The metabolic syndrome in overweight epileptic patients treated with valproic acid

Authors


Address correspondence to Alberto Verrotti, M.D., PhD., Department of Pediatrics, University of Chieti, Ospedale Policlinico, Via dei Vestini 5, Chieti, Italy. E-mail: averrott@unich.it

Summary

Purpose: To evaluate the presence of metabolic syndrome (MS) in children and adolescents treated with valproate (VPA).

Methods: One hundred fourteen patients (54 male and 60 female) were studied. These patients were followed from the beginning of therapy for at least 24 months; at the end of follow-up, 46 patients (40.4%) had a considerable increase in body weight, whereas the other patients (59.6%) remained with the same weight. The MS was defined as having at least three of the following: abdominal obesity, dyslipidemia, glucose intolerance, and hypertension.

Results: Forty-six patients developed obesity; 20 (43.5%) of 46 patients developed MS. Abnormal glucose homeostasis was identified in 45% of patients. High total serum cholesterol concentrations were noted in 10 (50%), high serum triglyceride concentrations in 7 (35%), and low high-density lipoprotein (HDL) in 15 (75%) of the 20 subjects with MS. However, there were no significant differences in the features of MS between boys and girls with MS.

Conclusions: Patients who gain weight during VPA therapy can develop MS with a possible risk of cardiovascular disease.

Valproic acid (VPA) is a broad spectrum antiepileptic drug (AED) that has been established as effective over the complete range of seizure types, with particular value for the idiopathic generalized epilepsies (Losher, 1981; Fariello et al., 1995). VPA can cause considerable increase in body weight, and VPA-induced obesity seems to be associated with many metabolic and endocrine disturbances (Isojarvi et al., 1993; Verrotti et al., 2002; Pylvanen et al., 2006a, 2006b;Verrotti et al., 2008). Obesity has a role in promoting the development of metabolic diseases including glucose intolerance, dyslipidemia, hypertension, and atherosclerosis (Steinberger & Daniels, 2003).

In 1988, Reaven and Colleagues described “the metabolic syndrome” (MS) as a link between insulin resistance and hypertension, dyslipidemia, type 2 diabetes, and other metabolic abnormalities associated with an increased risk of atherosclerotic cardiovascular disease in adults (Reaven and Banting Lecture 1988, 1988; Levitt & Lambert, 2002; Ozanne & Hales, 2002).

Obesity, which is the most common cause of insulin resistance in children, is also associated with dyslipidemia, type 2 diabetes, and long-term vascular complications (Arslanian, 2002; Caprio, 2002; Steinberger & Daniels, 2003). In a sample of adolescents in the United States who were included in the Third National Health and Nutrition Examination Survey (NHANES III), conducted between 1988 and 1994, the prevalence of MS was 6.8% among overweight adolescents and 28.7% among obese adolescents (Cook et al., 2003). The prevalence of MS in childhood was reported to be about 10%, whereas this rate climbs to 30–50% in overweight/obese children, 39% in moderately and 49% in severely obese youngsters (Cook et al., 2003; Sen et al., 2008). Despite many studies that have addressed the problem of obesity in VPA-treated patients (Egger & Brett, 1981; Covanis et al., 1982; Isojarvi et al., 1993; Novak et al., 1999; Verrotti et al., 1999; Pylvanen et al., 2002), there are no data about the prevalence of MS in VPA-treated epileptics.

This study was performed to determine the presence and risk factors of MS in obese patients treated with VPA monotherapy.

Methods

A total of 114 epileptic patients (54 male and 60 female) were referred to the Department of Pediatrics, University of Chieti. Their mean age was 10.1 ± 4.7 years.

All patients were treated with VPA monotherapy; the drug was prescribed at the usual dosage (median 30.4 mg/kg/day) and all patients showed plasma drug levels within the optimal range [mean ± standard deviation (SD) 61.9 ± 7 μg/ml).

Exclusion criteria were: abnormal neurologic examination; abnormal cerebral computed tomography (CT) scan or magnetic resonance imaging (MRI); thyroid disease; and other endocrinopathies; liver, or heart, or kidney diseases; cancer; any disease likely to affect lipid metabolism; any inflammatory disease; chromosomal abnormalities; and a body mass index (BMI) >25. All patients were Caucasian children.

Written informed consent was obtained from the parents of all subjects studied; the study was approved by the Ethical Committee of the University of Chieti.

Anthropometric data and fasting blood samples were obtained in obese children. Height and weight were measured with a wall-mounted stadiometer and calibrated scale. Quetelet index BMI was then calculated as weight (kg) divided by height (m2), and this was used an indirect measure of adiposity. Waist circumference was measured at the level midway between the lateral rib margin and the iliac crest. Hip circumference was measured at the levels of the major trochanters through the pubic symphysis, and the waist-to-hip ratio was calculated. Age- and sex-specific SD scores (SDS) for height, weight, and BMI were calculated according to Italian reference data (Cacciari et al., 2002). Obesity has been defined as BMI ≥95th percentile for age and sex according to the Italian cross-sectional growth charts (Cacciari et al., 2002). All subjects with diagnosis of MS underwent a complete physical examination with auxologic measurements, and were divided according to Tanner’s stages into prepubertal and pubertal patients (Marshall & Tanner, 1969, 1970). Prepuberty was defined as Tanner stage 1. Pubertal patients (Tanner stages 2–5) were defined by appearance of secondary sex characteristics, in particular the pubic and axillary hairs, appearance of breast in females, and testicular enlargement in males. Pubertal testicular volume was >4 ml and was established by comparison with ellipsoids of known volume (Prader’s orchidometer).

Diagnosis of MS is defined as having at least three of following abnormalities: obesity, hypertriglyceridemia (at least 150 mg/dl), low high-density lipoprotein (HDL) cholesterol (<40 mg/dl in men and <50 mg/dl in women), hypertension (at least 130/85 mm Hg), and impaired glucose tolerance; according to the Adult Treatment Panel III (National Institutes of Health, 2001). Because there is no current definition of the MS in children, we defined MS as the presence of at least three of the following criteria:

  •  Obesity:
    •  Abdominal obesity with BMI ≥95th percentile for age and gender;
    •  Waist circumference ≥95th percentile for age and gender.
  •  Dyslipidemia:
    •  High triglyceride (>105 mg/dl in children ≤10 years of age, >136 mg /dl in children >10 years of age);
    •  HDL cholesterol less than 45 mg/dl;
    •  Low-density lipoprotein (LDL) cholesterol ≥95th percentile for age and gender.
  •  Hypertension:
    •  Systolic/diastolic blood pressure ≥95th percentile adjusted for age, gender, and sex.
  •  Glucose homeostasis:
    •  Elevated fasting glucose (≥6.1 mmol/l);
    •  Impaired glucose tolerance (IGT), glucose at 120 min: 140–200 mg/dl;
    •  Hyperinsulinemia and insulin resistance.

Abnormal glucose homeostasis was defined following modified World Health Organization (WHO) criteria adapted for children (Alberti & Zimmel, 1998). Hyperinsulinism was defined from norms for pubertal stage: prepubertal ≥15 mU/L, mid-puberty (stages 2–4) ≥30 mU/L. Post-pubertal hyperinsulinism was defined as per adult WHO criteria (≥20 mU/L). Hypertension was defined as systolic blood pressure ≥95th percentile for age and sex; abnormal fasting lipids were defined from normative data (Hickman et al., 1998; Goran & Gower, 2001). The following formula was used to calculate homeostasis model assessment (HOMA) insulin resistance: (fasting insulin × fasting glucose)/22.5 (higher values indicate greater insulin resistance).

Blood samples were spun promptly and were collected before the beginning of VPA therapy and at the end of follow-up, and sera frozen at −70°C until assayed. Samples were drawn at 08.00 a.m., after an overnight fast and, at the second evaluation, before VPA administration. Serum glucose was determinate by a glucose oxidase method (Beckham Instruments, Fullerton, CA, U.S.A.). Plasma insulin was analyzed by a radioimmunoassay (RIA) technique with double antibody-polyethylene glycol (CIS Bio International, Gif-sur-Yvette, France). Plasma LDL fraction was isolated by single-vertical-spin ultra-centrifugation using a discontinuous NaCl/KBr density gradient. Then, it was dialyzed for 22 h in the dark against three changes of phosphate-buffered saline (PBS) containing ethylenediaminetetraacetic acid (EDTA) (2.7 mmol/L) (pH 7.4) at 4°C (Foreman et al., 1997). LDL cholesterol was measured by an enzymatic reagent (CHOD-PAP, MPR1; Boehringer, Mannheim, Germany).

Statistical Analysis

Data are expressed as mean ± SD. Chi-square test, Student’s t test, and analysis of variance (ANOVA) were used for comparisons, where appropriate. Simple and partial Pearson correlation analysis was used to establish associations between insulin sensitivity and features of MS. All analysis were performed using SPSS version 10.0 (SPSS Inc., Chicago, IL, U.S.A.) with a type I error set at p < 0.05.

Results

We followed all patients from the beginning of therapy for at least 24 months; at the end of follow-up, 46 patients (40.4%) had a considerable increase in body weight, whereas the other patients (59.6%) did not gain weight. Twenty-six of 46 patients (56.5%) were prepubertal and twenty (43.5%) were pubertal. Twelve subjects (26.1%) reported a family history of any MS component in the first or second generation. Pertinent data of the nonobese VPA-treated patients are reported in Table 1.

Table 1.   Physical and metabolic data of nonobese patients treated with VPA
 Non-obese VPA Patients (n = 68)
  1. BMI, body mass index; HOMA-IR, homeostasis model assessment of insulin resistance; LDL-cholesterol, low-density lipoprotein cholesterol; HDL-cholesterol, high-density lipoprotein cholesterol; VLDL-cholesterol, very low density lipoprotein cholesterol; VPA, valproate.

Sex (M/F)32/36
Age (yr)10.3 ± 4.1
Prepubertal/pubertal41/27
BMI (kg/m2)20.1 ± 1.9
Systolic blood pressure (mm Hg)95.5 ± 9.2
Diastolic blood pressure (mm Hg)70.2 ± 8.6
Waist-to-hip ratio0.81 ± 0.03
Fasting glucose (mmol/l)4.69 ± 1.50
Fasting serum insulin (pmol/ml)24.21 ± 16.94
HOMA-IR2.07 ± 1.5
Triglyceride (mg/dl)91.7 ± 22.5
Total cholesterol (mg/dl)110.5 ± 27.9
LDL-cholesterol (mg/dl)80.4 ± 10.5
HDL-cholesterol (mg/dl)59.9 ± 10.6
VLDL-cholesterol (mg/dl)16.4 ± 5.7

With respect to the 46 obese patients, pretreatment BMI values did not significantly predict the degree of increase in BMI and/or the development of MS. In the 46 obese patients prevalence of MS was 43.5% (20 of 46). Table 2 reports the main physical and metabolic characteristics of obese patients with and without MS. The presence of one or two or three or more than three components associated with MS was 21.8% (10 of 46), 34.7% (16 of 46), and 43.5% (20 of 46), respectively. There was a tendency for postpubertal children to have greater number of features of MS than pubertal and prepubertal children, but this did not reach statistical significance. Table 3 shows the prevalence of individual components of MS by sex. Hyperinsulinism was identified in five subjects (25%). Impaired fasting glucose was identified in only 3 subjects (15%), impaired glucose tolerance in 7 (35%), and abnormal HOMA insulin resistance (HOMA-IR) in 13 (65%) subjects. Overall, abnormal glucose homeostasis was identified in 45% of patients and was not associated with gender. High total serum cholesterol concentrations were noted in 10 (50%), high serum triglyceride concentrations in 7 (35%), and low HDL in 15 (75%) of the 20 subjects with MS. There were no significant differences by gender. Hypertension was noted in two patients (10%) with no significant differences by sex. Table 4 shows the distribution of the individual components of MS of the population study categorized by pubertal status; no significant differences between prepubertal and pubertal patients were evident.

Table 2.   Pertinent physical and metabolic data of the population study divided into two groups (with and without MS)
 Without MS (n = 26)With MS (n = 20) p-valuea
  1. MS, metabolic syndrome; BMI, body mass index; HOMA-IR, homeostasis model assessment of insulin resistance; LDL-cholesterol, low-density lipoprotein cholesterol; HDL-cholesterol, high-density lipoprotein cholesterol; VLDL-cholesterol, very low density lipoprotein cholesterol.

  2. ap-value without MS versus with MS.

Sex (M/F)13/139/11
Age (yr)10.9 ± 3.511.2 ± 3.1n.s.
Prepubertal/pubertal14/1212/8n.s.
BMI (kg/m2)27.1 ± 0.830.2 ± 1.1<0.001
Systolic blood pressure (mm Hg)101.7 ± 8.1119.8 ± 9.9<0.001
Diastolic blood pressure (mm Hg)69.4 ± 8.179.9 ± 8.8<0.001
Waist-to-hip ratio0.84 ± 1.80.87 ± 1.9n.s.
Fasting glucose (mmol/L)5.0 ± 1.17.1 ± 1.9<0.01
Fasting serum insulin (pmol/ml)31.2 ± 13.952.9 ± 19.2<0.001
HOMA-IR2.9 ± 2.64.4 ± 3.9<0.001
Triglycerides (mg/dl)122.8 ± 61.2143.8 ± 59.2<0.01
Total cholesterol (mg/dl)159.7 ± 36.9164.8 ± 39.9n.s.
LDL-cholesterol (mg/dl)94.6 ± 29.395.8 ± 32.6n.s.
HDL-cholesterol (mg/dl)57.9 ± 11.439.1 ± 9.5<0.001
VLDL-cholesterol (mg/dl)18.7 ± 12.236.2 ± 11.7<0.001
Table 3.   Prevalence of individual components of MS by sex
 Boys (n = 9)Girls (n = 11)Total (n = 20) p-valuea
  1. MS, metabolic syndrome; HOMA-IR, homeostasis model assessment of insulin resistance; HDL-cholesterol, high-density lipoprotein cholesterol; VLDL-cholesterol, very low density lipoprotein cholesterol.

  2. aBoys versus girls.

Abnormal glucose homeostasis
 Fasting hyperinsulinemia2 (22.2%)3 (27.3%)5 (25%)n.s.
 Impaired fasting glucose1 (11%)2 (18.1%)3 (15%)n.s
 Impaired glucose tolerance3 (33.3%)4 (36.3%)7 (35%)n.s
 Abnormal HOMA-IR6 (66.6%)7 (63.6%)13 (65%)n.s
Dyslipidemia
 High total cholesterol4 (44.4%)6 (54.5%)10 (50%)n.s.
 High triglycerides3 (33.3%)4 (36.3%)7 (35%)n.s.
 Low HDL cholesterol7 (77.7%)8 (72.7%)15 (75%)n.s.
 High VLDL cholesterol7 (77.7%)7 (63.6%)14 (70%)n.s.
Hypertension1 (11.1%)1 (9.09%)2 (10%)n.s.
Table 4.   Distribution of the individual components of MS of the population study categorized by pubertal status
 Prepubertal (n = 12)Pubertal (n = 8)Total (n = 20) p-value*
  1. *Prepubertal versus pubertal.

  2. MS, metabolic syndrome; HOMA-IR, homeostasis model assessment of insulin resistance; HDL-cholesterol, high-density lipoprotein cholesterol; VLDL-cholesterol, very low density lipoprotein cholesterol.

Abnormal glucose homeostasis
 Fasting hyperinsulinemia3 (25%)2 (25%)5 (25%)n.s
 Impaired fasting glucose2 (16.6%)1 (12.5%)3 (15%)n.s.
 Impaired glucose tolerance4 (33.3%)3 (37.5%)7 (35%)n.s.
 Abnormal HOMA-IR8 (66.6%)5 (62.5%)13 (65%)n.s.
Dyslipidemia
 High total cholesterol6 (50%)4 (50%)10 (50%)n.s.
 High triglycerides4 (33.3%)3 (37.5%)7 (35%)n.s.
 Low HDL cholesterol9 (75%)6 (75%)15 (75%)n.s.
 High VLDL cholesterol8 (66.6%)6 (75%)14 (70%)n.s.
Hypertension1 (8.3%)1 (12.5%)2 (10%)n.s.

When family medical histories of both groups were analyzed, the prevalence of cardiovascular disease and its risk factors were comparable in the families of cases with and without MS. Information about the level of physical activity was recorded using a self-reporting questionnaire, including level of daily activities, and time spent indoors watching TV, playing computers, or video games and studying. The questionnaire has been described previously (Verrotti et al., 1989).

In groups with and without MS, the median (min–max) birth weight [3450 g (1600–4500) and 3000 g (2400–4050)], duration of breast-feeding [9 months (0–21) and 10 months (0–24)], formula feeding [5 months (0–24) and 6 months (0–24)], and obesity [5 years (1–12) and 6 years (1–13)] were similar. Patients with MS spent similar time on daily activities (0.30 ± 0.04 vs. 0.32 ± 0.05 h), and they spent the same time on sedentary life activities, in comparison to children without MS.

In the group of 20 patients with MS, correlation analysis showed that log insulin sensitivity was correlated positively with log HDL cholesterol (p < 0.005) and negatively with log triglycerides (p < 0.05). However, log insulin sensitivity did not correlate with 2-h glucose. Log insulin sensitivity was not correlated with total or LDL cholesterol.

Discussion

The use of VPA in the treatment of epilepsy appears to cause considerable increase in body weight and can result in some hormonal and metabolic changes (Isojarvi et al., 1993; Fariello et al., 1995; Isojarvi et al., 1998; Novak et al., 1999; Verrotti et al., 1999, 2002; Pylvanen et al., 2002;Tan et al., 2005; Pylvanen et al., 2006a, 2006b; Verrotti et al., 2008). In our study, the percentage of VPA-treated patients who developed obesity was 40.4%; this percentage is similar to that reported in literature (Isojarvi et al., 1993; Verrotti et al., 1999; Pylvanen et al., 2002; Biton et al., 2003; Zimmermann et al., 2003; Meckling et al., 2004). In our patients, BMI at initiation of VPA therapy was not a predictor of the development of obesity and/or MS; this lack of predictivity is in agreement with previous studies (Biton et al., 2003; Wirrell, 2003; Pylvanen et al., 2006a, 2006b), which failed to establish a clear association between ascertained predictors of weight gain in patients treated with VPA. Many studies have demonstrated that obese patients on VPA therapy have high insulin level with insulin resistance (Isojarvi et al., 1998; Tan et al., 2005). On the other hand, it has been demonstrated that insulin resistance is a central component of MS in overweight subjects. Despite these data there are no studies that explored the prevalence of features of the MS in VPA-induced obesity. Therefore, we have decided to evaluate the presence of MS in the group of epileptic patients who become obese after VPA therapy. For the first time we have demonstrated that 40% of VPA-treated patients who became obese developed MS. Unfortunately, we did not evaluate an obese group of patients not taking VPA matched by weight, age, and sex and this is a limitation of this study; however, our results are similar to those reported in other studies carried out in obese patients without epilepsy where it has been demonstrated that childhood obesity is associated with several metabolic complications, such as insulin resistance (Cruz et al., 2003; Canete et al., 2007; Sen et al., 2008; Weiss & Kaufman, 2008).

The reason that VPA causes MS in some of the patients, but not in all patients, remains unclear and controversial. Because not all obese patients develop MS, there are likely other factors involved. Genetic factors that influence the several molecular pathways in energy homeostasis (e.g., insulin receptor signaling pathway, lipid metabolism) might represent a possible explanation; in fact, it has been suggested (Virkamaki et al., 1999; Kohen-Avramoglu et al., 2003; Weiss & Kaufman, 2008) that some mutations of genes responsible for insulin receptors, plasma cell membrane glycoprotein-1, glucose transporter 4, and peroxisomal proliferator-activated receptor-γ can explain the development of MS in a percentage of the obese subjects.

In our experience VPA per se does not seem to cause MS, because MS is not present in all VPA-treated patients. It is probable that the metabolic changes reported in epileptic patients treated with VPA are secondary to excess fat mass, because these changes are not present in those epileptic patients treated with VPA who do not gain weight. Our data suggest that MS is not caused by VPA medication but is due to the weight gain induced by VPA therapy.

In the literature there are diverse statements on the impact of gender and puberty (Cook et al., 2003; Cruz et al., 2003; Sen et al., 2008), with a higher prevalence of MS and its components in male compared to female subjects. We did not find a significant difference between these two groups. Moreover, distribution of the individual components of MS in the prepubertal and pubertal patients was similar, in agreement with previous data (Viner et al., 2005). Therefore, gender and/or pubertal status do not contribute to the development of MS.

We have evaluated the feeding habits and the daily physical activity of our patients, but we did not find any difference between VPA-treated patients with MS and patients without MS. Our experience suggests that these habits have apparently no effects on the development of MS.

The dyslipidemia of the MS may increase cardiovascular disease risk through different mechanisms from those associated with high total or LDL cholesterol. The relationship between insulin resistance and fasting lipids can be explained through the effect of insulin on lipoprotein metabolism. Insulin plays a central role in determining triglyceride clearance from the blood via activation of lipoprotein lipase and triglyceride output through effects on the synthesis and secretion of very LDL (VLDL) by the liver (Lewis & Steiner, 1996). Furthermore, insulin controls the output of free fatty acids from adipose tissue (Arner, 1995). It is possible that in the insulin-resistant state, triglyceride-rich lipoproteins accumulate in the circulation due to decreased activity of lipoprotein lipase (Pykalisto et al., 1975), increased lipolysis in adipose tissue (Arner, 1995), and increased output of VLDL particles from the liver (Lewis & Steiner, 1996). The delay in plasma lipoprotein triglyceride clearance allows for cholesterol esters to be passed on from HDL to triglyceride-rich particles, which results in potentially atherogenic lipoprotein particles (Patsch et al., 1992). The positive correlation between log insulin sensitivity with log HDL cholesterol and the negative correlation between log insulin sensitivity with log triglycerides found in our patients, confirm that insulin resistance is associated with an abnormal lipid profile and that hyperinsulinemia increases lipogenesis that can be responsible for the accumulation of triglycerides (Browing & Horton, 2004;Chiarelli & Marcovecchio, 2008).

The lack of association between insulin sensitivity and 2-h glucose in this cohort may result from the fact that the concentration of glucose in the blood is not only dependent on insulin sensitivity but also on β-cell secretory capacity (Lillioja et al., 1993; Buchanan et al., 2002). In adults, a failure of the β-cells to adequately compensate for the degree of insulin resistance underlies the transition from insulin resistance to overt type 2 diabetes (Lillioja et al., 1993; Buchanan et al., 2002). In children, these relationships are less well established, although early reports suggest a similar pathophysiology (Sinha et al., 2002).

In the study of Csàbi et al. (2000), duration of obesity in patients with three or four cardiovascular risk factors was demonstrated to be longer than those with a lower number of risk factors. However, in the same study, there was no correlation between the duration of obesity and MS. Still, higher mean chronologic age in cases with MS in their study may support the hypothesis that some time is required for the underlying pathologies to take effect (Viner et al., 2005).

In conclusion, the VPA-treated patients who become obese can develop MS; the prevalence of MS seems to be similar to that observed in overweight subjects. This observation raises the possibility that VPA-treated patients can be at risk for important metabolic disturbances that should be checked, especially if these patients became obese.

Acknowledgments

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Disclosure: None of the authors has any conflict of interest to disclose.

Ancillary