Address correspondence and reprint requests to Dr. K. Mikkonen at Department of Pediatrics, P.O. Box 5000, FIN-90014, University of Oulu, Finland. E-mail: firstname.lastname@example.org
Summary: Purpose: To assess growth and the serum lipid profile in girls with epilepsy receiving monotherapy at a mean age of 12.6 years and approximately 6 years later.
Methods: A population-based cohort of 77 girls with epilepsy and 49 healthy controls participated in this follow-up study including two cross-sectional evaluations (age range, 8–18.5 years on the first evaluation, and 12.5–25.8 years on the second evaluation). Forty of the patients were initially taking valproate (VPA), 19, carbamazepine (CBZ), and 18, oxcarbazepine (OXC). Growth data were compiled, body mass index (BMI) was calculated, and serum total (TC), and high-density lipoprotein (HDL-C) and low-density lipoprotein (LDL-C) cholesterol and triglyceride concentrations were analyzed.
Results: Linear growth and final height did not differ between the patients and the controls. At follow-up, the mean BMI of the patients who were off medication (61%) was similar to that of the controls, whereas the patients initially treated with VPA who were still taking any medication had a higher BMI. On the first evaluation, the patients taking VPA had low serum HDL-C, and those taking CBZ or OXC had high serum TC and LDL-C concentrations. At follow-up, serum lipid levels were similar in the patients off medication and the controls.
Conclusions: Neither epilepsy nor antiepileptic therapy affects linear growth or final height, but they may have unfavorable effects on body weight and serum lipid concentrations. Lipid-profile impairment seems to be transient if the medication is discontinued. Overweight is common in patients treated with VPA during puberty if epilepsy and medication continue into adulthood.
Valproate (VPA) is a widely used wide-spectrum antiepileptic drug (AED), which has been implicated to be associated with weight gain in patients with epilepsy (1). VPA-related obesity was reported to be common in women with epilepsy if the medication was initiated before age 20 years (2), and weight gain seemed to become more frequent, along with pubertal maturation (3). The long-term influence of VPA treatment on growth is poorly defined. VPA had no effect on linear growth (3,4) or pubertal development (3) in relatively short-term follow-up studies, but some contradictory findings also exist (5,6). Moreover, the reproductive endocrine disorders detected in female subjects with epilepsy, particularly those with VPA treatment (2,7), are associated with obesity and hyperinsulinemia (8,9).
Changes in serum lipid or lipoprotein levels also have been reported in children taking VPA (10,11) and carbamazepine (CBZ) (10–12), but one follow-up study indicated no permanent effects of these medications on lipid metabolism (11). Elevated serum lipid concentrations also were observed in children with epilepsy taking oxcarbazepine (OXC) (13). As a whole, the effects of OXC on serum lipid levels are poorly characterized, and until now, no controlled studies have been performed in children.
The long-term effects of medication with AEDs during childhood and adolescence on growth, weight, and lipid metabolism are not known. The purpose of the present study was to evaluate growth, especially final height and weight, and the long-term effects of VPA, CBZ, or OXC treatment during puberty on serum lipid profiles in girls and young women with epilepsy.
PATIENTS AND METHODS
This study was carried out in the Departments of Pediatrics and Adolescence and Neurology, Oulu University Hospital, Finland. The hospital is a primary care center for patients with epilepsy in an area with a total population of ∼300,000. The local Ethics Committee approved the protocol, and the study was performed according to the principles of the Declaration of Helsinki. Written informed consent was obtained from all study subjects or their parents before the study.
Subjects and study design
The study protocol included two cross-sectional evaluations with an average time interval of 5.8 years (SD, 1.0 year). A previously identified population-based cohort (3) of 77 girls and young women with epilepsy receiving monotherapy and 49 healthy control girls, 8 to 18 years old (mean, 12.6 years), took part in the first evaluation (Table 1). The growth of these subjects up to the first examination was reported earlier (3). Sixty-nine (90%) patients and 47 (96%) healthy control subjects agreed to take part also in the second assessment. Their ages at the follow-up examination ranged from 12.5 to 25.8 years (mean, 18.4 years). One patient had died, and nine subjects either refused to participate or could not be reached. During the follow-up period, one patient in the OXC group and one control subject had autoimmune thyroiditis, and both subjects were excluded from the statistical analysis on the second occasion.
Table 1. Clinical characteristics of the subjects
Medication on the first evaluation
The values are the numbers of subjects or means.
VPA, valproate; CBZ, carbamazepine; OXC, oxcarbazepine; LTG, lamotrigine; SD, standard deviation; Pubertal, B 2–4 PH 2–4; Postpubertal, B 5 PH 5–6/B 4 PH 4 + adult bone age; B, breast; PH, pubic hair; L, localization-related epilepsy; G, generalized epilepsy.
Number of subjects
Age, yr (SD)
Duration of medication, yr (SD)
Number of subjects
Age, yr (SD)
Age at menarche, yr (SD)
Pubertal status, no.
Off medication, no. (%)
Duration of initial medication, yr (SD)
Time without medication, yr (SD)
Type of epilepsy, no.
On medication, no. (%)
Duration of current medication, yr (SD)
Type of epilepsy, no.
All the subjects were clinically examined by one of the authors (K.M. or L.V.). Medical and menstrual histories were obtained by interviewing the study subjects and by examining their hospital records. Type of epilepsy was classified according to the international classification of epilepsies and epileptic syndromes (14). AEDs were initially started according to the type of epilepsy (e.g., VPA was used mainly for patients with primary generalized epilepsy). The methods used in the present study were previously described in detail (3). Subjects with a body mass index (BMI) exceeding 25 kg/m2 were considered obese (15). On the first examination 71 of the 77 patients and 46 of the 49 control girls, and on the second occasion of 64 (93%) of the 69 patients and 39 (83%) of the 47 control subjects agreed to undergo blood sampling and laboratory tests. A study subject was considered to have hypercholesterolemia if her serum total cholesterol (TC) concentration exceeded the mean serum TC + 2 SD in the control subjects at the corresponding stage of pubertal maturation. This reference limit was TC > 5.7 mM at all pubertal stages on the first evaluation.
The serum samples were kept frozen at −20°C until analyzed. The concentrations of serum TC, high-density lipoprotein cholesterol (HDL-C), and triglycerides (TGs) were analyzed on the first evaluation by the enzymatic colorimetric methods of Boehringer Mannheim GmbH (Mannheim, Germany) by using the Hitachi 911 Clinical Chemistry Analyzer (Boehringer Mannheim GmbH). The serum low-density lipoprotein cholesterol (LDL-C) concentrations were calculated by the Friedewald formula on the first evaluation (16). Respective analyses were carried out on the second evaluation by the enzymatic colorimetric methods of Roche Diagnostics (Rotkreuz, Switzerland) by using the Cobas Integra 700 analyzer (F. Hoffmann-La Roche Ltd, Basel, Switzerland).
The outcomes of the study were growth, final height, weight, and serum dyslipidemia. The PROC MIXED model in the SAS System for Windows (Release 8.02) was used to analyze the growth and weight charts (17). All the other data were analyzed with the Statistical Package for Social Sciences program (version 11.5; SPSS Inc., Chicago, IL, U.S.A.). One-way analysis of variance (ANOVA) with Tukey's post hoc test was used to compare parametric variables between more than two groups, and an unpaired t test between two groups. If ANOVA was not applicable because of nonhomogeneity of the variances, the results were confirmed by the Kruskal–Wallis test with the Mann–Whitney U test as a post hoc test. Bonferroni's correction was used when appropriate. The χ2 test was used to compare frequencies. Finally, to control for confounding factors, binary logistic regression analysis was performed.
At the follow-up visit, after ∼6 years, 42 (61%) of the 69 patients were off medication, whereas the others were still taking medication (see Table 1). All the patients taking medication had serum drug concentrations within the therapeutic window on both evaluations.
No difference in linear growth was found between the patients and the control subjects, but weight for height was higher in the patients initially taking VPA, but not in those taking CBZ or OXC (Figs. 1 and 2). On the second evaluation, no statistically significant difference was noted in the mean BMI between the patients off medication (n = 42; 21.2 kg/m2; SD, 4.1) and the control subjects (n = 47; 20.3 kg/m2; SD, 2.6; p = 0.46). In contrast, the patients still taking medication had a higher mean BMI (n = 27; 24.6 kg/m2; SD, 6.5; p < 0.002) than did the patients off medication or the control subjects. The prevalence of obesity did not differ between the patients taking different AEDs on the second evaluation. VPA medication and obesity at the first evaluation were the only predictive factors for obesity on the second occasion (Table 2).
Table 2. Results of binary logistic regression analysis
p Value to remove
Not in equation
Obesity, BMI > 25 mg/kg2; BMI, weight (kg) divided by the square of height (m); OR, odds ratio; CI, confidence interval; VPA, valproate; CBZ, carbamazepine; OXC, oxcarbazepine.
Obesity on the second evaluation
Obesity on the first evaluation
Medication on the first evaluation
Stage of puberty
VPA vs. Controls
Type of epilepsy
CBZ vs. Controls
(No obese patients taking OXC)
Forty-seven (69%) of the 68 patients and 30 (65%) of the 46 control subjects had reached their final height, which was confirmed by the growth chart and bone age. The mean final height was 164.8 cm (SD, 6.7) in the patients and 165.5 cm (SD, 5.6) in the control subjects. The difference between final height and midparental height did not differ in the patients (0.04 SDS; SD, 1.0) and the control subjects (–0.24 SDS; SD, 0.8; p = 0.21).
Serum lipid profile
On the first evaluation, the serum TC and LDL-C concentrations were increased in the patients taking CBZ or OXC, whereas the patients taking VPA had decreased serum HDL-C levels (Table 3). After the follow-up period, no differences were seen in the serum lipid levels between the postpubertal patients off medication and the control subjects (see Table 3). The number of patients still taking the same AED as on the first evaluation was small (n = 11, all AEDs together), and their results are not shown in detail.
Table 3. Serum lipid concentrations in girls with epilepsy and controls during pubertal maturation and after the follow-up period (mean, 5.8 years; SD, 1.0)
First evaluation Medication
p Value (ANOVA)
Second evaluation of postpubertal girls off medication; Original medication
p Value (ANOVA)
The values are means (SD). On the first evaluation, the groups were matched for age and pubertal stage; on the second evaluation, only the results of postpubertal subjects off medication are shown.
ap < 0.005; bp < 0.05 compared with control subjects; ANOVA with post hoc Tukey's test.
Number of subjects
No change was found in linear growth or final height in the patients with epilepsy who had been treated with AEDs during pubertal maturation. Most of the patients with epilepsy initially examined during puberty were off medication ∼6 years later. The weight of the patients off medication on the second evaluation was comparable to that of the same cohort of control subjects as in the first evaluation. However, the subjects treated with VPA during pubertal development had higher body weight on the second evaluation. Moreover, overweight was common in the patients still taking medication, especially among those who had gained weight while initially treated with VPA. An unfavorable lipid profile was observed in the patients taking any of the AEDs studied on the first evaluation, but the lipid changes leveled off when medication was discontinued.
Linear growth remained normal in the present patients taking long-term CBZ and VPA monotherapy, and also in patients taking OXC, although increased bone turnover in children receiving CBZ (19) and a reduction in bone mineral density in those taking VPA were reported previously (5,20). Untoward effects of VPA on growth have been implicated, but the target height of the patients was not assessed in those uncontrolled surveys (5,6). In contrast, a series of studies reported normal growth in patients taking VPA or CBZ (3,4). Thus our data support the latter observations that patients with epilepsy have normal linear growth and final height (4,21). To our knowledge, the first evaluation of the present cohort is the only report published thus far on the effects of OXC on growth in patients with epilepsy (3).
Obesity and VPA medication at the time of the first evaluation were independent predictors of obesity at the end of follow-up. This is in accordance with previous findings suggesting that VPA may induce weight gain in young women (22,23). Although the type of epilepsy was not found to a predictor of obesity in this study, further investigations are needed to assess its possible contribution. Overweight at the time of the initiation of VPA, normal neurocognitive status, and primary generalized seizure type were recently reported to be predictors of high BMI during VPA medication (23). However, it is difficult to predict which patients will gain weight with VPA therapy, and the weight gain is not unequivocally related to the patient's initial weight before starting medication (6,24). Interestingly, weight gain was seen among those VPA-treated patients who continued medication into adulthood, but the VPA-treated patients whose medication could be discontinued had not gained weight, not even while they were taking VPA. This may be explained by genetic factors, differences in the severity of epilepsy, or interaction between VPA and the seizure disorder. However, the observational design of this study and the fact that many patients had had their AED changed between the two evaluations complicated the interpretation of weight changes over time. The present results suggest that girls and young women who gain weight with long-term treatment with VPA during puberty remain obese if epilepsy continues into adulthood, even if VPA is replaced with some other AED. In accordance with this, BMI decreased in adult women during the first year after replacement of VPA with lamotrigine (LTG) medication, but despite the weight loss, they did not reach normal weight during the 1-year follow-up (9). Controlled long-term follow-up studies are needed to evaluate whether the patients who gain weight and are obese for several years during VPA treatment are able to restore and maintain normal weight after discontinuation of VPA.
Overweight not only is a cosmetic problem, but also may interfere with treatment compliance (8), and it also may be a factor predisposing to dyslipidemia, high blood pressure, hyperinsulinemia, diabetes, and cardiac problems (25,26). The mechanisms of VPA-related weight gain are still not known in detail, but it is suggested to be mediated by VPA-induced inhibition in β-oxidation of fatty acids (27). Hyperinsulinemia, which potentially enhances appetite and energy storage, also was reported during VPA treatment (8,28,29). Hyperinsulinemia was associated with obesity in adult patients with epilepsy taking VPA (8), but not in girls undergoing pubertal maturation (3,7). However, high serum insulin levels also have been reported in lean adults taking VPA (28). In general, obesity is associated with changes in the serum lipid profile, such as low HDL-C and high TG concentrations. Reduced serum HDL-C concentrations have been reported previously (9) and were confirmed in our series of patients taking VPA, and VPA-treated patients were also more often obese than the other patients.
Serum TC and LDL-C concentrations were increased in the patients taking CBZ or OXC. Increased serum TC, LDL-C, and also HDL-C levels have previously been observed in children with epilepsy taking CBZ (10–12), whereas serum TG concentrations remained unchanged in patients taking different AEDs. OXC has been used to treat epilepsy for >10 years in Europe, but this is the first controlled study on lipid metabolism in children and adolescents taking OXC. We observed that long-term treatment with OXC appears to have an unfavorable effect on the serum lipid profile in children and adolescents, which is in agreement with an earlier short-term study in children receiving OXC therapy (13), but differs from the short-term observations in adults (30). In the present study, all the lipid changes observed in patients taking different AEDs had reverted by the second assessment if the medication had been discontinued. These findings support previous reports suggesting that the effects of AEDs on lipid metabolism are transient (9,11). Hepatic induction of the P450 enzyme system is assumed to underlie many of the serum lipid–level changes observed in patients with epilepsy taking CBZ (10–12,30–32). The mechanism by which OXC may affect lipid metabolism is not known, but OXC also has been suggested to be associated with some degree of liver enzyme induction (3,32).
To summarize, epilepsy does not seem to affect linear growth or final height in patients receiving VPA, CBZ, or OXC monotherapy during pubertal maturation. These results also imply that epilepsy during puberty is not associated with obesity or serum dyslipidemia later in life if the medication is discontinued before adult age. However, an increased prevalence of obesity was observed in young women who were treated with VPA and were obese during puberty, if their epilepsy continued and they remained taking any AED until adult age. Moreover, all of the three AEDs studied may affect lipid metabolism unfavorably. Therefore it is important to pay attention early to weight gain and serum lipid profile in young patients with epilepsy, especially in those who require medication of long duration.
Acknowledgment: We thank Mrs. Anja Heikkinen and Mrs. Sirpa Anttila for technical assistance, Johanna Rättyä, PhD, for data collection, and statisticians Risto Bloigu, MSc, and Ahti Niinimaa, PhD, for analyses with the SAS System for Windows. The study was supported by the Arvo and Lea Ylppö Foundation, Finland, the Alma and K. A. Snellman Foundation, Oulu, Finland, and the Department of Pediatrics and Adolescence, Oulu University Hospital, and its research subsidy from the Ministry of Social Affairs and Health.