FULL-LENGTH ORIGINAL RESEARCH

Weight and fat distribution in patients taking valproate:

A valproate-discordant gender-matched twin and sibling
pair study
*†‡Sandra J. Petty, §Susan Kantor, ¶Kate M. Lawrence, ¶Samuel F. Berkovic, #Marnie Collins,
**††Keith D. Hill, ‡‡Joanna Makovey, ‡‡Philip N. Sambrook, *§§¶¶Terence J. O’Brien, and
§¶¶John D. Wark

Epilepsia, 55(10):1551–1557, 2014

doi: 10.1111/epi.12745

Dr. Sandra Petty is a
clinical and academic
neurologist at The
University of
Melbourne –
Melbourne Brain
Centre at The Royal
Melbourne Hospital,
Australia
SUMMARY
Objectives: Chronic treatment with valproate (VPA) is commonly associated with
weight gain, which potentially has important health implications, in particular
increased central fat distribution. We utilized a VPA-discordant same-sex, twin and
matched sibling pair study design to primarily examine for differences in fat distribu-
tion between patients with epilepsy treated with VPA compared to their matched twin
or sibling control. Weight, blood pressure, and leptin levels were assessed.
Methods: Height, weight, waist and hip measurements, exercise, blood pressure (BP),
and serum leptin levels were measured. Body composition was measured using dual-
energy x-ray absorptiometry (DXA). Abdominal fat was expressed as a percentage of
the abdominal region (AFat%); and of whole body fat (WBF); (AFat%WBF). Mean
within-pair differences were assessed (VPA-user and nonuser). Restricted maximum
likelihood (REML) linear mixed model analysis was fitted to examine associations of
anthropometrics, zygosity, gender, menopausal status, VPA dose and duration, with
weight and AFat%.
Results: We studied 19 pairs of VPA-discordant, gender-matched (five male, 14
female) twins and siblings. Mean (standard deviation, SD) duration of therapy for VPA
users was 11.0 (7.4) years. There were no statistically significant within-pair differences
in age, height, weight, body mass index (BMI), BP, leptin level, WBF, AFat%, or AFat%
WBF. For pairs in which VPA-user was treated for >11 years there were statistically
significant mean within-pair differences in AFat%, (+7.1%, p = 0.03, n = 10 pairs), mean
BP (+11.0 mm Hg, p = 0.006, n = 8 pairs); but not in AFat%WBF. VPA duration was
positively associated with weight (estimate +0.98 kg/per year of VPA, p = 0.03); VPA
treatment duration and dose were not significantly associated with AFat%.
Significance: This study demonstrated a relationship between long-term VPA use and
abdominal adiposity (AFat%), which could have significant health implications. We
recommend ongoing monitoring of weight, BMI, and blood pressure for patients
taking VPA.
KEY WORDS: Antiepileptic medication, Obesity, Medication side effects, Comorbidi-
ties of epilepsy.

The association between antiepileptic drug (AED) use
and increase in total body weight is well established.
Therapy with valproate (VPA) is often associated with
weight gain in adults with epilepsy.1 Although many studies
have examined the magnitude of weight gain,2–6 relatively
few have quantified the distribution of fat in patients taking
AEDs.7 Changes in the distribution of body fat have
important health implications, specifically the association
of increased relative abdominal fat with risk for cardiovas-
cular disease.8 Influences on abdominal fat include genetic
factors,9 gender and menopausal status,10 and environmen-
tal factors. There are limited data regarding cardiovascular
health in epilepsy, despite some studies that report reduction
in physical exercise,11 increased body fat percentage,12

1551
1552
S. J. Petty et al.
increased rates of smoking,11 increased cardiovascular
risk,13 and high rates of hypertension.14
VPA has been associated with the metabolic syndrome
in patients who gain weight on treatment.7 The mecha-
nisms by which VPA results in weight gain are unclear,15
although it has been shown to inhibit leptin secretion.16
However, increased leptin levels have also been observed
in obese patients taking VPA.17 It has been suggested that
hyper-insulinemia rather than changes in leptin levels may
play a more critical role.18
This study utilized dual energy x-ray absorptiometry
(DXA) in a gender-matched twin and sibling pair study
design to investigate the association of AED use, in particu-
lar VPA, with weight and body fat distribution. DXA is a
valuable and accurate tool used to noninvasively assess
body fat19,20 including regional fat analysis,21,22 but has not
been applied to study the effects of AEDs.
We hypothesized that VPA use would be associated with
increased weight and increased proportion of abdominal fat.
This work aimed to primarily examine body fat distribution
in patients taking VPA compared to their gender-matched
twin or sibling control, and to assess whether patients taking
VPA had increased weight, increased leptin levels, and/or
increased blood pressure (BP). We examined for significant
cofactors that may influence percentage of abdominal fat,
including gender, and menopausal status, and for associa-
tions with dose and duration of VPA therapy.
Methods
Participants
This was a substudy of the Longitudinal Twin and Sibling
AED Bone Health Study.23–25 All pairs were assessed at one
of two Australian study sites: (1) The Royal Melbourne Hos-
pital, Victoria, and (2) The Royal North Shore Hospitals,
Accepted July 2, 2014; Early View publication August 14, 2014.
*Department of Medicine, Melbourne Brain Centre, The Royal Mel-
bourne Hospital, The University of Melbourne, Parkville, Victoria, Austra-
lia; †Department of Medicine, St Vincent’s Hospital, The University of
Melbourne, Fitzroy, Victoria, Australia; ‡Ormond College, Parkville, Vic-
toria, Australia; §The Royal Melbourne Hospital Bone and Mineral Medi-
cine, Parkville, Victoria, Australia; ¶Epilepsy Research Centre, Austin
Health, The University of Melbourne, Heidelberg, Victoria, Australia;
#Statistical Consulting Centre, The University of Melbourne, Parkville,
Victoria, Australia; **National Ageing Research Institute, Parkville, Victo-
ria, Australia; ††Faculty of Health Sciences, School of Physiotherapy and
Exercise Science, Curtin University, Perth, Western Australia, Australia;
‡‡Kolling Institute of Medical Research, Royal North Shore Hospital, Uni-
versity of Sydney, Sydney, New South Wales, Australia; §§Department of
Neurology, The Royal Melbourne Hospital, Parkville, Victoria, Australia;
and ¶¶Department of Medicine, The Royal Melbourne Hospital, The Uni-
versity of Melbourne, Parkville, Victoria, Australia
Address correspondence to John D. Wark, Department of Medicine,
RMH, The University of Melbourne, C/- Post Office, Parkville, Vic. 3050,
Australia. E-mail: jdwark@unimelb.edu.au
Ms. S. Kantor passed away in December 2012.
Prof. P. Sambrook passed away in March 2012.
Wiley Periodicals, Inc.
© 2014 International League Against Epilepsy
Epilepsia, 55(10):1551–1557, 2014
doi: 10.1111/epi.12745
New South Wales, between 1997 and 2007. AED users with
a gender-matched, AED-discordant twin (or sibling within
4 years of age) were recruited from the twin and siblings
research databases of The University of Melbourne and The
Epilepsy Research Centre, The Australian Twin Registry,
The Sydney Twin Study at The Royal North Shore Hospital,
media releases, Epilepsy Clinics of the Royal Melbourne
Hospital and Austin Health, and affiliated medical practices.
The recruitment process included mail outs to participants in
the twin databases of each center, as well as media releases
and referrals from clinics and affiliated medical practices.
Potential participants were contacted by telephone, and writ-
ten and verbal information about the study was provided.
Subjects identified were surveyed to confirm and detail their
AED use. Participants were older than 18 years of age, and
all pairs were gender-matched.
Inclusion criteria were the following: same gender twins
or siblings where one twin was currently taking VPA for
>12 months, and the other was not taking VPA and did not
have a clinical indication for taking VPA or any other AED.
Exclusion criteria included corticosteroid use and concor-
dance for AED use.
DXA scanning
Conventional whole body analysis of nonbone lean mass,
fat mass, and bone mineral content (BMC) was undertaken
using DXA with Hologic QDR 1000W or 4500A densitome-
ters (Hologic, Inc. Bedford, MA, U.S.A.). Whole body fat
mass and whole body (nonbone) lean mass (minus head
region) were determined using HOLOGIC software (Version
5.73) (Hologic, Inc. Bedford, MA, U.S.A.). DXA scans were
uploaded to the same computer and then reanalyzed by a single
operator who remained blinded to the clinical history, utilizing
previously published methodology.9,19,26 The abdominal region
was defined and measured as the area bordered by the superior
surface of the second lumbar vertebra to the inferior surface of
the fourth lumbar vertebra and laterally to the inner aspect of
the ribcage. Abdominal fat was expressed as a percentage
of the abdominal region (AFat%), calculated by: [Abdominal
fat mass (grams)/(Abdominal fat + lean + bone mineral con-
tent mass (grams)]*100; and of whole body fat (AfatWB%),
calculated by: [(Abdominal fat mass/whole body fat)*100].
Blood pressure measurement
BP was measured three times at 1 min intervals while the
participants were sitting (Dinamap Monitor, Model 8100;
Critikon Inc.,Tampa, FL, U.S.A.), in a fasting state prior to
having venipuncture. Mean systolic, diastolic, and pulse
pressures were calculated from the three results. BP mea-
surements were available for the Royal Melbourne Site (12
pairs), but were not performed at the Sydney site.
Health and lifestyle
Medical history and lifestyle factors were assessed by
self-reported questionnaires.27,28 Medical history included

details of lifetime medication use, reproductive history, and
menstrual history. Usual physical activity was determined
by questionnaire at interview by surveying participants’ reg-
ular hours of weight-bearing exercise during the last
12 months (≤1 h per week; 2–3 h per week; 4–7 h per
week; and >7 h per week, and average hours were calcu-
lated from these categories). A short food frequency survey
was completed to obtain information for calculation of
calcium intake as part of the Bone Health Study; detailed
nutritional information was not included.
Laboratory measures
Serum storage and analysis: Samples were centrifuged
and serum was stored at 80°C. Samples were transported
frozen on dry ice to the analyzing laboratory in Sydney
(ANZAC Research Institute, The University of Sydney,
Australia). Samples were analyzed in a single assay run for
each test, using a single thaw cycle. Leptin levels were ana-
lyzed utilizing the DSL-23100 ACTIVE leptin Coated-Tube
immunoradiometric assay (IRMA) (Diagnostic Systems
Laboratories Inc., Webster, TX, U.S.A.). Reference range
for this assay was 0.25–120 ng/ml. Serum samples were
collected for those visits occurring after October 2004 when
an additional ethics approval for these blood samples was
obtained (12 pairs).
Statistical analysis
DXA whole body and abdominal fat data was assessed
for normal distribution using QQ plots—all data were
approximately normally distributed. The primary variable
of interest was AFat%. Weight, leptin value, and BP were
secondary outcome measures. Descriptive data included
age, height, exercise levels, body-mass index (BMI), waist
circumference, and whole body fat and lean mass DXA
measures. AFat%WBF was measured to examine for distri-
bution of the amount of fat in the abdominal region as a
proportion of whole body fat.
Within-pair differences (AED user and nonuser) were
calculated. The significant p-value was set at <0.05; 95%
confidence intervals (CIs) were utilized. Analysis was
performed using IBM SPSS V21.0 Statistics (IBM Corpora-
tion Software Group, Somers, NY, U.S.A.). Bonferroni is
applied where applicable to the secondary outcome mea-
sures (weight; leptin; BP systolic, diastolic, and pulse pres-
sure; p < 0.01); AFat% is the primary variable of interest,
therefore, p-value was set at <0.05. The remainder of the
variables are listed for descriptive purposes.
A restricted maximum likelihood (REML) general linear
mixed model analysis was applied using Genstat 14th
Edition (VSN International Ltd, Hemel Hempstead, UK.).
Univariate analyses were first undertaken to separately
examine the effects of age, height, weight, zygosity, scanner
type, gender, menopausal status, and VPA use on weight.
Pair was specified as a random effect, to take account two
levels of variation: between the pairs and within the pairs.
1553
Valproate, Weight and Fat Distribution
The REML linear mixed model analysis was then under-
taken using AFat% as the response variable. Due to the
distribution of data with all VPA nonusers having a zero
value for VPA dose and duration, a separate univariate
analysis of VPA users only was performed.
Ethics approvals
Ethics Approvals were obtained for the specific Twin and
Sibling AED Bone Health Study from The Royal Mel-
bourne Hospital including venipuncture in October 2004.
The study also received ethics approval from The Austin
Hospital, The Royal North Shore Hospital, and the Austra-
lian Twin Registry. Individuals in each pair gave written
informed consent.
Results
Participants
Nineteen pairs where the AED user was currently taking
VPA were studied, including five male (26%) and 14 female
pairs (74%). Six pairs were monozygous (MZ) twins, seven
pairs were dizygous (DZ) twins, and six pairs were nontwin
siblings. Twelve pairs were assessed at The Royal Mel-
bourne Hospital site, and seven pairs at The Royal North
Shore Hospital site. Members of each pair were examined on
the same densitometer, with 17 pairs measured on the same
day, and members of 2 pairs within 29 days of each other.
Of the VPA users’ BMI values, nine were in the healthy
weight range (47%), seven were in the overweight range
(37%), and three were in the obese range (16%). Of the non-
users, 12 were in the healthy weight range (63%), 5 were
overweight (26%), and 2 were in the obese range (11%).
Of the 14 female pairs, 13 pairs were concordant for men-
opausal status: 11 pairs were premenopausal, two were post-
menopausal. One VPA user (age 52) was premenopausal
while her twin sister was postmenopausal.
Medical history
Seventeen participants were taking VPA for epilepsy. Ten
VPA users had generalized epilepsy, three had focal epi-
lepsy, one described nocturnal seizures (unknown type), and
three had unknown seizure classification. Two took VPA for
other indications (listed as “severe headache” by one patient
and “neuralgia” by the other). Fifteen of the 17 AED users
taking VPA for epilepsy recorded an age of onset, the mean
(SD) age of onset of epilepsy was 18.6 (12.1) years, and
mean duration of taking any AED was 14.5 (11.3) years.
The mean (SD) duration of VPA use was 11.0 (7.4) years,
and mean current VPA dosage was 1,188 (627) mg per day.
Three VPA users were unable to supply their current VPA
dose at the time of the study (Royal North Shore); removing
these pairs from the REML analysis requiring VPA dose
did not affect the results (data not shown). Of the seven
VPA users taking AED polytherapy, two used VPA and
carbamazepine; one VPA and lamotrigine (with past use
Epilepsia, 55(10):1551–1557, 2014
doi: 10.1111/epi.12745
1554
S. J. Petty et al.

of carbamazepine and phenytoin); one VPA and phenyt-
oin; one VPA, levetiracetam, and topiramate; one VPA,
carbamazepine, and levetiracetam; and one used VPA and
levetiracetam.
Descriptive data
The mean (SD) characteristics for VPA users and non-
users were the following: age of VPA users 37.5 (14.7), and
nonusers 37.4 (14.8) years; height of VPA users 168.1 (9.8)
cm, and nonusers 169.3 (12.1) cm; BMI of VPA users 25.6
(4.1), and nonusers 24.8 (4.2); and exercise hours per week
for VPA users 1.74 (2.23) h, and nonusers 1.68 (2.24) h. For
the 12 pairs where waist circumference was available: waist
circumference for VPA users was 90.0 (15.6) cm, and non-
users 88.1 (16.3) cm. There were no statistically significant
within-pair differences in the descriptive measures.
The mean (SD) for the primary variable of interest, AFat
%, for VPA users was 35.4 (10.6) and nonusers 31.9 (13.6);
the mean within-pair difference (VPA-user – nonuser) +3.5
(95% confidence interval [CI] 1.2, 8.2), p = 0.13, was not
statistically significant.
For the secondary outcome measures, there were no sta-
tistically significant within-pair differences in mean (SD)
weight (VPA users 72.6 [14.36] kg, nonusers 71.1 [13.4]
kg); leptin levels (VPA user 7.78 [5.96], nonuser 5.73 [3.79]
ng/ml n = 12 pairs); and mean (SD) BP (VPA users 117.9
[11.2] systolic/66.1 [9.0] diastolic, and nonusers 119.7
[15.6] systolic/65.0 [11.6] diastolic mmHg; and mean pulse
pressure for VPA users 70.8 [10.6] mmHg and nonusers
68.9 [10.9] mm Hg).
With restriction of the within-pair analysis to those 10
pairs with greater than the mean duration of VPA use
(11 years), there was a mean (SD) within-pair difference in
the primary variable of interest AFat% of +7.1% (95%CI
0.7, 13.5, p = 0.03) in AFat%. There was no within-pair dif-
ference in abdominal fat as a proportion of whole body fat
(AFat%WBF).
Of these pairs, seven were premenopausal females and
three were male. Five were monozygotic twins, two were
dizygotic twins, and three were siblings. Four VPA users
were currently taking AED polytherapy. The descriptive
data in this group showed a trend for increased mean (SD)
BMI (VPA user 26.3 [3.7], nonuser 24.0 [2.8], mean within-
pair difference +2.3 [95% CI 0.4,5.0], p = 0.08, which
was not statistically significant. There was a trend for
increased weight in the VPA user 76.3 [14.6], vs. nonuser
70.0 [13.3] kg, mean within-pair difference [VPA user –
nonuser] +6.3 kg [95% CI 1.3,13.9], p = 0.09), which
was not statistically significant. Although the mean BP sys-
tolic and diastolic results were not in the hypertensive range,
a within-pair difference in pulse pressure was noted for
VPA users (+11 mm Hg, [95% CI 4.3, 17.7 p = 0.006])
compared to nonusers. No statistically significant within-
pair differences were seen in the longer-term VPA user pairs
in the descriptive variables age, height, waist circumfer-
ence, exercise levels, or the secondary outcome measure
leptin.
REML analysis
A significant association between female menopausal
status and higher weight was seen (estimate: postmeno-
pausal +18.78 kg, p = 0.004) (Table 1). On multivariate
analysis, there was a significantly significant association
between whole body fat and higher AFat% (+1.06% per kg
whole body fat, 95% CI 0.83, 1.30, p < 0.0001); VPA use
was not statistically significantly associated with AFat%
(Table 2).
Univariate analysis of VPA users
Three participants were unable to provide their current
VPA dose for the study, and were therefore excluded from
this analysis. VPA duration was statistically significantly
associated with higher weight but not of AFat% (Table 3).
A higher dose of VPA was statistically significantly

Table 1. Restricted maximum likelihood (REML) analyses examining for associations of anthropometric data, VPA
use, zygosity, gender, and menopausal status with weight in all participants
Variable Reference level Comparison level Estimate 95% CI p-Value
Univariate analyses for
Weight (kg) x x+1 0.69 0.26, 1.11 0.003
Height (cm) x x+1 0.32 0.09, 0.73 0.11
Age (years) Monozygous twins Dizygous twins 11.93 26.8, 2.93 0.26
Zygosity Monozygous twins Nontwin siblings 6.04 21.5, 9.38
VPA use No Yes 1.49 3.70, 6.68 0.55
Gender Male Female 15.94 28.1, 3.83 0.01
Menopausal status Premenopause Postmenopause 14.95 2.62, 27.3 0.003
Male 18.61 8.11, 29.1
Multivariate analyses
Menopausal status Premenopause Postmenopause 18.78 8.32, 29.2 0.004
Male 8.57 2.12, 19.3
Height x x+1 0.66 0.23, 1.09 0.006
kg, kilogram; cm, centimeter; x, reference level; VPA, valproate.

Epilepsia, 55(10):1551–1557, 2014

doi: 10.1111/epi.12745
 1555
Valproate, Weight and Fat Distribution

Table 2. Restricted maximum likelihood (REML) analyses examining for associations of anthropometric data, VPA
use, zygosity, gender, and menopausal status with AFat% in all participants
Variable Reference level Comparison level Estimate 95% CI p-Value
Univariate analyses – AFAT%
Height (cm) x x+1 0.22 0.68, 0.25 0.34
Weight (kg) x x+1 0.63 0.40, 0.87 <0.0001
Whole body fat (kg) x x+1 1.06 0.83, 1.30 <0.0001
Age (years) x x+1 0.45 0.15, 0.76 0.006
Zygosity Monozygous twins Dizygous twins 3.78 17.4, 9.80 0.62
Monozygous twins Nontwin siblings 6.52 20.6, 7.58
VPA use No Yes 3.53 1.15, 8.22 0.13
Gender Male Female 1.06 11.6, 13.7 0.86
Menopausal status Premenopause Postmenopause 22.00 11.2, 32.8 0.002
Male 2.87 6.07, 11.8
Multivariate analysis
Whole body fat (kg) x x+1 1.06 0.83, 1.30 <0.0001
AFat%, abdominal fat expressed as a percentage of the abdominal region; cm, centimeter; x, reference level; kg, kilogram; VPA, valproate.

Table 3. Univariate restricted maximum likelihood (REML) analysis in VPA users only, examining for effects of VPA
dose and duration on (A) whole body fat, (B) AFat%, and (C) weight (univariate analyses)
Variable Reference level Comparison level Estimate 95% CI p-Value
(A) REML analysis for whole body fat – VPA users only
VPA dose (g) x x+1 8.51 15.7, 1.45 0.02
VPA duration (years) x x+1 0.26 0.34, 0.86 0.38
(B) REML analysis for AFat% – VPA users only
VPA dose (g) x x+1 2.82 12.7, 7.11 0.56
VPA duration (years) x x+1 0.15 0.57, 0.87 0.67
(C) REML analysis for Weight (kg) – VPA users only
VPA dose (g) x x+1 9.39 21.2, 2.41 0.12
VPA duration (years) x x+1 0.98 0.12, 1.84 0.03
REML, restricted maximum likelihood analysis; VPA, valproate; x, reference level; VPA dose (g), valproate (g/day); AFat%, abdominal fat expressed as a percent-
age of the abdominal region.

associated with lower whole body fat ( 8.51 kg/g of
VPA, 95%CI 15.7, 1.45, p = 0.02) (Table 2); how-
ever, only three VPA users were currently taking a dose of
2,000 mg/day or greater.
Discussion
To our knowledge this is the first twin and sibling study to
utilize DXA to examine fat distribution, and its relationship
to blood pressure and anthropometry, in pairs discordant for
VPA use. Utilizing within-pair differences, this methodol-
ogy is highly advantageous for association studies, enabling
a threefold to fivefold reduction in study sample size com-
pared with observations in unrelated individuals.29,30
In this cross-sectional analysis the overall group of VPA
users showed no statistically significant differences in AFat
% or weight compared to their gender-matched non–VPA-
using twin or sibling controls, nor in the secondary outcome
measures of BP, leptin, and weight. However, for those pairs
where the VPA user had a longer than mean duration of
VPA treatment in this study (11 years), the VPA users had a
mean 7% higher abdominal fat content compared to their
non–AED-using twin or sibling. There was no significant
difference in abdominal fat as a proportion of whole body
fat, suggesting that the increase in abdominal fat is likely to
be associated with a more global increase in body fat. Fur-
ther study into the amount, distribution, and fat type, includ-
ing assessment for further attributes of the metabolic
syndrome, is warranted in patients who gain weight while
taking VPA. Clinically, the metabolic syndrome comprises
at least three of the following: increased waist circumfer-
ence, elevated triglycerides, reduced high-density lipopro-
tein cholesterol (HDL-C), hypertension, and elevated
fasting glucose,31 and is a marker for patients being at
increased risk of developing atherosclerotic cardiovascular
disease. There are limited available data regarding the prev-
alence of metabolic syndrome in patients taking VPA for
epilepsy.7,32
Although there were no significant within-pair differ-
ences in BMI, 53% of VPA users and 37% of nonusers were
Epilepsia, 55(10):1551–1557, 2014
doi: 10.1111/epi.12745
1556
S. J. Petty et al.
overweight or obese in this study, according to World
Health Organization (WHO) classifications.33 There are
well-established associations of excess abdominal adiposity
with longer-term illness, all-cause mortality, cardiovascular
disease, and some malignancies associated with increased
body weight.34 The findings of this study would support the
use of clinical measures of weight, BMI, and blood pressure
in patients with epilepsy receiving longer-term VPA treat-
ment. The levels of physical activity were similar between
AED users and nonusers in our study, which is similar to
rates for moderate exercise in the California Health Inter-
view Survey.35 This indicates that differences in physical
activity are unlikely to explain the increased abdominal adi-
posity seen in the longer term VPA users. In contrast, other
reports have noted lower rates of sports participation,36
although walking levels may be increased due to driving
restrictions.35 There were also no significant differences in
systolic and diastolic BP between AED users and nonusers;
however, the higher pulse pressure in the VPA users
requires further investigation with respect to its potential
clinical significance.
A limitation of this study is that we assessed patients
cross-sectionally after long-term use of VPA, rather than
prospectively at onset of therapy for any acute weight gain,
which has been seen in other studies.15 On the basis of the
results of this study, a prospective longitudinal study of
body composition and cardiovascular risk in epilepsy
patients is indicated to further clarify the relative roles of
weight gain, gender, menopausal effects, and fat distribu-
tion in patients taking VPA; inclusion of detailed dietary
history would be of interest.
A limitation of utilizing DXA scanning for measuring fat
mass is that it gives a two-dimensional image, which does
not allow for determination of amount of subcutaneous fat
on the anterior and posterior abdominal walls versus the vis-
ceral fat, and also has other limitations in assessing soft tis-
sue.37 Although visceral fat has been associated with
cardiovascular risk factors including abnormal serum lipid
profile and hypertension,38 the role of subcutaneous fat39
and any contribution to cardiovascular risk remains to be
fully established.
This study demonstrated that long-term users of VPA had
an increased mean percentage of abdominal fat compared to
the matched twin and sibling non-VPA users. There was a
statistically significant association between duration of
VPA therapy and weight in the univariate analysis in the
group of VPA users, highlighting a need to monitor patients
taking VPA for weight gain not only at initiation of treat-
ment, but to continue monitoring in the longer term. The
association of VPA use with increased weight is clinically
important, and strategies to assess and limit weight gain and
to examine useful methods of weight loss in this population
will be important in the design of further clinically relevant
research to reduce comorbidity in epilepsy. We recommend
that clinicians continue to monitor weight, BMI, and blood
Epilepsia, 55(10):1551–1557, 2014
doi: 10.1111/epi.12745
pressure as part of the long-term management of patients
with epilepsy taking VPA. The long-term health implica-
tions of increased abdominal fat percentage in VPA users
requires further study, including full cardiovascular risk
assessment profile and metabolic syndrome assessment.
Such investigations would help to determine whether
screening routinely for these factors should be recom-
mended, and to identify whether weight loss and cardiovas-
cular preventative health strategies will provide benefit for
patients.
Acknowledgments
The authors wish to thank the twins and siblings who kindly participated
in the study. We would like to thank Prof. M.J. Seibel for the leptin assays
performed at the ANZAC Research Institute, The University of Sydney,
Concord NSW, Australia. We would like to thank the research staff of the
Wark Group at Department of Medicine RMH, and at the Kolling Institute
of Medical Research for conducting participant study visits.
This research was facilitated through access to the Australian Twin Reg-
istry, a national resource supported by an Enabling Grant (ID 628911)
from the National Health & Medical Research Council (NHMRC). The
Bone Research Program in Twins (University of Melbourne) was funded
by the NHMRC. The study was partially funded by an NHMRC pro-
ject grant (ID 400089). Dr. Sandra Petty was supported by NHMRC
Postgraduate Scholarship during this project, and is currently supported
by an MBC Postdoctoral Medical Research Fellowship at The Mel-
bourne Brain Centre RMH (NHMRC Centre for Research Excellence
Grant 1001216).
Disclosure
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. Dr. Sandra Petty has received support from UCB Pharma for a
project not related to this research. Prof. John D. Wark has received support
from UCB Pharma for a project not related to this research. Prof. Sam
Berkovic has received support from UCB Pharma, Janssen Cilag, and Sano-
fi Aventis for a project not related to this research. The remaining authors
have no potential conflicts of interest to disclose.
References
1. Morrell MJ, Isojarvi J, Taylor AE, et al. Higher androgens and weight
gain with valproate compared with lamotrigine for epilepsy. Epilepsy
Res 2003;54:189–199.
2. Dinesen H, Gram L, Andersen T, et al. Weight gain during treatment
with valproate. Acta Neurol Scand 1984;70:65–69.
3. El-Khatib F, Rauchenzauner M, Lechleitner M, et al. Valproate,
weight gain and carbohydrate craving: a gender study. Seizure
2007;16:226–232.
4. Gaspari CN, Guerreiro CA. Modification in body weight associated
with antiepileptic drugs. Arq Neuropsiquiatr 2010;68:277–281.
5. Isojarvi JI, Laatikainen TJ, Knip M, et al. Obesity and endocrine
disorders in women taking valproate for epilepsy. Ann Neurol
1996;39:579–584.
6. Pickrell WO, Lacey AS, Thomas RH, et al. Weight change associated
with antiepileptic drugs. J Neurol Neurosurg Psychiatry 2013;84:796–
799.
7. Verrotti A, Manco R, Agostinelli S, et al. The metabolic syndrome in
overweight epileptic patients treated with valproic acid. Epilepsia
2010;51:268–273.
8. Rexrode KM, Carey VJ, Hennekens CH, et al. Abdominal adiposity
and coronary heart disease in women. JAMA 1998;280:1843–1848.

9. Carey DG, Nguyen TV, Campbell LV, et al. Genetic influences on
central abdominal fat: a twin study. Int J Obes Relat Metab Disord
1996;20:722–726.
10. Wells JC. Sexual dimorphism of body composition. Best Pract Res
Clin Endocrinol Metab 2007;21:415–430.
11. Bjorholt PG, Nakken KO, Rohme K, et al. Leisure time habits and
physical fitness in adults with epilepsy. Epilepsia 1990;31:83–87.
12. Nakken KO, Kornstad S. Do males 30–50 years of age with chronic
epilepsy and on long-term anticonvulsant medication have lower-than-
expected risk of developing coronary heart disease? Epilepsia
1998;39:326–330.
13. Elliott JO, Jacobson MP, Haneef Z. Cardiovascular risk factors and
homocysteine in epilepsy. Epilepsy Res 2007;76:113–123.
14. St Germaine-Smith C, Liu M, Quan H, et al. Development of an
epilepsy-specific risk adjustment comorbidity index. Epilepsia
2011;52:2161–2167.
15. Verrotti A, D’Egidio C, Mohn A, et al. Weight gain following
treatment with valproic acid: pathogenetic mechanisms and clinical
implications. Obes Rev 2011;12:e32–e43.
16. Lagace DC, McLeod RS, Nachtigal MW. Valproic acid inhibits leptin
secretion and reduces leptin messenger ribonucleic acid levels in
adipocytes. Endocrinology 2004;145:5493–5503.
17. Verrotti A, Basciani F, Morresi S, et al. Serum leptin changes in
epileptic patients who gain weight after therapy with valproic acid.
Neurology 1999;53:230–232.
18. Pylvanen V, Knip M, Pakarinen A, et al. Serum insulin and
leptin levels in valproate-associated obesity. Epilepsia 2002;43:514–
517.
19. Pritchard JE, Nowson CA, Strauss BJ, et al. Evaluation of dual energy
X-ray absorptiometry as a method of measurement of body fat. Eur J
Clin Nutr 1993;47:216–228.
20. Haarbo J, Gotfredsen A, Hassager C, et al. Validation of body
composition by dual energy X-ray absorptiometry (DEXA). Clin
Physiol 1991;11:331–341.
21. Peppa M, Koliaki C, Hadjidakis DI, et al. Regional fat distribution and
cardiometabolic risk in healthy postmenopausal women. Eur J Intern
Med 2013;24:824–831.
22. Kamel EG, McNeill G, Han TS, et al. Measurement of abdominal fat
by magnetic resonance imaging, dual-energy X-ray absorptiometry
and anthropometry in non-obese men and women. Int J Obes Relat
Metab Disord 1999;23:686–692.
23. Shiek Ahmad B, Hill KD, O’Brien TJ, et al. Falls and fractures in
patients chronically treated with antiepileptic drugs. Neurology
2012;79:145–151.
24. Petty SJ, Hill KD, Haber NE, et al. Balance impairment in chronic
antiepileptic drug users: a twin and sibling study. Epilepsia
2010;51:280–288.
1557
Valproate, Weight and Fat Distribution
25. Petty SJ, Paton LM, O’Brien TJ, et al. Effect of antiepileptic
medication on bone mineral measures. Neurology 2005;65:1358–
1365.
26. Guthrie JR, Dennerstein L, Taffe JR, et al. Central abdominal fat and
endogenous hormones during the menopausal transition. Fertil Steril
2003;79:1335–1340.
27. Nowson CA, Green RM, Hopper JL, et al. A co-twin study of the
effect of calcium supplementation on bone density during adolescence.
Osteoporos Int 1997;7:219–225.
28. MacInnis RJ, Cassar C, Nowson CA, et al. Determinants of bone
density in 30- to 65-year-old women: a co-twin study. J Bone Miner
Res 2003;18:1650–1656.
29. Christian JC, Kang KW. Efficiency of human monozygotic twins in
studies of blood lipids. Metabolism 1972;21:691–699.
30. El Haber N, Hill KD, Cassano AM, et al. Genetic and environmental
influences on variation in balance performance among female twin
pairs aged 21–82 years. Am J Epidemiol 2006;164:246–256.
31. Grundy SM, Cleeman JI, Daniels SR, et al. Diagnosis and
management of the metabolic syndrome: an American Heart
Association/National Heart, Lung, and Blood Institute scientific
statement. Curr Opin Cardiol 2006;21:1–6.
32. Fang J, Chen S, Tong N, et al. Metabolic syndrome among Chinese
obese patients with epilepsy on sodium valproate. Seizure
2012;21:578–582.
33. Report of a WHO consultation. Obesity: preventing and managing the
global epidemic. World Health Organ Tech Rep Ser 2000; Vol. 894,
pp. i–xii, 1–253.
34. Zhang C, Rexrode KM, van Dam RM, et al. Abdominal obesity and
the risk of all-cause, cardiovascular, and cancer mortality: sixteen
years of follow-up in US women. Circulation 2008;117:1658–1667.
35. Elliott JO, Lu B, Moore JL, et al. Exercise, diet, health behaviors, and
risk factors among persons with epilepsy based on the California
Health Interview Survey, 2005. Epilepsy Behav 2008;13:307–315.
36. Howard GM, Radloff M, Sevier TL. Epilepsy and sports participation.
Curr Sports Med Rep 2004;3:15–19.
37. Clasey JL, Kanaley JA, Wideman L, et al. Validity of methods of body
composition assessment in young and older men and women. J Appl
Physiol 1999;86:1728–1738.
38. Liu KH, Chan YL, Chan WB, et al. Sonographic measurement of
mesenteric fat thickness is a good correlate with cardiovascular risk
factors: comparison with subcutaneous and preperitoneal fat thickness,
magnetic resonance imaging and anthropometric indexes. Int J Obes
Relat Metab Disord 2003;27:1267–1273.
39. Smith SR, Lovejoy JC, Greenway F, et al. Contributions of total body
fat, abdominal subcutaneous adipose tissue compartments, and visceral
adipose tissue to the metabolic complications of obesity. Metabolism
2001;50:425–435.
Epilepsia, 55(10):1551–1557, 2014
doi: 10.1111/epi.12745