3.3.1. Prenatal exposure to antiseizure medications and the Risk of ASD in children
The studies by Wiggs et al, [12] and Bromley et al, [19] comprehensively focused on 3 major ASM: Valproic acid, lamotrigine, and carbamazepine. Wiggs et al, had a relatively robust result because it had a large sample size (14,614 children), accounted for various possible confounders (e.g., bipolar disorders, psychiatric diagnoses, maternal use of other additional medications, etc.), and used high-quality data source; the Swedish register data. In their study, children born to mothers who used valproic acid during pregnancy had a 2.3-fold increase in the risk of ASD (hazard ratio [HR] 2.3, [95% CI: 1.53–3.47]) compared to those whose mothers reported no use of ASM. Carbamazepine usage had a weak non statistically significant association, while Lamotrigine was not associated with increased risk of ASD after adjusting for all possible confounders. Since all three drugs are also used to treat mood swings, the study ensured that bipolar disorder and maternal psychiatric diagnoses as confounders were accounted for. It however did not assess exposure by trimester of pregnancy or dosage of the drugs administered. Similarly, Bromley et al, observed that monotherapy with valproic acid had a significantly increased risk for ASD, 12.0% (OR: 6.05 [95% CI: 1.65% − 24.53%]) compared to the control children (1.87%). A weak non statistically significant association was noted between the increase in dosage of valproic acid and ASD, however, the study had a very low sample size which could have greatly affected its power to determine such associations. Again, no increased risk for ASD was seen among those that used lamotrigine or carbamazepine monotherapy, while exposure by trimester of pregnancy was not assessed.
The other studies that looked at a combination of ASM were Rasalam et al, [14], the Nordic countries cohort by Bjørk et al.[15] and the Norwegian mother and Baby cohorts by Veiby et al, [20] and Bjørk et al, [18]. In all these studies, in utero exposure to ASM, in general, showed an increased risk of ASD. Unlike other studies, Bjørk et.al [15] found topiramate-exposed children of epileptic mothers have a higher risk of developing ASD than valproate-exposed children with HRs of 2.8 (95%CI, 1.4–5.7) and 2.4 (95%CI, 1.7–3.3). in their study a weak association HR, 1.3; [95% CI, 1.1–1.5], was also observed for lamotrigine exposure and no increased risk for levetiracetam, gabapentin, pregabalin, or phenobarbital. Meanwhile, Veiby et al[20], reported that exposure to lamotrigine and not valproate had the highest risk of ASD, however, the sample size for exposure to valproate (n = 19), was much lower than that for lamotrigine (n = 44) exposure, which could explain the difference. They further analyzed children whose fathers used ASM prior to the conception of the baby and found that these children were more likely to show neurodevelopmental delays than the controls, but this was not specific for ASD. It is thought that having epilepsy, these fathers could possess inheritable traits that predispose their babies to neurodevelopmental delays just like epileptic mothers. It would be interesting to know the risk of ASD in children from non-epileptic fathers exposed to ASM prior to conception.
Two studies, Christensen et al, [10] and Stadelmaier et al, [11] looked at the risk of exposure to Valproic acid as monotherapy in detail. Christensen et al assessed a very large number of children including (508, mean age 8.84 years) exposed to Valproic acid in utero from 1996 to 2006. The study reported that the absolute risk of developing ASD in children exposed to Valproic acid in utero was 4.42% [95% CI: 2.59% – 7.46%], for autism spectrum disorder and 2.50% [95% CI: 1.30% − 4.81%] for childhood autism, compared to 2.44% [95% CI: 1.88% − 3.16%] and 1.02% [95% CI: 0.70% − 1.49%] for autism spectrum disorder and childhood autism respectively in those not exposed to valproic acid. This result indicates that in utero exposure to valproic acid increases the risk of developing ASD by 2-folds. A similar finding was observed in the study by Stadelmaier et al. where out of 47 children exposed to valproic acid in utero, 13 were diagnosed with autism spectrum disorder, significantly higher than the unexposed children. These results are generally consistent with findings from other studies that valproic acid use in pregnancy is indeed a risk factor for autism [21, 22].
3.3.2. Risk of ASD According to trimester of exposure to ASM
According to Bjørk et al,[18], Wood et al[13], Stadelmaier et al[11] and Wiggs et al [12], exposure to antiseizure medications (especially valproic acid) in utero in the first trimester of pregnancy appeared to present relatively higher risks for developing ASD compared to second and third trimesters .However, Christensen et al [10], found that the risk of autism spectrum disorder among children whose mothers were exposed to valproic acid in the first trimester (adjusted HR, 2.9 [95% CI, 1.6–5.1]) were relatively lower compared to the risk in those exposed in the second and third trimesters, [3.1 (95% CI, 0.8–12.2)]. However, the confidence intervals of the second and third trimester exposures varies widely, probably due to the smaller sample size used (67 children), and so was not powered enough to detect a meaningful association. The findings of Christensen was supported by Bjørk et al[15] who found that exposure to valproate during the second and third trimesters without exposure during the first trimester was still associated with an increased risk of ASD with a HR of 1.94 (1.04–3.60). Bjørk et al[18], and Wood et al [13] studied the usefulness of folic acid supplements in the first trimester in preventing ASD, so the supposed increase in the risk of ASD in first-trimester exposure to ASM is extrapolated from the fact that women who did not take folic acid early in pregnancy had increased risks of producing autistic children. Given the observed benefits of folic acid supplements in decreasing ASD, it may not be possible on ethical grounds to conduct a study purely comparing exposure to ASM in the first trimester of pregnancy and the risk of ASD.
3.3.3 Risk of ASD according to dosage of ASM administered
The Nordic countries cohort study by Bjork et al [15], assessed the relationship between dosage of ASM used and the risk of ASD and noticed positive associations. In their study they found HR of 1.7 (95% CI, 1.0-2.8) for those exposed to topiramate doses less than 100mg per day, and HR of 2.9 (95% CI, 1.3–6.7) for doses greater than 100mg per day. Meanwhile, the HR for those exposed to less than 750 mg per day of valproate was found to be 2.3 (95%CI, 1.9–2.8) and 5.6 (95%CI, 4.7–6.8) for those exposed to more than 750 mg per day. Two studies, Wood et.al. [13], and Bromley et.al [19] also found positive associations. However, the associations were not statistically significant probably because of the small sample sizes in each study. In contrast, Bjork et al [16] revealed that there is no significant association between ASM doses and the risk of ASD. This is supported by Christensen et al (Christensen & Gr, n.d.) who assessed a large number of women and found that those that used a high dosage of Valproic acid (VPA) did not have a significantly different risk for producing autistic children (adjusted HR, 2.5 [95% CI, 1.03–6.1] and 3.2 [95% CI, 1.4–13.4] for high and low dose VPA respectively). The question of what dose of ASM (especially VPA) is safe for pregnant women with epilepsy is very important because if a safe dose exists, then women whose epilepsy can only be controlled by VPA will not have to discontinue its use in case they decide to get pregnant. It is therefore important that further studies are pursued in this direction.