Weight gain following treatment with valproic acid: pathogenetic mechanisms and clinical implications


A Verrotti, Department of Paediatrics, University of Chieti, Ospedale policlinico, Via dei vestini 5, 66100 Chieti, Italy. E-mail: averrott@unich.it


In the last years, a growing body of literature indicates an association between valproic acid therapy and weight gain. Weight gain during valproate treatment can be observed within the first 3 months of therapy and women seem to be more susceptible than men. The mechanism through which valproic acid may induce a weight gain is still controversial. The scope of this paper is to investigate the possible causal link between treatment and weight gain in epileptic patients. Systematic review of published epidemiological studies has been done in order to evaluate the real extent of this side effect of valproic acid and its clinical implications, such as an increased risk of insulin resistance and other secondary metabolic abnormalities. The knowledge of the potential of valproic acid to cause significant changes in body weight will help in appropriate selection and modification of antiepileptic therapy to minimize the risk for weight abnormalities. Measurements of body weight before initiation of valproic acid therapy should be done as part of the monitoring of patients with epilepsy to detect changes before there are serious adverse consequences; an increase of 2 kg of body weight after 1 month of treatment should imply considerations to change antiepileptic drug therapy.


Valproic acid (VPA) is an antiepileptic drug (AED) that is effective for treatment of a variety of seizure types both in adults and in children (1), and it is increasingly used for other indications, such as bipolar psychiatric disorder and migraine prophylaxis (2). Among the side effects of VPA, weight gain is frequently reported (3), although the real incidence and magnitude of this problem is unknown: the reported frequency of weight gain is between 10% and 70% (3–35). First, Egger and Brett (17) noted weight gain in 44 of the 100 children treated with VPA. Increased appetite and excessive weight gain were reported in 23 boys and 21 girls. A similar percentage of weight gain was found also in adult patients by several authors (4,8,13).

The main studies and the percentage of weight gain observed in adults and children are presented in Table 1 and Table 2, respectively.

Table 1.  Obesity, serum insulin and insulin resistance in valproic acid-treated adults in recent published studies
AuthorPatientsAge, yearsDaily dosageDuration of therapy, yearsPatients who gained weight, %Fasting serum insulinInsulin resistance
  • *

    Range values.

  • Median values.

  • Mean values.

  • +, present; −, absent; NR, not reported.

Dinesen et al. (14)63 (M/F)19.1–79.0*900–3.500 mg*2.557.0NRNR
Isojärvi et al. (13)22 (F)29.0 ± 7.0*1.01 ± 331 mg7 ± 5.050.0NR
Sthephen et al. (6)23 (F)33.0 (F)400–2.500 mg*≥2.029.2NR
17 (M)29.0 (M)800–4.500 mg*16.4
Biton et al. (7)68 (M/F)30.1 ± 14.0633 ± 1.82 mg0.662.2NRNR
Pylvänen et al. (8)81 (M/F)29.6 ± 10.91.12 ± 311 mm6.5 ± 4.949.0NR
Luef et al. (4)22 (F)31.2 ± 4.2975 mg13.5 ± 6.545.4NR
Pylvänen et al. (9)37 (M)28.7 ± 2.1413 ± 1.21 mg0.5–15.0*54.0NR
Pylvänen et al. (10)31 (M)30.4 ± ± 352 mg75.8 ± 3.952.9+
20 (F)32.8 ± 12.8997.4 ± 358 mg7.4 ± 6.2
El-Khatib et al. (11)51 (M)35.0 ± 13.11125 ± 526 mg1.8 ± 0.527.5Normal
Prabhakar et al. (12)25 (F)18.3 ± 3.7724 ± 150 mg1.072.0NRNR
11 (F)  2.054.5  
6 (F)  2.7–3.0*50.0  
Table 2.  Obesity, serum insulin and insulin resistance in valproic acid-treated children in recent published studies
AuthorPatientsAge, yearsDaily dosageDuration of therapy, yearsPatients who gained weight, %Fasting serum insulinInsulin resistance
  • *

    Range values.

  • Median values.

  • Mean values.

  • +, present; −, absent; NR, not reported.

Egger et al. (17)100 (M/F)1.2–18.0*30–50 mg kg−1*0.5–5.0*44.0NRNR
Easter et al. (26)103 (M/F)4.0–15.0*30 mg kg−13.010.7NRNR
Verrotti et al. (18)40 (F)17.3–1.230–50 mg kg−1*1.037.5+
Novak et al. (19)25 (M)1.8–16.9*375–2.500 mg*0.7–2.8*10.9NRNR
30 (F)250–2.000 mg*
Rättyäet al. (20)40 (F)8.4–18.5*8.7–27.5 mg kg−1*0.8–8.0*50.0+
Demir and Aysun(21)20 (M/F)3.5–15.0*10–20 mg kg−1*0.2*55.0+
Verrotti et al. (22)20 (F)8.5–11.2*30 mg kg−11.035.1+
Biton et al. (24)20 (M/F)16.0 ± 3.0100–500 mg*0.6770.0NRNR
El-Khayat et al. (32)15 (F)13.4 ± 3.5NR1–5*27.0NR
Verrotti et al. (28)20 (F)8.5–11.2*30 mg kg−11.035.0NR
Mikkonen et al. (25)40 (F)12.5 ± 3.1NR5.821.0NRNR
de Vries et al.(37)43 (F)14.9–3.312.2 ± 4.5 mg kg−11–9.5*25.5Normal
Rauchenzauner et al. (31)84 (M/F)6.0–18.6*NR>6 months44.0+
Hamed et al. (36)35 (M/F)10.3 ± 3.450–200 mg dL−1*6.5 ± 2.460.0NR
Sharpe et al. (29)94 (M/F)9.218.3 mg kg−12.124.0NRNR
Verrotti et al. (111)114 (M/F)10.1 ± 4.730.4 mg kg−11.040.4+

Results from clinical studies have suggested that some risk factors may contribute to VPA-induced weight gain (i) Gender: women seem to be more prone to weight gain during VPA therapy than men although very few studies have analysed the patients by sex. In particular, a possible effect of gender on the extent of weight gain has been studied by El-Khatib et al. (11), who compared the incidence and extent of weight gain associated with VPA monotherapy in male and female epileptic patients. This study demonstrated a more pronounced and more frequent weight gain in women receiving VPA monotherapy, compared with men. The authors reported a significant weight gain (≥5 kg) in 43.6% of women compared with 23.5% of men on VPA therapy; furthermore, percentage of body fat and waist-to-hip ratio differed statistically between genders with women having higher percentage of body fat and a lower waist/hip ratio. In agreement with these authors, other studies have suggested a risk of weight gain in female patients (6,36). Results of the papers that have studied weight gain in female and male patients are reported in Table 3. (ii) Puberty: the increase in body weight appears to occur most frequently in post-pubertal girls taking VPA (5,7,12,17,18,20,24,32,37), suggesting that a mature endocrine system of the adult type may be necessary for the development of VPA-related obesity; furthermore, a recent population-based study reported that increase in body weight is more common in patients treated with VPA during puberty if epilepsy and medication continue into adulthood (25). (iii) Duration of treatment: a long duration of therapy is associated with significant weight gain (36); in fact, across retrospective or prospective studies in children and adults with various seizures disorders, initiation of VPA therapy was consistently associated with an increase in weight, usually observed within the first 3 months of therapy and peaking by the sixth month (5,10,12,17,25). Another study (38) has confirmed the peak at the sixth month and, recently, we have demonstrated a major increase in body size and insulin resistance during the first year of therapy (39). (iv) Daily dosage: although not all studies have analysed the role of VPA dosage, the literature (10,19–21,40), as well as our experience (18), demonstrates that there is no correlation between the degree of weight gain and the daily VPA dosage and a serum VPA concentrations. (v) Seizure type: there are few evidences that weight gain problems are more common in patients with psychogenic seizures (41), and in patients with generalized vs. partial seizures (12,23,33) In contrast, many studies have not found a difference of incidence between generalized and partial epilepsy (4,6,11,33,42). (vi) Pre-treatment overweight: some studies (19,20,23–25) showed that the risk of obesity is magnified in those patients who had a greater weight at the start of treatment.

Table 3.  Weight gain in female and male patients during valproic acid treatment
AuthorPatientsAge, yearsDaily dosageDuration of therapy, yearsPatients who gained weight, %
  • *

    Girls demonstrated higher body mass index (F: 27.8 ± 1.6, M: 24.5 ± 0.8, P < 0.01) vs. boys.

  • Range values.

  • Median values.

  • §

    Mean values.

Egger et al. (17)40 (M)1.2–18.030–50 mg kg−10.5–5.057.5
60 (F)35.0
Dinesen et al. (14)20 (M)19.1–79.0900–3.500 mg2.5§60.0
43 (F)55.8
Sthephen et al. (6)17 (M)29.0 (M)800–4.500 mg≥2.016.4
23 (F)33.0 (F)400–2.500 mg29.2
Biton et al. (7)31 (M)30.1 ± 14.0633 ± 1.82 mg0.661.3
37 (F)51.3
Pylvänen et al. (8)46 (M)29.6 ± 10.91126 ± 311 mM6.5 ± 4.952.1
35 (F)45.7
El-Khatib et al. (11)51 (M)35.0 ± 13.11125 ± 526 mg1.8 ± 0.527.5
55 (F)34.0 ± 10.01141 ± 498 mg1.5 ± 0.756.4
Hamed et al. (36)35 (M/F)10.3 ± 3.450–200 mg dL−16.5 ± 2.460.0*

In recent years, many studies have revealed metabolic and endocrine abnormalities in patients who gained weight after therapy with VPA. In fact, it is known that weight gain is associated with pathologic consequences related to obesity as reproductive disorders, dyslipidaemia, hypertension, insulin resistance, diabetes mellitus and atherosclerosis and its related vascular implications (36).

This article reviews the data from the existing literature, regarding the effect of VPA on weight, analyses the main mechanisms of VPA-induced obesity and discusses the clinical implications of this effect for the management of patients with epilepsy. The MEDLINE and PsycLit databases were used to identify studies in children and adults, using the following keywords and truncated versions: ‘valproate and weight gain’, ‘insulin resistance’, ‘leptin’, ‘adiponectina’, ‘ghrelin’, ‘visfatin’ and ‘metabolic syndrome’. The date of our last search was February 2010.

Mechanisms relating valproic acid and weight gain

In spite of the studies mentioned above, unsolved problems remain. In fact, the real pathogenetic mechanism underlying VPA-induced weight gain is still unclear. It is most likely multifactorial because control of food intake and energy expenditure are complex and are regulated at peripheral and central levels by various appetite-regulating neuropeptides and cytokines that act within the hypothalamus. However, various hypotheses have been submitted to explain an effect of VPA on weight increase: dys-regulation of the hypothalamic system, effect on adipokine levels, hyperinsulinaemia, insulin resistance and genetic susceptibility (mechanisms of induced weight gain by VPA are presented schematically in the Fig. 1).

Figure 1.

Mechanisms of induced weight gain by VPA. β-OX, beta-oxidation; CEBPα, CCAAT/enhancer binding protein-α; FFA, free fatty acids; HDAC, histone deacetylase activity; RSTN, resistin; SOCS-3, cytokine signalling-3; VPA, valproic acid.

The effect of valproic acid on the hypothalamus

Valproic acid has been suggested to affect the hypothalamus. This hypothesis is supported by the observation that VPA-treated epileptic patients who reported weight gain developed increased appetite, thirst and quenching with calorie-rich beverages (14,18,34,43). All these behaviours can indicate hypothalamic stimulation. Moreover, experimental data demonstrate that VPA can cause dys-regulation of the hypothalamic system (44).

Although this theory may be explained by the enhancement of γ-aminobutyric acid (GABA) transmission within the hypothalamic axis (24), this is not considered the only mechanism because other AEDs that increase GABA (e.g. tiagabine) are not reported to induce weight gain (45). VPA may induce weight gain by the modifying expression of adipokine genes that are expressed in the brain and pituitary (cephalokines); these genes codify for neuropeptides involved in central energy metabolism, such as resistin (RSTN) and fasting-induced adipose factor (FIAF) also known as angiopoietin-like protein 4, which have become major targets implicated in the aetiology of obesity and development of leptin and insulin resistance (46). RSTN is a member of the newly discovered family of cysteine-rich secretory proteins called ‘RSTN-like molecules’(47) and is up-regulated in obesity, participating in the pathogenesis of insulin resistance (48). However, RSTN implication in the control of fatness and insulin sensitivity is still a matter of debate. In humans, serum RSTN was reported to be related to fat mass and to insulin resistance (49), while other reports revealed any correlations with body mass index (BMI), percentage body fat or with insulin sensitivity (50).

Angiopoietin-like protein 4 was first identified in adipose tissue as an FIAF, substantially up-regulated during fasting (51), raising the possibility that it could play a major role in signalling nutritional deprivation.

Recently, RSTN and FIAF expression was detected in the novel N-1 hypothalamic neuronal cell line. Subsequent RNAi studies confirmed the existence of an adipokine autocrine/paracrine loop in N-1 neurons, as the silencing of endogenous RSTN induced the expression of suppressor of cytokine signalling-3 (SOCS-3), an intracellular inhibitor of leptin signalling implicated in the development of obesity (52). Further studies in N-1 neurons and 3T3-L1 adipocytes have revealed that endogenous RSTN also activates adenosine 5′-monophosphatase-activated protein kinase, a cellular energy sensor implicated in the hypothalamic regulation of appetite and glucose metabolism (53). RSTN, FIAF and their associated downstream signalling molecules, such as SOCS-3, can have a profound impact on development of leptin and insulin resistance (54).

Interestingly, VPA seems to be capable of increasing RSTN and SOCS-3, and suppressing FIAF gene expression in N-1 hypothalamic neurons, probably through a CCAAT/enhancer binding protein-α (CEBPα)-dependent mechanism (46). Perhaps, VPA modifies hypothalamic gene expression in humans, such as the expression of SOCS-3, which may lead to leptin and insulin resistance, unintended weight gain and glucose intolerance, which has been described in patients treated with VPA (see Effect of valproic acid on adipokines). Although VPA may modify hypothalamic gene expression in vitro, it remains to be determined whether it has similar effects in vivo; in fact, in same experiments, VPA failed to have any effect on body weight or gene expression in vivo, although high doses of VPA has been used (55,56). Furthermore, Qiao et al. (57) reported a modest and unexpected weight loss and decrease in epididymal pad mass in C57BL/6J mice that were treated over week with multiple daily injections of VPA. By analysing the data of these studies, regard to the effect of VPA on the body weight, some discrepancies between animals and humans are evident. A possible explanation may be lie in the different metabolism of animals compared with that of humans; e.g. rodents are much more resistant to the metabolic effects associated with VPA treatment than humans. Perhaps, they are particularly efficient at metabolizing and clearing VPA from their systems, which might explain why substantially higher doses are required to achieve comparable serum concentrations to those measured in humans (57,58).

Interestingly, VPA seems to be capable of significantly increasing RSTN and SOCS-3, and suppressing FIAF gene expression in N-1 hypothalamic neurons, probably through a CEBPα-dependent mechanism (46). Although VPA may modify hypothalamic gene expression in vitro, it remains to be determined whether it has similar effects in vivo. Perhaps, VPA modifies hypothalamic gene expression in humans, such as the expression of SOCS-3, which may lead to leptin and insulin resistance, unintended weight gain and glucose intolerance, which has been described in patients treated with VPA (see Effect of valproic acid on adipokines).

Effect of valproic acid on adipokines

Adipose tissue is an endocrine organ, which releases various biologically active mediators (e.g. adiponectin, leptin, soluble leptin receptor, ghrelin, visfatin). The so-called adipokines have recently been considered as an exciting new link between obesity and insulin resistance but also obesity and cardiovascular disease, hypertension, as well as hyperlipidaemia (59).

Valproic acid and hypoadiponectinaemia

Adiponectin is the most abundant adipose tissue protein expression of adiponectin gene transcript-1 (60). Adiponectin has been found to prevent a high-fat diet-induced weight gain despite unaffected food consumption: it plays an important role in the modulation of insulin sensitivity and its plasma concentrations are negatively correlated with BMI (61,62). A longitudinal study in primates suggests that adiponectin decreases with weight gain as animals become obese (63). In contrast, weight loss results in significant increases in circulating adiponectin levels (64–67).

In fact, hypoadiponectinaemia is associated with insulin resistance in animal and human studies (66).

At the molecular level, VPA was shown to suppress adiponectin gene expression in adipocytes, through the inhibition of histone deacetylase activity, one of the several enzymatic activities involved in the transcriptional regulation of DNA; this inhibitory effect is dose- and time-dependent (57). However, the inhibition of adiponectin expression by VPA could be also due to its inhibitory effects on the adipogenic transcription factor CEBPα(68). Rauchenzauner et al. (69) evaluated in vitro the influence of VPA on adiponectin-binding receptors, adipoR1 and adipoR2, in human hepatoma cell line, HepG2. To date, adiponectin exerts its biological effects by interacting with adipoR1 and adipoR2, mediating increased cyclic adenosine monophosphate-activated protein kinase and peroxisome proliferator-activated receptor alpha ligand activities as well as fatty acid oxidation and glucose uptake (70). VPA increased adipoR1 mRNA expression in HepG2 cells. As adiponectin mRNA expression is known to be down-regulated following VPA treatment in vivo(71) and in vitro(57), increased adipoR1 mRNA expression in liver cells possibly represents a favoured reaction balancing suppressed adiponectin secretion from adipocytes. Hypothetically, up-regulation of adipoR1 mRNA expression increases adiponectin-binding capacity leading to an insulin-sensitizing effect and an increased β-oxidation of free fatty acids (FFA) mediated through increased adiponectin action. Changes concerning this balance of receptor/ligand expression might contribute to changes in fatty acid oxidation and insulin resistance in VPA-related obesity. This hypothesis is supported by the differential regulation of serum adiponectin concentrations, in overweight and lean VPA-treated children (31). Other authors (69) found low adiponectin levels in overweight children during VPA therapy; hypoadiponectinaemia in association with high low-density lipoprotein-cholesterol levels and high body fat reflect a greater cardiovascular burden of these children. These findings are in accordance with previous reports showing lower concentrations of adiponectin in patients with obesity and type 2 diabetes, providing a biological link between overweight and overweight-related disorders (71).

Valproic acid and hyperleptinaemia

Leptin is a product of the OB gene and it is a signal factor that regulates body weight. Leptin regulates body weight through neuropeptide Y (NPY), which stimulates food intake and decreases thermogenesis in the hypothalamus (27). Leptin increases fatty acid metabolism within the adipocytes, coordinates partitioning of fatty acid away from storage towards oxidation and decreases esterification of fatty acid into muscle triacylglycerol (72,73). Leptin plasma concentration (74) and mRNA expression in adipose tissue (75) are directly related to obesity severity, as an increase of fat mass is associated with an increase of leptin, which makes leptin an indicator of total fat mass. Leptin is also associated with insulin resistance independently of fat mass (76). Serum leptin concentrations are not only positively correlated with BMI but also with fasting insulin concentrations (4,8,18,19,22). Leptin production is regulated by insulin-mediated glucose and adipocyte metabolism and not by insulin per se. Leptin influences the action of insulin at various levels: in adipocytes, leptin inhibits insulin's actions on glucose metabolism by decreasing glucose transport (77). On the other hand, insulin increases leptin mRNA expression and its secretion by adipocytes (73). Leptin facilitates the action of insulin on hepatic glucose metabolism, and it can enhance insulin sensitivity and inhibition of hepatic glucose production in rats (77).

Hyperleptinaemia and leptin resistance can explain the VPA-induced weight gain. Many clinical studies reported increased serum levels of leptin in children and adults who gain weight during VPA treatment (8,11,18,19,22,27,36,37,78).

The mechanisms of VPA-induced changes in leptin are still controversial. Overfeeding and hyperleptinaemia associated with human obesity have been suggested to be due to state of decreased sensitivity to leptin (72). Leptin resistance is defined as reduced sensitivity or complete insensitivity to leptin action that probably contributes to altering leptin signalling and decreases negative feedback that may be situated at least at two distinct levels: saturable transport of leptin across the blood brain barrier or abnormalities at the level of leptin receptors activation or signal transduction (72). However, increase in serum leptin levels in VPA-induced weight gain may be a consequence of increase in adipose tissue (68). Furthermore, it is probable leptin behaviour in VPA-induced obesity is similar to that in any other obesity situation (8,18). We demonstrated that after 1 year of VPA treatment, only obese VPA-treated patients had higher levels of leptin, while non-obese VPA-treated patients continued to show normal leptin levels (18). Finally, some studies have failed to demonstrate a significant correlation between BMI and leptin levels in epileptic patients (4,9,11,18).

However, it is also possible that VPA causes direct secretion of leptin from adipocytes or alters leptin signalling and decreases negative feedback. In fact, VPA has been shown to have direct effects on hormone secretion from other endocrine cells (79), such as pancreatic islet cells (see below) and ovarian theca cells (80).

Interestingly, a cross-sectional cohort study (78) determined the influence of VPA treatment on leptin, soluble leptin receptor (sOB-R), the sOB-R/leptin ratio, body composition and insulin resistance in epileptic children: overweight VPA-treated children showed lower sOB-R concentrations and sOB-R/leptin ratio as well as higher body fat, leptin levels, compared with lean VPA-treated children. The majority of leptin circulates as free leptin, whereas in lean subjects, the majority of leptin circulates in bound form. sOB-R is considered the major form binding protein in human circulation and is generated by cleavage of the membrane-bound form of OB-R (81). sOB-R concentrations are negatively correlated with body weight. A reduction of body weight significantly increases sOB-R concentrations and the fraction of bound leptin, respectively (82). Therefore, the sOB-R might act as a modulating factor of leptin action and plays an important role in leptin resistance. Because sOB-R could acts as a potential reservoir of bioactive leptin, and thus prolongs the bioavailability of its ligand leptin in blood by preventing it from degradation and clearance, high sOB-R concentrations could contribute to weight maintenance in lean VPA-treated (83). Conversely, low sOB-R concentrations were found to be independently associated with insulin resistance and obesity in the MS(84). MS is a cluster of potent risk factors for atherosclerotic cardiovascular disease and type 2 diabetes mellitus in adults, and it is composed of insulin resistance, obesity, hypertension and hyperlipidemia.

Finally, the effects of VPA on leptin biology and fatty acid metabolism in 3T3-L1 adipocytes have been tested(69): in vitro VPA paradoxically reduces leptin mRNA levels and secretion of the leptin protein in a dose- and time-dependent manner. Probably, the inhibition of leptin secretion by VPA induces enhanced appetite in patients, resulting in enhanced adiposity and an increase in leptin secretion.

In conclusion, VPA can modify leptin levels through the increase of body weight, but therapy itself does not appear to cause the alteration of leptin levels in those patients who remain non-obese.

Valproic acid and ghrelin

Ghrelin, a polypeptide hormones, is an endogenous ligand for the growth hormone, and has an orexigenic effect (30). Plasma ghrelin level increases before meals and decreases post-prandially (71); moreover, ghrelin level decreases after increased calorie intake in patients with obesity and increases during the fasting state and in patients with anorexia nervosa. Ghrelin regulates the secretion of leptin and insulin and stimulates food intake through metabolic effects contradictory to the effects of leptin, and increases the consumption of carbohydrates while decreasing fat consumption (85). Its orexigenic action is mediated via hypothalamic arcuate nucleus, which coexpresses NPY and agouti-related protein. It increases the appetite and food intake through this mechanism and leads to weight gain, in addition to stimulating hyperinsulinaemia (86).

There are very few data concerning the association between ghrelin levels and body-weight changes, and it seems that ghrelin levels are reduced in VPA-induced obesity (30,71). Gungor et al. (30) reported a significant increase in serum ghrelin levels associated with weight gain in pre-pubertal children, but not in pubertal children, at month 6 of the VPA treatment; in this study there was no significant correlation between leptin and ghrelin levels, and no patients had leptin or insulin resistance. These results are not consistent with those of a previous study (71) showing low levels of ghrelin and high levels of leptin and insulin in post-pubertal epileptic women developing obesity at the end of 2 years of VPA therapy. These conflicting results may be explained by the lack of obesity in Gungor's patients; furthermore, in the early period of treatment, VPA may activate NPY pathway by increasing ghrelin levels, consequently stimulating appetite and food intake, thereby leading to weight gain.

Valproic acid and visfatin

Visfatin is a newly discovered adipokine that mimics the action of insulin via a distinct binding site on the insulin receptor, activating the intracellular signalling cascade for insulin (87). The molecule is identical to pre-YB cell colony-enhancing factor and has been implicated in the development of obesity-associated insulin resistance and diabetes mellitus in mice (87). However, recent results demonstrate increased visfatin plasma concentrations in obese subjects or patients with type 2 diabetes mellitus (88). In particular, a cross-sectional study by Haider et al. (89) suggested that the elevation of plasma visfatin concentrations in obese children may be independently involved in the development of MS and precedes abnormal glucose tolerance.

There are also conflicting data on visfatin circulating levels in obese humans. Some studies confirmed the increased levels of circulating visfatin (90,91), but there is also a study that showed reduced plasma visfatin levels in obese subjects (92). Paradoxically, both weight reduction (93,94) as well as overnutrition, down-regulated circulating visfatin concentrations in humans (95). The controversial findings on visfatin levels, reaching the increased (87), unchanged (96) or decreased levels (97) during the obesity and MS were also reported in various rat or mouse models of obesity.

There is only one study (31) evaluated the influence of VPA treatment on visfatin levels in epileptic children; VPA-induced overweight was associated with lower adiponectin and higher leptin concentrations, but not with visfatin concentration. It is reasonable that visfatin does not have a role in VPA-associated metabolic alterations. Therefore, visfatin may not be considered as a potential regulator of glucose and fat metabolism during VPA treatment.

Valproic acid-induced hyperinsulinaemia and insulin resistance

Many studies have reported hyperinsulinaemia among patients on VPA treatment who gained weight (5,6,8,10,13,18,20–22,31,32,36). In general, hyperinsulinaemia is known to be associated with obesity, dyslipidaemia and insulin resistance. Hyperinsulinaemia in obese persons taking VPA is not merely a consequence of insulin resistance induced by weight gain but the development of insulin resistance may be one of the factors leading to weight gain in some patients (8,22,66). This is supported by the observation that weight gain during VPA treatment is related to increase in insulin concurrent with decrease in glucose level, which can stimulate appetite and may cause weight gain (21). Replacement of VPA with lamotrigine in women with epilepsy resulted in decreased insulin levels within 2 months and body weight within 12 months (43). This hyperinsulinaemia has been attributed to the primary VPA-induced metabolic changes. Furthermore, we showed that insulin resistance was more severe in VPA-related obesity than in obesity in general (22). In the same period, Pylvänen et al.(9) founded that both obese and lean patients taking VPA had hyperinsulinaemia, suggesting a development of insulin resistance as the leading factor to weight gain during VPA treatment. It is difficult to explain the presence of hyperinsulinaemia in lean patients. For this reason, more recently, the same authors (10) strengthened these results by studying 51 adult patients on VPA monotherapy and comparing them with 45 healthy control subjects with respect to fasting plasma glucose, serum insulin, proinsulin and C-peptide concentrations after overnight fast. The VPA patients had fasting hyperinsulinaemia, although the fasting serum proinsulin and C-peptide concentrations were not significantly higher compared with the control. Therefore, probably VPA does not induce insulin secretion but may interfere with insulin metabolism in the liver, resulting in higher insulin concentrations in peripheral circulation. These changes were also seen irrespective of concomitant weight gain, suggesting that increased insulin concentrations induce weight gain whereas the reverse is not true.

However, the exact mechanism that can explain VPA-induced hyperinsulinaemia and a state of insulin resistance is still not fully clarified. Several mechanisms have been proposed including:

  • 1Increased availability of local FFA. This hypothesis is supported by the observation that insulin resistance was found to be correlated with increased plasma levels of FFA caused by VPA (98). In general, insulin resistance state is associated with increased lipolysis and reduced re-esterification of FFA in adipose tissue resulting in increased serum levels of FFA (98). VPA is a branched chain fatty acid; therefore, it can compete with FFA for albumin binding, increasing their local availability and thus their physiological modulation of insulin secretion (4). Elevated FFA induced by VPA may impair insulin biosynthesis and increase the proinsulin/insulin ratio of secretion. In an insulin-resistant state, increased lipolysis and reduced re-esterification of FFA in adipose tissue results in increased serum levels of FFA, and prolonged elevation of them further enhances insulin resistance by suppressing insulin-mediated peripheral glucose uptake and also affects pancreatic insulin secretion (99). Furthermore, VPA may inhibit β-oxidation of fatty acids resulting in increased level of non-esterified fatty acid and inhibit gluconeogenesis because of decreased serum carnitine, a metabolite involved in the transfer of fatty acids across the inner membrane of mitochondria for β-oxidation (100). Deficiency of carnitine may result in reduction of fatty acid metabolism and increase in glucose consumption (16).
  • 2Effect on pancreatic β-cells. VPA is a GABA-ergic agonist, it increases plasma levels of GABA. This neurotransmitter is known to be involved in pancreatic β-cells regulation and insulin secretion; therefore, stimulation of GABA receptors by GABA-ergic drug causes membrane depolarization and insulin release (101) In support, it has been demonstrated that VPA can directly stimulate the pancreatic β-cells in ex vivo. In fact, Luef et al. (79) investigated the effect of VPA on insulin secretion in pancreatic islet cells from pancreases of multi-organ donors: the incubation with VPA caused a time- and dose-dependent increase of insulin concentration in cell supernatant, suggesting that VPA can directly induce hyperinsulinaemia.
  • 3VPA affects the sympathetic nervous system by acting on hypothalamic neurons. In support: after initiation of VPA treatment, the catecholamine response to glucose load was found to be decreased (102). In obese people, weight gain that is not associated with increased energy intake has been explained by defective sympathetic activity (103).
  • 4VPA impair insulin signal transduction pathway. Wong et al. (104) tested the hypothesis that VPA can inhibit GLUT-1, one of the five proteic glucose carriers on the cell membrane activated at the end of the insulin transduction signal. VPA inhibited GLUT-1 activity in normal astrocytes and fibroblasts, with an important reduction of glucose transport accompanied by an up to 60% down-regulation of GLUT-1 mRNA expression.
  • 5Influence of VPA on oxidative stress. Recently, several studies have reported that oxidative stress might play a role in causing insulin resistance and β-cell dysfunction (105,106). Moreover, a possible association between high levels of oxidative stress and VPA therapy has been suggested, as a consequence of either of VPA biotransformation or a deficiency of antioxidant defence (107,108); in particular, an alteration in antioxidant enzymes resulting in a reduction of glutathione peroxidase and elevated glutathione peroxidase and elevated glutathione reductase has been demonstrated in children and adults receiving VPA therapy (108,109).

Clinical implications

Overweight and obesity are of special concern in all fields of medicine because they contribute negatively to the overall health. Drug-induced weight may represent a potential safety issue, especially in patients with pre-existing health risk that could be aggravated by added weight.

Managing patients with epilepsy, physicians should be aware of the different metabolic changes associated with long-term antiepileptic therapy as many of these adverse complications are latent and become manifest with time. VPA, compared with other AEDs, significantly increases weight. The consequences of this adverse effect, which limits the drug use in clinical practice, are the increased risk of metabolic disorders, such as MS, abnormal lipid profile and cardiovascular disorders (10,110). In fact, we have demonstrated that patients who gain weight during VPA therapy can develop MS, showing high total serum cholesterol and triglyceride concentrations and low high-density lipoprotein and might contribute to endothelial dysfunction later in life (111).

Recently, intrahepatic fat accumulation in the course of VPA therapy was demonstrable in animal studies (112) and in cell cultures exposed to VPA (113). Furthermore, in three clinical studies (114,115,116), ultrasonic examination of the liver in epileptic patients revealed that 61% of patients on VPA and 21% with carbamazepine had non-alcoholic fatty liver disease.

Finally, it has been also reported that VPA-related obesity is associated to a very high occurrence of polycystic ovary syndrome and hyperandrogenism in women with epilepsy (13,43), particularly when the medication is started before the patients reached the age of 20 years (40).

Therefore, the first step in treatment of epilepsy is to register body weight of all patients prior to begin VPA treatment and to weigh patients at each visit, calculate BMI and react to weight changes; an increase of 2 kg of body weight after 1 month of treatment should imply considerations to change AED therapy. Furthermore, weight gain and related risk of reproductive endocrine disorders should be taken into consideration when choosing antiepileptic therapy for younger women.

The importance of the effects of AEDs on weight in treatment selection depends largely upon the individual patient's needs and the risks and benefits of therapy for that patient. Unless seizure control is satisfactory, weight effect may play a relatively minor role in determining choice of therapy. On the other hand, effects on weight become more salient determinants of treatment choice for patients in whom drug-induced changes in weight are particularly likely to have detrimental consequences (e.g. the overweight or obese patient who should not gain additional weight).


Valproic acid treatment seems to be clearly associated with significant weight gain and related metabolic disorders both in adult and children with epilepsy.

States of insulin and leptin resistances are suggested as principal mechanisms underlying weight gain with VPA, although the mechanisms behind the development of leptin and insulin resistances in patients taking VPA should be a subject of intensive research.

Marking patients at risk for weight gain, particularly women, will help clinicians to select patients for appropriate treatment and to counsel them against weight gain.

Finally, as iatrogenic obesity may contribute to reduced adherence to the medication regimen especially in adolescents, the possibility of weight changes should be discussed with patients before medication therapy begins.

Conflict of Interest Statement

We have no conflicts of interest.