Serum vitamin D level and its association with anti-seizure medications in children with seizure disorder

Abstract

BACKGROUND

Vitamin D deficiency is frequently reported in children with epilepsy and may be influenced by antiseizure medications (ASMs). However, prospective longitudinal data evaluating this relationship remain limited.

AIM

To assess longitudinal changes in serum vitamin D levels and examine their association with seizure control in children receiving ASMs.

METHODS

In this prospective longitudinal study, children aged 1-14 years with seizure disorders on ASM therapy were enrolled and followed for 18 months. Serum 25-hydroxyvitamin D levels were measured at baseline, 6 months, and 18 months using chemiluminescent immunoassay. Seizure control was defined as seizure freedom for ≥ 6 months prior to evaluation. Longitudinal trends were analyzed using linear mixed-effects models, and multivariable regression was performed to adjust for potential confounders.

RESULTS

A total of 150 children were included (44.7% aged 1-5 years; 64.7% male). Mean serum vitamin D levels showed a declining trend from 21.33 ng/mL at baseline to 18.61 ng/mL at 18 months; however, this change was not statistically significant (P = 0.65). Longitudinal analysis also did not demonstrate a significant trend (P = 0.41). Low vitamin D levels (insufficiency + deficiency) were observed in 80.7% of participants. No independent association was found between vitamin D levels and seizure control (P = 0.054).

CONCLUSION

Vitamin D insufficiency is highly prevalent among children with epilepsy receiving ASMs. Although a non-significant declining trend in vitamin D levels was observed over time, no independent association with seizure control was identified. These findings are associative and hypothesis-generating.

Key Words: Pediatric epilepsy; Vitamin D; Antiseizure medications; Seizure control; 25-hydroxyvitamin D

Core Tip: In this prospective longitudinal study of 150 children with epilepsy, vitamin D insufficiency was highly prevalent, affecting over 80% of participants. Although serum vitamin D levels showed a declining trend during follow-up, the change was not statistically significant, and no independent association with seizure control was observed. These findings suggest that while hypovitaminosis D is common in this population, its direct impact on seizure outcomes remains uncertain and warrants further controlled studies.

INTRODUCTION

Epilepsy affects approximately 0.5%-1% of the pediatric population worldwide[1]. Long-term therapy with antiseizure medications (ASMs) remains the cornerstone of management but is associated with metabolic adverse effects, including disturbances in bone and mineral metabolism[2].

Beyond its established role in skeletal health, vitamin D has been increasingly implicated in neuronal function and seizure susceptibility[3]. Vitamin D receptors are widely expressed in the brain, particularly in regions such as the hippocampus that are involved in seizure generation[4]. Mechanistically, vitamin D may modulate neuronal excitability through regulation of calcium homeostasis, enhancement of inhibitory GABAergic activity, suppression of excitatory glutamatergic transmission, and reduction of neuroinflammation and oxidative stress[5]. Experimental and clinical studies have suggested that vitamin D deficiency may increase neuronal excitability, while supplementation has been associated with reduction in seizure frequency in some patient populations[6].

Vitamin D deficiency is highly prevalent among children with epilepsy, with reported rates ranging from approximately 45% to over 60% across different populations[7,8]. These variations are based on geographic region, age group, and ASM exposure[7,8]. Studies from Asian and Middle Eastern populations have demonstrated even higher prevalence rates, often exceeding 60%, likely reflecting differences in nutritional status, sunlight exposure, and socioeconomic factors. The reasons are multi-factorial including reduced outdoor activity, dietary insufficiency, comorbid neurological conditions, and the metabolic effects of ASMs. Several ASMs, particularly enzyme-inducing agents such as carbamazepine, phenytoin, and phenobarbital, accelerate hepatic cytochrome P450 metabolism of vitamin D, leading to reduced serum 25-hydroxyvitamin D [25(OH)D] levels and altered calcium metabolism[9]. In addition to drug-related effects, the duration of therapy, polytherapy, limited sun exposure, and nutritional status may influence vitamin D levels in children with epilepsy[10,11].

Given the potential implications of hypovitaminosis D on skeletal health and possibly seizure control, evaluation of vitamin D status in children receiving ASMs has important clinical relevance. However, prospective data assessing the relationship between serum vitamin D levels and ASM therapy in children with seizure disorders remain limited, particularly in developing countries. Therefore, the present prospective longitudinal study was conducted to determine the serum vitamin D levels in children with seizure disorders and to evaluate their association with the use of ASMs.

MATERIALS AND METHODS

Study design

This prospective longitudinal study was conducted in the pediatrics department (a 200 bedded hospital dedicated for children), which is attached to a tertiary care teaching institute in Eastern India. The study duration was from July 2023 to June 2025. The study was approved by the Institutional Ethics Committee. Written informed consents were taken from the parents or care givers, and assents were taken from children > 7 years age. The study is reported as per the declaration of Helsinki.

Patients

Inclusion criteria included children of 1 to 14 years of age presenting either to the outpatient department or admitted to the inpatient department (IPD) with history of seizure and on antiseizure medication (ASM) for at least 3 months. They were followed over 18 months with serial measurements of serum vitamin D at predefined intervals (baseline, 6 months, and 18 months). These time points (baseline, 6 months, 18 months) were selected to capture short-term and intermediate longitudinal changes. Children with known metabolic bone disease or chronic renal or hepatic disorders, endocrine disorders affecting vitamin D metabolism, receiving vitamin D or calcium supplementation during the preceding three months, or medications known to influence bone metabolism such as systemic corticosteroids, and refusal to participate were excluded.

Case management

After history and clinical examinations, the cases underwent a series of investigations that included complete blood count, arterial or venous blood gas analysis, liver and renal function tests, blood glucose level, magnetic resonance imaging brain, electroencephalogram, and evaluation of underlying neurometabolic diseases (in selected cases). Eye, hearing and psychological evaluation were done as part of the work up.

Venous blood samples were collected from all participants under aseptic conditions. Serum levels of [25(OH)D] were measured using a standardized chemiluminescent immunoassay method in the hospital biochemistry laboratory. Serum calcium, phosphorus, and alkaline phosphatase levels were also measured to assess bone metabolism parameters.

Vitamin D status was classified according to widely accepted criteria: Deficiency: Serum 25(OH)D < 20 ng/mL; insufficiency: 20-29 ng/mL; sufficiency: ≥ 30 ng/mL. These cut-offs are commonly used in pediatric studies evaluating vitamin D status[8].

The children were started on ASM as per the International League Against Epilepsy guideline 2017[12]. The primary outcome measure was the prevalence of vitamin D deficiency among children with seizure disorders receiving ASMs. The secondary outcomes included association of vitamin D levels with type of antiseizure medication, duration of therapy, and number of drugs used (monotherapy vs polytherapy).

Seizure control was defined as complete seizure freedom for at least 6 months preceding evaluation, while uncontrolled seizures were defined as persistence of ≥ 1 seizure episode during this period[12]. Seizure outcomes were assessed prospectively during scheduled follow-up visits.

Data collection

Detailed demographic and clinical information were recorded using a structured proforma. Detailed information regarding ASM exposure was recorded, including type, duration, and regimen (monotherapy or polytherapy). Duration of exposure to individual ASMs was calculated in months from initiation to last follow-up. Where available, dose ranges were recorded; however, due to variability in documentation, dose was not included in adjusted analyses. Patients transitioning between therapies were classified based on predominant exposure during the follow-up period. In cases of polytherapy, exposure was categorized based on the presence of enzyme-inducing vs non-enzyme-inducing ASMs, irrespective of combination. ASMs were categorized into two groups: Enzyme-inducing (phenytoin, phenobarbitone, carbamazepine, and oxcarbazepine), and non-enzyme-inducing (sodium valproate, levetiracetam, lamotrigine, and benzodiazepines). Information regarding sunlight exposure, dietary intake, and physical activity was also documented where possible. Socio-economic status was classified as per Indian guideline[13].

Sample size calculation

Previous studies have reported that the prevalence of vitamin D deficiency among children receiving ASMs ranges from 30% to 50%[7,8]. For sample size estimation, an expected prevalence (p) of 40% was assumed based on available literature. The required sample size was calculated using the standard formula for prevalence studies: Where “n” represents the required sample size, “Z” is the standard normal deviate corresponding to a 95% confidence level (1.96), p is the expected prevalence (0.40), “q = 1 - p (0.60)”, and “d” is the allowable absolute precision, set at 8% (0.08).

Substituting these values, the calculated minimum sample size was approximately 144 participants. To account for potential exclusions, incomplete laboratory data, or loss to follow-up, the sample size was increased slightly to 150. This sample size also provided adequate statistical power (approximately 80%) to detect clinically meaningful differences in serum vitamin D levels between important clinical subgroups, such as children receiving monotherapy vs polytherapy or enzyme-inducing vs non-enzyme-inducing ASMs, assuming a moderate effect size and a significance level (α) of 0.05. Therefore, the final sample size of 150 participants was considered sufficient to achieve the objectives of the study.

Statistical analysis

The data were entered into Microsoft excel sheet. The analysis was conducted using MS Excel, SPSS version 22 (IBM SPSS Statistics, Somers NY, United States). Continuous variables were expressed as mean ± SD or median with interquartile range depending on data distribution. Categorical variables were presented as frequencies and percentages. Comparisons between groups were performed using Student’s t-test or Mann-Whitney U test for continuous variables and χ² test or Fisher’s exact test for categorical variables. Analysis of variance was used to compare vitamin D levels among different medication groups where appropriate. For longitudinal analysis of serum vitamin D levels measured at baseline, 6 months, and 18 months, linear mixed-effects models were used to account for within-subject correlations and repeated measurements over time. Time was treated as a fixed effect, and individual participants were included as random effects. Multivariable linear regression analysis was performed to evaluate independent predictors of serum vitamin D levels. The model was adjusted for age, sex, seizure type, epilepsy etiology (where available), nutritional status, developmental delay, socioeconomic status, type of antiseizure medication (enzyme-inducing vs non-enzyme-inducing), and duration of ASM exposure. A P < 0.05 was considered as statistically significant.

RESULTS

Demography & clinical features

A total of 178 children were screened, 28 were excluded, and 150 children were included. Among them, 67 (44.7%) children were between 1 years and 5 years of age, indicating that a substantial proportion of the study participants were young children in early childhood. and 97 (64.7%) were male reflecting a male predominance among children presenting with seizure disorders in our cohort. The details of the study flow have been mentioned in Figure 1. The baseline characteristics of the included children have been described in Table 1. The details regarding seizure and ASM are shown in Table 2. The mean ± SD age at onset of seizure (year) was 6.5 ± 3.9. The median (IQR) duration of seizure prior to start of ASM (month) was 1 (0-6). The mean ± SD duration of ASM (month) was 4.3 ± 1.1. Monotherapy was given to 61 (40.7%) children, and 62 (41.3%) children were on enzyme inducing ASM.

Figure 1 Study flow diagram.

Table 1 Baseline characteristics of the included children (n = 150).

Characteristics	n (%)
Age sub-groups
1-5 years	67 (44.7)
> 5 years	83 (55.3)
Gender
Male	97 (64.7)
Female	53 (35.3)
Birth asphyxia	28 (18.7)
Developmental delay	19 (12.7)
Severe malnutrition	8 (5.3)
Rural residence	84 (56)
Low socioeconomic status	51 (34)

Table 2 Details regarding seizure, antiseizure medication, and baseline vitamin D level.

Characteristics	Values [mean ± SD or median (IQR)]	Values, n (%)
Age at onset of seizure (year)	6.5 ± 3.9	-
Duration of seizure prior to start of ASM (month)	1 (0-6)	-
Duration of ASM (month)	4.3 ± 1.1	-
Baseline vitamin D [25(OH)D] level (ng/mL) in different treatment groups
Monotherapy	21.84 ± 10.1	61 (40.7)
Polytherapy	20.12 ± 9.3	89 (59.3)
Enzyme-inducing ASM1	19.97 ± 9.6	62 (41.3)
Non-enzyme-inducing ASM2	21.15 ± 9.8	88 (58.7)

¹Phenytoin, phenobarbitone, carbamazepine, and oxcarbazepine.

²Sodium valproate, levetiracetam, and benzodiazepines.

ASM: Antiseizure medication; IQR: Inter-quartile range; [25(OH)D]: 25-hydroxyvitamin D.

Laboratory investigations

Vitamin D status: The prevalence of vitamin D deficiency was 19.3% (n = 29). Low vitamin D levels (insufficient + deficient) were observed in 80.7% (n = 121) of participants (Table 3).

Table 3 Association of serum vitamin D categories with seizure control.

Vitamin D category	Controlled, n (%)	Uncontrolled, n (%)	Total, n	P value
Sufficient (≥ 30 ng/mL)	29 (53.7)	25 (46.3)	54
Insufficient (20-29 ng/mL)	28 (41.8)	39 (58.2)	67
Deficient (< 20 ng/mL)	8 (27.6)	21 (72.4)	29
Total	65	85	150	0.08a

^aP > 0.05.

Vitamin D trends: Serum vitamin D levels at different time points are described in Figure 2 and Table 4. At baseline, the mean serum vitamin D level was 21.33 ng/mL, with a range of 10.6 ng/mL to 40.24 ng/mL. At the 6-month follow-up, the mean serum vitamin D level decreased slightly to 19.75 ng/mL, with values ranging from 10.2 ng/mL to 38.65 ng/mL. At the 18-month follow-up, the mean serum vitamin D level further declined to 18.61 ng/mL, with a range of 8.3 ng/mL to 30.1 ng/mL. Although a gradual decline in mean vitamin D levels was observed over the course of follow-up, statistical analysis demonstrated that the differences in vitamin D levels across the three time points were not statistically significant. In addition, longitudinal analysis using linear mixed-effects modelling did not demonstrate a statistically significant change in serum vitamin D levels over time (β = -0.18 ng/mL per time point, P = 0.41).

Figure 2 Trend of mean serum vitamin D levels at baseline, 6 months, and 18 months of follow-up.

Table 4 Serum vitamin D level at baseline, and follow-up (6 months and 18 months).

Time points of measurement	Vitamin D level (ng/mL)		P value
	mean ± SD	Range
Baseline	21.33 ± 9.6	10.6-40.24
6 months	19.75 ± 8.9	10.2-38.65
18 months	18.61 ± 8.7	8.3-30.1	0.651

¹Comparison performed using repeated measures analysis (Analysis of Variance); differences were not statistically significant.

Clinical course, treatment and outcomes

Data completeness was ensured at all follow-up time points, and no missing data were observed for key study variables. Multivariable logistic regression analysis was performed to identify independent predictors of low serum vitamin D levels (Table 5). After adjusting for gender, age group, type of antiseizure medication, and seizure control status, none of the variables were found to be significantly associated with low vitamin D levels. Although children with controlled seizures had lower odds of low vitamin D compared to those with uncontrolled seizures (adjusted odds ratio: 0.55, 95%CI: 0.27-1.12), this association did not reach statistical significance (P = 0.09).

Table 5 Multivariable logistic regression analysis showing association of low vitamin D status with clinical characteristics antiseizure medication therapy and seizure control.

Variables	Adjusted OR (95%CI)	P value
Gender
Male	0.72 (0.34-1.51)	0.38
Female	1.0
Age group
1-5 years	1.03 (0.52-2.04)	0.93
> 5 years	1.0
ASM therapy
Monotherapy	0.68 (0.33-1.39)	0.29
Polytherapy	1.0
Seizure control
Controlled	0.55 (0.27-1.12)	0.09
Uncontrolled	1.0

OR: Odds ratio; CI: Confidence interval; ASM: Antiseizure medication.

The grouping of ‘low vitamin D’ by combining insufficiency and deficiency may reduce discriminatory power and obscure potential dose response relationships between vitamin D levels and outcomes. So, additional exploratory analyses were performed using three categories (deficient, insufficient, sufficient), however, no significant trend was observed (P = 0.08 for the trend) (Table 3).

Multivariable regression analysis adjusting for demographic and clinical variables, including seizure type, nutritional status, developmental delay, socioeconomic status, and ASM exposure characteristics was conducted. Standardized beta coefficients (β*) demonstrated that none of the variables had a strong independent effect on serum vitamin D levels, further supporting the absence of significant predictors (Table 6).

Table 6 Multivariable linear regression analysis of factors associated with serum vitamin D levels.

Variables	β coefficient	Standardized β (β1)	95%CI	P value
Age (years)	-0.05	-0.06	-0.2 to 0.1	0.52
Male sex	0.3	0.08	-0.8 to 1.4	0.59
Malnutrition	-1.1	-0.14	-2.8 to 0.6	0.20
Developmental delay	-0.9	-0.12	-2.5 to 0.7	0.26
Low socioeconomic status	-0.7	-0.09	-2.1 to 0.7	0.32
Enzyme-inducing ASM	-0.8	-0.11	-2.0 to 0.4	0.19
Duration of ASM (months)	-0.12	-0.13	-0.3 to 0.06	0.18

β: Unstandardized regression coefficient; β1: Standardized regression coefficient; CI: Confidence interval; ASM: Antiseizure medication.

DISCUSSION

The present prospective longitudinal study evaluated the trend of serum vitamin D levels and their association with ASM and seizure control in children with seizure disorders. A key finding was the high prevalence of low vitamin D (deficiency and insufficiency) levels (80.7%), along with a non-significant declining trend over 18 months and the absence of an independent association between vitamin D status and seizure control. These findings contribute to the growing but inconclusive body of evidence regarding the role of vitamin D in pediatric epilepsy.

A gradual decline in mean serum vitamin D levels was observed over time, although this did not reach statistical significance. This trend is consistent with prior studies reporting modest reductions in vitamin D levels in children receiving long-term ASM therapy[14-17]. Studies have demonstrated that children on ASMs frequently exhibit low vitamin D levels, although longitudinal decline may be variable[16-19]. In the present study, the observed trend should be interpreted cautiously and does not confirm a true longitudinal decline.

Long-term use of enzyme-inducing ASMs has been shown to accelerate hepatic metabolism of vitamin D through induction of cytochrome P450 enzymes[17-20]. This results in increased conversion of vitamin D into inactive metabolites and may lead to reduced circulating levels of [25(OH)D]. However, the magnitude of this effect varies among studies, and some investigations have reported only modest reductions in vitamin D levels during treatment. But, in the present study, no statistically significant difference in vitamin D levels between children receiving enzyme-inducing and non-enzyme-inducing ASM was found. This finding is consistent with reports suggesting that vitamin D deficiency in epilepsy may not be solely drug related. Studies have demonstrated that both enzyme-inducing and non-enzyme-inducing ASMs can affect bone metabolism, indicating a more complex interaction[21-23]. In addition, a prolonged use of ASM (for > 2 years) causes osteoclast inhibition followed by osteoblast dysfunction eventually causing reduced bone formation and osteoporosis. Though these effects are pronounced with enzyme-inducing ASMs and polytherapy[24]. In contrary to previous studies, no significant difference was observed between children receiving monotherapy and those receiving polytherapy in the present study[11]. These findings suggest that a low vitamin D status in children with epilepsy has multifactorial causations.

The prevalence of low vitamin D (80.7%) in our cohort is higher than general pediatric Indian population estimates (approximately 40%-70%), suggesting a potentially higher burden in children with epilepsy[25,26]. In addition, the prevalence is higher than earlier studies that reported vitamin D deficiency among children receiving ASMs[7,8]. Multiple factors may contribute to this high prevalence. In addition to drug-induced alterations in vitamin D metabolism, children with epilepsy may have limited outdoor activity, reduced sunlight exposure, poor nutritional intake, or associated neurodevelopmental conditions that predispose them to vitamin D deficiency[14-20].

An important finding of the present study was the observation of a graded increase in the proportion of uncontrolled seizures across worsening vitamin D categories without the lack of statistical significance. This means, the finding should be interpreted cautiously. The lack of a statistically significant association in our study may be explained by several factors. Seizure control is multifactorial and influenced by epilepsy type, etiology, treatment adherence, and pharmacological response, which may overshadow the contribution of vitamin D status. Additionally, the study may have been underpowered to detect subtle dose-response relationships. The borderline P value observed in our study suggests that a larger sample size may be required to clarify this relationship. Similar observations have been reported in several clinical studies, which failed to demonstrate a clear relationship between serum vitamin D levels and seizure outcomes[19]. However, some clinical trials have shown a beneficial effect of vitamin D supplementation in children with epilepsy in form of reduction in seizure frequency[6,27].

Another notable finding was the absence of significant predictors of vitamin D levels in multivariable regression analysis. Factors such as age, gender, nutritional status, socioeconomic status, ASM type, and duration of therapy were not independently associated with vitamin D levels. This supports the concept that vitamin D status is influenced by a complex interplay of environmental, nutritional, and disease-related factors[28]. The exploratory analysis using three categories of vitamin D status revealed a non-significant trend toward poorer seizure control with worsening deficiency. Although not statistically significant, this observation may indicate a potential dose-response relationship. Studies have suggested that vitamin D effects may be threshold-dependent, which could explain the absence of clear associations in smaller studies[11,29].

The strengths of the present study include its prospective design, relatively large sample size, and serial longitudinal measurement of vitamin D levels at multiple time points. These features allowed us to evaluate trends in vitamin D levels over time and assess potential associations with antiseizure medication therapy and seizure outcomes. However, several limitations should also be acknowledged. First, the study was conducted in a single tertiary care centre, which may limit the generalizability of the findings. Second, important determinants such as sunlight exposure, dietary intake, and seasonal variation were not quantitatively measured and therefore could not be included in adjusted models. Third, Bone-related outcomes such as bone mineral density, parathyroid hormone levels, or fracture history were not assessed. Fourth, due to absence of a healthy control or untreated epilepsy group, the findings should be interpreted as descriptive and associative, rather than causal. Future studies incorporating matched controls and multivariable longitudinal modelling would be necessary to clarify the independent contribution of antiseizure therapy to hypovitaminosis D risk.

Given these findings, routine monitoring of vitamin D levels in children with epilepsy may be justified, particularly in regions with high baseline prevalence of deficiency. While our study does not demonstrate a direct causal relationship with seizure control, the broader health implications of vitamin D deficiency, including effects on bone health and overall development, support its clinical relevance. Recent literature also advocates for proactive screening and supplementation strategies in high-risk pediatric populations receiving long-term ASM therapy.

CONCLUSION

Vitamin D insufficiency is highly prevalent in children with epilepsy receiving ASM. Although a non-significant declining trend was observed, no independent association with seizure control was demonstrated. These findings should be interpreted as associative and hypothesis-generating.