Seizure: European Journal of Epilepsy 71 (2019) 60–65

Contents lists available at ScienceDirect

Seizure: European Journal of Epilepsy

journal homepage: www.elsevier.com/locate/seizure

Review

Valproate decreases vitamin D levels in pediatric patients with epilepsy T

a,1 a,1 a,b,2 a a a a
Zejun Xu , Xia Jing , Guizhou Li , Jieyu Sun , Hongli Guo , Yahui Hu , Fang Sun ,
Xiaoyi Wena, Feng Chena, , Tengfei Wangc, , Xiao-Peng Lud
⁎ ⁎⁎

a
Department of Pharmacy, Children’s Hospital of Nanjing Medical University, Nanjing, China
b
School of Basic Medicine and Clinical Pharmacy, China Pharmaceutical University, Nanjing, China
c
Department of Pharmacology, University of Tennessee Health Science Center, Memphis, TN, USA
d
Department of Neurology, Children's Hospital of Nanjing Medical University, Nanjing, China

ARTICLE INFO ABSTRACT

Keywords: Purpose: To compare Vitamin D (Vit D) levels in children with epilepsy on valproate monotherapy with healthy
Meta-analysis controls.
Valproate Methods: A meta-analysis performed on articles identified from PubMed and Web of Science online databases
Carbamazepine evaluated using National Institute of Health National Heart, Lung, and Blood Institute Study Quality Assessment
Monotherapy
Tools. Subgroup analyses and publication bias assessments were also performed.
Pediatric
Epilepsy
Results: Eleven publications were eligible based on inclusion/exclusion criteria for the meta-analysis. Results
25-OH-Vit D noted a decrease in the mean Vit D level in children with epilepsy on valproate monotherapy compared with
healthy children with a Standard Mean Difference = -0.313 [-0.457, -0.169]. Cumulative meta-analysis showed
progressive negative effect of valproate therapy on Vit D levels across time. Other antiepileptic medications
caused a similar effect on Vit D status. There was no evidence of publication bias in the analyses. Type of study
design and country of origin introduced heterogeneities into the meta-analyses.
Conclusion: This meta-analysis provides evidence that long-term therapy with valproate causes a decrease in Vit
D levels in children. Therefore, in children with a seizure disorder on long-term valproate therapy, 25-OH-Vit D
levels should be monitored and appropriate supplementation implemented if levels are deficient.

1. Introduction periods for bone development. The role of Vit D is not limited to bone
health as it also has important roles in many extra-skeletal targets
Individuals with epilepsy being treated with chronic antiepileptic throughout the body, such as the muscles, immune system, and the
drug (AED) therapy represent approximately 1% of the general popu- cardiovascular system [12,13]. Thus, Vit D deficiency due to chronic
lation [1,2]. Vitamin D (Vit D) deficiency is commonly reported in adult AED therapy has additional risks beyond the decreased calcium ab-
patients with epilepsy on AED therapy [3], particularly in patients on sorption, secondary hyperparathyroidism, decreased bone density, and
older AEDs that induce hepatic metabolism such as carbamazepine and elevated serum alkaline phosphatase associated with bone homeostasis
phenobarbital [4]. Valproate and other newer AEDs (e.g. lamotrigine, [14].
oxcarbazepine) are generally considered to have minimal effects on There are two main forms of Vit D, Vit D2, a plant and yeast sterol
hepatic metabolic enzymes, and thus, considered to have a lower po- product, and Vit D3, the form synthesized by mammals. In humans, Vit
tential to affect Vit D levels [5–7]. However, several studies have shown D3 is synthesized through UVB-radiation-induced photosynthesis in the
that valproate induces CYP3A4 and CYP24A1, which are involved in Vit skin with less than 10% derived from dietary sources (mostly D3 and
D catabolism [8–11], making valproate a candidate AED to cause Vit D minimal amounts of D2) [15]. Total 25-hydroxyvitamin D (25-OH-Vit
deficiency. D) concentration in serum is the major index used to monitor the Vit D
Vit D is essential for calcium and bone homeostasis, especially in level, which includes both 25-OH-Vit D2 and 25-OH-Vit D3 [14,16].
children because childhood and adolescence are the most critical Serum concentrations of 25-OH-Vit D less than 20 ng/mL are generally

⁎
Corresponding author at: the Children's Hospital of Nanjing Medical University, 72 Guangzhou Road, Nanjing, 210008, China.
⁎⁎
Corresponding author at: University of Tennessee Health Science Center, 874 Union Ave., Memphis, TN, 38163, USA.
E-mail addresses: [email protected] (F. Chen), [email protected] (T. Wang).
1
These authors contributed equally to this work.
2
Visiting graduate student from China Pharmaceutical University.

https://doi.org/10.1016/j.seizure.2019.06.009
Received 20 February 2019; Received in revised form 6 June 2019; Accepted 8 June 2019
1059-1311/ © 2019 British Epilepsy Association. Published by Elsevier Ltd. All rights reserved.
Z. Xu, et al. Seizure: European Journal of Epilepsy 71 (2019) 60–65

considered as the breakpoint for Vit D deficiency [12]. SE 11.0 (Stata Corporation, College Station, TX, USA). The “Metan”
Currently, the risk of Vit D deficiency in children with seizure dis- [24] command was used to combine the continuous data of Vit D
orders treated chronically with AEDs is unclear and specific clinical concentrations (mean ± SD). The SEM values in studies were con-
guidelines on monitoring Vit D levels in children do not exist. Given verted into SD values based on the formula SD= SEM× n . Due to the
that valproate, carbamazepine and phenobarbital are of proven efficacy measurement of Vit D using different methods and the use of different
and frequently administrated to children with epilepsy [17], under- units (ng/mL, μg/L or nmol/L), the pooled results are expressed in
standing the long-term risk of Vit D deficiency imposed by chronic AED terms of the standardized mean difference (SMD) and their 95% con-
therapy in children is critical to avoiding deleterious effects on healthy fidence interval (CI). Cochran’s Q-statistic was used to estimate within
development. and between study variations. If a significant Q-statistic (I2 > 50%,
Numerous studies have reported on the effect of valproate therapy p < 0.1) occurred indicating heterogeneities across studies, then the
on Vit D levels, but due to the varying cut-offs for 25-OH Vit D blood random effects model was used for the meta-analysis [25]. Otherwise
concentration defining deficiency (< 10 ng/mL [18,19], < 12 ng/mL the fixed effects model was the default. A Cumulative meta-analytic
[20], < 20 ng/mL [21–23]), small sample size, and inconsistent meth- method [26] was applied to identify how estimates of the negative ef-
odology conflicting conclusions on the effect of long-term valproate fect on Vit D levels evolve over time. Subgroup analyses were con-
therapy in children are reported. To bring clarity to this issue we ducted by stratifying according to variables such as countries of sub-
conducted a pooled analysis of the available literature using the meta- jects to determine potential sources of heterogeneities. Publication bias
analysis method. was assessed by Begg’s and Egger’s tests with no publication bias de-
fined by both p values > 0.05 [27].
2. Materials and methods
3. Results
2.1. Literature searching strategy
3.1. Identification of eligible studies
PubMed and Web of Science (WOS) were searched thoroughly from
January 1980 to July 2018. Combined terms “valproate vitamin D”,
Our search strategy produced 332 and 160 publications from
“valproic acid vitamin D”, “VPA vitamin D”, “valproate vit D”, “val-
PubMed and WOS respectively. Non-English publications and dupli-
proic acid vit D”, “VPA vit D” were used to search PubMed. For WOS
cates were excluded and yielded 90 and 121 papers in PubMed and
the combined terms used were “valproate vitamin D”, “valproic acid
WOS. During title and abstract screening, 64 PubMed derived and 97
vitamin D” and “VPA vitamin D”. Non-English publications were ex-
WOS extracted publications were recognized as either animal/cellular
cluded from this analysis. Retrieved articles were processed in succes-
research, reviews, letters to the editor, or not related to Vit D testing in
sion and duplicate publications were excluded. All remaining literature
children. Consequentially, 26 original articles retrieved from PubMed
was reviewed by title and abstract to eliminate reviews, letters to the
and 24 obtained from WOS were integrated into a pool to undergo
editor, comments, case reports, animal studies, molecular and cellular
further evaluation. Two more articles [28,29] were added by manual
studies, and previous meta-analyses. Relevant clinical and epidemio-
screening. After 18 duplicates were removed from the pool, full context
logic studies were combined into a pool. Articles in the pool underwent
evaluation was considered among the remaining 34 publications (refer
context screening in accordance with defined inclusion/exclusion cri-
to Table S1). The context of these 34 articles was evaluated thoroughly
teria. All qualified studies were individually reviewed to search for
to distinguish the studies were in accordance with the inclusion and
additional relevant studies within the articles’ references.
exclusion criteria. Finally, 11 eligible articles were selected after re-
moval of unqualified literature for various reasons (procedure as shown
2.2. Criteria of study selection and exclusion
in Fig S1). We summarized the reasons that individual articles were
included or excluded in Table S1. Despite efforts to contact the authors,
Eligible studies were original research-based articles that: 1) mea-
unavailability of raw data from primary articles was the most common
sured Vit D level in blood and included the mean and standard devia-
cause of exclusion from enrollment.
tion (SD), or standard error of the mean (SEM), or this data was ob-
tainable from the authors, 2) included only children whose epilepsy was
treated with valproate therapy, 3) had a control group of healthy 3.2. Classification, characteristics and risk of bias in selected publications
children or patients on other medication treatment that did not include
an AED. Studies were excluded from the analysis if they were: 1) con- More than half of the reviewed 11 articles were of good or fair
ducted in adult patients, 2) not related to either valproate or Vit D, 3) quality according to NIH quality assessment criteria (details are shown
not a “case-control” study design, or 4) lacked Vit D data. in Table S2). Major studies (6 out of 11) were conducted in Mid-eastern
countries such as Turkey [14,22,30–32] and Iran [19]. Cross-sectional
2.3. Data extraction and study quality evaluation study design was the most common method applied among the 11
studies. Only one article [22] did not provide the gender information of
We collected information from each selected article, e.g., first au- its subjects. Most of the studies included both male and female subjects,
thor, year of publication, countries of patients, study design, Vit D le- but they did not report Vit D status in each gender group treated with
vels in patients and control groups and details associated with valproate valproate monotherapy. These studies all included a control group
therapy. The methodological quality of each study was rated using the consisting of subjects not on valproate medication treatment. Two
NIH quality assessment tool for observational cohort and cross-sectional studies [21,33] did not include healthy children as control group. An-
studies (https://www.nhlbi.nih.gov/health-topics/study-quality- other study [19] recruited untreated patients with epilepsy to be con-
assessment-tools). All assessments were completed by two people trols. The duration of valproate treatment was more than six months in
working separately. If any inconsistent result for a particular article all the studies including years of valproate therapy in some reports.
occurred, it was resolved by discussions among the entire group con- Most of the eligible studies had more than two “case-control” com-
ducting this analysis. parisons, e.g., valproate treatment vs. healthy controls, carbamazepine
treatment vs. valproate treatment groups. All the studies measured 25-
2.4. Statistical analysis OH-Vit D levels, except for two [21,32] which reported 25-OH-Vit D3
and non-specified Vit D, respectively. The detailed characteristics of
The statistical analyses were conducted using the software STATA individual study are shown in Table 1.

61
 Z. Xu, et al.

Table 1
Characteristics of the 11 eligible studies selected to perform meta-analysis.

First Author Publication Country of Study type M/F (n) Age (years, mean ± SD) VPA Duration VPA dosage (mean ± SD)
date patients in VPA group of VPA group (mean ± SD)

Gabriele Rieger-Wettengl 2001 Germany CS 9/10 12.5 ± 3.7 ≥ 1 year NA

Sefer Kumandas 2006 Turkey CS 17/16 8.8 ± 2.0 33.72 ± 15 months NA
Retro-
Arzu Babayigit 2006 Turkey CS 15/16 11.18 ± 4.07 3.32 ± 1.09 years 15-40 mg/kg/day
Markus Rauchenzauner 2010 Austria CS cohort 38/ 47 12.42 ± 3.33 > 6 months 869 ± 380 mg / day
Ayse Aksoy 2011 Turkey CS 28/25 8.42 ± 2.03 3.23 ± 1.09 years 19.23 ± 4.42 mg/kg/day
Mehmet Ibrahim Turan 2014 Turkey CS NA 8.2 ± 3.8 ≥6 months NA
Jung-Hyun Baek 2014 Korea Retro- 90/53 11.21 ± 4.49 4.92 ± 3.68 years NA
Omid Yaghini 2015 Iran Retro- 68/52 7.4 ± 2.4 ≥ 6 months NA
Hepsen Mine Serin 2015 Turkey CS 32/27 8.6 ± 4.6 ≥ 2 years NA
Teodoro Durá-Travé 2018 Spain CS 18/41 School age (n = 30) and 2.5 ± 1.4 years 20.7 ± 4.7 mg/kg/day
Adolescent (n = 29)
Mini Sreedharan 2018 India CS 13/25 8.2 ± 2.9 ≥ 6 months NA

First Author BMI Assay for Vit D in VPA- Vit D in healthy Vit D in CAR- Vit D in other Other medication Note Ref

62
(kg/m2) VitD treated group(N; control group(N; treated group(N; medication treated
Mean(SD) ng/mL) Mean(SD) ng/mL) Mean(SD) ng/mL) group(N; Mean
(SD) ng/mL)

Gabriele Rieger-Wettengl NA RI 19;25(14) NA 20;18(7) NA NA [33]

Sefer Kumandas NA HPLC 33;15.1(3.5) 22;16.6(4.7) 33;9.8(3.2) NA NA [30]
Arzu Babayigit 17.49 ± 2.85 NA 31;17.87(7.27) 30;22.3(7.12) 23;20.47(11.6) 14;24.07(14) Oxcarbazepine [14]
Markus Rauchenzauner 0.4 ± 1.3 CEI 85;43.2(25.6) 41;49.9(33.8) NA 40;52.6(31.5) Oxcarbazepine, Sulthiame, or Lamotrigine BMI (SD [18]
score)
# # # #
Ayse Aksoy 17.44 ± 2.92 RI 53;37.68(30.73) 50;38.82(27.33) 23;26.12(24.19) NA NA nmol/L [31]
Mehmet Ibrahim Turan NA ECL 51;22.6(77.8) 44;23.1(78.3) 45;16.8(44.9) 48;20.1(88.7) phenobarbital [22]
Jung-Hyun Baek NA NA 49;36.83(11.66) NA NA 30;30.41(9.53) Oxcarbazepine [21]
Omid Yaghini NA IRA 30;14.51(1.82) NA* NA 60;10.88(1.14) Phenobarbital, Carbamazepine, and Primidone *Epileptic patients [19]
withouttreatment
as control
Hepsen Mine Serin NA NA 28;19.5(8.1) 20;21.9(7) 11;20.19(7.02) 20;22.8(11.08) Levetiracetam [32]
Teodoro Durá-Travé 0.02 ± 0.59 CL 59;23.37(9.11) 244;26.97(7.09) NA 31;22.64(9.08) Levetiracetam BMI (Z-score) [23]
Mini Sreedharan 15.1 ± 2.8 ELISA 28;23.4(11.7) 109;31.2(17.5) 28;26.3(21.9) NA NA [20]

Note: CAR, carbamazepine; CEI, Competitive enzyme immunoassay; CL, Chemiluminescence; CS, cross-sectional; ECL, Electrochemiluminescence; IRA, Immunoradiometric assay; M/F, Male/Female; Retro-, retro-
spective; RI, Radioimmunoassay; NA, not available; SD, standard deviation;
Seizure: European Journal of Epilepsy 71 (2019) 60–65
Z. Xu, et al. Seizure: European Journal of Epilepsy 71 (2019) 60–65

Fig. 1. Forest plot regarding valproate decreased Vit D level in epileptic children compared with healthy children.

3.3. Difference of Vit D levels among valproate- or carbamazepine-treated medication therapy, as well as that in children who received carba-
children with epilepsy and healthy controls mazepine monotherapy vs. healthy children. All the results of subgroup
analyses are shown in Table 2.
We first pooled comparison of Vit D level in valproate treated pa-
tients versus healthy controls. A negative SMD of Vit D level was noted
between valproate treated group and healthy controls (SMD = -0.313, 3.5. Publication bias
[-0.457, -0.169]), see Fig. 1.), indicating a significant difference across
these two groups in terms of a negative effect of valproate treatment on No potential publication bias was shown when comparing Vit D level
Vit D level. No evidence of heterogeneity was present across the related in children treated by valproate monotherapy vs. healthy controls based
studies (I2 = 7.1%, p = 0.376). on evidence from the Begg’s and Egger’s tests (Begg’s statistic p = 0.902,
Cumulative analytic technology was utilized to address the effect of Egger’s statistic p = 0.963). Similar results were observed in the analyses
time on the difference in Vit D status. As shown in Fig S2, cumulative of Vit D status in children with epilepsy treated by carbamazepine
estimates were progressively more robust as evidenced by the nar- therapy vs. healthy children (Begg’s statistic p = 0.260, Egger’s statistic
rowing 95% confident intervals by ascending year of publication. p = 0.311); valproate monotherapy vs. children with epilepsy treated by
We also combined data from carbamazepine treated children com- carbamazepine (Begg’s statistic p = 0.764, Egger’s statistic p = 0.739;
pared to healthy children, and a negative pooled (SMD = -0.484 valproate vs. other medication treatments in subjects with epilepsy
[-0.919, -0.049]) on Vit D level was observed in the carbamazepine (Begg’s statistic p = 0.548, Egger’s statistic p = 0.516)
treated group compared with healthy participants. However, significant
heterogeneity was present across studies (I2 = 76%, p = 0.001). In Table 2
total, both valproate and carbamazepine were associated with a de- Potential source of heterogeneities.
creased Vit D level in children with epilepsy when compared with
Variables I-squared p value
healthy children. Interestingly, carbamazepine and valproate had a si-
milar effect on the Vit D level as shown by directly comparing Vit D Vitamin D in patients treated with carbamazepine vs. healthy controls
concentration in valproate-treated to carbamazepine-treated children Turkey 80.20% 0.000
India – –
(SMD = -0.309 [-0.155, -0.773]) with significant heterogeneity among
Cross-sectional 0 0.85
studies (I2 = 80.8%, p = 0.000). Cross-sectional retrospective – –
Furthermore, there was no difference in Vit D level in valproate- Overall 76.00% 0.001
treated children with epilepsy from the level in other anti-epileptic Vitamin D in patients treated with valproate vs. carbamazepine
medication groups (SMD = 0.277, 95%CIs [-0.407, 0.961]), however, Germany – –
Turkey 85.30% 0.000
heterogeneity was suggested (I2 = 93.0%, p = 0.000). Two articles India – –
[21,32] reported 25-OH-Vit D3 and non-specified Vit D status instead of Cross-sectional 30.70% 0.205
25-OH-Vit D data. We performed pooled analysis de novo by excluding Cross-sectional retrospective – –
these two references. The results showed that using 25-OH-Vit D data Overall 80.80% 0.000
Vitamin D in patients treated with valproate vs. other medications
alone did not change our conclusions, producing only minor insignif-
Cross-sectional 31.70% 0.222
icant changes in the estimations of the pooled SMDs and their 95% CIs. Cross-sectional cohort – –
Retrospective 96.40% 0.000
Turkey 39.20% 0.193
3.4. Potential sources of heterogeneity Austria – –
Korea – –
Iran – –
The study design and country of origin of children was used as Spain – –
stratifying variables in conducting subgroup analyses to reveal poten- Overall 93.00% 0.000
tial heterogeneities. The results showed heterogeneities in both sub-
groups regarding comparison of Vit D levels in children who underwent Note: "-" means due to the limited study, I-squared and p value cannot be cal-
valproate vs. carbamazepine monotherapy, valproate vs. other culated.

63
Z. Xu, et al. Seizure: European Journal of Epilepsy 71 (2019) 60–65

4. Discussion prevent Vit D deficiency is warranted in children on chronic valproate

therapy.
We collected data from 11 eligible studies with a total sample size of
640, resulting in the most comprehensive meta-analysis to the best of Author contributions
our knowledge estimating Vit D status in children with epilepsy related
to valproate monotherapy. Several publications [6,28,29,34–43] were F. Chen and T. Wang designed this paper. J.Y. Sun, H.L. Guo, Y.H.
excluded from the meta-analysis due to the unavailability of retrievable Hu, X.Y. Wen, F. SuDepartment of Clinical Pharmacy and X.P. Lu dis-
raw data on Vit D status. Some of them [35,37] concluded that cussed this topic and provided comments. Z.J. Xu, X. Jing and G.Z. Li
valproate monotherapy had a negative effect on Vit D level while others read through all references and summarized the data. T. Wang and F.
concluded valproate had no effect the Vit D levels. Since the meta- Chen wrote the manuscript.
analysis combined weighted estimates from each study, it is not pos-
sible to estimate the effect inclusion of these studies would have on our Funding
meta-analysis.
Our results demonstrated a consistent decrease in the Vit D level in None.
children with epilepsy treated with valproate. Similar findings were
found in children treated with carbamazepine. The difference in the Acknowledgements
decline in the Vit D level between the valproate treatment group and
the carbamazepine group was not significant. Our findings on the Vit D The authors thank the UTHSC Office of Scientific Writing (Memphis,
decrease in carbamazepine monotherapy was in agreement with a TN, USA) for copy editing service. We are also grateful to Prof. S. Casey
previously published meta-analysis [44]. When comparing results on Laizure from the Department of Clinical Pharmacy, UTHSC (Memphis,
Vit D level in children on valproate therapy, the previous meta-analysis TN, USA) for revising our manuscript.
[44] did not collect all reasonable relevant primary articles but re-
ported that valproate treated patients had similar Vit D level as their Appendix A. Supplementary data
control group. Due to the inclusion of only three primary articles, their
analysis is deemed less robust than our meta-analysis. Supplementary data associated with this article can be found, in the
Multiple factors, like diet, daily activities and sunlight exposure can online version, at https://doi.org/10.1016/j.seizure.2019.06.009.
affect Vit D status. These factors were not available as variables to
perform combined analysis because of insufficient information from References
individual studies. Other factors, such as duration of valproate treat-
ment, also influenced Vit D status. For example, 25-OH Vit D3 levels [1] Ihle M, Feldwisch-Drentrup H, Teixeira CA, Witon A, Schelter B, Timmer J, et al.
were significantly lower among children who were on AEDs therapy for EPILEPSIAE - a European epilepsy database. Comput Methods Programs Biomed
2012;106:127–38.
longer than 2 years [21] and longer duration of therapy was associated [2] Klatt J, Feldwisch-Drentrup H, Ihle M, Navarro V, Neufang M, Teixeira C, Adam C,
with a greater negative impact on vit D levels [45]. We were able to Valderrama M, Alvarado-Rojas C, Witon A, Le Van Quyen M, Sales F, Dourado A,
collect length of valproate treatment and all were long-term treatment Timmer J, Schulze-Bonhage A, Schelter B. The EPILEPSIAE database: an extensive
electroencephalography database of epilepsy patients. Epilepsia 2012;53:1669–76.
in the eligible studies. However, most of the data were noted as “more [3] Pack A. Bone health in people with epilepsy: is it impaired and what are the risk
than” a certain months or years, making classification of subjects by factors? Seizure 2008;17:181–6.
duration of therapy unavailable for analysis. [4] Christakos S, Dhawan P, Verstuyf A, Verlinden L, Carmeliet G. Vitamin D: meta-
bolism, molecular mechanism of action, and pleiotropic effects. Physiol Rev
According to the United States Endocrine Society criteria for clas- 2016;96:365–408.
sification of Vit D status, Vit D deficiency is defined by a 25-OH Vit D [5] Sato Y, Kondo I, Ishida S, Motooka H, Takayama K, Tomita Y, Maeda H, Satoh K.
level less than 20 ng/ml (< 50 nmol/L) [23]. In the current meta-ana- Decreased bone mass and increased bone turnover with valproate therapy in adults
with epilepsy. Neurology 2001;57:445–9.
lysis, Vit D deficiency was not assessed among half (5 in 11) of the
[6] Farhat G, Yamout B, Mikati MA, Demirjian S, Sawaya R, El-Hajj Fuleihan G. Effect
primary articles. In the remaining 6 articles, inconsistent criteria were of antiepileptic drugs on bone density in ambulatory patients. Neurology
used to determine Vit D deficiency in children, including 25-OH-Vit 2002;58:1348–53.
D < 10 ng/mL [18,19], < 12 ng/mL [20] and < 20 ng/mL [21–23]. [7] McCorry DJ. Effect of antiepileptic drugs on bone density in ambulatory patients.
Neurology 2004;62:342. author reply 342.
Therefore, we did not perform meta-analysis on whether the valproate [8] Perucca E. Clinically relevant drug interactions with antiepileptic drugs. Br J Clin
therapy contribute to deficiency of Vit D in children with epilepsy. Pharmacol 2006;61:246–55.
Study design and country of origin may be a source introducing [9] Vrzal R, Doricakova A, Novotna A, Bachleda P, Bitman M, Pavek P, et al. Valproic
acid augments vitamin D receptor-mediated induction of CYP24 by vitamin D3: a
heterogeneities to the meta-analysis of Vit D levels in children on possible cause of valproic acid-induced osteomalacia? Toxicol Lett
valproate or carbamazepine therapy as our results show. Potential ex- 2011;200:146–53.
planations such as seasonally related sun exposure, genetic variation, [10] Sawada N, Kusudo T, Sakaki T, Hatakeyama S, Hanada M, Abe D, Kamao M, Okano
T, Ohta M, Inouye K. Novel metabolism of 1 alpha,25-dihydroxyvitamin D3 with
differing diets, and variations in study methodology could cause het- C24-C25 bond cleavage catalyzed by human CYP24A1. Biochemistry
erogeneity [46]. In particular, our clinical observation (data not shown) 2004;43:4530–7.
and data from a separate research group [23] noted that gender of [11] Cerveny L, Svecova L, Anzenbacherova E, Vrzal R, Staud F, Dvorak Z, Ulrichova J,
Anzenbacher P, Pavek P. Valproic acid induces CYP3A4 and MDR1 gene expression
participants was a contributable factor to decreased Vit D in patients by activation of constitutive androstane receptor and pregnane X receptor path-
with epilepsy treated with valproate. Further study is warranted to ways. Drug Metab Dispos 2007;35:1032–41.
determine other important factors that contribute to low vitamin D [12] Bouillon R. Comparative analysis of nutritional guidelines for vitamin D. Nat Rev
Endocrinol 2017;13:466–79.
levels in children on valproate therapy.
[13] Bouillon R, Marcocci C, Carmeliet G, Bikle D, White JH, Dawson-Hughes B, Lips P,
Munns CF, Lazaretti-Castro M, Giustina A, Bilezikian J. Skeletal and extra-skeletal
5. Conclusions actions of vitamin D: current evidence and outstanding questions. Endocr Rev 2018.
[14] Babayigit A, Dirik E, Bober E, Cakmakci H. Adverse effects of antiepileptic drugs on
bone mineral density. Pediatr Neurol 2006;35:177–81.
This meta-analysis found a decrease of Vit D level in children with [15] Cebeci AN, Ekici B. Epilepsy treatment by sacrificing vitamin D. Expert Rev
epilepsy receiving valproate monotherapy compared with healthy Neurother 2014;14:481–91.
children. The decrease in Vit D level in valproate treated pediatric [16] Munns CF, Shaw N, Kiely M, Specker BL, Thacher TD, Ozono K, Michigami T,
Tiosano D, Mughal MZ, Makitie O, Ramos-Abad L, Ward L, DiMeglio LA, Atapattu
patients was similar to that found in children on carbamazepine N, Cassinelli H, Braegger C, Pettifor JM, Seth A, Idris HW, Bhatia V, Fu J, Goldberg
therapy. In view of the impact of Vit D deficiency on childhood de- G, Savendahl L, Khadgawat R, Pludowski P, Maddock J, Hypponen E, Oduwole A,
velopment, routine monitoring of Vit D levels and supplementation to Frew E, Aguiar M, Tulchinsky T, Butler G, Hogler W. Global consensus

64
Z. Xu, et al. Seizure: European Journal of Epilepsy 71 (2019) 60–65

recommendations on prevention and management of nutritional rickets. J Clin receiving anticonvulsant monotherapy with valproic acid or carbamazepine.
Endocrinol Metab 2016;101:394–415. Pediatrics 2001;108:E107.
[17] Nevitt SJ, Sudell M, Weston J, Tudur Smith C, Marson AG. Antiepileptic drug [34] Nettekoven S, Strohle A, Trunz B, Wolters M, Hoffmann S, Horn R, Steinert M,
monotherapy for epilepsy: a network meta-analysis of individual participant data. Brabant G, Lichtinghagen R, Welkoborsky HJ, Tuxhorn I, Hahn A. Effects of anti-
Cochrane Database Syst Rev 2017;12:CD011412. epileptic drug therapy on vitamin D status and biochemical markers of bone
[18] Rauchenzauner M, Griesmacher A, Tatarczyk T, Haberlandt E, Strasak A, turnover in children with epilepsy. Eur J Pediatr 2008;167:1369–77.
Zimmerhackl LB, Falkensammer G, Luef G, Hogler W. Chronic antiepileptic [35] Nicolaidou P, Georgouli H, Kotsalis H, Matsinos Y, Papadopoulou A, Fretzayas A,
monotherapy, bone metabolism, and body composition in non-institutionalized Syriopoulou V, Krikos X, Karantana A, Karpathios T. Effects of anticonvulsant
children. Dev Med Child Neurol 2010;52:283–8. therapy on vitamin D status in children: prospective monitoring study. J Child
[19] Yaghini O, Tonekaboni SH, Amir Shahkarami SM, Ahmad Abadi F, Shariat F, Neurol 2006;21:205–9.
Abdollah Gorji F. Bone mineral density in ambulatory children with epilepsy. [36] Borusiak P, Langer T, Heruth M, Karenfort M, Bettendorf U, Jenke AC. Antiepileptic
Indian J Pediatr 2015;82:225–9. drugs and bone metabolism in children: data from 128 patients. J Child Neurol
[20] Sreedharan M, Devadathan K, Mohammed Kunju PA, Sasidharan B, Pillai JP, 2013;28:176–83.
Vasumathy Amma MA, et al. Vitamin d deficiency in Ambulant Children on car- [37] Chaudhuri JR, Mridula KR, Rathnakishore C, Balaraju B, Bandaru VS. Association of
bamazepine or sodium valproate monotherapy. Indian Pediatr 2018;55:307–10. 25-Hydroxyvitamin D deficiency in pediatric epileptic patients. Iran J Child Neurol
[21] Baek JH, Seo YH, Kim GH, Kim MK, Eun BL. Vitamin D levels in children and 2017;11:48–56.
adolescents with antiepileptic drug treatment. Yonsei Med J 2014;55:417–21. [38] He X, Jiang P, Zhu W, Xue Y, Li H, Dang R, Cai H, Tang M, Zhang L, Wu Y. Effect of
[22] Turan MI, Cayir A, Ozden O, Tan H. An examination of the mutual effects of val- antiepileptic therapy on serum 25(OH)D3 and 24,25(OH)2D3 levels in epileptic
proic acid, carbamazepine, and phenobarbital on 25-hydroxyvitamin D levels and children. Ann Nutr Metab 2016;68:119–27.
thyroid function tests. Neuropediatrics 2014;45:16–21. [39] Guo CY, Ronen GM, Atkinson SA. Long-term valproate and lamotrigine treatment
[23] Dura-Trave T, Gallinas-Victoriano F, Malumbres-Chacon M, Moreno-Gonzalez P, may be a marker for reduced growth and bone mass in children with epilepsy.
Aguilera-Albesa S, Yoldi-Petri ME. Vitamin D deficiency in children with epilepsy Epilepsia 2001;42:1141–7.
taking valproate and levetiracetam as monotherapy. Epilepsy Res 2018;139:80–4. [40] Vera V, Moran JM, Barros P, Canal-Macias ML, Guerrero-Bonmatty R, Costa-
[24] Chaimani A, Mavridis D, Salanti G. A hands-on practical tutorial on performing Fernandez C, Lavado-Garcia JM, Roncero-Martin R, Pedrera-Zamorano JD. Greater
meta-analysis with Stata. Evid Based Ment Health 2014;17:111–6. calcium intake is associated with better bone health measured by quantitative ul-
[25] Wang T. The link between Parkinson’s disease and breast and prostate cancers: a trasound of the phalanges in pediatric patients treated with anticonvulsant drugs.
meta-analysis. Int J Neurosci 2015;125:895–903. Nutrients 2015;7:9908–17.
[26] Wang T. Glutathione S-transferases variants as risk factors in Alzheimer’s disease. [41] Onoe S, Mimaki T, Seino Y, Yamaoka K, Yabuuchi H, Yamashita G. Delayed bone
Neurol Sci 2015;36:1785–92. development in epileptic children assessed by microdensitometer. Dev Pharmacol
[27] Wang T, Marei HE. Landscape of NAT2 polymorphisms among breast cancer. Ther 1988;11:24–31.
Biomed Pharmacother 2016;77:191–6. [42] Davie MW, Emberson CE, Lawson DE, Roberts GE, Barnes JL, Barnes ND, et al. Low
[28] Fong CY, Riney CJ. Vitamin D deficiency among children with epilepsy in South plasma 25-hydroxyvitamin D and serum calcium levels in institutionalized epileptic
Queensland. J Child Neurol 2014;29:368–73. subjects: associated risk factors, consequences and response to treatment with vi-
[29] Fong CY, Kong AN, Poh BK, Mohamed AR, Khoo TB, Ng RL, Noordin M, Nadarajaw tamin D. Q J Med 1983;52:79–91.
T, Ong LC. Vitamin D deficiency and its risk factors in Malaysian children with [43] Verrotti A, Agostinelli S, Coppola G, Parisi P, Chiarelli F. A 12-month longitudinal
epilepsy. Epilepsia 2016;57:1271–9. study of calcium metabolism and bone turnover during valproate monotherapy. Eur
[30] Kumandas S, Koklu E, Gumus H, Koklu S, Kurtoglu S, Karakukcu M, et al. Effect of J Neurol 2010;17:232–7.
carbamezapine and valproic acid on bone mineral density, IGF-I and IGFBP-3. J [44] Zhang Y, Zheng YX, Zhu JM, Zhang JM, Zheng Z. Effects of antiepileptic drugs on
Pediatr Endocrinol Metab 2006;19:529–34. bone mineral density and bone metabolism in children: a meta-analysis. J Zhejiang
[31] Aksoy A, Sonmez FM, Deger O, Hosver I, Karaguzel G. The effects of antiepileptic Univ Sci B 2015;16:611–21.
drugs on the relationships between leptin levels and bone turnover in prepubertal [45] Lee YJ, Park KM, Kim YM, Yeon GM, Nam SO. Longitudinal change of vitamin D
children with epilepsy. J Pediatr Endocrinol Metab 2011;24:703–8. status in children with epilepsy on antiepileptic drugs: prevalence and risk factors.
[32] Serin HM, Koc ZP, Temelli B, Esen I. The bone mineral content alterations in pe- Pediatr Neurol 2015;52:153–9.
diatric patients medicated with levetiracetam, valproic acid, and carbamazepine. [46] Wang T. Meta-analysis of PvuII, XbaI variants in ESR1 gene and the risk of
Epilepsy Behav 2015;51:221–4. Alzheimer’s disease: the regional European difference. Neurosci Lett
[33] Rieger-Wettengl G, Tutlewski B, Stabrey A, Rauch F, Herkenrath P, Schauseil-Zipf 2014;574:41–6.
U, et al. Analysis of the musculoskeletal system in children and adolescents

65