of ID, and 1.5 times the risk of ADHD. In some studies, there appeared to be a dose-dependent relationship, where higher doses increased the risk. However, this wasn’t a consistent finding. Six of the studies had a sample size too small to analyze, and 10 studies failed to account for confounding variables. However, this leaves dozens of well-conducted studies that found a harmful effect of valproate on neurodevelopmental outcomes, so this is considered the strongest evidence.
Topiramate: 11 studies included topiramate. One study found a doubling of the risk of ASD and 3.5 times the risk of ID; another found 2.38 times the risk for ADHD. A small study linked the drug to poor adaptive functioning. The studies that analyzed doses found a connection: higher doses appeared to be linked to worse functioning. Five of the studies had a sample size too small to analyze.
Drugs with mixed results
Phenytoin: 13 studies included phenytoin, with mixed results. Several found no effect, but at least one found a dose-dependent effect on adaptive functioning. Seven of the studies had a sample size too small to analyze, and seven failed to account for confounding variables.
Carbamazepine: 34 studies included carbamazepine. Several showed that the drug was linked to behavioral and emotional problems, and one study found the drug linked to 1.8 times the risk of tic disorders. Much of this comes from small, preliminary studies. Three of the studies had a sample size too small to analyze, and nine of the studies failed to account for confounding variables.
Levetiracetam: 13 studies included levetiracetam. One population study linked the drug to 1.8 times the risk of ADHD and 2.2 times the risk of anxiety diagnoses. The researchers write that the evidence was not consistent, though. Six of the studies had a sample size too small to analyze, and two of the studies failed to account for confounding variables.
Phenobarbital: Eight studies included phenobarbital. One study found a large increase of 7.6 times the risk for behavioral and emotional disorders, and another found an increased risk of learning disorders. However, these were small preliminary studies and are not considered solid evidence. Three of the studies had a sample size too small to analyze, and three of the studies failed to account for confounding variables.
Drugs with a lack of evidence to draw conclusions
Oxcarbazepine: 11 studies included oxcarbazepine. Although there were some inconsistent results, most of the studies found no connection between the drug and neurodevelopmental problems. Five of the studies had a sample size too small to analyze, and one study failed to account for confounding variables.
Clonazepam: 10 studies included clonazepam. Although there were some inconsistent results, most of the studies found no connection between the drug and neurodevelopmental problems. Four of the studies had a sample size too small to analyze.
Gabapentin: Seven studies included gabapentin. The included studies found no connection between the drug and neurodevelopmental problems. Three of the studies had a sample size too small to analyze.
Lamotrigine: No evidence for increased risk
Lamotrigine: 28 studies included lamotrigine. Although two studies found an increase in “autistic traits,” five large population studies found no connection. Four of the studies had a sample size too small to analyze, and six studies failed to account for confounding variables. The researchers conclude that there is no connection between lamotrigine and neurodevelopmental problems.
Ultimately, the researchers write that these findings should impact treatment decisions in pregnancy.
“The adverse neurobehavioral outcomes linked to prenatal [antiseizure drug] exposure in this study have significant implications for the treatment and counseling of women of childbearing age because they often have lifelong flow on impacts for both the child and carer’s lifestyle and quality of life.”
****
Honybun, E., Cockle, E., Malpas, C. B., O’Brien, T. J., Vajda, F. J., Perucca, P., & Rayner, G. (2024). Neurodevelopmental and functional outcomes following in utero exposure to antiseizure medication: A systematic review. Neurology, 102, e209175. doi:10.1212/WNL.0000000000209175 (Link)
Autism and ADHD are neurodivergent differences, not “problems”, “disorders”, “harms”, or “risks”. I agree that chemicals can attribute to these conditions. I was exposed to diazanon insecticide in utero, and am the most obviously autistic member of my immediate family. I was also exposed to whooping cough from my mom and was born premature. But genetics play the biggest role. If a mom is on a medication for bipolar or epilepsy, it is because she herself is neurodivergent. So the baby (whose mitochondrial DNA is identical to Mom’s), will have Neurodivergent genes.
Report comment