Epilepsy is among the most common chronic neurological disorders of childhood, with active prevalence estimated at 4-10 per 1000 children and a cumulative incidence of approximately 7-8 per 1000 by 15 years of age.^1,2 Children constitute nearly a quarter of the global epilepsy burden, and low and middle-income countries, including India, account for a disproportionate share of cases due to perinatal insults, central nervous system infections, and delayed diagnosis.^3,4 A regional observational study of 191 neonates likewise identified birth asphyxia, septicaemia, hypoglycaemia and hypocalcaemia as leading seizure aetiologies, illustrating the importance of preventable perinatal and metabolic insults.⁵

Antiepileptic drugs (AEDs) remain the mainstay of treatment, achieving adequate seizure control in 60-70% of children, but often require continuation for several years.^6,7 Prolonged AED exposure has been increasingly linked to endocrine adverse effects, particularly disturbances of the hypothalamic-pituitary-thyroid axis.⁸ Enzyme-inducing AEDs such as carbamazepine and phenobarbital accelerate hepatic clearance of thyroid hormones, while valproate has been associated with elevated thyroid-stimulating hormone (TSH) through effects on hormone synthesis and peripheral conversion.^9,10 The resulting biochemical picture is most often subclinical hypothyroidism—elevated TSH with normal circulating thyroid hormones—a definition also used in contemporary clinical cohorts.¹¹ Although usually asymptomatic, undetected thyroid dysfunction may affect growth, cognition and metabolic health during critical developmental years.^12,13

Existing literature on AED-associated thyroid dysfunction is heterogeneous, with most data derived from older AEDs and limited paediatric data from India, particularly regarding newer agents such as levetiracetam.^14,15 Given the essential role of thyroid hormones in neurodevelopment and the widespread use of AEDs in childhood epilepsy, we undertook this study to evaluate thyroid function in children aged 2-18 years receiving AED therapy for at least one year, to correlate thyroid function with the specific AED used, and to assess the frequency of seizure recurrence among children adherent to therapy.

MATERIALS AND METHODS

This case-control study was conducted in the Department of Paediatrics, Government Medical College, Kota, at the J.K. Lone Mother and Child Hospital, over a period of one year (March 2025 to March 2026), after approval from the Institutional Ethics Committee. Written informed consent was obtained from parents/guardians of all participants.

Study population and sample size

Sixty children aged 2-18 years who had received AED therapy for at least one year (cases) and 120 age-and sex-matched healthy children (controls) attending the outpatient and inpatient departments were enrolled. The sample size was calculated using an expected prevalence of thyroid dysfunction of 13.2% among AED-treated children, a 95% confidence level (Z=1.96) and an 8% margin of error, yielding a minimum requirement of 60 cases; a case:control ratio of 1:2 was used.¹⁶

Inclusion and exclusion criteria

Cases included all children aged 2-18 years on AED therapy for ≥1 year; controls were healthy children of the same age range with parental consent. Children with pre-existing thyroid disease, those on drugs known to affect thyroid function (e.g., dopamine agonists, amiodarone, cycloserine, rifampicin), those with a family history of thyroid or other endocrine dysfunction, non-adherent patients, and syndromic children (e.g., Down syndrome) were excluded.

Sample collection and biochemical analysis

A fasting morning venous blood sample was collected from each participant under aseptic precautions. Serum TSH and free T4 (fT4) were estimated using the DiaSorin LIAISON automated chemiluminescent immunoassay analyser and interpreted against age-specific paediatric reference ranges. Subclinical hypothyroidism was defined as an elevated TSH with a normal fT4 in the absence of overt clinical symptoms.¹⁷ Clinical details including seizure type, age of seizure onset, AED regimen, dose, duration of therapy, adherence, and seizure recurrence were recorded on a structured proforma; EEG and MRI findings, where available, were also documented.

Statistical analysis

Data were entered and analysed using SPSS software. Continuous variables were expressed as mean±SD and compared using the independent-samples t-test (two groups) or one-way ANOVA with Tukey's post-hoc test (more than two groups). Categorical variables were compared using the chi-square test. Correlation between quantitative variables was assessed by correlation analysis. Univariate and multivariate logistic regression were used to identify predictors of subclinical hypothyroidism, with results expressed as odds ratios (OR) with 95% confidence intervals (CI). A two-tailed p-value <0.05 was considered statistically significant.

RESULTS

A total of 60 cases and 120 controls were enrolled. The mean age was comparable between cases (10.0±3.8 years) and controls (9.2±3.9 years; p=0.19), as was the sex distribution (60.0% male, 40.0% female in both groups), confirming adequate baseline matching (Table 1).

Table 1. Baseline age and sex distribution of study participants

Figure 1. Baseline age and sex distribution of study participants

Mean TSH was significantly higher in cases than controls (4.50±2.80 vs 2.10±1.20 mIU/L; p<0.001), and mean fT4 was significantly lower in cases (1.10±0.33 vs 1.28±0.28 ng/dL; p=0.001) (Table 2). Subclinical hypothyroidism was present in 13/60 cases (21.7%) compared with 3/120 controls (2.5%), a highly significant difference (χ²=18.6, p<0.001) (Table 3).

Table 2. Comparison of thyroid function test profile between cases and controls

Figure 2. Comparison of thyroid function test profile between cases and controls (mean ± SD)

Table 3. Prevalence of subclinical hypothyroidism among cases and controls

Figure 3. Prevalence of subclinical hypothyroidism among cases and controls

Among cases, seizure onset was most frequent between 6–10 years (40.0%), and generalized tonic–clonic seizures (GTCS) were the commonest type (46.7%), followed by focal seizures (33.3%). Nearly half the children (46.7%) had received AED therapy for 24–59 months, and 90.0% were adherent to treatment; seizure recurrence occurred in 33.3% despite this. Valproate was the most frequently prescribed AED (36.7%), followed by carbamazepine (23.3%) and levetiracetam (20.0%); polytherapy was used in 10.0% and had the longest mean duration of exposure (68±40 months). EEG abnormalities (63.3%) were considerably more frequent than MRI abnormalities (20.0%).

Thyroid function varied significantly across AED regimens (one-way ANOVA, p=0.003 for TSH). Mean TSH was highest with polytherapy (6.00±3.40 mIU/L) and valproate (5.10±2.90 mIU/L), and lowest with levetiracetam (2.30±1.10 mIU/L); mean fT4 showed the inverse pattern. Subclinical hypothyroidism was most prevalent with polytherapy (50.0%) and valproate (27.3%), and was not observed with levetiracetam (0.0%) (Table 4).

Table 4. AED-wise thyroid function and prevalence of subclinical hypothyroidism among cases

One-way ANOVA across AED regimens, overall p=0.003.

Figure 4. AED-wise thyroid function (mean TSH and mean fT4) among cases

Figure 5. Prevalence of subclinical hypothyroidism by AED regimen

A clear dose–response relationship was observed between AED dose category and raised TSH: 15.0% in the low-dose group, 29.2% in the medium-dose group, and 50.0% in the high-dose group (one-way ANOVA, p=0.01). Seizure recurrence was significantly more frequent among children with subclinical hypothyroidism than euthyroid children (61.5% vs 25.5%; p=0.01), and the median number of seizure episodes was also higher in the subclinical hypothyroid group (2 [IQR 1–3] vs 1 [IQR 0–2]; Mann–Whitney U test, p=0.02).

Figure 6. Dose–response relationship between AED dose category and raised TSH

Figure 7. Seizure recurrence according to thyroid status

On univariate logistic regression, longer duration of AED therapy (OR 1.15/year, 95% CI 1.02–1.30, p=0.02), high-dose therapy (OR 2.40, 95% CI 1.10–5.20, p=0.03), valproate use (OR 3.20, 95% CI 1.45–7.20, p=0.005) and polytherapy (OR 2.80, 95% CI 1.05–7.40, p=0.03) were significantly associated with subclinical hypothyroidism; age, sex and age of seizure onset were not significant predictors (Table 5).

Table 5. Univariate logistic regression for predictors of subclinical hypothyroidism

On multivariate analysis, valproate use (adjusted OR 2.90, 95% CI 1.24–7.00, p=0.015), duration of AED therapy (adjusted OR 1.13/year, 95% CI 1.01–1.27, p=0.035) and high-dose therapy (adjusted OR 2.05, 95% CI 1.01–4.20, p=0.046) remained independent predictors of subclinical hypothyroidism; polytherapy (adjusted OR 2.40, p=0.06) and carbamazepine use (adjusted OR 1.45, p=0.45) were not independently significant (Table 6).

Table 6. Multivariate logistic regression for predictors of subclinical hypothyroidism

Figure 8. Forest plot: multivariate logistic regression for predictors of subclinical hypothyroidism

DISCUSSION

This case–control study demonstrates a significant alteration in thyroid function among children receiving long-term AED therapy, characterised by elevated TSH, reduced fT4, and a markedly higher prevalence of subclinical hypothyroidism compared with healthy controls (21.7% vs 2.5%). These findings are in agreement with Zhang et al., whose meta-analysis of 35 studies showed a consistent rise in TSH and fall in T4/fT4 among AED-treated patients, and with Elshorbagy et al., who reported elevated TSH and reduced fT4 specifically among children on conventional AEDs.^10,18 The prevalence of subclinical hypothyroidism in our cohort is comparable to that reported by Sarker et al. (nearly one-third of AED-treated children) and by Karatoprak et al. (16% among children on valproate).^19,20

Drug-specific analysis showed that valproate and polytherapy carried the greatest risk of thyroid dysfunction, while levetiracetam was not associated with any case of subclinical hypothyroidism. This pattern mirrors previous reports describing valproate-induced elevation of TSH through effects on hormone synthesis and metabolism, in contrast to the favourable endocrine profile of levetiracetam.^20,21 A significant dose–response relationship between AED dose and raised TSH further supports a cumulative-exposure-related mechanism, consistent with the findings of Sepahi et al. and Rafik et al. regarding endocrine effects of valproate.^21,22

An important and clinically relevant observation in our study was the association between subclinical hypothyroidism and seizure recurrence: children with subclinical hypothyroidism had more than twice the recurrence rate of euthyroid children (61.5% vs 25.5%) and a higher seizure-episode burden. This finding is biologically plausible given the essential roles of thyroid hormones in brain maturation and neurological function.^12,13,23,24 However, the present case–control study cannot establish the direction of causality; prospective longitudinal studies are required before thyroid dysfunction can be considered a determinant of seizure recurrence.

On multivariate analysis, valproate use, longer duration of therapy, and high-dose treatment were independent predictors of subclinical hypothyroidism, findings that closely parallel those of Karatoprak et al. and Mohankumar et al., who similarly identified valproate as an important drug-related determinant of thyroid dysfunction in children with epilepsy.^20,25 EEG abnormalities were more frequent than MRI abnormalities in our cohort (63.3% vs 20.0%).

Our study has certain limitations. The case–control design precludes assessment of pre-treatment thyroid status and therefore cannot establish a temporal or causal relationship between AED exposure and thyroid dysfunction. The study was conducted at a single tertiary care centre with a relatively modest sample size, which may limit generalisability. Thyroid autoantibodies, iodine status, and detailed nutritional profiling were not assessed and could be potential confounders. Finally, cross-sectional assessment does not capture longitudinal evolution of thyroid function with drug dose escalation or changing adherence over time.

CONCLUSION

Children on long-term antiepileptic drug therapy, particularly those receiving valproate, high drug doses, or polytherapy, show a significantly higher burden of subclinical hypothyroidism than healthy peers, and this thyroid dysfunction is associated with a higher rate of seizure recurrence. These findings support incorporating periodic thyroid function testing (TSH and fT4) into the routine follow-up of children on AED therapy exceeding one year, with closer surveillance for those on valproate, high-dose, or combination regimens. Levetiracetam appears to have a relatively favourable thyroid safety profile and may be preferred in susceptible children where clinically appropriate. Larger, prospective, multicentric studies with longitudinal thyroid assessment are needed to confirm causality and to guide evidence-based monitoring protocols.

ACKNOWLEDGEMENTS

The authors wish to thank the faculty and staff of the Department of Paediatrics, Government Medical College, Kota, for their opportunity and constant support. The authors gratefully acknowledge Dr. Shailendra Vashistha (Assistant Professor, Transplant Immunology and HLA Laboratory, Department of IHTM, Government Medical College, Kota) for his valuable guidance in scientific manuscript preparation. The authors also sincerely thank the VAssist Research Team (www.thevassist.com) for assistance with manuscript formatting, and technical support during manuscript submission. The authors wholeheartedly thank all participating children and their parents/guardians.

CONFLICT OF INTEREST: None declared.

SOURCE OF FUNDING: Nil.

REFERENCES

1. Beghi E. The epidemiology of epilepsy. Neuroepidemiology. 2020;54(2):185-91. doi:10.1159/000503831.
2. Xu K, Qian W, Xu W, et al. Global, regional, and national epidemiology of childhood epilepsy from 1990 to 2021: a systematic study based on the Global Burden of Disease Study 2021. Seizure. 2025;130:51-60.
3. Feigin VL, Vos T, Nair BS, et al. Global, regional, and national burden of epilepsy, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet Public Health. 2025;10(3):e203-27.
4. Patel M, Goel AD, Saini L, et al. Prevalence of pediatric and adolescent epilepsy in India: a systematic review and meta-analysis. Seizure. 2025;127:36-43.
5. Haridas C, Meena CP, Thomas S, Mohan C, Gupta RK. Evaluation of clinical profile of neonatal seizures and outcome interventions. Biomed Biopharm Res. 2024;21(2):11-21.
6. Park KM, Kim SE, Lee BI. Antiepileptic drug therapy in patients with drug-resistant epilepsy. J Epilepsy Res. 2019;9(1):14-26. doi:10.14581/jer.19001.
7. Jones KE, Howells R, Mallick AA, Paul SP, Dey I. NICE guideline review: Epilepsies in children, young people and adults NG217. Arch Dis Child Educ Pract Ed. 2023;108(6):416-21.
8. Dalkilic EB. Effects of antiepileptic drugs on hormones. Neurosci Lett. 2021;754:135800. doi:10.1016/j.neulet.2021.135800.
9. Zevenbergen C, Korevaar TI, Schuette A, et al. Association of antiepileptic drug usage, trace elements and thyroid hormone status. Eur J Endocrinol. 2016;174(4):425-32.
10. Zhang YX, Shen CH, Lai QL, et al. Effects of antiepileptic drug on thyroid hormones in patients with epilepsy: a meta-analysis. Seizure. 2016;35:72-9.
11. Choudhary N, Kom TT, Khuman V, Singh LBC, Sharma P, Kumar P. Maternal outcomes in pregnant women with subclinical hypothyroidism: an observational cohort study at a tertiary care hospital in Northeast India. Int J Med Pharm Res. 2025;6(6):1619-24.
12. Hanley P, Lord K, Bauer AJ. Thyroid disorders in children and adolescents: a review. JAMA Pediatr. 2016;170(10):1008-19. doi:10.1001/jamapediatrics.2016.0486.
13. Schiera G, Di Liegro CM, Di Liegro I. Involvement of thyroid hormones in brain development and cancer. Cancers (Basel). 2021;13(11):2693. doi:10.3390/cancers13112693.
14. Comert E, Celebi U, Piri Cinar B, et al. Thyroid functions as a parameter in monitoring of antiepileptic drugs. Neurol Res. 2022;44(7):614-21.
15. Ilia TS, Dragoumi P, Papanikolopoulou S, Goulis DG, Pavlou E, Zafeiriou D. Is the prevalence of thyroid disease higher in children receiving antiepileptic medication? A systematic review and meta-analysis. Seizure. 2022;94:117-25. doi:10.1016/j.seizure.2021.11.010.
16. Daniel WW. Biostatistics: A Foundation for Analysis in the Health Sciences. 7th ed. New York: John Wiley & Sons; 1999.
17. Biondi B, Duntas LH. Subclinical hypothyroidism. In: Luster M, Duntas LH, Wartofsky L, editors. The Thyroid and Its Diseases: A Comprehensive Guide for the Clinician. Cham: Springer; 2019. p. 255-63.
18. Elshorbagy HH, Barseem NF, Suliman HA, et al. The impact of antiepileptic drugs on thyroid function in children with epilepsy: new versus old. Iran J Child Neurol. 2020;14(1):31-40.
19. Sarker S, Saha N, Mistry B, Begum SA, Saha D, Hoque SA. Thyroid hormone status in children taking antiepileptic drugs: an experience in a tertiary care hospital of Bangladesh. Bangladesh J Child Health. 2022;46(2):58-64.
20. Karatoprak E, Paksoy S. Thyroid functions in children on levetiracetam or valproic acid therapy. J Pediatr Epilepsy. 2021;10(1):22-6.
21. Sepahi S, Riahi-Zanjani B, Ghorani-Azam A. Effect of valproic acid on metabolic status and endocrine system in pediatric patients with epilepsy: systematic literature review. Rev Clin Med. 2017;4(1):7-13.
22. Rafik A, El-Din NS, El Khayat NM, Nada M, Abushady EM. Effect of anti-epileptic drugs usage on thyroid profile in Egyptian epileptic children. Egypt J Neurol Psychiatr Neurosurg. 2024;60:3.
23. Nandi-Munshi D, Taplin CE. Thyroid-related neurological disorders and complications in children. Pediatr Neurol. 2015;52(4):373-82.
24. Nickels KC, Zaccariello MJ, Hamiwka LD, Wirrell EC. Cognitive and neurodevelopmental comorbidities in paediatric epilepsy. Nat Rev Neurol. 2016;12(8):465-76.
25. Mohankumar S, Rameshkumar R, Selvan T, Nandeesha H, Dhanalakshmi CG. A prospective cohort study on the effect of antiseizure medications on thyroid function in children aged 6 months to 12 years with epilepsy. Sultan Qaboos Univ Med J. 2024;24(4):528-35.