Kaohsiung Journal of Medical Sciences (2017) 33, 130e136

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: http://www.kjms-online.com

ORIGINAL ARTICLE

Drug interaction between valproic acid and

carbapenems in patients with epileptic seizures
Chi-Ren Huang a, Chih-Hsiang Lin a, Shu-Chen Hsiao b, Nai-Ching Chen a,
Wan-Chen Tsai a, Shang-Der Chen a, Yan-Ting Lu a, Yao-Chung Chuang a,*

a
Department of Neurology, Kaohsiung Chang Gung Memorial Hospital and Chang Gung
University College of Medicine, Kaohsiung, Taiwan
b
Department of Pharmacy, Kaohsiung Chang Gung Memorial Hospital and Chang Gung
University College of Medicine, Kaohsiung, Taiwan

Received 4 August 2016; accepted 13 December 2016

Available online 11 January 2017

KEYWORDS Abstract Valproic acid (VPA) is a widely used antiepileptic drug (AED). When carbapenems
Antiepileptic drugs; are concomitantly used with VPA, the serum levels of VPA may decrease and aggravate sei-
Carbapenems; zures. The aim of this study was to evaluate the risk factors associated with decreased serum
Drug interaction; VPA levels and clinical outcome in patients being treated with a combination of carbapenems
Epileptic seizures; and VPA. Fifty-four adult patients who were treated with VPA for epileptic seizures concomi-
Valproic acid tant with carbapenems for the treatment of infections were evaluated in this study. Serum VPA
levels were measured before and during combination therapy with VPA and carbapenems, and
the change in serum VPA levels was calculated. The risk factors related to the decrease in
serum VPA levels and clinical outcomes were evaluated. Our results show that VPA concentra-
tions were reduced to subtherapeutic levels after the introduction of carbapenems. The reduc-
tion in VPA concentrations was found within 24 hours of the start of treatment with
carbapenems. VPA levels continuously declined while the combination of treatments was used,
which aggravated epileptic seizures in 48% of the patients. Renal disease and enzyme-inducing
AEDs were risk factors that contributed to the severity of reduced serum VPA levels during
combined treatment with carbapenems. This study suggests that clinicians need to be aware
of the reduction of VPA concentrations to subtherapeutic levels and the aggravation of seizures
while patients are treated with a combination of carbapenems and VPA.
Copyright ª 2017, Kaohsiung Medical University. Published by Elsevier Taiwan LLC. This is an
open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/
by-nc-nd/4.0/).

Conflicts of interest: All authors declare no conflicts of interest.

* Corresponding author. Department of Neurology, Kaohsiung Chang Gung Memorial Hospital, Kaohsiung City 83301, Taiwan.
E-mail address: ycchuang@adm.cgmh.org.tw (Y.-C. Chuang).

http://dx.doi.org/10.1016/j.kjms.2016.12.001
1607-551X/Copyright ª 2017, Kaohsiung Medical University. Published by Elsevier Taiwan LLC. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Interaction between valproic acid and carbapenems
Introduction
Patients with epilepsy may have other medical comorbid-
ities that require various medications, other than antiepi-
leptic drugs (AEDs), and thus are at high risk of the effects
of drug interactions [1]. Many neurological or systemic in-
sults also provoke epileptic seizures, such as stroke, trau-
matic brain injury, brain tumor, central nervous system
(CNS) infections, electrolytic and metabolic disorders, and
anoxic encephalopathy [2,3]. Moreover, patients with epi-
lepsy may be admitted to hospital for infections, or develop
infections during hospitalization. Thus, drug interactions
between AEDs and antibiotics are important issues for the
management of infectious disease comorbidities in patients
with epilepsy [4]. For example, erythromycin and clari-
thromycin increase serum levels of carbamazepine, rifam-
picin increases the clearance of lamotrigine, and isoniazid
increases serum levels of carbamazepine, ethosuximide,
and phenytoin [4,5].
Valproic acid (VPA) is an AED commonly used for the
management of both partial and generalized seizures [6].
The drug is extensively metabolized in the liver and is
known to interact with other AEDs and non-AEDs [6]. Car-
bamazepine, phenytoin, and phenobarbital can decrease
serum levels of VPA [5]. Moreover, serum levels of VPA in-
crease with the use of chlorpromazine and decrease with
the use of oral contraceptives or carbapenems [5].
Carbapenems are a group of antibiotics that have a
broad spectrum of activity for the treatment of infections
caused by multidrug-resistant bacteria [7]. Carbapenems
have been shown to reduce serum levels of VPA in case
reports [8e14] and retrospective studies [15,16]. The
mechanisms of the interaction between VPA and carbape-
nems are complex and have still not been fully investigated
[17]. However, the reduction in serum VPA levels induced
by carbapenems may lead to the aggravation of seizures,
and it has been advised that the combined use of these
medications is best avoided [8e12,15].
This study evaluated the interaction between carbape-
nems and VPA in patients who were being treated for
epileptic seizures and comorbid infections. Furthermore,
we evaluated the risk factors that might contribute to
decreased serum VPA concentrations, and the conse-
quences of reduced VPA levels on clinical outcome.
Methods
This is an observational study from a single tertiary medical
center. The study was conducted by the Epilepsy Branch of
the Neurological Department of Kaohsiung Chang Gung
Memorial Hospital, Kaohsiung, Taiwan. Ethics approval for
this study was provided by the Chang Gung Medical Foun-
dation Institutional Review Board. All procedures per-
formed in studies involving human participants were in
accordance with the 1964 Helsinki declaration and its later
amendments or comparable ethical standards.
Study design
Between January 2005 and December 2013, 54 adult pa-
tients with epilepsy aged 17 years or over who were treated
131
with VPA for at least 3 months for seizure control were
included in this study. All of these patients were addition-
ally treated with carbapenems (imipenem, ertapenem, or
meropenem) for comorbid infections. Only one type of
carbapenem was selected for the control of infection. Pa-
tients who developed liver function impairments, defined
by elevations in liver function enzymes by more than
threefold their normal values, during the combined use of
carbapenems and VPA were excluded from this study. We
recorded the demographic data of the patients, including
the dose of VPA administered prior to the use of carbape-
nems, serum concentrations of VPA before and during the
use of carbapenems, seizure frequency before and after
the combination therapy, medical disease and infections
associated with the use of carbapenems, and the type of
AED currently being used. We also reviewed the literature
to look for clinical reports of drug interactions between VPA
and carbapenems, including imipenem, ertapenem, and
meropenem.
Blood samples obtained from patients were sent to the
central laboratory in Kaohsiung Chang Gung Memorial Hos-
pital for biochemical analyses. A fluorescence polarization
immunoassay was used to measure the serum levels of VPA
in conjunction with an AxSYM analyzer (Abbott Labora-
tories, Abbott Park, IL, USA).
The comorbidities in our patients included renal disease,
liver disease, diabetes mellitus, hypertension, and CNS le-
sions (cerebral infarct, cerebral hemorrhage, and brain
tumor). Comorbidity with renal disease was defined as an
insufficient renal function with a glomerular filtration rate
of 90 mL/min/1.73 m2 or less. Patients with a history of
chronic viral hepatitis, liver cirrhosis, and liver tumor were
considered to have comorbidity with liver diseases. For
outcome analyses, patients who experienced an increase in
seizure frequency during the combined use of both drugs or
died during admission were defined as the poor outcome
group. Those who survived without experiencing any in-
crease in seizure frequency were defined as the good
outcome group. For AED analyses, patients who used
phenytoin, carbamazepine, or phenobarbital with VPA were
defined as the enzyme inducer group. Patients using
oxcarbazepine or topiramate concurrently with VPA were
defined as the weaker enzyme inducer group. Patients who
used levetiracetam or gabapentin alongside VPA were
defined as the non-enzyme inducer group. All patients had
steady usage of combined AEDs before and during the study
period, particularly with enzyme inducers (phenytoin,
carbamazepine, or phenobarbital).
Statistical analysis
We collected serum VPA levels before the use of carbape-
nems as the baseline serum VPA level. The lowest serum
VPA level during the combination therapy was recorded.
The duration of time for which serum VPA levels dropped to
their lowest levels was also recorded. The change in serum
VPA level (D[VPA]) was defined as follows:
½Baseline VPA level  ½Lowest VPA level
D½VPAZ  100%
½Baseline VPA level
132 C.-R. Huang et al.

Statistical analyses were performed using SPSS software

Table 1 Demographic data of the 54 patients treated with
(version 11.0 for Windows; IBM, Chicago, IL, USA). An in-
a combination of valproic acid and carbapenems.
dependent sample t test was used for the analysis of cat-
egorical factors. For factor analysis, a correlation test was Characteristics
used to analyze the relationship between risk factors and D Sex, n (%)
[VPA]. The relationships between three different groups of Male/female 32/22
add-on AEDs, including the enzyme inducer, weaker Age (y) 65.2  16.6
enzyme inducer, and non-enzyme inducer groups, with the Age at seizure onset (y) 61.8  19.9
D[VPA] were analyzed. The relationships of the different Type of infection, n (%)
enzyme-inducing AEDs, including phenytoin, carbamaze- Central nervous system infection 5 (9.3)
pine, and phenobarbital, with D[VPA] were also analyzed. A Pneumonia 24 (44.4)
one-way analysis of variance was conducted to evaluate the Complicated urinary tract infection 7 (13.0)
contribution of the three different carbapenems to the D Abdominal infection 1 (1.9)
[VPA], and also to examine the different clinical conditions Other infection 17 (31.5)
that may affect the D[VPA]. For outcome analyses, we used Associated medical diseases, n (%)
the Chi-square test to explore the factors that may relate Diabetes mellitus 15 (27.8)
to poor outcome. Differences with a p value less than 0.05 Hypertension 24 (44.4)
were deemed to be statistically significant. Central nervous system lesion 35 (64.8)
Renal disease, n (%) 13 (24.1)
Liver disease, n (%) 7 (13.0)
Results Type of carbapenems, n (%)
Imipenem 5 (9.3)
Demographic data and drug interactions in 54 Meropenem 30 (55.6)
patients Ertapenem 19 (35.2)
Antiepileptic drug use, n (%)
The demographic data of all patients are listed in Table 1. Valproic acid use alone 22 (41.5)
There were 32 male and 22 female patients in this study. Combination therapy 32 (58.5)
The mean age of the patients (mean  standard deviation) Combined AEDs, n (%)
was 65.2  16.6 years (range, 20e92 years). Underlying Phenytoin 24 (44.4)
medical diseases included diabetes mellitus (n Z 15, Topiramate 7 (13.0)
27.8%), hypertension (n Z 24, 44.4%), CNS lesion (n Z 35, Phenobarbital 2 (3.7)
64.8%), renal diseases (n Z 13, 24.1%), and liver diseases Levetiracetam 2 (3.7)
(n Z 7, 13.0%). The infections for which carbapenems were Oxcarbazepine 1 (1.9)
administered included CNS infection (n Z 5, 9.3%), pneu- Carbamazepine 1 (1.9)
monia (n Z 24, 44.4%), sepsis (n Z 17, 31.5%), complicated Gabapentin 1 (1.9)
urinary tract infection (n Z 7, 13.0%), and abdominal Outcome, n (%)
infection (n Z 1, 1.9%). The etiology of epileptic seizures Increase seizure frequency 26 (48.1)
included symptomatic epilepsy (n Z 52, 96.3%) and idio- Death 14 (25.9)
pathic epilepsy (n Z 2, 3.7%). The types of epileptic sei- Values are expressed in mean  standard derivation unless
zures included partial seizures (n Z 41, 75.9%) and otherwise indicated.
generalized convulsive seizures (n Z 13, 24.1%). AED Z antiepileptic drug.

The use of carbapenems leads to decreased serum
levels of VPA
The median duration of concomitant VPA and carbapenems
use was 7 days (range, 1e34 days). The average dose of VPA
used was 1600 mg/d (range, 600e3200 mg/d) and the
baseline serum VPA concentration was 63.10 mg/mL (range,
21.00e110.84 mg/mL). All patients experienced a decrease
in VPA serum concentrations while being concomitantly
treated with carbapenems. The average low concentration
of VPA was 16.9 mg/mL (range, 0.13e49.00 mg/mL) and the
median duration during which VPA serum levels dropped to
their lowest was 5 days (range, 1e20 days). The carbape-
nems used in this study included imipenem (n Z 5, 9.3%),
meropenem (n Z 30, 55.6%), and ertapenem (n Z 19,
35.2%). The mean durations during which VPA levels
decreased to levels that were subtherapeutic and the
lowest recorded were 4.0 days (range, 1e16 days) and 5.8
days (range, 1e20 days), respectively. On the 1st day of
concomitant treatment, eight patients experienced a drop
in VPA levels to subtherapeutic levels and three of them to
the lowest levels recorded.
The D[VPA] of the patients was calculated, which may
reflect the severity of the interaction between VPA and
carbapenems. The mean D[VPA] of all patients was 73%
(range, 38e98%). In the three different carbapenem
groups, the mean D[VPA] was 80% (range, 49e99%) in the
meropenem group, 68% (range, 38e88%) in the ertapenem
group, and 51% (range, 31e73%) in the imipenem group. We
calculated the Pearson product-moment correlation coef-
ficient to investigate the relationship between D[VPA] and
the dose of carbapenems being used. There was no corre-
lation between these two factors.
The serum levels of VPA were significantly decreased
from the baseline levels in all patients when carbapenems
were administered, which is demonstrated in Figure 1. VPA
serum levels over subsequent days were all significantly
lower than baseline levels. When different carbapenems
Interaction between valproic acid and carbapenems
Figure 1. Trend of changes in serum VPA levels during
combination therapy with carbapenems. The serum levels of
VPA were regularly monitored at baseline (before the use of
carbapenems; (B), and 5 days, 10 days, 15 days, 20 days, and
25 days (D5, D10, D15, D20, and D25) after combination ther-
apy with carbapenems. Values represent mean  standard
error of the mean. * p < 0.05 versus baseline VPA levels (B) as
assessed using a paired-samples t test. VPA Z valproic acid.
were compared, as illustrated in Figure 2, they were also
found to decrease VPA serum levels from those of baseline
levels over subsequent days, which showed a trend toward
significance.
Importantly, 48.1% of patients experienced aggravation
of seizures, and 25.9% of patients died. The VPA serum
levels of those experiencing an aggravation of seizures
were 17.7  9.8 mg/mL and of those without aggravation
were 17.9  12.6 mg/mL. The difference was not statisti-
cally significant (p Z 0.944). Those patients who experi-
enced an aggravation of seizures were using phenytoin
Figure 2. Decrease in serum VPA levels due to combination
therapy with different carbapenem antibiotics. The serum
levels of VPA were regularly monitored at baseline (before the
use of carbapenems; B), and 5 days and 10 days (D5 and D10)
after combination therapy with three different carbapenems.
Values represent mean  standard error of the mean. * p < 0.05
versus baseline VPA levels (B) as assessed using a paired-
samples t test. VPA Z valproic acid.
133
(n Z 11), carbamazepine (n Z 1), topiramate (n Z 4), and
gabapentin (n Z 1) concurrently. Patients who died from
infection experienced pneumonia (n Z 8), sepsis (n Z 5),
and complicated urinary tract infections (n Z 1).
The drug interactions between VPA and
carbapenems
Table 2 shows the drug interaction between VPA and three
different carbapenems. These data were obtained from the
results of this study (54 patients) and a review of the
literature.
In the meropenem group, the total analysis of 185 cases
included 31 events from 13 case reports [8,14,18e24], 124
patients from four serial studies [15,16,25,26], and 30 pa-
tients from this study (Table 2). The VPAemeropenem drug
interaction analysis showed a mean reduction of 77%
(range, 34e99%) in VPA concentrations. VPA levels signifi-
cantly decreased from baseline levels (60.5 mg/mL) to the
lowest levels recorded (15.2 mg/mL), and 36% of patients
experienced an aggravation of seizures when VPA levels
were at subtherapeutic levels (Table 2).
In the VPA and ertapenem group, 19 patients from this
study and 11 patients from seven reports
[9,13,14,24,25,27,28] in the literature were included in the
analysis reported in Table 2. The analysis of the VPA-
eertapenem drug interaction showed a mean reduction of
71% (range, 38e98%) of VPA co-concentrations. VPA con-
centrations were significantly reduced (from a baseline
level of 63.9 mg/mL to 15.3 mg/mL), despite the increased
VPA dose. An aggravation of seizures was noted in 48% of
patients when VPA concentrations were at subtherapeutic
levels.
Reports about the interaction between VPA and imipe-
nem are limited in the literature. Based on an analysis of
five patients in this study and a report on seven patients
from the literature (Table 2) [13,21,22,24,29], the inter-
action between VPA and imipenem was found to reduce VPA
serum levels by 52% (29e73%). VPA concentrations were
also significantly reduced (65.2 mg/mL to 28.9 mg/mL) and
33% of patients experienced new seizures.
Risk factors for interaction between VPA and
carbapenems
A one-way analysis of variance was conducted to explore
the clinical conditions related to the D[VPA] and the results
are listed in Table 3. The results show that “renal disease”
(p Z 0.036) and “combination therapy of AEDs” (p Z 0.038)
were significantly associated with the D[VPA]. On analysis
of the three different groups of AEDs, the enzyme inducer
group was found to be significantly associated with D[VPA].
On further analysis of the different types of AEDs in the
enzyme inducer group, phenytoin was significantly associ-
ated with D[VPA] (p Z 0.028).
Of the 32 patients placed in the poor outcome group, 14
patients died during admission while 18 patients survived.
Of the 14 patients who died during admission, eight pa-
tients experienced an increase in the frequency of seizures.
All of the 18 patients who survived also had an increase in
the frequency of their seizures. The Chi-square test was
134 C.-R. Huang et al.

Table 2 The interaction between valproic acid and carbapenem.

Cases Dosage (mg) Baseline (mg/mL) Interaction (mg/mL) Effect (reduction) Seizure
Meropenem
a
Case reports 31 600e3600 61.1 (28e112) 12.6 (<1e28) 79% (49e99%) 10
[15] 29 400e4000 64.3 (40e111) 22.5 (<3e43) 66% (34e92%) 7
[16] 36 2040  70 50.8  4.5 9.9  2.1 82% (70e91%) ND
[30] 32 ND ND (1.63e43.87) (49e88%) 13
[25] 27 29.1e55.3/kg 52.0 (19.0e123.0) 11.0 (6.2e36.0) ND 10
This study 30 600e3200 63.8 (32.55e110.84) 14.52 (0.13e36.90) 78% (49e99%) 13
Total 185 400e4000 60.5 (19.0e123.0) 15.2 (<1e43) 77% (34e99%) 36% (53/149)
Ertapenem
b
Case reports 10 800e2000 69.1 (44e130) 8.4 (<1e22.5) 74% (64e98%) 4 (40%)
This study 19 1200e3200 61.17 (32.88e85.42) 18.91 (3.38e37.73) 70% (38e88%) 10 (52%)
Total 29 800e3200 63.9 (32.88e130) 15.3 (<1e37.73) 71% (38e98%) 14 (48%)
Imipenem
c
Case reports 7 400e1600 54.4 (36e80) 19.5 (12.6e24) 52% (29e73%) 1 (14%)
This study 5 1200e2400 71.69 (57.22e85.50) 34.5 (20.00e53.80) 52% (31e73%) 3 (60%)
Total 12 400e2400 65.2 (36e85) 28.9 (12.6e53.80) 52% (29e73%) 4 (33%)
Data are presented as mean  standard derivation; median (range).
ND Z no description.
a
References [8,14,18e24].
b
References [9,13,14,24,25,27,28].
c
References [13,21,22,24,29].

used to explore the clinical conditions that may be linked to Discussion

this group. No statistically significant association was noted
between infection type, comorbidities, different carbape- In this study, we found that serum VPA concentrations in all
nems used, and AEDs in this group. patients fell to subtherapeutic levels during combined
treatment with carbapenems. Among the carbapenems
used, a stronger interaction was observed between mer-
Table 3 Factors related to the percentage of serum val-
openem and VPA than that between ertapenem and imi-
proic acid level decrease.
penem. An increase in seizure frequency was experienced
Factors p Effect 95% CI by 48% of patients, which implies poor outcomes during the
size use of this combination therapy. Furthermore, renal dis-
Sex 0.109 0.05 0.15 to 0.02 ease and enzyme-inducing AEDs were found to be inde-
Type of infection pendent risk factors that contributed to the reduction of
CNS infection 0.196 0.03 0.05 to 0.24 serum VPA levels during combination therapy with carba-
Pneumonia 0.760 0.30 0.10 to 0.07 penems, especially when phenytoin was coadministered.
Sepsis 0.397 0.01 0.13 to 0.05 The interaction between carbapenems and VPA may
Complicated urinary 0.564 0.01 0.09 to 0.16 result in a significant decrease in serum VPA concentra-
tract infection tions. The concomitant use of meropenem and VPA caused
Abdominal infection 0.873 0.00 0.34 to 0.29 a 78% reduction in VPA concentrations in our patients, and
Associated medical disease 77% in the pooled analysis of the results from the litera-
Diabetes mellitus 0.997 0.00 0.10 to 0.10 ture [8,14e16,18e25,30]. The reduction of VPA concen-
Hypertension 0.118 0.05 0.15 to 0.02 trations occurred within 24e48 hours after the initiation
CNS lesions 0.093 0.05 0.16 to 0.01 of treatment with carbapenems. Our results also show
Renal disease 0.036* 0.08 0.20 to 0.01 that the VPA concentrations remained at subtherapeutic
Liver disease 0.420 0.01 -0.18 to 0.08 levels despite the use of increased doses of VPA, with
Mode of AED use most of the VPA concentrations recorded being less than
Valproic acid use alone 0.093 0.05 0.16 to 0.01 20 mg/mL (therapeutic level, 50e100 mg/mL). The reduc-
Combination therapy 0.042* 0.05 0.01 to 0.16 tion in VPA concentrations was not significantly different
Enzyme inducing AED 0.038* 0.08 0.01 to 0.17 between the high-dose meropenem (CNS infection treat-
Weak enzyme inducing 0.186 0.03 0.04 to 0.20 ment) group and the low-dose (non-CNS infection treat-
AED ment) group, and also showed no correlation with VPA
Non-enzyme inducing 0.839 0.00 0.11 to 0.13 dosage. After discontinuation of meropenem, the serum
AED VPA levels recorded within the first 7 days were less than
20 mg/mL and returned to the therapeutic range after
* p < 0.05.
8e14 days following discontinuation of meropenem
AED Z antiepileptic drug; CI Z confidence interval;
CNS Z central nervous system.
[15,16].
Interaction between valproic acid and carbapenems
The interactions between VPA and ertapenem or imi-
penem are less well studied than between VPA and mer-
openem. We also found a reduction in VPA concentrations
in patients using ertapenem or imipenem concomitantly.
The median reduction of VPA concentrations was 68% when
used in combination with ertapenem, and the combined use
of VPA and imipenem reduced VPA concentrations by 50%.
The reduction in VPA concentrations during ertapenem or
imipenem treatment also showed no dose-dependent ef-
fect, which was similar to the findings with meropenem.
Subtherapeutic VPA concentrations were found within 1e3
days after ertapenem use, and 2e5 days after the use of
imipenem. Among the three different carbapenems used in
our study, the imipenem group showed a lesser reduction in
VPA levels. These results suggest that imipenem may
interact less with VPA than meropenem and ertapenem.
The mechanisms related to the interactions between VPA
and carbapenems remain unclear. Possible mechanisms
related to the interaction between VPA and carbapenem have
been proposed based on the results from animal studies
[17,31]. The drug interactions may take place in the liver
where carbapenems enhance the glucuronidation of VPA by
increasing uridine diphosphateeglucuronic acid levels,
resulting in decreased serum levels of VPA [17,31,32]. More-
over, after intravenous administration of carbapenems and
VPA, VPA and carbapenems may interact in the erythrocytes.
This potential mechanism suggests that the distribution shift
related to multidrug-resistance proteins on adenosine
triphosphate-binding cassette transporters on erythrocyte
membranes may be inhibited by carbapenems [17,33].
Therefore, VPA is not effluxed out of the erythrocytes, which
results in decreased serum levels of VPA. Finally, when VPA is
given orally, its absorption into the intestinal lumen may be
restricted by intravenously administered carbapenem anti-
biotics [34]. This may relate to the inhibition of the mem-
brane transporter in intestinal cells [17,34].
Phenytoin is a strong cytochrome P450 enzyme inducer
and may interact with VPA [4]. However, the interaction
between VPA and phenytoin is complex and depends on the
protein binding activity and renal clearance of phenytoin
[4]. In general, when using a combination of phenytoin and
VPA, dosage adjustments are usually not necessary, and
clinically relevant changes in serum concentrations are not
expected [4]. In our patients, phenytoin had been used with
a steady-state dose before carbapenems were introduced;
thus, the effect of phenytoin on decreasing VPA levels might
be less significant than the interaction with carbapenems.
This study showed that renal disease was an indepen-
dent risk factor of decreased VPA serum levels after the
introduction of carbapenems. While the elimination of
carbapenems occurs mainly through the kidney [7], renal
diseases may impair their clearance and result in a pro-
longed interaction between VPA and carbapenems. There-
fore, caution should be exercised when treating patients
with renal disease with a combination of VPA and carba-
penems. However, the safety profile requires further study.
We are aware that there were multiple factors that may
result in worsening of seizures, including the epi-
leptogenicity of carbapenems [30], infections [35], and
reduced VPA serum levels. Throughout the literature, it has
been shown that different carbapenems have different rates
of epileptogenicity. The rate of imipenem-related seizures
135
ranges from 0.5% to 3.3% [36]; the range is 0.5e6.0% with
regard to meropenem-related seizures [37,38], and 0.4%
with regard to ertapenem-related seizures [39,40]. More-
over, the rate of seizures reported to occur because of in-
fections ranges from 22% to 32% [35]. While VPA serum levels
and seizures did not show statistical significance in this study,
we found that 48.1% of patients had increased seizure fre-
quency and 25.9% of patients died, which were higher than
the rates caused by carbapenems or infections. Because all
patients were already being treated with VPA for existing
seizures prior to the use of carbapenems, a higher percent-
age of seizures and poor outcomes were seen. Thus, we
suggest that the worsening of seizures may relate to the
interaction between VPA and carbapenems, and the
decrease in VPA serum levels, beyond the side effect of sei-
zures caused by carbapenems. Furthermore, the increased
percentage of poor outcomes in our patients probably stems
from the greater percentage of critical patients in the study,
including those with multiple organ failure, neurosurgical
conditions, and status epilepticus. Further studies with large
case numbers are needed to exclude the role of reduced VPA
levels in seizure aggravation.
Most reports suggest that combination therapy with VPA
and carbapenems should be avoided, unless necessary
[8,11]. When simultaneous use is inevitable, other AEDs
should be temporarily added during the coadministration of
VPA and carbapenems. The addition of another AED should
be maintained for at least 2 weeks after the discontinuation
of carbapenems because it takes 1e2 weeks for serum VPA
levels to return to therapeutic concentrations. The effect
of increasing VPA dose, via “oral or intravenous” routes or
“loading or maintenance,” did not help in achieving ther-
apeutic levels. Levetiracetam could be a good choice
because of the low risk of drug interactions and the rapid
titration protocol. Based on our analysis, to prevent further
decreases in VPA levels, phenytoin and other enzyme-
inducing AEDs, including phenobarbital and carbamaze-
pine, were unfavorable. When the dose of VPA was
increased during the interaction period, close monitoring of
VPA concentrations and dose reductions should be consid-
ered within 7e14 days of the discontinuation of the
carbapenems.
In conclusion, this study showed that serum VPA levels
decrease to subtherapeutic concentrations after the
introduction of carbapenems. The reduction of VPA levels
was seen within 24 hours of the initiation of treatment with
either high or low doses of carbapenems. The combination
of carbapenems and VPA should be avoided unless there is
no alternative available. Clinicians should be aware of the
possibility of a reduction in VPA concentrations to sub-
therapeutic levels and the worsening of seizures when
these two drugs are combined. Add-on non-enzyme-
inducing AEDs might be a better choice than enzyme-
inducing AEDs for preventing further decreases in VPA
levels, especially in patients with renal disease.
Acknowledgments
This work was supported in part by research grant
CMRPG8E0791 awarded to Y.C.C. from Kaohsiung Chang
Gung Memorial Hospital, Taiwan.
136 C.-R. Huang et al.

References [22] Nacarkucuk E, Saglam H, Okan M. Meropenem decreases

serum level of valproic acid. Ped Neurol 2004;31:232e4.
[23] Lee SG, Kim JH, Joo JY, Kwon OH. Seven cases of decreased
[1] Ruiz-Gimenez J, Sanchez-Alvarez JC, Canadillas-Hidalgo F,
serum valproic acid concentration during concomitant use of
Serrano-Castro PJ. Antiepileptic treatment in patients with
carbapenem antibiotics. Korean J Lab Med 2007;27:338e43
epilepsy and other comorbidities. Seizure 2010;19:375e82.
[Article in Korean].
[2] Beleza P. Acute symptomatic seizures: a clinically oriented
[24] Lee MC, Sun YH, Lee CH, Wu AJ, Wu TW. Interaction between
review. Neurologist 2012;18:109e19.
valproic acid and carbapenems: cases series and literature
[3] Kim MA, Park KM, Kim SE, Oh MK. Acute symptomatic seizures
review. Tsu Chi Med J 2012;24:80e4.
in CNS infection. Eur J Neurol 2008;15:38e41.
[25] Miranda Herrero MC, Alcaraz Romero AJ, Escudero Vilaplana V,
[4] Johannessen SI, Landmark CJ. Antiepileptic drug inter-
Fernandez Lafever SN, Fernandez-Llamazares CM, Barredo
actionsdprinciples and clinical implications. Curr Neuro-
Valderrama E, et al. Pharmacological interaction between
pharmacol 2010;8:254e67.
valproic acid and carbapenem: what about levels in pediatrics?
[5] Perucca E. Clinically relevant drug interactions with antiepi-
Eur J Paediatr Neurol 2015;19:155e61.
leptic drugs. Br J Clin Pharmacol 2006;61:246e55.
[26] Vélez Dı́az-Pallarés M, Delgado Silveira E, Alvarez Dı́az AM, Pérez
[6] Perucca E. Pharmacological and therapeutic properties of
Menéndez-Conde C, Vicente Oliveros N, Bermejo Vicedo T.
valproate: a summary after 35 years of clinical experience.
Analysis of the valproic acid-meropenem interaction in hospi-
CNS Drugs 2002;16:695e714.
talised patients. Neurologia 2012;27:34e8 [Article in Spanish].
[7] Papp-Wallace KM, Endimiani A, Taracila MA, Bonomo RA.
[27] Cabanes Mariscal MA, Sánchez López P, Alvarez Herranz P,
Carbapenems: past, present, and future. Antimicrob Agents
Chamorro Merino G. Pharmacokinetic interaction between
Chemother 2011;55:4943e60.
valproic acid and ertapenem. Farm Hosp 2006;30:313e5.
[8] Coves-Orts FJ, Borras-Blasco J, Navarro-Ruiz A, Murcia-
[28] Liao FF, Huang YB, Chen CY. Decrease in serum valproic acid
Lopez A, Palacios-Ortega F. Acute seizures due to a probable
levels during treatment with ertapenem. Am J Health Syst
interaction between valproic acid and meropenem. Ann
Pharm 2010;67:1260e4.
Pharmacother 2005;39:533e7.
[29] Nagai K, Shimizu T, Togo A, Takeya M, Yokomizo Y, Sakata Y,
[9] Lunde JL, Nelson RE, Storandt HF. Acute seizures in a patient
et al. Decrease in serum levels of valproic acid during treat-
receiving divalproex sodium after starting ertapenem therapy.
ment with a new carbapenem, panipenem/betamipron. J
Pharmacotherapy 2007;27:1202e5.
Antimicrobial Chemother 1997;39:295e6.
[10] Fudio S, Carcas A, Pinana E, Ortega R. Epileptic seizures
[30] Miller AD, Ball AM, Bookstaver PB, Dornblaser EK, Bennett CL.
caused by low valproic acid levels from an interaction with
Epileptogenic potential of carbapenem agents: mechanism of
meropenem. J Clin Pharm Ther 2006;31:393e6.
action, seizure rates, and clinical considerations. Pharmaco-
[11] Santucci M, Parmeggiani A, Riva R. Seizure worsening caused
therapy 2011;31:408e23.
by decreased serum valproate during meropenem therapy. J
[31] Mori H, Takahashi K, Mizutani T. Interaction between valproic
Child Neurol 2005;20:456e7.
acid and carbapenem antibiotics. Drug Metabol Rev 2007;39:
[12] Clause D, Decleire PY, Vanbinst R, Soyer A, Hantson P. Phar-
647e57.
macokinetic interaction between valproic acid and mer-
[32] Yamamura N, Imura-Miyoshi K, Naganuma H. Panipenum, a
openem. Intensive Care Med 2005;31:1293e4.
carbapenem antibiotic, increases the level of hepatic UDP-
[13] Tobin JK, Golightly LK, Kick SD, Jones MA. Valproic acid-
glucuronic acid in rats. Drug Metab Dispos 2000;28:1484e6.
carbapenem interaction: report of six cases and a review of
[33] Ogawa K, Yumoto R, Hamada N, Nagai J, Takano M. Interac-
the literature. Drug Metabol Drug Interact 2009;24:153e82.
tion of valproic acid and carbapenem antibiotics with multi-
[14] Park MK, Lim KS, Kim TE, Han HK, Yi SJ, Shin KH, et al.
drug resistance-associated proteins in rat erythrocyte
Reduced valproic acid serum concentrations due to drug in-
membranes. Epilepsy Res 2006;71:76e87.
teractions with carbapenem antibiotics: overview of 6 cases.
[34] Torii M, Takiguchi Y, Izumi M, Fukushima T, Yokota M. Carba-
Ther Drug Monit 2012;34:599e603.
penem antibiotics inhibit valproic acid transport in Caco-2 cell
[15] Spriet I, Goyens J, Meersseman W, Wilmer A, Willems L, Van
monolayers. Int J Pharm 2002;233:253e6.
Paesschen W. Interaction between valproate and meropenem:
[35] Ziaja M. Septic encephalopathy. Curr Neurol Neurosci Rep
a retrospective study. Ann Pharmacother 2007;41:1130e6.
2013;13:383.
[16] Haroutiunian S, Ratz Y, Rabinovich B, Adam M, Hoffman A.
[36] Verwaest C. Meropenem versus imipenem/cilastatin as
Valproic acid plasma concentration decreases in a dose-
empirical monotherapy for serious bacterial infections in the
independent manner following administration of meropenem:
intensive care unit. Clin Microbiol Infect 2000;6:294e302.
a retrospective study. J Clin Pharmacol 2009;49:1363e9.
[37] Klugman KP, Dagan R. Randomized comparison of meropenem
[17] Mancl EE, Gidal BE. The effect of carbapenem antibiotics on
with cefotaxime for treatment of bacterial meningitis. Mer-
plasma concentrations of valproic acid. Ann Pharmacother
openem Meningitis Study Group. Antimicrob Agents Chemo-
2009;43:2082e7.
ther 1995;39:1140e6.
[18] Taha FA, Hammond DN, Sheth RD. Seizures from valproate-
[38] Nichols RL, Smith JW, Geckler RW, Wilson SE. Meropenem
carbapenem interaction. Ped Neurol 2013;49:279e81.
versus imipenem/cilastatin in the treatment of hospitalized
[19] Gu J, Huang Y. Effect of concomitant administration of mer-
patients with skin and soft tissue infections. Southern Med J
openem and valproic acid in an elderly Chinese patient. Am J
1995;88:397e404.
Geriatr Pharmacother 2009;7:26e33.
[39] Solomkin JS, Yellin AE, Rotstein OD, Christou NV, Dellinger EP,
[20] De Turck BJ, Diltoer MW, Cornelis PJ, Maes V, Spapen HD,
Tellado JM, et al. Ertapenem versus piperacillin/tazobactam
Camu F, et al. Lowering of plasma valproic acid concentra-
in the treatment of complicated intraabdominal infections:
tions during concomitant therapy with meropenem and ami-
results of a double-blind, randomized comparative phase III
kacin. J Antimicrob Chemother 1998;42:563e4.
trial. Ann Surg 2003;237:235e45.
[21] Llinares Tello F, Bosacoma Ros N, Hernandez Prats C, Climent
[40] Ortiz-Ruiz G, Caballero-Lopez J, Friedland IR, Woods GL,
Grana E, Selva Otaolaurruchi J, Ordovas Baines JP. Pharma-
Carides A. A study evaluating the efficacy, safety, and toler-
cokinetic interaction between valproic acid and carbapenem-
ability of ertapenem versus ceftriaxone for the treatment of
like antibiotics: a discussion of three cases. Farm Hosp 2003;
community-acquired pneumonia in adults. Clin Infect Dis
27:258e63 [Article in Spanish].
2002;34:1076e83.