Folate transfer across the placenta during late pregnancy in women with epilepsy: A cross‐sectional, two‐center study

Abstract

Objective

In women with epilepsy who are treated with antiseizure medications (ASMs), folate concentrations in maternal serum may not be a good indicator of fetal folate supply, because ASMs can interfere with folate handling by the placenta. We aimed to assess the transplacental folate transfer at birth in persons with epilepsy in comparison to controls with no known epilepsy and identify factors affecting it.

Methods

This was a cross‐sectional, two‐center study, involving 22 pregnant women with epilepsy treated with ASMs, 10 nontreated pregnant women with epilepsy, and 19 control pregnant individuals with no known epilepsy. Maternal venous blood and umbilical cord blood samples were collected at delivery. Folate concentrations were measured using a validated chemiluminescence assay. ASM concentrations were analyzed in serum and umbilical cord blood.

Results

Maternal and neonatal characteristics were generally comparable across the study groups, but cesarian sections were more frequent among controls (p < .001). Folate concentrations in maternal serum were lower than recommended levels in nine of 26 (35%) of women with epilepsy and eight of 19 (42%) controls. Median cord/maternal serum folate ratios were 1.51 (95% confidence interval [CI] = .92–6.35; n = 17) in women with epilepsy treated with ASMs and 2.02 (95% CI = 1.69–3.71; n = 19) in controls. In women with epilepsy, placental folate transfer was saturable (Cmax = 131.1 ng/mL, 95% CI = 96.4–190.5 ng/mL; Michaelis‐Menten constant, Km = 32.6 ng/mL, 95% CI = 12.5–76.9 ng/mL; adjusted R ² = .70). Predictability of cord values improved upon exclusion of preterm births (Cmax = 125.2 ng/mL, 95% CI = 97.4–164.7 ng/mL; Km =41.1 ng/mL, 95% CI = 22.4–71.1 ng/mL; adjusted R ² = .78).

Significance

Folate delivery to the fetus during late pregnancy is primarily explained by its concentrations in maternal circulation. In persons with epilepsy, it may be saturable. The frequent apparent folate deficiency observed in our cohort highlights the importance of folate supplementation and monitoring of its levels in maternal circulation throughout pregnancy.

Key points.

• Our study aimed to assess folate transfer across the placenta at term in persons with epilepsy and identify factors that affect it.

• Maternal folate levels explained 79% of the variation in folate delivery to the fetus, regardless of epilepsy and antiseizure medication treatment.

• In individuals with epilepsy, the late‐pregnancy capacity of placental folate transfer was limited and did not further increase at high maternal folate concentrations.

• In 35% of women with epilepsy and 42% of controls, serum folate concentrations were below recommended values.

• Folate levels in maternal serum should be monitored throughout pregnancy to guide supplementation.

1. INTRODUCTION

Periconceptional folate supplementation reduces the risk of neural tube defects (NTDs) and other congenital abnormalities. ¹ , ² For most women, the daily recommended dose of folic acid is .4 mg. Higher folate doses, up to 5 mg/day, are often advised in the presence of risk factors for congenital malformations. ¹ This practice is based on a study from 1991 demonstrating that 4 mg/day of folic acid reduces NTD risk in patients with a history of NTD compared with other vitamins, both, or neither. ³ Recommendations for high‐dose folate apply to persons with epilepsy because in utero exposure to some antiseizure medications (ASMs) has been associated with NTDs and other adverse pregnancy outcomes. ⁴ , ⁵ In addition, ASMs including valproic acid, ⁶ carbamazepine, ¹ , ⁷ and lamotrigine ⁸ can interfere with folate metabolism.

Even when folate concentrations in maternal serum are adequate, they might not be a good indicator of fetal supply, because ASMs can modulate the expression or activity of placental mechanisms of folate transport. Valproic acid, levetiracetam, and phenytoin downregulate the expression of a principal placental folate carrier, folate receptor α (FRα). ⁹ , ¹⁰ Valproic acid also directly inhibits this carrier. ¹¹

The current study was conducted to test the hypothesis that ASMs could disrupt folate transfer to the fetus. Accordingly, our main objective was to assess folate transfer across the placenta at birth in persons with epilepsy compared to controls without epilepsy and identify factors affecting it.

2. MATERIALS AND METHODS

2.1. Study design and setting

The study was conducted at two tertiary‐care centers in Israel, the Soroka and Meir Medical Centers, between April 2022 and May 2024. It adhered to the ethical principles outlined in the Declaration of Helsinki and was approved by the Helsinki Committees of Soroka (approval code 0257‐21‐SOR, October 2021) and Meir (approval code 0240‐21‐MMC, February 2022) Medical Centers. The privacy rights of human subjects have been observed, and all participants were fully informed about the study's purpose and procedures. Written informed consent was obtained from all participants before delivery. Blood and tissue samples of additional participants, along with their clinical information, were obtained through the Negev BioBank (https://negevbiobank.org.il/). Data collection involved secure storage in a pseudonymized Microsoft Excel database accessible only to the research team.

Participants were eligible for enrollment if they were between 18 and 45 years old. Our approved protocol included comparisons among persons treated with lamotrigine, levetiracetam, lacosamide, carbamazepine, or valproic acid, or not treated with ASMs, and participants were recruited accordingly. Exclusion criteria were persons younger than 18 and older than 45 years, those who could not provide informed consent, use of folate antagonists, multiple gestations, known genetic disease of the newborn, and known mutations in the methylenetetrahydrofolate reductase (MTHFR) gene. The data were obtained from the participant medical records at the medical centers. By the end of the predefined study period, 22 women with epilepsy who received ASMs, 10 women with epilepsy not treated with ASMs, and 19 control participants without epilepsy or ASM treatment were enrolled.

2.2. Serum and tissue collection

Immediately after delivery, we collected 20‐mL samples of venous cord blood and maternal venous blood according to the World Health Organization Guiding Principles on Human Cell, Tissue and Organ Transplantation. ¹² Maternal and cord blood samples were collected simultaneously, up to 15 min apart. Samples were allowed to clot at room temperature for 30 min, then centrifuged at 2500 × g for 10 min at 4°C. The serum was aliquoted and stored at −80°C until analysis. Additional maternal and cord serum samples were obtained from the Negev BioBank. ¹³

2.3. Folate quantification

Total folate concentrations in serum were measured by the Atellica IM Folate chemiluminescence assay (Siemens Healthcare Diagnostics; Supporting Methods).

2.4. Determination of ASM concentrations in serum

Blood samples were centrifuged, and serum was stored at −40°C, transported on dry ice, and analyzed in one run at the Section for Clinical Pharmacology, National Center for Epilepsy, Department of Pharmacology, Oslo University Hospital. Serum concentration measurements of ASMs were analyzed by ultra‐high‐pressure liquid chromatography with mass spectrum detection on a Vanquish Quantis UPLC‐MS/MS instrument from Thermo Fisher Scientific, with the ClinMass TDM Platform from RECIPE (MS9000) with add‐on sets for ASMs and benzodiazepines. ¹⁴ The serum concentrations were compared with recently updated reference ranges. ¹⁵ The conversion factors from μmol·L⁻¹ to mg/L were calculated from Patsalos et al. ¹⁶

2.5. Outcome measures

The primary outcome variable was the cord/maternal folate concentration ratio of total folates, indicating the placental folate transfer. The primary analysis was a cross‐group comparison of the placental folate transfer and its kinetics between ASM‐treated women with epilepsy and those without epilepsy.

2.6. Secondary and sensitivity analyses

As secondary analyses, we compared across groups the folate concentrations in maternal and cord serum. We additionally assessed the effects of maternal age, weight, body mass index, birth weight, the newborn sex, and ASM concentrations on placental folate transfer, and compared between groups the maternal concentration‐normalized folate ratios.

Potential confounders were the presence of pregestational or gestational diabetes, treatment with the highly teratogenic ASM valproic acid, delivery by cesarian section, folate levels in maternal serum higher than 60.9 ng/mL (the highest value obtained in controls), and nontreated epilepsy. Sensitivity analyses were conducted by excluding individuals with each of these conditions. Missing data were not included in the analyses.

2.7. Statistical analysis

The nonparametric tests for comparing distributions between two or more than two groups, Mann–Whitney and Kruskal–Wallis, respectively (the latter followed by Dunn multiple comparisons test), and Fisher exact test (for frequency comparisons) were applied to demographic variables. Folate concentrations and placental transfer were compared between groups by the Mann–Whitney or Kruskal–Wallis tests. The relationships between maternal factors and cord folate or cord/maternal ratios were described by linear regression, Michaelis–Menten kinetics (for cord serum–maternal serum relationships), and one‐ and two‐phase decay models (for gestational age–folate transfer relationships; Supporting Methods).

All analyses were performed using Prism 10.4.1 (GraphPad). p < .05 was considered significant. Unless otherwise stated, data are shown as median with 95% confidence interval (CI).

3. RESULTS

3.1. Participant and sample disposition

Fifty‐one women were enrolled in the study. Samples of 11 participants were obtained from the Nege BioBank, of which two maternal serum and two cord serum samples of ASM‐treated women, as well as four samples each of maternal and cord serum of nontreated women with epilepsy, were missing (Table 1). Accordingly, cord‐to‐maternal ratios of folate concentrations could be calculated for 19 controls, three women with nontreated epilepsy, and 18 women with pharmacologically treated epilepsy.

TABLE 1.

Participant disposition.

Treatment group	No known epilepsy	Epilepsy without ASM treatment	Epilepsy with ASM treatment
Total number	19	10	22
Participants whose samples were obtained from the Negev Biobank	0	7	4
Maternal samples	19	6	20
Cord samples	19	6	20
Folate cord‐to‐maternal ratios, n values based on documentation in the medical records a	19	3	18
Folate cord‐to‐maternal ratios, n values based on measurements of ASM concentrations (after allocation of Participant 13 to the nontreated group) b	19	4	17
Preterm births	2	1	4
Folate cord‐to‐maternal ratios, term births after allocation of Participant 13 to the nontreated group	17	3	13

Abbreviation: ASM, antiseizure medication.

^a The samples obtained from six participants with nontreated epilepsy included either maternal or fetal serum (three each). Both maternal and cord samples of an additional participant were missing. Accordingly, fetal‐to‐maternal ratios could be calculated only for three participants. The samples obtained from four participants with ASM‐treated epilepsy included either maternal or cord serum (two each). Therefore, the cord‐to‐maternal ratios could be calculated for only 18 participants.

^b The measurement of ASM concentrations in serum demonstrated that Participant 13 did not receive the documented lamotrigine treatment, and this individual was reallocated to the nontreated group.

3.2. Participant and pregnancy characteristics

The majority of clinical characteristics and pregnancy outcomes did not significantly differ across treatment groups, but caesarian deliveries were significantly more prevalent among controls due to difficulties in recruitment (Table 2). In the entire cohort, 13 women had medical conditions other than epilepsy, including diabetes mellitus type 2 (two women), gestational diabetes mellitus (two), hypertension (one), and hypo‐ or hyperthyroidism (one each; Table S1). Data on periconceptional folate consumption, epilepsy, and seizures were missing.

TABLE 2.

Participant and pregnancy characteristics.

Treatment group	No known epilepsy	Epilepsy without ASM treatment	Epilepsy with ASM treatment	p
n	19	10	22
Maternal age, years	31 (28.7–34.0)	30 (26.2–33.3)	29 (26.2–31.2)	.25
Maternal height, cm	163.0 (161.3–167.5)	159.5 (155.8–167.0)	161.5 (158.0–163.5)	.20
Maternal weight, kg	80.0 (73.5–86.7)	68.5 (63.3–88.7)	76.0 (69.2–82.0)	.55
Body mass index, kg/m2	27.0 (26.5–31.4)	27.6 (25.6–32.3)	29.3 (26.4–31.2)	.99
Arab or Bedouin ethnicity, n (%)	14 (74)	4 (40)	9 (41)	.09
Gestational age, weeks	38.6 (37.9–39.2)	38.9 (37.6–39.9)	38.4 (37.2–39.3)	.83
Preterm delivery, number of cases (%)	2 (11)	1 (10)	4 (18)	.87
Birth weight, g	3305 (2998–3493)	2875 (2653–3339)	3123 (2736–3274)	.47
Cesarean section, number of cases (%)	14 (74)	1 (10)	3 (14)	<.001
Female infant, number of cases (%)	7 (37)	6 (60)	13 (59)	.30
Apgar at 1 min	9.0 (8.3–9.2)	9.0 (8.7–9.1)	9.0 (8.2–9.1)	.89
Apgar at 5 min	10.0 (9.8–10.1)	10.0 (10.0–10.10)	10.0 (9.6–10.1)	.58

Note: The categorization is based on the medical records. The folate data of one participant originally classified as ASM treated were handled as ASM nontreated based on measurement of lamotrigine concentrations (see Tables 3 and S1). Values are median with 95% confidence interval, unless otherwise stated. Ethnicity and incidence of preterm delivery were compared using Fisher exact test. The Kruskal–Wallis test was used for all other comparisons.

Abbreviation: ASM, antiseizure medication.

ASM doses were recorded in the medical files for 17 (77.3%) of the 22 treated women (Table S1). Among women with epilepsy receiving ASMs, 16 were treated with monotherapy, and six received dual therapy. Drug concentrations were measured in 11 maternal and cord sample pairs and indicated that one woman listed as lamotrigine‐treated was not taking lamotrigine (Participant 13; Tables 3 and S1). Accordingly, Participant 13 was recategorized into the nontreated epilepsy group.

TABLE 3.

Concentrations of antiseizure medications (ng/mL) in a subset of maternal and cord serum samples.

No.	Carbamazepine			Lamotrigine			Levetiracetam			Valproic acid
	M	C	R	M	C	R	M	C	R	M	C	R
3				3.1	2.5	.80	18.6	17.4	.94
9	1.7	1.8	1.03
10				3.2	3.3	1.01	10.1	10.6	1.05
11	2.4	1.4	.60
12	7.0	5.0	.72
13				0	0
15				6.0	4.6	.77	30.2	27.4	.91
39				2.0	1.9	.96
42				5.7	6.0	1.05				31.5	49.4	1.57
43	3.4	2.5	.73
44				2.2	2.3	1.06

Note: Conversion factors from μmol·L−1 to ng/mL are based on Patsalos et al. ¹⁶

Abbreviations: C, cord serum; M, maternal serum; R, cord/maternal serum concentration ratio.

3.3. Folate delivery to the fetus in women with epilepsy is comparable to that in controls

Folate concentrations in maternal sera did not differ between women with ASM‐treated epilepsy and controls, although values in the epilepsy group were more variable (Figure 1A). Nine (35%) maternal folate concentration values in individuals with epilepsy (including epilepsy not treated with ASMs) and eight (42%) in controls were below the well‐studied levels associated with protection against folic acid‐related NTDs in the general population (13–14 ng/mL). ¹ The values in cord serum were higher in the ASM‐treated women than in controls (p = .04; Figure 1B,E). The placental folate transfer (indicated by cord/maternal serum ratios) in ASM‐treated women with epilepsy did not differ from that of controls, and nearly all values were within those observed in women without epilepsy (Figure 1C). The only exception was a ratio of .48, obtained in the woman whose serum folate concentration was the highest (241.0 ng/mL). The values of all participants are presented in Figure 1D–F.

FIGURE 1.

Cross‐group comparisons of folate transfer across the placenta. (A–C) Comparisons between antiseizure medication (ASM)‐treated women with epilepsy and controls, Mann–Whitney test. (D, E) Comparisons across the ASM‐treated women with epilepsy, nontreated women with epilepsy, and controls by Kruskal–Wallis analysis. (A, D) Total folate concentrations in maternal serum; n = 19 controls, 7 nontreated women with epilepsy (D only), 19 ASM‐treated women with epilepsy. (B, E) Total folate concentrations in cord blood serum; n = 19 controls, 7 nontreated women with epilepsy (E only), 19 ASM‐treated women with epilepsy. (C, F). Cord/maternal ratio of folate concentrations in serum; n = 19 controls, 4 nontreated women with epilepsy (F only), 17 ASM‐treated women with epilepsy. The cohort of nontreated women included three individuals initially identified as such and an additional individual (No. 13) who was assigned to this group based on measurements of serum ASM concentrations (Table 1). Horizontal bars denote medians and 95% confidence intervals. Split symbols, women with epilepsy; blue squares, early preterm; blue triangles, late preterm; orange inverted triangles, valproate‐treated women. *p < .05. In panel E, the populations were differently distributed, but pairwise comparisons did not identify statistically significant differences.

3.4. Maternal folate concentrations explain the major portion of the variation in folate transfer to the fetus

Across all participants, folate concentrations in cord serum depended on the levels in maternal serum. This was a saturable process, with a Cmax of 137.1 ng/mL (95% CI = 103.0–188.3 ng/mL; R ² = .69; Figure 2A). Values obtained at early preterm (blue squares) were outside the 95% CI. Saturability was also demonstrated by the exponential decrease in placental transfer with increasing maternal folate concentrations (Figure 2B). Folate transfer decreased as pregnancy progressed toward term, with an initial half‐life of .68 weeks (Figure 2C). Cord folate concentrations did not correlate with the time of delivery (Figure 2D). The cord–maternal concentration relationships in the subgroup of women with epilepsy were better described by Michaelis–Menten kinetics than by linear regression (Figure 2E, Table S2), whereas control values presented linearity (Figure 2F, Table S2).

FIGURE 2.

Factors affecting folate transfer across the placenta. (A) Association between cord folate concentrations and maternal concentrations, Michaelis–Menten kinetics, n = 40. (B) Cord/maternal serum folate concentrations versus maternal concentrations, two‐phase decay, n = 40. The inset is a zoom‐in for narrower ranges of maternal concentrations and cord/maternal ratios. (C) Association between cord/maternal folate concentrations and gestational age, one‐phase decay, n = 40. (D) Association between cord folate concentrations and gestational age, linear regression, n = 40. (E, F) Associations between cord folate concentrations and maternal folate concentrations in the subgroups of women with epilepsy (E; antiseizure medication treated and nontreated; Michaelis–Menten kinetics, n = 21) and women without epilepsy (F; linear regression, n = 19). Cmax, maximal concentration; Km, Michaelis‐Menten constant. Shown are the curve fits with 95% confidence intervals (CIs). Split symbols, women with epilepsy; blue squares, early preterm; blue triangles, late preterm; orange inverted triangles, valproate‐treated women.

Placental folate transfer in the entire cohort and among women with epilepsy did not correlate with maternal age, weight, or body mass index (Figure S1A–C). The relationships with newborn weight became nonsignificant when preterm values were excluded (Figure S1D,E). Similar results were obtained for women with epilepsy (Figure S1F–J). Values were not affected by the newborn sex (p > .05; Figure S1K).

The differences between the effects of individual ASMs on cord/maternal ratios of folate were not assessed due to the small numbers in each treatment group (Figure S2A). Across all lamotrigine‐treated women, the cord/maternal ratios of folate did not appear to be associated with lamotrigine concentrations (Figure S2B). ASM transfer to cord blood in Participants 11 and 42, who exhibited the highest cord/maternal ratios of folate concentrations (22.3 and 8.2, respectively), was within the ranges observed in the other participants (Table 3) or those reported in the literature. ¹⁷ This includes valproic acid, which (unlike other ASMs) tends to accumulate in its protein‐bound form in the fetal compartment, likely due to higher third‐trimester concentrations of albumin in fetal than in maternal serum. ¹⁷

3.5. Sensitivity analyses

When preterm births were excluded, the cross‐group differences in maternal folate concentrations became significant (Figure 3A), but the cord/maternal folate ratios remained comparable across groups (Figure 3C). In addition, the predictability of cord folate concentrations improved (R ² = .78 vs. .69; narrower 95% CI; Figure 3D). We did not proceed to a multivariate model that incorporates gestational age given the paucity of preterm values.

FIGURE 3.

Folate transfer across the placenta in women with and without epilepsy, preterm births excluded. (A) Total folate concentrations in maternal serum; n = 17 controls, 15 antiseizure medication (ASM)‐treated women with epilepsy. (B) Total folate concentrations in cord serum; n = 17 controls, 15 ASM‐treated women with epilepsy. (C) Cord/maternal ratio of folate concentrations in serum; n = 17 controls, 13 ASM‐treated women with epilepsy. (A–C) Mann–Whitney test. Shown are medians with 95% confidence intervals. (D) Association between cord folate concentrations and maternal concentrations, Michaelis–Menten kinetics, n = 33. See Table 1 for details on the number of values in each treatment group. Cmax, maximal concentration; Km, Michaelis‐Menten constant. Curve fits with 95% confidence intervals. *p < 0.05. Split symbols, women with epilepsy; orange inverted triangles, valproate‐treated women.

Values obtained upon exclusion of women with diabetes, those who underwent cesarian sections, were treated with valproic acid, were not ASM‐treated, or had folate concentrations higher than 60.9 ng/mL (for comparability in modeling with the control group) were consistent with the primary results (Figure S3).

4. DISCUSSION

Maternal folate level was the major parameter explaining the variability in folate transfer to the fetus at term, regardless of epilepsy or ASM treatment. Unexpectedly, nearly all folate concentration values of ASM‐treated women with epilepsy and their newborns were within the range obtained in controls. The findings did not change when potential confounders were considered. In women with epilepsy, folate transfer to cord blood was saturable, suggesting that very high folate concentrations in maternal circulation do not further increase the fetal exposure. Another key finding was the high incidence of apparent maternal folate deficiency, in the presence of ASM concentrations in maternal serum below or within the reference ranges in women with epilepsy, as well as in controls. ¹⁵

To our knowledge, this is the first report of folate transfer across the placenta in persons with epilepsy. Similar to previous in vivo ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ and ex vivo ²⁶ studies, we found higher folate concentrations in fetal serum than in maternal circulation in most women, regardless of the treatment group. The relationships in women with epilepsy were characterized by Michaelis–Menten kinetics, which could be revealed as a result of the wide range of folate concentrations in maternal serum. This finding is in contrast to the linearity in the transfer of 5‐methyl tetrahydrofolate (5‐MTHF; 83% of total folates in maternal circulation ²⁵ ) across perfused placentas from healthy women at concentrations of up to 222 ng/mL ²⁷ or 444 ng/mL. ²⁶ In vivo, the kinetics of folate transfer to the fetus were previously reported to be linear at maternal folate levels of up to ~60 ng/mL. ¹⁸ , ¹⁹ , ²⁰ , ²³ Similar results were obtained in our study across women without epilepsy, but not in the epilepsy group.

4.1. Suggested mechanisms of saturable folate transport

A primary placental carrier that mediates folate uptake from maternal blood is FRα. ²⁸ This receptor transfers across membranes both folic acid (the synthetic form of folate used in supplements and fortified foods) and its metabolically active product, 5‐MTHF. ² Folates, particularly at their reduced forms, are also transported across both the maternal‐ and fetal‐facing membranes of placental syncytiotrophoblasts by the reduced folate carrier. ²⁹ Downregulation or blockade of these carriers may restrict folate uptake from maternal serum and/or folate transfer from the placenta to the fetal circulation. In our previous studies in perfused human placentas, valproic acid ¹⁰ and lacosamide ³⁰ downregulated the expression of FRα by up to 72% and up to 50%, respectively, but the duration of these studies was limited to 3 h. More recently, we found that valproic acid does not impair folate transfer across monolayers of trophoblast cells, ³¹ in line with the results observed in the two valproate‐treated participants in the current study. Carbamazepine, lamotrigine, and levetiracetam can also affect folate carrier expression, although to a lesser extent than valproic acid. ⁹

Another placental folate‐transporting protein is the proton‐coupled folate receptor, whose role is mostly in modulating intracellular folate kinetics. Saturation of this carrier is therefore less likely to underlie the observed Michaelis–Menten kinetics. ²⁸ The placenta additionally expresses folate‐metabolizing enzymes. Their downregulation may interfere with the conversion of folic acid (whose placental transfer is saturable at low concentrations ²¹ , ³² ) to 5‐MTHF, thereby potentially reducing the transfer of total folates to the fetus.

Most cord/maternal folate values obtained in women with epilepsy were within those of controls, and the low extremities were obtained at high maternal levels as a consequence of the aforementioned saturation kinetics. However, given the effects of some ASMs on placental transport or folate metabolism at their therapeutic concentrations, ¹ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ we cannot rule out potential interference of these ASMs with folate delivery to the fetus. Yet our results suggest that appropriate folate supplementation throughout pregnancy may overcome this issue.

4.2. Clinical implications

Most guidelines relating to pregnant persons with epilepsy recommend high‐dose folate during the first trimester of pregnancy, but referral to late pregnancy is conflicting and often vague. ¹ In Israel, recommendations for high‐dose folate do not extend beyond the first 3 months of pregnancy, ³³ and low‐dose folic acid is consumed as over‐the‐counter preparations. The high frequency of apparent folate deficiency in maternal serum at term identified in our study calls for greater attention to folate supplementation during late pregnancy, when folate levels tend to decline ²⁴ and awareness might be lower than during the first trimester of pregnancy.

The World Health Organization published recommendations for folate concentrations in red blood cells in women of reproductive age for preventing NTDs, ³⁴ but has not referred to serum levels. In addition, the recommendations apply to the general population and not necessarily to women with epilepsy. There is also a controversy about the optimal dosing of folic acid. ³⁵ , ³⁶ High‐dose periconceptional folic acid (≥1 mg daily) has been linked with an increased risk of cancer in women who have given birth ³⁷ and in children (although the latter has been debated), ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ and neither observational study has yet been reproduced. In utero exposure to high‐dose folate in the general population has additionally been associated with an increased risk of congenital heart disease, ⁴² autism spectrum disorder, ⁴³ and adverse neurocognitive outcomes. ⁴⁴ Some of these effects have been attributed to unmetabolized folic acid, which accumulates in plasma when given at excessive doses. ¹ On the other hand, folate deficiency during pregnancy can result in adverse outcomes other than structural teratogenicity, such as maternal anemia. ¹ Specifically in populations of persons with epilepsy, folate supplementation has been associated with reduced risk of preterm birth ⁴⁵ and improved cognitive and behavioral outcomes of children exposed in utero to ASMs. ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹

Although our findings provide reassuring information on the effects of epilepsy and its treatments on folate transfer across the placenta, they relate only to the last weeks of gestation and are based on small participant numbers. Whether epilepsy and ASMs affect folate handling by the placenta during the first trimester, when folate can prevent NTDs, has yet to be established. Future studies should also assess the relationships between folate doses, concentrations, and effects before further folate dosing recommendations can be drafted.

4.3. Limitations and strengths

The study has several limitations. First, it was underpowered for detecting epilepsy and ASM effects on placental folate transport. Yet modeling provided important information on the kinetics of folate transfer to the fetus. A larger sample size and exposure to a wider ASM selection could have helped establish the saturation phenomenon with greater confidence, allow comparisons between ASMs, and improve generalization. Second, participants were not recruited systematically. However, selection bias is less likely because the medical teams were not aware of the folate status of the participants. Third, although documented information on folate intake was absent or minimal, the primary parameter was placental folate transfer, which is driven by its measured concentrations in maternal plasma. Fourth, data on epilepsy were scarce, calling for efforts to increase the awareness of medical teams of this issue. Nevertheless, parameters such as seizure type or frequency cannot currently be considered confounders of placental folate transfer in the absence of support in the literature, and ASM treatment confirmed the diagnosis of epilepsy in the main treatment group. Fifth, all folates can bind the folate‐binding protein used in the biochemical assay, making the assay nonspecific for individual folates, as described above. ⁵⁰ , ⁵¹ Lastly, the studied population was heterogeneous. Specifically, preterm deliveries were not excluded, based on the previously demonstrated comparable folate transfer to the fetus from week 30 to term. ²⁴ Other variables, including diabetes and mode of delivery, did not affect the results (Figure S3).

Strengths of our study include the representation of individuals in a realistic setting with concomitant diseases and medications, which allowed the analysis of distinct sources of variability; the simultaneous collection of maternal and cord blood, which increased the confidence in the cord/placental ratios; and the analysis of ASM concentrations in a patient subset. The latter allowed demonstrating that enhanced preterm transfer was folate‐specific and not related to a nonselective malfunction of the placental barrier.

5. CONCLUSIONS

At birth, folate concentrations in cord serum are primarily driven by the levels in maternal serum in both ASM‐treated women with epilepsy and those without epilepsy. Maternal folate concentrations below the recommended values were frequent, whereas very high maternal folate levels only marginally added to fetal exposure. These findings underscore the need for careful monitoring and personalized folate treatment strategies in pregnant persons with epilepsy. Our findings also call for future research and collaborations with larger participant numbers and women at earlier gestational ages.

AUTHOR CONTRIBUTIONS

Erez Berman: Study concept and design; sample collection and processing; data curation; methodology; writing. Gali Pariente: Patient recruitment. Natalia Erenburg: Sample processing. Roa'a Hamed: Sample processing. Cecilie Johannessen‐Landmark: Sample processing; data curation; methodology; writing; critical revision of manuscript for intellectual content. Michal Kovo: Study concept and design; sample collection; patient recruitment; critical revision of manuscript for intellectual content. Sara Eyal: Study concept and design; critical revision of manuscript for intellectual content.

FUNDING INFORMATION

The study was funded by the Israel Science Foundation (Award No. 2054/18, S.E, M.K, and C.J.L). The funding agency had no role in the study design, collection, analysis and interpretation of data, writing of the report and decision to submit the article for publication.

CONFLICT OF INTEREST STATEMENT

S.E. has served as a consultant for Dexcel Pharma, Israel. C.J.L. has received speaker honoraria from Angelini, Eisai, Jazz, and UCB Pharma and scientific advisory board honoraria from Jazz Pharma Nordic. The other authors declare no competing interests. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Supporting information

Data S1.

Appendix S1.

Berman E, Pariente G, Erenburg N, Hamed R, Landmark CJ, Kovo M, et al. Folate transfer across the placenta during late pregnancy in women with epilepsy: A cross‐sectional, two‐center study. Epilepsia. 2025;66:4752–4763. 10.1111/epi.18575

DATA AVAILABILITY STATEMENT

The available anonymized participant data are available in Table S1.