JLE

Epileptic Disorders

MENU

Differential impact of antenatal exposure to antiseizure medications on motor and mental development in infants of women with epilepsy Volume 24, issue 3, June 2022

Figures


  • Figure 1

Tables

 

Women constitute approximately half of the 50 million people with epilepsy globally [1]. Most women with epilepsy (WWE) require use antiseizure medication (ASM) in order to remain seizure-free during the pregnancy period. Women with childbearing potential and epilepsy have concerns about the risk of developmental impairment in infants with exposure to ASMs during pregnancy. Several prospective studies have demonstrated that antenatal exposure to ASMs can lead to impaired cognitive development at one-year [2], three-year [3], six-year [4-6] and tenyear follow-up time-points [7, 8]. Further, valproate (VPA) exposure carries the highest risk for language and intelligence deficits [4-6].

Prenatal and postnatal care, including good nutrition, protection against infections and psychological stimulation, are critical for development in infancy. Deficiencies in nutrition and psychological stimulation may jeopardise the physical, cognitive and emotional development of infants [9]. Some WWE face challenges in providing optimal care and support to their babies during the critical period of pregnancy and early infancy [10]. It is critical to identify children of WWE who are at risk of developmental problems in early infancy, as early detection of developmental delay provides opportunities to offer appropriate remedial therapies to the infants [11]. In this study, we aimed to assess the motor and mental developmental outcomes of infants of WWE according to antenatal exposure to ASM.

Methods

Study settings

The Kerala Registry of Epilepsy and Pregnancy (KREP) was started in 1998 as a prospective observational registry to monitor the maternal and foetal outcomes of pregnancies in WWE. The protocol of this registry has been published previously [2, 12, 13]. Briefly, KREP enrolled WWE in the pre-pregnancy stage or in the first trimester of pregnancy. All WWE were under the care of experienced epileptologists. Each pregnant woman in KREP maintained a pregnancy diary to record the daily use of ASMs, folic acid and seizure count. The data for each month of pregnancy and three postpartum months were transferred from the pregnancy diary to the clinical records of the registry during their clinic visits. Screening was carried out in four phases: screening for anomalies before 18 weeks of pregnancy, physical examination at birth, echocardiography, and abdomen ultrasonography at three months and physical review at 12 months. All infants were scheduled for evaluation at six years, 10-12 years and before 18 years.

Study population

We included all babies in the registry who were over 12 months of age on 31st December, 2020. Infants older than 24 months of age were not included in this study. The demographic and clinical characteristics of the mothers were extracted from the medical records. Further, we reviewed the clinical charts and extracted the ASMs during pregnancy.

Outcome measures

An independent team of developmental paediatricians and developmental therapists, who were blinded to the ASMexposure, had examine dall infant safter 12 months of age in their clinics. Infants who did not participate optimally were excluded from the study. The developmental assessment scale for Indian infants (DASII), which is an adaptation of the Bayley Scale of Infant Development Version I, was used for assessing motor andmentaldevelopment[14]. The DASII is validated and widely used in India to assess infant development [15]. The DASII scale provided raw scores separately for motor and mental development based on a set of tests. The motor cluster included several test items for neck control (seven items), body control (23 items), locomotion Icoordinated movements(10 items), locomotion II skills (13 items), and manipulations (14 items). Further, the mental cluster included test items for cognizance visual (25 items), cognizance auditory (seven items), manipulating and exploring (36 items), memory (11 items), social interaction and imitative behaviour (22 items), language-vocalization, speech and communication (11 items), language vocabulary and comprehension (18 items), understanding relationships (18 items), differentiating between use, shapes and movements (eight items) and manual dexterity (seven items). The motor and mental age of the child were computed separately by matching the respective raw scores with the standard score provided in the DASII manual. Themotordevelopmentquotient(MODQ) and mental developmentquotient (MEDQ) were computed as the ratio of motor and mental age to chronological age. Infants of WWE who were not exposed to any ASM during pregnancy were classified as the internal comparison (reference) group. Infants exposed to ASM who had MODQ or MEDQ corresponding to less than one standard deviation of the mean MODQ or MEDQ of unexposed infants (reference group) were categorised with delayed development.

Exposure

The main exposure variable of interest was the use of ASM anytime during the antenatal period. Infants were categorised into mono or polytherapy groups based on number of ASMs used during pregnancy. The highest dose of each ASM used anytime during pregnancy was taken as the representative dosage of that ASM. Those infants with no exposure to ASM during the entire period of pregnancy were considered as the comparison group.

Confounders

We extracted data of potential confounding variables regarding mother (age, seizure type, epilepsy classification) and baby (birth weight, malformation status) from the clinical records of the KREP.

Statistical analysis

All data were transcribed to a spreadsheet (Microsoft Excel) and analysed using the Stata packages (STATA/ MP 16.1). Continuous variables were summarized as mean with their standard deviations and discrete variables as proportions. Group comparisons were made with one-way ANOVA or chi square test, as appropriate. We employed linear mixed models to account for measured confounders while computing the adjusted motor and mental development quotients (aMODQ and aMEDQ). The deviations of the aMODQ and aMEDQ for each ASM group from the reference group (unexposed infants) were expressed as mean with 95% confidence interval.

We also examined the dose dependency of the development quotient (DQ) by comparing the DQ of infants with exposure tolow, medium and high doses of ASMs. The dosage of ASMs were categorized as low, medium or high, as follows: carbamazepine (CBZ; low 800 mg); lamotrigine (LTG; low 100 mg); levetiracetam (LEV; low 1000 mg); oxcarbazepine (OXC; low 1000 mg); phenobarbitone (PB; low 60 mg); phenytoin (PHT; low 200 mg); and valproate (VPA; low800 mg). We employed similar linear mixed models to account for the effect of the measured confounding variables, as in our main model.

Ethical oversight

The KREP is approved by the Institutional Ethics Committee of Sree Chitra Tirunal Institute for Medical Sciences and Technology (Approval No. SCTIEC/17/ 2003 dated 11 June 2003). Written informed consent was obtained from all WWE at the time of enrolment.

Results

In total, 1,902 infants were eligible for this study (those who had completed the evaluation at three months and who were older than 12 months), of whom 1,485 (78.1%) were assessed. For other infants, the mothers declined, did not respond, could not keep their appointment, or could not be contacted. The developmental assessment was completed satisfactorily in 1,357 infants (71.2%). Tests were not completed for 128 infants, although these were included in the evaluation. There were 705 males and 652 females; 1,323 singleton pregnancies and 17 twin pregnancies. Their mean age (SD) was 15.3 (3.9) months. Congenital malformations were seen in 94 of the 1,357 infants included in the study. More than one third (37%) of WWE reported educational level corresponding to below the 10th grade (approximately 10 years of schooling). Their epilepsy was classified as generalized epilepsy in 39%. Generalised seizures were reported in 38% of WWE before the pregnancy period. The congenital malformation rate, age at examination of the infants and maternal education were comparable between those exposed to ASMs as monotherapy or polytherapy, or unexposed (table 1).

The mean birth weight in the study population was 2.9±0.5 Kg. The birth weight of infants unexposed to ASMs (3.0±0.5 Kg) was significantly greater (p=0.03) than that of infants exposed to monotherapy (2.9±0.4 Kg), and those exposed to polytherapy (2.8±0.5 Kg,). The proportion of WWE who had generalized seizures during the index pregnancy was significantly higher (p<0.001) for the monotherapy group (360/840) than the polytherapy (125/407) and no ASM group (30/110) (table 1). None of the mothers reported smoking or alcohol habits during pregnancy. Infants exposed to ASM in utero as monotherapy or polytherapy had lower MODQ and MEDQ than infants unexposed to ASM (supplementary table 1). The MODQ ranged from 88.9±13.4 in the clobazam monotherapy group to 110.4±14.3 in the LEV monotherapy group (supplementary table 2). Similarly, the MEDQ ranged from 90.2±17.7 in the polytherapy group with VPA to 104.3±9.1 in the LEV monotherapy group. In the unexposed group, the mean MODQ and MEDQ were 97.7±16.6 and 99.8±12.5, respectively. In the polytherapy group, the corresponding mean MODQ and MEDQ were 91.1±19.3 and 92.9±14.9, respectively.

Infants exposed to ASM were classified in the delayed development group if their developmental quotients were less than one standard deviation of the mean MODQ (84) and MEDQ (88) of unexposed infants. Accordingly, the proportions of infants with delayed motor development for different ASM monotherapies were: LEV 1/62 (1.6%), LTG 1/26 (3.9%), OXC 6/42 (14.5%), PB 15/83 (18.1%), PHT 15/81 (18.5%), CBZ 58/317 (18.3%), VPA 69/211 (32.7%), other ASMs 2/18 (11.1%), polytherapy excluding VPA 92/315 (29.2%) and polytherapy including VPA 45/92 (48.9%). The proportions of infants with delayed mental development according to ASM exposure were: LEV 1/62 (1.6%), LTG 1/26 (11.5%), OXC 2/42 (4.7%), PB 21/83 (25.3%), PHT 21/80 (26.3%), CBZ 49/317 (15.5%), VPA 62/210 (29.5%), other ASMs 6/18 (33.3%), polytherapy without VPA 94/314 (29.9%) and polytherapy including VPA 34/92 (36.7%). Delay in motor and mental development was greater in the VPA monotherapy, VPA polytherapy and other polytherapy groups.

Infants exposed to VPA, compared to other ASMs, had significantly lower MODQ in the monotherapy (-8.1; 95% CI: -12.2 to -4.0) and polytherapy groups (-12.1; 95% CI: -16.9 to -7.2), after adjustment for maternal education, seizure classification, infant age at the time of examination and malformation status (figure 1A, table 2). In contrast, the mean adjusted MODQ was higher for the LEV monotherapy group (10.1; 95% CI: 4.6 to 15.6), compared to the unexposed group. The adjusted MEDQ of the VPA monotherapy group was lower than that of the unexposed group by -7.4 (95% CI: -10.4 to -4.3, p=<0.001) (see figure 1B, table 3). The aMEDQ of the polytherapy group with VPA and without VPA was significantly (p<0.001) lower than that of unexposed infants (see table 3, figure 1). Further, those exposed to PB also showed lower adjusted mean MEDQ (-3.8; 95% CI: -7.5 to -0.1) than the unexposed group. These results remained consistent based on models adjusted for all the abovementioned confounders and birth weight of infants (supplementary table 3, 4).

In the dose-response analysis, the dose dependent changes in aMODQ and aMEDQ were significant for VPA (aMODQ and aMEDQ), PB (aMODQ and aMEDQ) and CBZ (aMEDQ only). PB and LEV (aMEDQ only) showed a trend of lower aMODQ and aMEDQ with medium and high doses, compared to low dose of the same drug (table 4).

Discussion

The KREP is a prospective register that aims to capture the developmental trajectories of children of WWE from early infancy to late adolescence. This evaluation of children in the registry at one year is the first of four assessments up to 18 years. The instrument that we used had a fair distribution of tests, aimed at assessing receptive and expressive language development as well as other cognitive functions, although it was a generic test.

We demonstrated that motor and mental development of infants of WWE varied widely according to antenatal ASM exposure. Within the monotherapy subgroups, VPA stood out with the lowest developmental quotients followed by PB and PHT. Poly-therapy that included VPA as well as polytherapy without VPA was associated with significant developmental delay. Further, the dose-dependent decrease in development quotients with exposure to both VPA and phenobarbitone that was observed in this study indicates a causal relationship between the two factors. In contrast, infants exposed to LEV and LTG did not show any developmental impairment. There are very few prospective studies that have examined developmental outcome at one year for infants exposed to ASM. A prospective study of two-year-old children did not show a significant difference in motor and mental (cognitive) development between children of healthy women and children of WWE with antenatal exposure to LEV or LTG [16]. The present study demonstrates that ASMs differ widely regarding adverse developmental effects; viz VPA and PB showed the greatest adverse effect, while most other ASMs showed minimal to mild adverse effects. According to an earlier study, prenatal exposure to PHT is associated with lower mental development scores in infants at one year [17]. However, other studies on exposure to CBZ have yielded variable results ranging from normal intelligence [18] to low intelligence quotients [19,20]. The negative impact of VPA exposure on neurocognitive development is well established in older children [3-6, 18, 21, 22].

The dose response relationship for VPA further strengthens the causal association between intrauterine exposure and developmental outcomes. It is debatable whether exposure to a single ASM at high dose is more harmful than two drugs at low doses in terms of developmental outcome. In this cohort, polytherapy including and excluding VPA was associated with a substantial risk of developmental devel-opmental impairment. A network meta-analysis also demonstrated significant impairment of intelligence, psychomotor development and behavioural problems with polytherapy including and excluding VPA [23]. More observational data are required to support the hypothesis of the effectiveness of two ASMs at low dose for the management of seizures in women of reproductive age.

The precise mechanisms that mediate the developmental adverse effects of ASMs have been investigated. In a population-based study from Sweden, newborns exposed to VPA and CBZ in utero had a smaller head circumference [24]. Reduced grey matter volume and abnormal network connectivity for language nodes were also reported in children exposed to ASMs in utero [25, 26]. Further, voxel-based morphometric magnetic resonance imaging (MRI) showed focal thinning of the cerebral cortex over language areas in exposed children [27]. Neuronal apoptosis is an important pathway for ASM-induced brain changes that eventually lead to neurocognitive deficits. In animal experimental studies, mice that were exposed to VPA, PB or PHT in utero showed a relatively higher level of neuronal apoptosis [28], whereas those exposed to LEV, LTG and CBZ did not show similar changes [29].

The strength of our study is the large prospective cohort of children with well characterized antenatal ASM exposure status and detailed documentation of early developmental outcome measures. In our study, we were able to compare between exposure to newer ASMs, older ASMs and no exposure to any ASMs. The developmental outcomes determined at one year, however, may not be entirely representative of delays in several domains, especially language-related developmental delays. Although the tool used for assessment was validated in Indian settings, the scores generated from this tool for different domains may not be directly comparable to scores generated from other international tools. Non-availability of ASM plasma concentration in pregnancy is an important limitation in interpreting observations related to dose response.

Conclusion

Our study strengthens the hypothesis that the negative impact of antenatal exposure to ASM is specific to certain ASMs, such as VPA and PB. Newer ASMs, such as LEV and LTG, are relatively safer in preventing developmental impairments in children. It is important for children of WWE exposed to VPA, PB and polytherapy to be examined at 12 months in order to detect any developmental delay, as the negative developmental impact of such exposure can be detected as early as 12 months. Our findings offer an opportunity to make pregnancies safer by avoiding the use of VPA and other high-risk ASMs in women of childbearing potential.

Key points
  • Exposure to valproate, phenobarbitone and polytherapy of antiseizure medications during pregnancy is associated with a relatively high risk of developmental delay.
  • Infants exposed to lamotrigine or levetiracetam in utero did not show developmental impairment.
  • Developmental quotients showed a dose-dependent association with valproate and phenobarbitone with a similar trend for other antiseizure medications.

Supplementary material.

Supplementary tables and summary slides accompanying the manuscript are available at www.epilepticdisorders.com.

TEST YOURSELF
  • (1)Is the motor and mental development of infants from pregnancies of women with epilepsy exposed to ASM in utero inferior to that of unexposed infants?
  • (2)Which ASMs may impact infant development?
  • (3)What is the developmental outcome of infants exposed to newer ASMs, such as lamotrigine or levetiracetam?
Note: Reading the manuscript provides an answer to all questions. Correct answers may be accessed on the website, www.epilepticdisorders.com.

Acknowledgements and disclosures.

The corresponding author (SVT) received research grants from Government of India through Indian Council for Medical Research, Department of Science and Technology and Department of Biotechnology. SVT received royalty / honorarium from BMJ India and Wiley India for lectures that had no bearing on this study. None of the other authors have any conflicts of interest to disclose.

Financial support

The Kerala Registry of Epilepsy and Pregnancy received a research grant from Indian Council of Medical Research (ICMR; approval No. ICMR 5/4-5/152/Neuro/2015-NCD-1) for its maintenance.


This work has not been presented at any meetings. A preliminary report -corresponding to approximately a third of the present data set- was published in Epilepsy and Behaviour in 2008 (Thomas SV, Ajaykumar B, Sindhu K, Nair MK, George B, Sarma PS. Motor and mental development of infants exposed to antiepileptic drugs in utero. Epilepsy Behav. 2008 Jul;13(1):229-36.)