Acta Neurol Scand 2011: 124 (Suppl.

191): 23–27  2011 John Wiley & Sons A ⁄ S

ACTA NEUROLOGICA
SCANDINAVICA

Haematological side effects of antiepileptic

drug treatment in patients with epilepsy
Bachmann T, Bertheussen KH, Svalheim S, Rauchenzauner M, Luef G,
Gjerstad L, Taubøll E. Haematological side effects of antiepileptic drug
treatment in patients with epilepsy.
Acta Neurol Scand: 2011: 124 (Suppl. 191): 23–27.
 2011 John Wiley & Sons A ⁄ S.
Objectives – Little is known about the haematological side effects of the
newer antiepileptic drugs (AEDs), but recent case reports have raised
concerns regarding the possibility of altered thrombocyte counts or
function in some patients during levetiracetam (LEV) treatment. The
aim of our study was to investigate haematological changes in patients
treated with the newer AEDs, LEV and lamotrigine (LTG), compared
with the older AEDs, valproate (VPA) and carbamazepine (CBZ).
Methods – This cross-sectional study included 251 patients with
epilepsy of both genders, aged 18–45 years, using AED monotherapy:
52 patients on LEV (31 men, 21 women), 80 on LTG (37 men, 43
women), 90 on CBZ (61 men, 29 women), 29 on VPA (15 men, 14
women), and 79 healthy controls (36 men, 43 women). Haemoglobin
(Hb), white blood cells (WBC) and platelet (thrombocyte) counts were
estimated. The subjects were recruited from hospitals in south-eastern
Norway and Innsbruck, Austria. Results – Significantly lower platelet
counts were recorded in both men and women on LEV monotherapy.
In the LEV group, platelets were 14% lower (40.68 · 109 ⁄ l lower) than
in the control group. There was no difference according to sex or age of
the patients. Only minor changes in haematological parameters were
observed for the other drugs investigated. Conclusions – Both men and
women treated with LEV monotherapy have lower blood platelet
counts than healthy controls, with no difference in Hb or WBC.
Haematological changes observed with the other AEDs were minor.
T. Bachmann1, K. H. Bertheussen1,
S. Svalheim1, M. Rauchenzauner2,
G. Luef3, L. Gjerstad1,4, E. Taubøll1,4
1
Department of Neurology, Oslo University Hospital –
Rikshospitalet, Oslo, Norway; 2Department of
Paediatrics IV, Medical University Innsbruck, Innsbruck,
Austria; 3Department of Neurology, Medical University
Innsbruck, Innsbruck, Austria; 4Faculty of Medicine,
University of Oslo, Oslo, Norway
Key words: carbamazepine; haematology;
haemoglobin; lamotrigine; leucocytes; levetiracetam;
thrombocytes; valproate
K. H. Bertheussen, Department of Neurology, Oslo
University Hospital – Rikshospitalet, 0027 Oslo, Norway
Tel.: +4741566423
Fax: +4723074891
e-mail: bachmann_tanja@hotmail.com
Accepted for publication April 10, 2011

reported in which patients develop changes in

Introduction
Antiepileptic drugs (AEDs) are the cornerstone of
epilepsy treatment. The goals of treatment are
freedom from seizures, no side effects and a good
quality of life for the patient. However, AEDs are
known to cause a wide range of adverse effects,
including haematological changes. These have been
well described for the older AEDs, but clinical
experience for the more recent AEDs is currently
insufficient to draw robust conclusions.
Levetiracetam (LEV) is a relatively new, broad-
spectrum AED that has seen extensive use during
recent years (1). Its mechanism of action is not
completely known, but differs from those of other
AEDs. LEV is described as being well tolerated (2),
and the most common adverse effects are generally
mild (3). Nevertheless, several cases have been
thrombocyte count and function during LEV
treatment (2, 4–6).
So far, the effect of LEV on thrombocyte count
has not been investigated in larger cross-sectional
or prospective studies. Recently, Sahaya et al.
(2010) reviewed medical records of 758 patients,
aged 18 years or older, who received LEV during
their stay in a university hospital. In patients with
thrombocytopenia (29 patients) medical records
were examined to identify possible causes. For
most cases, alternative causes were identified, but
in one patient, a clear association between throm-
bocytopenia and LEV therapy was found (7).
Few haematological changes have been
described during lamotrigine (LTG) treatment.
When these changes have been observed, concom-
itant AEDs or other medication, high LTG doses,

23
Bachmann et al.
or rapid dose escalation have been considered to be
contributory (8).
Valproic acid (VPA) can cause several haema-
tological changes. Onset and severity may vary, but
haematological adverse effects normally occur
when VPA serum levels are high and can be
reversed by reducing the VPA dose. The most
common change is thrombocytopenia, which most
often is mild and transient (9, 10). VPA can also
affect the bone marrow directly and cause cytope-
nia in one or more cell lines (11).
Carbamazepine (CBZ) is known to cause a
transient reduction in white blood cells (WBC)
during the first 3 months of treatment (12). In only
2% of cases, the leucocytopenia persists until CBZ
discontinuation (13). It has also been shown that
CBZ can cause thrombocytopenia (13, 14), and
some cases have shown clinical symptoms such as
petechiae and bleeding tendency. The most haz-
ardous and potentially lethal adverse effect is CBZ-
induced aplastic anaemia. (12, 13). There is no
known association between dosage, duration of
CBZ use or patient age (12, 14).
The aim of our study was to investigate haema-
tological changes, especially the possible occur-
rence of thrombocytopenia, in patients using the
more recently developed AEDs, LEV and LTG.
Additionally, we investigated whether the haema-
tological side effects of the older AEDs, VPA and
CBZ, which have previously been reported,
occurred in our population of patients with
epilepsy.
Materials and methods
This cross-sectional study included 251 patients
with epilepsy (women ⁄ men: 107 ⁄ 144) and 79 con-
trols (women ⁄ men: 43 ⁄ 36). The patients were
treated with LEV (n = 52; women ⁄ men: 21 ⁄ 31),
LTG (n = 80; women ⁄ men: 43 ⁄ 37), VPA (n = 29;
women ⁄ men: 14 ⁄ 15) or CBZ (n = 90; women ⁄
men: 29 ⁄ 61) in monotherapy for at least 6 months.
Morning blood samples were drawn before AED
doses. In women, blood samples were taken 5–
7 days after their last menstrual period (follicular
phase).Haematological parameters, Hb, WBC and
thrombocyte counts, were measured in the patients
using AEDs, and the results were compared with
corresponding measurements from healthy con-
trols.
The participants were aged between 18 and
45 years (mean age 31.5 years;, SD = 8.397) with
no differences in age between the four study groups
(patients treated with either LEV, LTG or CBZ or
controls) or sexes and were recruited from hospi-
tals in the south-eastern part of Norway and the
24
Department of Neurology of the University Hos-
pital Innsbruck, Austria. The control group was
recruited among hospital staff, students and people
living in close proximity to the hospitals. Pregnant
women and subjects with concomitant treatment,
including warfarin or oral contraception, were
excluded from this study, as were severely mentally
retarded patients. The control group used no med-
ications at the time the blood samples were drawn,
had never used AEDs or had any epileptic seizures.
Statistics
Data are presented as means and standard devia-
tions (SD). Statistical analysis was performed using
Statistical Package for Social Sciences for Win-
dows (version 16.0; SPSS Inc., Chicago, IL, USA).
Data were normally distributed. Independent
sample t-tests were used for the analyses of
continuous data, comparing treatment groups
with controls. A two-sided P £ 0.05 was consid-
ered statistically significant.
Results
Thrombocytes
Significantly lower thrombocyte levels were mea-
sured in all patients taking LEV compared with
controls (240.0 vs 280.7 · 109 ⁄ l; P-value < 0.001)
(Table 1). The difference was also significant if the
results from the men and women were analysed
separately (P-values; women 0.008, men 0.015).
Among patients on LEV monotherapy, mean
thrombocyte levels were 14.5% lower (40.7 ·
109 ⁄ l lower) than in the control group. For
women, the mean reduction in platelets was
12.5%, whereas the average reduction among
men was 14.4%. There were no changes in throm-
bocyte levels for patients using LTG, VPA or CBZ
compared with the control group.
Leucocytes (WBC)
There was a small, but statistically significant,
increase in WBC counts in women on either LEV
(P = 0.017), LTG (P = 0.006), VPA (P = 0.030)
or CBZ (P = 0.039) compared with controls
(Table 2). For men alone, and for women and
men together, no increase could be identified.
Haemoglobin (Hb)
Hb was slightly, but significantly, higher in women
using LEV (P = 0.011) and lower in men taking
CBZ (P = 0.034) compared with that in controls
 Haematological side effects of AEDs

Table 1 Effects of antiepileptic drugs on thrombocyte counts (measured as number in the LEV group was 240 · 109 plate-
109 ⁄ l)
lets ⁄ l, which is well above the level at which an
increased bleeding tendency might be anticipated
Mean SD 95% CI P-value Range
in otherwise healthy patients, a decrease in platelets
Levetiracetam may turn out to be clinically significant in certain
Women 253.7 49.7 10.2, 62.7 0.008 167–361 cases. For example, this decrease could be of
Men 230.7 65.6 7.8, 69.7 0.015 126–407
All 240.0 50.3 20.2, 61.2 0.000 126–407
clinical significance in patients who are using
Lamotrigine additional medications, as well as LEV, which
Women 282.2 79.7 )20.6, 61.2 0.597 92–500 might affect the platelets and ⁄ or the blood clotting
Men 261.1 45.1 )16.5, 33.4 0.503 159–379 system (i.e. Warfarin, non-steroidal anti-inflamma-
All 272.6 66.5 )10.9, 27.0 0.402 92–500
tory drugs or selective seretonin reuptake inhibi-
Valproate
Women 266.6 70.7 )19.1, 66.1 0.261 136–411 tors) (15). As the use of concomitant medications
Men 270.0 69.1 )42.9, 41.7 0.976 128–388 increases with age and our population was an
All 268.4 68.7 )16.2, 40.8 0.389 128–411 otherwise healthy population aged 18–45 years, we
Carbamazepine recommend that this study should be extended to
Women 279.7 98.7 )29.5, 50.2 0.602 177–605
Men 258.4 71.7 )16.0, 38.1 0.420 58–559
include elderly people and patients with polyphar-
All 265.3 81.4 )5.4, 36.1 0.145 58–605 macy. It could also be interesting to know the
Controls levels of haematological parameters in patients
Women 290.1 60.3 N ⁄A N ⁄A 204–384 with epilepsy who are not taking any AEDs or
Men 269.5 46.6 N ⁄A N ⁄A 151–376
All 280.7 53.9 N ⁄A N ⁄A 151–384
other medications, to evaluate whether the changes
in Hb, WBC or tbc might be associated with the
N ⁄ A, not applicable. P-values are from comparison with controls. disease itself. However, this comparison may not
be valid because patients taking no AEDs at all are
normally far healthier than those needing contin-
Table 2 Effects of antiepileptic drugs on white blood cell counts (measured as
109 ⁄ l) uous AED treatment to control seizure frequency.
There is to our knowledge no discrepancy in Hb,
Mean SD 95% CI P-value Range WBC or tbc levels between patients with epilepsy
before commencing AED treatment and otherwise
Levetiracetam
Women 6.5 1.8 )2.1, )0.2 0.017 3.4–10.8
healthy persons, and haematological changes are
Men 6.1 1.8 )0.8, 0.9 0.847 3.7–12.5 observed immediately after start of AED treatment
All 6.3 1.8 )1.2, 0.1 0.081 3.4–12.5 (12). It is therefore no reason to believe that the
Lamotrigine epilepsy itself affects haematological parameters,
Women 6.4 2.0 )1.9, )0.3 0.006 3.7–12.1
although it cannot be excluded.
Men 6.3 2.7 )1.2, 0.9 0.834 3.4–14.5
All 6.4 2.3 )1.3, )0.0 0.046 3.4–14.5
Valproate Table 3 Effects of antiepileptic drugs on haemoglobin concentrations (Hb)
Women 6.9 2.4 )3.0, )0.2 0.030 2.7–11.1 (measured as g ⁄ 100 ml)
Men 6.1 1.9 )1.1, 1.2 0.878 4.0–11.3
All 6.5 2.1 )1.7, 0.1 0.079 2.7–11.3 Mean SD 95% CI P-value Range
Carbamazepine
Women 6.3 2.2 )2.0, )0.1 0.039 3.6–11.5 Levetiracetam
Men 5.8 1.6 )0.2, 1.2 0.174 2.5–10.5 Women 13.9 0.9 )1.1, )0.2 0.011 11.7–15.6
All 5.9 1.8 )0.1, 0.3 0.435 2.5–11.5 Men 15.3 1.0 )0.4, 0.5 0.847 12.9–17.0
Controls All 14.8 1.2 )1.0, )0.1 0.024 11.7–17.0
Women 5.3 1.5 N ⁄A N ⁄A 3.0–10.2 Lamotrigine
Men 6.6 1.7 N ⁄A N ⁄A 3.4–10.2 Women 13.4 1.5 )0.7, 0.4 0.557 6.7–15.5
All 5.7 1.6 N ⁄A N ⁄A 3.0–10.2 Men 15.1 0.9 )0.1, 0.8 0.145 12.6–17.3
All 14.2 1.5 )0.4, 0.5 0.806 6.7–17.3
N ⁄ A, not applicable. P-values are from comparison with controls. Valproate
Women 13.1 1.0 )0.5, 0.7 0.663 10.5–14.3
Men 14.8 1.6 )0.1, 1.3 0.110 10.3–17.1
(Table 3). There were no other statistically signif- All 14.0 1.6 )0.4, 0.9 0.496 10.3–17.1
icant changes in Hb for the other AEDs. Carbamazepine
Women 13.5 0.7 )0.6, 0.1 0.149 12.5–15.6
Men 15.0 0.9 0.0, 0.8 0.034 11.9–17.4
Discussion All 14.5 1.1 )0.7, 0.1 0.146 11.9–17.4
Controls
The most important finding of our study was that Women 13.2 0.8 N ⁄A N ⁄A 11.6–14.5
Men 15.4 1.0 N ⁄A N ⁄A 13.3–17.2
blood platelet counts were significantly lower in
All 14.2 1.4 N ⁄A N ⁄A 11.6–17.2
patients treated with LEV compared with those in
healthy controls. Although the mean platelet N ⁄ A, not applicable. P-values are from comparison with controls.

25
Bachmann et al.
Lower platelet counts could also be clinically
important in patients with low pretreatment plate-
let numbers or with other haematological diseases.
A further reduction in platelet count might result
in a critically low thrombocyte level, below which
the patient develops clinical symptoms with a
tendency to bleeding.
In addition, it is important to emphasise that in
some patients, the reduction in platelet number will
be greater than the mean decrease of 40 · 109
platelets ⁄ l. Although this was not found in this
study of otherwise healthy young adults, it would
concur with previous reports (16, 17). For individ-
ual patients with low pretreatment platelet counts,
a greater than average reduction in thrombocytes
may lead to clinically significant bleeding.
It should be noted that not only the total
number of platelets might be affected by the
drug, but also their function. In one case report
(6), a woman developed ecchymoses and prolonged
bleeding time directly after starting LEV treatment,
and this persisted until discontinuation of the drug
13 months later. She did not use any other med-
ication, had no known haematological disorder
herself or in her family, and apart from the
prolonged bleeding, all her haematological param-
eters were normal (including thrombocyte counts,
clotting factors and von Willebrand factor) (6).
The authors concluded that the ecchymoses and
prolonged bleeding time were probably caused by
LEV-induced thrombocytopathia (normal number,
but altered function of the thrombocytes).
Levetiracetam is usually relatively well tolerated,
and the use of the drug is increasing (1, 18). The
concentrations used have been fairly low. In
Norway, the therapeutic drug level has been 30–
130 lm until recently, but this has now been
increased to 40–240 lm. In our study, mean
serum concentrations based on a subset of the
patients were 43.3 lm in women and 58.1 lm in
men. Higher concentrations might increase the risk
of adverse effects, but clinical experience with such
adverse effects is so far derived mainly from
patients taking the drug according to earlier
recommendations. Unfortunately, our data did
not allow a further analysis of the relation between
LEV concentrations and haematological changes.
If LEV is going to be used in higher concentrations
in the future, and possibly also more frequently
intravenously in treatment of status epilepticus (1),
it is important that its effects on platelet number
and function are investigated more thoroughly in
these situations.
There is no information on the mechanism by
which LEV might affect thrombocyte count or
function. The mode of action of LEV has not been
26
determined in detail, but it is known that LEV binds
to SV2A vesicle proteins, which are widely distrib-
uted throughout the CNS and also in endocrine cells
(19). To date, no information on a possible throm-
bocyte expression of SV2A vesicle protein is avail-
able. But if so, this could be a possible mechanism
by which LEV affects thrombocytes. Another
possibility was proposed by Sahaya et al. (2010),
who speculated that there could be an immune-
mediated pathogenesis, based on a few recent cases
of LEV-induced thrombocytopenia in immune-
competent patients (i.e. patients not taking immu-
nomodifying drugs or having immunocompromiz-
ing illnesses) and also a case report of LEV-induced
thrombocytopenia in an immunocompromised
individual. However, this theory currently has no
further support in the literature.
Other statistically significant findings in this
study were small and therefore considered to be
without clinical significance. They comprise a
decrease in Hb for men using CBZ and an increase
in women on LEV. It also includes increased
leucocytes in women using either of LEV, LTG,
VPA or CBZ. Such findings have not previously
been reported. We also did not identify the
haematological changes previously described for
older AEDs. This could be attributed to the fact
that one of the inclusion criteria for the study was
monotherapy for at least 6 months. Patients who
might have experienced haematological adverse
effects within the first 6 months of use (which is
often the case) would probably have discontinued
the treatment and hence would not have been
included in this study. This result may also reflect
properties of the study population.
Additionally, platelet reduction as a major
adverse effect in patients treated with VPA (9)
was not identified in the current study. This may be
because a decrease in thrombocytes on VPA
treatment seems to occur mainly when serum
levels exceed 700 lm in women and 900 lm in
men (16). In addition, only patients on long-term
VPA treatment were included in the present study;
patients who develop lower platelet counts after
starting VPA treatment would probably have
discontinued the treatment long before eligibility
for the study, which included only chronically-
treated VPA patients.
In conclusion, haematological side effects may
occur in patients with epilepsy using the newer
AEDs. LEV seems to have an effect on thrombo-
cyte counts and perhaps on their function also.
Further investigations are needed with regard to
the clinical implications of this observation in
selected patient groups, especially elderly people
and patients using polypharmacy.

Acknowledgments
The first two authors have contributed equally to this article.
We thank GlaxoSmithKline, USA, for supporting the study
with an investigator-initiated grant.
Conflict of interest
This study was supported by an unrestricted research grant
from GlaxoSmithKline, USA.
References
1. Crepeau AZ, Treiman DM. Levetiracetam: a comprehensive
review. Expert Rev Neurother 2010;10:159–71.
2. Kimland E, Höjeberg B, Von Euler M. Levetiracetam in-
duced thrombocytopenia. Epilepsia 2004;45:877–8.
3. Ben-Menachem E. Levetiracetam: treatment in epilepsy.
Expert Opin Pharmacother 2003;4:2079–88.
4. Meschede A, Runge U, Sabolek M. Thrombocytopenia
during levetiracetam therapy. Epilepsy Res 2008;80:91–2.
5. Hartmann H, Schumacher U, Seeck M. Verschlechterung
einer Immunthrombopenie unter Monotherapie mit Lev-
etiracetam. Akt Neurol 2002;29:467–8.
6. Herrman H, Taubøll E, Gjerstad L Levetiracetam-induced
thrombocytopathia. American Epilepsy Society Meeting,
San Diego, 2006. http://www.aesnet.org/go/publications/
aes-abstracts/abstract-search/mode/display/st/herrman%20/
sy/2006/sb/All/id/6556
7. Sahaya K, Goyal MK, Sarwal A, Singh NN. Levetirace-
tam-induced thrombocytopenia among inpatients: a ret-
rospective study. Epilepsia 2010;51:2492–5.
8. Ural AU, Avcu F, Gökcil Z, Nevruz O, Cetin T. Leu-
copenia and thrombocytopenia possible associated with
Haematological side effects of AEDs
lamotrigine use in a patient. Epileptic Disord 2005;7:33–
5.
9. Allan RW. Myelodysplastic syndrome associated with
chronic valproic acid therapy: a case report and review of
the literature. Hematology 2007;12:493–6.
10. Acharya S, Bussel J. Hematologic toxicity of Sodium
Valproate. J Pediatr Hematol Oncol 2000;22:62–5.
11. Handoko KB, Souverein PC, Van Staa TP et al. Risk of
aplastic anemia in patients using antiepileptic drugs. Epi-
lepsia 2006;47:1232–6.
12. Harden CL. Therapeutic safety monitoring-what to look
for and when to look for it. Epilepsia 2000;41:37–44.
13. Hart RG, Easton JD. Carbamazepine and hematological
monitoring. Ann Neurol 1982;11:309–12.
14. Sobotka JL, Alexander B, Cook BL. A review of Carba-
mazepine‘s hematologic reactions and monitoring recom-
mendations. DICP 1990;24:1214–9.
15. Patsalos PN, Fröscher W, Pisani F, Van Rijn CM. The
importance of drug interactions in epilepsy therapy. Epi-
lepsia 2002;43:365–85.
16. Nasreddine W, Beydoun A. Valproate-induced thrombocy-
topenia: a prospective monotherapy study. Epilepsia
2008;49:438–45.
17. Tohen M, Castillo J, Cole JO, Miller MG, De Los
HEROS R, Farrer RJ. Thrombocytopenia associated with
carbamazepine: a case series. J Clin Psychiatry
1991;52:496–8.
18. Bauer J, Ben-Menachem E, Krämer G, Fryze W, Da Silva
S, Kasteleijn-Nolst Trenité DG. Levetiracetam: a long-
term follow-up study of efficacy and safety. Acta Neurol
Scand 2006;114:169–76.
19. Lynch BA, Lambeng N, Nocka K et al. The synaptic vesicle
protein SV2A is the binding site for the antiepileptic drug
levetiracetam. Proc Natl Acad Sci 2004;101:9861–6.
27
Copyright of Acta Neurologica Scandinavica is the property of Wiley-Blackwell and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.
Copyright of Acta Neurologica Scandinavica is the property of Wiley-Blackwell and its content may not be
copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written
permission. However, users may print, download, or email articles for individual use.
Copyright of Acta Neurologica Scandinavica. Supplementum is the property of Wiley-
Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv
without the copyright holder's express written permission. However, users may print,
download, or email articles for individual use.