If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Unit of Reproductive Endocrinology, 1st Department of Obstetrics and Gynecology, Medical School, Aristotle University of Thessaloniki, Thessaloniki, Greece
Valproate and carbamazepine are associated with thyroid hormone alterations in children.
•
Elevated TSH concentration is the most common thyroid abnormality.
•
Clinical thyroid disease is not associated with child antiepileptic treatment.
•
Monitoring/treatment of subclinical thyroid dysfunction in children should be assessed.
Abstract
Purpose
Antiseizure medications (ASM) have long been examined for their potential to induce thyroid dysfunction. The aim of this systematic review and meta-analysis was to assess the prevalence of thyroid disease in children up to 16 years receiving monotherapy with valproate (VPA), carbamazepine (CBZ) and levetiracetam (LEV).
Methods
PubMed/MEDLINE, Cochrane/CENTRAL databases and the gray literature were searched to identify observational studies providing the prevalence of thyroid dysfunction in the target population under VPA, CBZ, or LEV monotherapy schemes. The results were pooled using a random-effects model, and additional subgroup analyses were performed for the three ASM groups.
Results
Fifteen and thirteen studies met inclusion criteria for the qualitative and the quantitative analysis, respectively, with a total of 945 pediatric patients with prevalence data. Only VPA and CBZ were associated with thyroid dysfunction. The overall prevalence of thyroid abnormality was higher in children receiving ASM [odds ratio (OR) 6.82, 95% confidence interval (CI) 3.96–11.75]. In the subgroup analysis, the prevalence of biochemical thyroid abnormality with increased TSH was higher in the VPA (OR 9.54, 95%CI 5.25–17.34) and the CBZ group (OR 4.08, 95%CI 1.84–9.04) compared with controls.
Conclusion
This study confirms the higher prevalence of biochemical thyroid abnormality in children under VPA and CBZ monotherapy, whereas no such evidence is present for LEV. In children with a predisposition for thyroid disease, LEV should be considered over VPA and CBZ, if appropriate for seizure type and epilepsy syndrome. More studies are needed to reach a consensus on monitoring and management of thyroid dysfunction in children receiving ASM therapy.
Epilepsy affects a large number of the worldwide pediatric population. The overall prevalence of active epilepsy in childhood varies among different countries and age groups, with an estimate of 4–5 per 1000 children in population-based studies [
Antiseizure medication (ASM) remains the treatment of choice for epilepsy in childhood, in the absence of a structural brain abnormality amenable to epilepsy surgery, and treatment for a period no shorter than two to three years is often required [
], highlighting the significance of early ASM initiation and the need for substantial knowledge on the possible adverse events associated with their use. These are more commonly mild adverse events in clinical practice, such as gastrointestinal problems [
], which only exceptionally constitute reasons for drug discontinuation.
Children who require the administration of ASM undergo close monitoring for potential endocrinological abnormalities, thyroid abnormalities included. Thyroid hormones play a pivotal role in brain development and central myelination, especially during the first 3 years of life and in various metabolic pathways, bone maturation, and growth [
The spectrum of thyroid diseases in childhood and its evolution during transition to adulthood: natural history, diagnosis, differential diagnosis and management.
] and increased cardiovascular risk in adults, with similar data for children being to date scarce.
The occurrence of thyroid dysfunction post ASM initiation and the various biochemical abnormalities regarding the thyroid hormone and TSH concentrations have been widely speculated, reported and reviewed in the literature. However, no cumulative evidence exists regarding the prevalence of thyroid dysfunction in children to whom ASM has been administered. The aim of this systematic review and meta-analysis was to assess whether the prevalence of thyroid dysfunction is higher in children with epilepsy, aged one to 16 years, to whom monotherapy of valproate (VPA), carbamazepine (CBZ), or levetiracetam (LEV) has been administered compared with the general pediatric population not receiving any medication. The choice to focus the analysis on these three ASM was based on the prevalence of their utilization in the pediatric epilepsy population worldwide.
2. Methods
2.1 Protocol and registration
The present systematic review protocol and meta-analysis were determined based on the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) statement guidelines [
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration.
] with the Unique Identification Number CRD42020214012.
2.2 Eligibility criteria
Retrospective and prospective observational studies, comprising of cohort, case-control, and cross-sectional studies were considered for inclusion in fulfillment of the following eligibility criteria:
1
Children aged one to 16 years included in the study population
2
Administration of monotherapy with valproate (VPA), carbamazepine (CBZ), or levetiracetam (LEV)
3
Prevalence outcome of either overt or subclinical thyroid abnormality, following ASM administration.
The types of thyroid dysfunctions that were included were the following: hypothyroidism, hyperthyroidism, autoimmune thyroid disease, and goiter. Subclinical thyroid dysfunction was defined biochemically with abnormal TSH concentrations >5 mIU/L and <0.4 mIU/L in the presence of normal free-T3 and free-T4 concentrations in subclinical hypothyroidism (SCH) and subclinical hyperthyroidism, respectively. No secondary outcomes were evaluated.
2.3 Information sources and search strategy
The systematic literature search was conducted in PubMed/MEDLINE, Cochrane Library (Cochrane Central Register of Controlled Trials (CENTRAL)) and "gray" literature, in the following online databases: ProQuest Dissertations and Theses, Open gray and New York Academy of gray literature. A manual search for relevant conference abstracts from the European Pediatric Neurology Society (EPNS) Congress and the American Neurology Association (ANA) Annual Meeting was also conducted.
The search strategy for the PubMed database (Appendix A) included free-text and MeSH terms, and it was accordingly adapted for the rest of the databases. The language of the retrieved literature was restricted to English, and only studies published later than 1990 were considered for inclusion. The search was initially conducted on October 21, 2020, and was re-run before final analysis to retrieve additional relevant studies. The results are up to date as of December 4, 2020.
2.4 Study selection
Duplicate records were excluded before study selection. Two independent reviewers (TSI and SP), each blinded to the other's selection, initially checked the retrieved literature titles and abstracts for eligibility. In the second eligibility check, the full-text articles were screened. In cases of missing full text, the reviewers contacted the study investigators. Conflicts and uncertainties were discussed, and a mutual consensus was achieved. Both investigators used Rayyan online software for the study selection process [
The two reviewers (TSI and SP) independently extracted the following data (where available):
1
Study design
2
Patients’ baseline characteristics [age, gender, Body Mass Index (BMI), seizure type, seizure etiology, and age of seizure onset]
3
Antiseizure medication (VPA, CBZ, or LEV)
4
Prevalence and type of thyroid dysfunction.
Thyroid abnormalities were defined by abnormal TSH concentrations, or abnormal autoantibody titers, in cases of autoimmune thyroid disease. Therefore, only information on the prevalence of such abnormal values was considered, and this could be expressed in odds ratio (OR), relative risk (RR) or frequency (%).
2.6 Risk of bias assessment
The Newcastle-Ottawa Quality Assessment Scale (NOS) [
] was used to assess the quality of the included cohort and case-control studies. The NOS was further adapted for the assessment of cross-sectional studies (Appendix B). This tool evaluates three major domains: selection, comparability, and outcome, rated with a maximum of four, two, and three stars, respectively. Collectively, each study was considered as good (3 or 4 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome domain), fair (2 stars in selection domain AND 1 or 2 stars in comparability domain AND 2 or 3 stars in outcome domain), and poor (0 or 1 star in selection domain OR 0 stars in comparability domain OR 0 or 1 stars in outcome domain) quality. Both reviewers (TSI and SP) independently assessed the quality of the included studies, and any disagreements were discussed until mutual consensus was reached.
2.7 Data analysis
A narrative synthesis of the extracted data was initially performed, based on the three antiseizure medications (VPA, CBZ, LEV) and the prevalence of thyroid dysfunction, along with the type of thyroid dysfunction. The meta-analysis that followed was completed using the Review Manager software (RevMan, version 5.4, The Cochrane Collaboration 2020). Analysis of heterogeneity was conducted using the Chi2 and I2 statistics. An I2 of <25%, 25–50% and >50% was suggestive of low, moderate and high heterogeneity, respectively. A random-effects statistical model was used to pool the results since the expected heterogeneity was high. For the outcome of interest, the OR with a 95% confidence interval (CI) and the respective p-values were calculated. The level of statistical significance was 0.05. Subgroup analyses were planned for the three ASM groups when adequate data were present.
3. Results
The study selection process is presented in the flowchart of the systematic review (Fig. A). After eliminating duplicate studies, 3466 studies were screened, among which 3427 were excluded based on the title and/or the abstract. A full-text assessment was conducted in the remaining 39 records. Following the second eligibility check, 24 studies were excluded, and the reasons for exclusion are summarized in Fig. A. Overall, 15 studies [
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
] fulfilled the eligibility criteria for the narrative synthesis and the meta-analysis, respectively.
Fig. APRISMA flow diagram showing the selection process of the included studies for qualitative and quantitative synthesis. CBZ: Carbamazepine; LEV: Levetiracetam; VPA: Valproate.
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
] included patients within the age reference range, and only the results of the patients between one and 16 years old were extracted. No patient had an underlying chronic disease, including chronic neurological, metabolic or endocrine disease, and, in some studies [
], a clear family history of thyroid disease was included in the eligibility criteria. ASM monotherapy was prescribed in all patients, and at least one of VPA, CBZ or LEV was included in every study; some of the selected studies included the administration of additional ASM [
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
], rather than monotherapy of VPA, CBZ or LEV. In these studies, only data concerning the three per-protocol determined ASM were extracted. In addition to various outcomes, all studies assessed the prevalence of biochemical thyroid abnormality, namely prevalence of abnormal TSH and/or abnormal autoantibody titers, in the intended population. In five studies [
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
]; recognized as idiopathic in all. The age at seizure onset was unreported in all selected studies, and, thus, this information was not included in Table A.
Table APatient baseline characteristics in the 15 eligible studies, presented in ascending chronological order.
Reference
Design
ASM
Number
Patient baseline characteristics
Age (years) (mean ± SD)
Gender (M/F)
BMI (kg/m2) (mean ± SD)
Seizure type (G/P)
Eiris Punal et al. 199921
Cross-sectional
VPA
51
9.5 ± 2.9
NA
NA
NA
CBZ
61
11.2 ± 2.5
NA
NA
NA
Verrotti et al. 200122
Cross-sectional
VPA
16
8.5 ± 2.8
8/8
NA
NA
CBZ
12
8.3 ± 2.5
7/5
NA
NA
Cansu et al. 200623
Cohort
VPA
30
10.1 ± 4.6
NA
NA
14/16
Attilakos et al. 200724
Cohort
CBZ
18
8.9 ± 3.4
11/7
NA
NA
Castro-Cago et al. 200725
Cross-sectional
VPA
32
9.7 ± 3.7
18/14
NA
NA
CBZ
20
11.8 ± 1.5
12/8
NA
NA
Mikati et al. 200726
Cross-sectional
VPA
104
NA*
NA
NA
NA
Attilakos et al. 200927
Cohort
VPA
30
9.0 ± 3.7
15/15
NA
NA
Verrotti et al. 200935
Cohort
VPA
14
7.0 ± 1.8
7/7
NA
NA
CBZ
18
7.3 ± 3.1
8/10
NA
NA
Aggarwal et al. 201128
Cross-sectional
VPA
34
8.5 ± 2.6
24/10
15.0 ± 2.4
27/7
CBZ
30
7.3 ± 0.2
13/17
14.8 ± 2.2
0/30
Sahu et al. 201229
Cross-sectional
VPA
57
8.5 ± 3.0
33/24
NA
21/36
Turan et al. 201430
Cross-sectional
VPA
51
7.7 ± 3.5
NA
NA
NA
CBZ
45
9.8 ± 2.8
NA
NA
NA
Yilmaz et al. 201431
Cohort
VPA
129
6.5 ± 4.2
76/53
NA
89/40
CBZ
36
8.6 ± 4.1
18/18
NA
18/17 U:1
LEV
11
8.8 ± 3.8
4/7
NA
4/7
El-Farahaty et al. 201532
Cross-sectional
LEV
12
11.6 ± 1.9
10/2
23.7 ± 2.4
0/12
Nishiyama et al. 201933
Cohort
CBZ
8
8.8 ± 3.7
3/5
NA
0/8
LEV
12
9.2 ± 2.8
6/6
NA
0/12
Güngör et al. 202034
Cohort
VPA
114
8.7 ± 3.2
60/54
NA
NA
ASM: Antiseizure medications; BMI: Body Mass Index; CBZ: Carbamazepine; F: Female; G: Generalized; LEV: Levetiracetam; M: Male; NA: Not available; P: Partial; SD: Standard deviation; U: Unknown; VPA: Valproate.
*Age of all study participants regardless of the administered ASM, 8.5 ± 6.7 years.
Table B depicts the results of the individual studies on the prevalence of thyroid dysfunction in the study and control groups, where available. The majority of eligible studies assessed the prevalence of biochemical thyroid abnormality with elevated TSH concentrations and a few among these assessed the prevalence of subclinical hypothyroidism in the intended pediatric population. Additionally, the overall prevalence of unclassified thyroid disease was recorded in two studies [
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
]. The results of the former outcome are summarized in Table C for the three ASM groups. The prevalence of biochemical thyroid abnormality varied from 0% to 52.9% in the VPA groups, with most studies recording a prevalence of approximately 25%. The prevalence of biochemical thyroid abnormality varied from 0% to 33.3% in the CBZ groups, and the prevalence of any thyroid disease was 0% in all LEV groups. In the control groups of healthy age- and gender-matched children, the prevalence of biochemical thyroid dysfunction varied from 0% to 7.7%. Finally, the prevalence of autoimmune thyroid disease was 0% in both studies assessing anti-TPO titers in children under VPA administration [
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
]. None of the selected studies reported biochemically and/or clinically confirmed hyperthyroidism or clinical and/or ultrasonographic evidence of goiter as adverse events post ASM administration in the reference population.
Table BResults of the data extraction process and the risk of bias assessment in the 15 eligible studies in ascending chronological order.
The quality assessment of the 15 selected studies was conducted per protocol using the NOS scale and the NOS adapted for cross-sectional studies. Four of the included studies [
] were deemed of poor quality, and all of them were cross-sectional studies. The remaining 11 studies were good-quality studies. The poor quality of the studies was attributed mainly to the absence of any control for possible confounders in the comparator (control) groups. Moreover, one of the poor-quality studies [
] lacked a representative pediatric population, sample size justification and information on the ascertainment method of the prescribed ASM. The results of the study quality assessment are presented in Table B.
] were excluded from the meta-analysis due to the lack of information on the outcome of interest for the control groups. Thus, the remaining 13 studies were included in the quantitative analysis. Two further analyses were conducted before the final analysis: sensitivity analysis and subgroup analysis to compare the studies with healthy control groups with the studies that used the baseline evaluation of the pediatric patients as the control group. The two analyses are presented as supplementary material. These analyses aim to reveal whether the results remain robust after the addition of all studies, or the final analysis should include only studies of good quality and studies with a well-defined control group, which are less likely to introduce any source of bias in the meta-analysis. Firstly, the quantitative synthesis of all study results, in comparison to the synthesis of good-quality studies, was conducted (sensitivity analysis). Secondly, a subgroup analysis was conducted. Both analyses revealed no difference between the studied groups (p = 0.5 and p = 0.76, respectively), and, therefore, it was concluded that the results remain robust after the inclusion of all 13 eligible studies.
The overall prevalence of thyroid dysfunction in children aged one to 16 years receiving monotherapy of CBZ, VPA or LEV is evaluated in the forest plot (Fig. B). The study of Sahu et al. [
]. appears twice in the forest plots because the prevalence of thyroid disease was examined separately as the prevalence of SCH and prevalence of autoimmune thyroid disease. Moreover, the indication "not estimable" refers to the studies in which zero events were reported in the ASM and the control groups. This, however, is unlikely to alter the study results.
Fig. BForest plot of the meta-analysis of the 13 included studies. The prevalence of thyroid abnormality in the antiseizure medication groups (valproate, carbamazepine and levetiracetam) in comparison to the prevalence of thyroid abnormality in the control groups (healthy children without epilepsy and children prior to ASM initiation) is analyzed using a random-effects model with 95% CI. ASM: Antiseizure medications; CI: Confidence interval.
Nearly all studies with estimable results report a higher prevalence of thyroid dysfunction in the ASM groups, compared with the prevalence of thyroid dysfunction in the control groups. The only study to present a non-significantly higher prevalence of autoimmune thyroid disease in the control group is the study of Sahu et al. [
] (OR 0.3, 95% CI 0.01–7.49). The majority of the remaining studies with estimable results present a significantly higher prevalence of biochemical thyroid abnormality in the ASM pediatric groups, with only one study [
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
] reporting non-significantly higher prevalence rates (OR 11.48, 95% CI 0.57–230.99). Based on the cumulative results, the overall prevalence of thyroid dysfunction in children aged up to 16 years under ASM monotherapy with VPA, CBZ or LEV is significantly higher, compared with children of the same age under no drug therapy, with OR 6.82 and 95% CI 3.96 to 11.75 (p-value <0.001). The heterogeneity among the included studies is low, with an I2 value of 21%.
3.4 Additional analyses
Additional analysis for biochemical thyroid abnormality with elevated TSH concentrations in the three ASM subgroups was conducted since this outcome was present in the vast majority of the studies. Among the eligible studies for the quantitative analysis, only one study reported the prevalence of SCH in children receiving LEV. Consequently, the subgroup analysis was limited to VPA and CBZ groups. The results from this subgroup analysis are presented in the following forest plot (Fig. C). Based on the forest plot, the prevalence of biochemical thyroid dysfunction is significantly higher in children receiving either VPA or CBZ compared with controls (OR 7.28, 95% CI 4.57–11.59). Moreover, an approximately 2.5 times higher overall prevalence is reported in the VPA group (OR 9.54, 95% CI 5.25–17.34) compared with the CBZ group (OR 4.08, 95% CI 1.84–9.04), although this difference is not statistically significant (p = 0.09). A lack of significant results is reported in five out of six studies to assess the prevalence of biochemical thyroid abnormality in the CBZ group, whereas all studies examining pediatric patients under VPA treatment report significant results for this outcome of interest. In the subgroup analysis, the within-group heterogeneity varies from low for the CBZ group (I2=0%) to moderate for the VPA group (I2=27%). The heterogeneity remains low (I2=16%) in the analysis of the total events but is considered high in the comparison of the two ASM subgroups (I2=64.3%).
Fig. CForest plots of the ASM subgroup analysis for the prevalence of biochemical thyroid abnormality with elevated TSH concentrations. Valproate and carbamazepine subgroups are pooled using a random-effects model with 95% CI. ASM: Antiseizure medications; CI: Confidence interval.
The present systematic review of the literature confirms that VPA and CBZ can influence thyroid function in children causing mainly biochemical thyroid dysfunction and, more specifically, SCH. With the additional meta-analysis, it is shown that the prevalence of thyroid abnormality in children up to 16 years under VPA or CBZ monotherapy is significantly higher compared with children receiving no antiseizure medication. Based on the current study results, LEV is associated with no thyroid alterations in the pediatric population.
The effects of ASM have long been studied in the pediatric and adult populations. Many original studies have assessed thyroid dysfunction post ASM therapy in children without reporting prevalence outcomes and, thus, were excluded per protocol from the present systematic review. Most of the studies and narrative reviews agree with the present systematic review and meta-analysis results in that both VPA and CBZ are associated with thyroid dysfunction, whereas treatment with LEV runs a very low risk of this adverse effect. More specifically, many studies have reported an increase in the TSH concentrations of children under VPA treatment [
Journal of trace elements in medicine and biology serum thyroid hormone profile and trace elements in children receiving valproic acid therapy : a longitudinal and controlled study.
] prospectively assessing the thyroid alterations in children under LEV therapy and reporting no significant alterations.
Similar results have been reported in adults or adolescents outside the eligible age range for this systematic review, namely children over 16 years. In the adult study of Lossius et al. [
], alterations in the thyroid function were reported in patients under CBZ, and VPA treatment and CBZ administration were associated with overt hypothyroidism in patients with organic encephalopathies. In another adult cross-sectional study [
], a logistic regression model showed that age, female gender, CBZ therapy and ASM polytherapy schemes are possible risk factors for thyroid dysfunction. High prevalence rates of thyroid disease have also been reported in adults with drug-resistant epilepsy (DRE) and subsequent polypharmacy. Aparicio et al. [
] have reported an overall prevalence of 25% in this population, with VPA being one of the prescribed ASM in most reported cases. Such high prevalence is comparable with the prevalence of biochemical thyroid abnormality in children of the target population of the current study under VPA therapy. Older children up to 18 years have been assessed in the recent study of Choi et al. [
]. In this study, all ASM resulted in approximately the same prevalence of thyroid dysfunction in the target population, with the prevalence in the VPA subgroup (26.9% in 130 thyroid tests) being comparable to the respective prevalence in the LEV subgroup (24.3% in 140 thyroid tests). The results concerning the LEV subgroup differ significantly from the results of our systematic review and meta-analysis. However, this difference may be attributed either to a former ASM administration or an underlying endocrine abnormality, which was not formally excluded from the study mentioned above. The prevalence of subclinical thyroid dysfunction, specifically SCH, has been assessed in the general adult and child population ranging between 5 and 17% [
], respectively, similar to the extracted prevalence in the control groups of the target population of the current study (0–7.7%).
The present systematic review provides evidence that the prevalence of biochemical thyroid abnormality with elevation of TSH concentrations is significantly higher in children receiving VPA and CBZ compared with children under no therapy. The implications of such thyroid dysfunction have been examined in the current literature. The alterations in thyroid hormone concentrations of patients under ASM treatment are often of minimal clinical significance but may further precipitate subclinical or even overt hypothyroidism in patients with a history of thyroid disease or levothyroxine-supplemented individuals [
] indicated that thyroid abnormalities post-ASM can further cause long-term memory abnormalities, cognitive impairment or subtle neuromuscular disorders. Finally, in their review on the causes and implications of ASM on thyroid function, Hamed et al. [
] concluded that SCH caused by ASM could lead to metabolic and lipid disturbances and neurodevelopmental adverse events, which, however, can be reversible either upon ASM discontinuation or upon levothyroxine supplementation, as additionally supported by the review of Crisafulli et al. [
The presence of biochemical thyroid alterations and SCH and their implications in the pediatric population under ASM therapy has raised many questions regarding the necessity of monitoring and treatment, for which international consensus is currently lacking. Some physicians propose regular monitoring of the thyroid functions in children under VPA or CBZ monotherapy, especially in children under ASM polytherapy [
]. Likewise, a lack of consistent data exists in terms of the treatment strategy for SCH. To date, most investigators agree that TSH concentrations over 10 mIU/L require levothyroxine supplementation, even in the absence of clinical findings of hypothyroidism [
]. Other specialists suggest that TSH concentrations within the range of 5–10 mIU/L require therapy initiation when certain indications are present, including goiter, hyperlipidemia, young age and/or TSH concentrations >8 mIU/L in two measurements [
There are certain strengths and limitations in this review. On the one hand, this is the first meta-analysis to assess thyroid disease prevalence in children receiving ASM. To date, only one meta-analysis of the alterations on the thyroid hormone profile in individuals receiving ASM, adults and children included, has been published. However, the meta-analysis of Zhang et al. [
] assesses a different outcome and a different population compared to the present systematic review. Secondly, the current study is the first to explicitly examine children's population up to 16 years. From the authors’ perspective, the choice of this age group is of great importance, not least because young age is considered a risk factor for thyroid disorders post-ASM treatment, but mostly because the younger the age, the higher the risk of implications due to overt or subclinical thyroid disease. Finally, the assessment of three of the most frequently used antiepileptic agents provides significant information to many patients, caregivers and health providers.
On the opposite side, the present systematic review and meta-analysis have certain limitations. Firstly, the sample size of many studies was small, and only a few studies presented a sample size calculation. The studies assessing VPA administration were mainly the ones to present adequate sample size, whereas the respective studies assessing LEV administration had particularly small sample sizes. Secondly, the lack of proper control groups is considered an additional limitation of the studies; many studies included healthy age-matched control groups, with no controlling for further confounders and other studies, mainly prospective cohorts, considered the baseline measurements thyroid hormones and TSH concentrations as the control groups. The sensitivity analysis and the subgroup analysis of the studies with a proper control group and baseline values as the control groups minimized the present limitations. Thirdly, studies with zero events in both the ASM and the control groups were not included in the meta-analysis results (“not estimable” results in Figs. B, C). However, based on the Cochrane Handbook for Systematic Reviews of Interventions, such exclusion is justified because these studies provide no valuable information on either the direction or the magnitude of the effect of the studied ASM on thyroid function [
]. Finally, although the choice of this narrow population of children aged one to 16 years was intended to achieve more focused results, some high-quality studies with ages up to 18 years had to be excluded.
In conclusion, the prevalence of biochemical thyroid abnormality is higher in children receiving VPA and CBZ, but not in children receiving LEV, compared to children under no ASM therapy. No substantial evidence of overt thyroid disease, hyperthyroidism, goiter or autoimmune thyroid disease exists for children up to 16 years under ASM therapy with VPA, CBZ or LEV. The lack of international consensus calls for additional studies to establish practice guidelines for the monitoring and managing subclinical thyroid dysfunction in children taking ASM.
Declaration of Competing Interest
None of the authors has any conflict of interest to disclose. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
The spectrum of thyroid diseases in childhood and its evolution during transition to adulthood: natural history, diagnosis, differential diagnosis and management.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration.
The evaluation of thyroid functions, thyroid antibodies, and thyroid volumes in children with epilepsy during short-term administration of oxcarbazepine and valproate.
Thyroid dysfunction associated with increased low-density lipoprotein cholesterol in epileptic children treated with carbamazepine monotherapy: a causal relationship?.
Journal of trace elements in medicine and biology serum thyroid hormone profile and trace elements in children receiving valproic acid therapy : a longitudinal and controlled study.
The studies use the baseline values of the study groups, as the control groups.
00370-8&id=gr1.jpg){kind=link}
00370-8&id=gr2.jpg){kind=link}
00370-8&id=gr3.jpg){kind=link}