ARTICLE

Accuracy of Calculated Free Valproate Levels in

Adult Patients With Status Epilepticus
Urs Fisch, MD, PhD, Sira M. Baumann, Saskia Semmlack, MD, Stephan Marsch, MD, PhD, Stephan Rüegg, MD, Correspondence
and Raoul Sutter, MD Dr. Fisch

®
urs.fisch@usb.ch
Neurology 2021;96:e102-e110. doi:10.1212/WNL.0000000000011000

Abstract
Objective
To test the accuracy of an equation in adult patients with status epilepticus that calculates the
free concentration of serum valproic acid (fVPA) from the total concentration of serum valproic
acid (tVPA) and serum albumin.

Methods
All adult patients with status epilepticus who were treated at a Swiss academic medical center
between 2005 and 2018 with concurrent measurements of tVPA, fVPA, and serum albumin
were included. fVPA was categorized as subtherapeutic, therapeutic (5–10 mg/L), or supra-
therapeutic. Agreement was defined as the proportion of measured and calculated fVPA falling
within the same category.

Results
Of 676 patients with status epilepticus, 104 had 506 measurements, with a median of 3
(interquartile range [IQR] 1.5–6.5) per patient. The median tVPA was 43.5 mg/L (27.4–63.6),
with measured fVPA 9.1 mg/L (4.5–14.7) and calculated fVPA 10.1 mg/L (7.0–13.0), re-
spectively. The median deviation of calculated from measured fVPA was −0.8 mg/L (−3.2 to
2.5) with 336 measurements >1 mg/L. While the association between measured and calculated
fVPA was linear (regression coefficient 1.1, 95% confidence interval 0.9–1.2, p < 0.0001), the
agreement on effective drug levels did not match in 39.8% of measurements regardless of serum
albumin levels, with calculated fVPA overestimating measured fVPA in 30.4%. tVPA and serum
albumin independently influenced the accuracy of the calculated fVPA in the multivariable
model.

Conclusions
Calculated fVPA is inaccurate when using the proposed equation in adult patients with status
epilepticus, calling for drug monitoring based on measured fVPA in this context.

From the Department of Neurology (U.F., S.R.) and Clinic for Intensive Care Medicine (S.M.B., S.S., S.M., R.S.), University Hospital Basel; and Medical Faculty of the University of Basel
(S.M., S.R., R.S.), Switzerland.

Go to Neurology.org/N for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.

e102 Copyright © 2020 American Academy of Neurology

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Glossary
ASD = antiseizure drug; BUN = blood urea nitrogen; CI = confidence interval; fVPA = free valproic acid serum concentration;
ICU = intensive care unit; IQR = interquartile range; STEP UP = Status Epilepticus Unicenter Population; tVPA = total
valproic acid serum concentration; VPA = valproic acid.
Status epilepticus is a neurologic emergency with high mor-
bidity and mortality.1–3 After first-line treatment with ben-
zodiazepines, rapid escalation with second-line antiseizure
drugs (ASDs) is recommended.4–6 Of these, valproic acid
(VPA) has been shown to be equally effective as levetiracetam
or fosphenytoin to treat convulsive status epilepticus.7
Drug monitoring ensures optimal ASD treatment. Dose-
dependent toxicities have mostly been described with supra-
therapeutic VPA levels.8,9 Measurement of the non-protein-
bound free VPA serum concentration (fVPA) reflects bi-
ologically active drug levels, but this assay is not readily available
in many institutions. VPA drug monitoring is complicated by
its complex pharmacokinetics with a saturable high binding to
plasma proteins, mainly albumin.8–10 Subsequently, the ratio of
fVPA to total VPA serum concentration (tVPA) is nonlinear,
with a relatively constant free fraction of approximately 10%
until tVPA of 60 mg/L.9,11 In addition, multiple drug–drug
interactions, serum urea, bilirubin, free fatty acids, and renal or
liver insufficiency have been reported to modify fVPA.10,12–14
Due to the limited access to fVPA measurements, indirect
methods for an accurate calculation of fVPA were proposed by
an equation that corrects for hypoalbuminemia.13,14 An accu-
rate assessment of fVPA is essential, as clinicians may in-
appropriately exclude possible VPA toxicity with low to normal
tVPA, or attribute lack of therapeutic effectiveness to low tVPA.
However, data are lacking regarding the accuracy of the
aforementioned equation in critically ill patients with status
epilepticus, of whom many have hypoalbuminemia.15 There-
fore, we investigated the accuracy of calculated fVPA by the
proposed equation in adult patients with status epilepticus.
Methods
This observational study was performed at the University
Hospital of Basel, a Swiss tertiary academic medical care
center. The Strengthening the Reporting of Observational
Studies in Epidemiology (STROBE) guidelines were fol-
lowed to enhance the quality and standardization for the
reporting of observational studies.16
Standard Protocol Approvals, Registrations,
and Patient Consents
The study was approved by the local ethics committee
(Ethikkommission Nordwest-und Zentralschweiz) in com-
pliance with the ethical standards laid down in the 1964
Declaration of Helsinki and its later amendments and waived
patients’ consent accordingly.
Neurology.org/N
Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Data Collection
Data analyzed in this study are part of the Status Epilepticus
Unicenter Population (STEP UP) study (NCT04204863) at the
University Hospital of Basel. The STEP UP study collects data on
patients with status epilepticus. From January 1, 2005, to De-
cember 31, 2018, clinical, laboratory, and epileptologic data of all
consecutive adult patients (≥18 years old) with status epilepticus
were collected. Data were assessed with digital institutional EEG
and medical databases. Patients without concurrent measurements
of fVPA, tVPA, and serum albumin concentration were excluded.
Types, treatment response, severity, and duration of status epi-
lepticus (as described in previous studies17–20) were assessed
from the digital EEG database. Status epilepticus types were
defined as recommended by the current guidelines (i.e., focal
nonconvulsive status epilepticus without coma [with or without
altered consciousness and absence status epilepticus], status
epilepticus with motor symptoms [myoclonic and convulsive
status epilepticus], and nonconvulsive status epilepticus with
coma).21 Status epilepticus severity was quantified by the Status
Epilepticus Severity Score as previously described.22–25
The following clinical characteristics were systematically col-
lected: sex, age, status epilepticus etiology (categorized as po-
tential nonfatal and fatal as previously defined26), comorbidities
quantified by the Charlson Comorbidity Index,27 duration of
mechanical ventilation, ASD and continuous administration of
anesthetics, and length of intensive care unit (ICU) and hos-
pital stay. Complications emerging during status epilepticus
were assessed, including infections, arterial hypotension re-
quiring continuous administration of vasopressors, and renal or
liver insufficiency/failure. Renal insufficiency/failure was de-
fined according to the Kidney Disease Improving Global
Outcomes (KDIGO) guidelines.28 Liver insufficiency/failure
was defined as the presence of clinical signs associated with
severe hepatic dysfunction, such as ascites, hepatic encepha-
lopathy, esophageal varices, or coagulopathy, usually Child-
Pugh score C and advanced Child-Pugh score B.
Outcomes
The primary outcome was defined as the proportion of cal-
culated fVPA measurements in agreement, underestimating
or overestimating the actual measured fVPA. Secondary
outcomes were the association of serum albumin levels, tVPA,
or renal or liver insufficiency with calculated or measured
fVPA.
Statistics
fVPA was categorized as subtherapeutic (<5 μg/L), thera-
peutic (5–10 μg/L), or supratherapeutic (>10 μg/L). For
Neurology | Volume 96, Number 1 | January 5, 2021 e103
 VPA measurements from patients with serum albumin levels
between 18 and 42 g/L (which are considered in the pro-
posed equation),13,14 linear regression models were per-
formed to assess the association between calculated and
measured fVPA and serum albumin levels. Linear regression
models were repeated in subgroups of patients with and
without renal or liver insufficiency. Subsequently, multivar-
iable linear regression models were performed to identify
variables independently associated with the accuracy of
calculated fVPA. For measurements with serum albumin
levels below 18 g/L or above 42 g/L, the free VPA fraction
was calculated using the original equation y = A・e −BX (y =
free VPA fraction; A = 130.69; B = 4.96・10−3; r = −0.82; X
= serum albumin [μmol/L]) by Parent et al.14 Two-sided p
values ≤0.05 were considered significant. Statistical analysis
was performed with STATA 15.1 (Stata Corp., College
Station, TX).
Data Availability
The corresponding author has full access to all of the data in the
study. He takes full responsibility for the integrity of the data, the
accuracy of the data analysis and interpretation, and the conduct
Figure 1 Flow Chart of Patient Inclusion
e104 Neurology | Volume 96, Number 1 | January 5, 2021
Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
of the research. The authors have the right to publish any and all
data, separate and apart from the guidance of any sponsor.
Results
Baseline Characteristics and VPA
Serum Measurements
Of 676 patients with status epilepticus, 106 had measurements of
serum albumin, tVPA, and fVPA available. Two patients without
parallel measurements were removed, resulting in 104 patients
with a total of 506 concurrent measurements of serum albumin,
tVPA, and fVPA (figure 1). Demographic and clinical character-
istics, course, and outcome of the 104 patients are presented in
table 1. The median number of concurrent measurements in
patients with status epilepticus was 3 (interquartile range [IQR]
1.5–6.5). The median serum albumin concentration was 23 g/L
(IQR 19–28). There were 91 albumin measurements less than 18
g/L or more than 42 g/L (i.e., outside the range for which the
equation is claimed to be valid14). Of the 506 measurements, the
median tVPA was 43.5 mg/L (IQR 27.4–63.6) and the median
measured fVPA 9.1 mg/L (IQR 4.5–14.7). In 12 patients, 21
fVPA = free valproate; IQR = interquartile range; SE
= status epilepticus; tVPA = total valproate.
Neurology.org/N
Table 1 Characteristics of Patients with Measurements of
Free and Total Valproic Acid Serum
Concentrations During Status Epilepticus (104
Patients with 506 Measurements)
Demographics and admission characteristics
Table 1 Characteristics of Patients with Measurements of
Free and Total Valproic Acid Serum
Concentrations During Status Epilepticus (104
Patients with 506 Measurements) (continued)
Values Demographics and admission characteristics Values

Age, y 63.2 Complications during status epilepticus

(47–72)
Infections 37 (35.6)
Female 52 (50)
Arterial hypotension requiring vasopressors 14 (13.5)
Prehospital status epilepticus onset 70 (67.3)
Renal insufficiency/failure 16 (15.4)
Clinical features
Liver insufficiency/failure 9 (8.7)
Status epilepticus etiology
Outcomes
Confirmed potentially fatal etiology 42 (40.4)
(not mutually exclusive) No return to premorbid neurologic function at 69 (66.4)
discharge (including death)
Hypoxic-ischemic encephalopathy 17 (16.4)
Death at hospital discharge 22 (21.2)
Infectious meningitis/encephalitis 10 (9.6)
Abbreviations: ICU = intensive care unit; NCSE = nonconvulsive status epi-
Rapidly growing brain tumors 8 (7.7)
lepticus; STESS = Status Epilepticus Severity Score.23,24,45
Values are median (interquartile range) or n (%).
Acute autoimmune encephalitis 5 (4.8) a
Status epilepticus types according to the task force of the International
League Against Epilepsy.21
Acute intracranial hemorrhage 2 (1.9)

Acute ischemic stroke 2 (1.9)

Acute severe traumatic brain injury 1 (1.0) measurements of fVPA were less than 3 mg/L despite tVPA of
more than 20 mg/L, most likely due to preanalytical errors. The
Status epilepticus typea
median calculated fVPA was 10.1 mg/L (IQR 7.0–13.0).
Focal NCSE without coma 45 (43.3)

With altered consciousness 42 (40.4)

Without altered consciousness
Status epilepticus with motor symptoms
(convulsive, myoclonic)
Convulsive status epilepticus
Myoclonic status epilepticus
NCSE with coma
NCSE with coma (nonsubtle)
Subtle status epilepticus
Charlson Comorbidity Index27
STESS23,24,45
Status epilepticus duration, d
Treatment characteristics during status epilepticus
In-hospital treatment, d
ICU treatment, d
Patients with benzodiazepines
Patients with second-line antiseizure drugs
Patients with anesthetic drugs
Mechanical ventilation
Association of Measured and Calculated Serum
Concentrations of VPA
3 (2.9) Measured fVPA was significantly associated with tVPA (re-
20 (19.2) gression coefficient 1.7, 95% confidence interval [CI] 1.5–1.9, p
< 0.0001) with less fitting with increasing fVPA serum concen-
10 (9.6) trations. The results were similar in patients with (regression
coefficient 1.69, 95% CI 1.48–1.90, p < 0.0001) or without renal
10 (9.6)
or liver insufficiency (regression coefficient 2.10, 95% CI
39 (37.5) 1.87–2.33, p < 0.0001) (figure 2A).
29 (27.9)
Measured fVPA was significantly associated with calculated fVPA
10 (9.6) (regression coefficient 1.1, 95% CI 0.9–1.2, p < 0.0001) with a
2 (1–3) median deviation of −0.8 mg/L (IQR −3.2 to 2.5) (figure 2B).
However, 336 (66%) fVPA measurements differed from calcu-
3 (2–4)
lated fVPA by more than 1 mg/L. Most often, the calculated fVPA
2 (1–4) overestimated the measured fVPA (figure 2B).
Factors Influencing the Accuracy of the
19 (11–35) Calculated Serum Concentration of VPA
7 (3–16) The difference between calculated and measured fVPA sig-
nificantly increased with higher albumin levels (regression
102 (98.1)
coefficient 0.3, 95% CI 0.2–0.4, p < 0.0001) in patients with
104 (100.0) and without liver or renal insufficiency (figure 2C), and
57 (54.8)
likewise with higher tVPA (regression coefficient 0.1, 95% CI
0.03–0.1, p < 0.0001) (figure 2D).
66 (63.5)

Duration of mechanical ventilation, d 9 (5–16) For fVPA measurements with corresponding serum albumin
between 18 and 42 g/L, the multivariable linear regression
Patients with vasopressors 14 (13.5)
showed that both tVPA (regression coefficient 0.04, 95% CI

Neurology.org/N Neurology | Volume 96, Number 1 | January 5, 2021 e105

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
 Figure 2 Measured and Calculated Serum Valproate Levels in Patients With Status Epilepticus With Corresponding Linear
Regression Models

Comparison of (A) measured total vs measured free serum valproate levels (smaller plots: with and without renal or liver insufficiency), (B) measured vs
calculated free serum valproate levels, (C) difference between measured and calculated free serum valproate vs serum albumin levels (smaller plots: with and
without renal or liver insufficiency), and (D) difference between measured and calculated free serum valproate vs total serum valproate levels (104 patients
with 506 measurements). All linear regression models were calculated for patients with albumin serum levels of ≥18 g/L and ≤42 g/L according to Hermida
et al.13 CI = confidence interval.

0.01–0.1, p = 0.009) and serum albumin (regression co-
efficient 0.2, 95% CI 0.1–0.4, p < 0.0001) independently
influenced the accuracy of the calculated fVPA, with in-
creasing inaccuracy with higher levels of albumin or tVPA,
respectively.
Agreement Between Therapeutic Levels of
Measured and Calculated Free VPA
Serum Concentrations
When categorizing the fVPA measurements into sub-
therapeutic (<5 mg/L), therapeutic (5–10 mg/L), and
supratherapeutic (>10 mg/L) levels, 305 (60.3%) of 506
measurements showed an agreement between calculated and
measured fVPA, while calculated fVPA underestimated mea-
sured fVPA in 44 (8.7%) cases, and overestimated in 157
(31.0%) cases, respectively (table 2). The equation was vali-
dated for serum albumin levels from 18 to 42 g/L.13 If only
measurements with corresponding serum albumin within this
range were considered, the proportion of agreement between
the categories remained essentially unchanged (table 2).
Carbapenems are known to lower VPA levels.12,29 A total of 4
patients (3.8%) treated with meropenem or imipenem had 13
(2.6%) concurrent fVPA measurements. In 2 measurements, the

e106 Neurology | Volume 96, Number 1 | January 5, 2021 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
Table 2 Measured and Calculated Free Fractions of Valproic Acid (VPA) Serum Levels
In all 506 measurements (including patients with serum albumin <18 g/L or >42 g/L)

Measured total VPA category N %

Supratherapeutic (>100 mg/L) 14 2.8

Therapeutic (50–100 mg/L)a 196 38.7

Subtherapeutic (<50 mg/L) 296 58.5

Calculated fVPA category

Supratherapeutic (>10 mg/L) Therapeutic (5–10 mg/L)b Subtherapeutic (<5 mg/L)

Measured fVPA category N % N % N %

Supratherapeutic (>10 mg/L) 181 35.8 33 6.5 2 0.4

b
Therapeutic (5–10 mg/L) 69 13.6 69 13.6 9 1.8

Subtherapeutic (<5 mg/L) 16 3.2 72 14.2 55 10.9

b
Total agreement or disagreement between measured and calculated fVPA according to categorization N %

Total agreements 305 60.3

Total underestimations 44 8.7

Total overestimations 157 31.0

In 415 measurements with serum albumin ≥18 g/L and ≤42 g/L

Calculated fVPA category

Supratherapeutic (>10 mg/L) Therapeutic (5–10 mg/L)b Subtherapeutic (<5 mg/L)

Measured fVPA category N % N % N %

Supratherapeutic (>10 mg/L) 146 35.2 31 7.5 2 0.5

Therapeutic (5–10 mg/L)b 53 12.8 64 15.4 6 1.4

Subtherapeutic (<5 mg/L) 10 2.4 63 15.2 40 9.6

b
Total agreement or disagreement between measured and calculated fVPA according to categorization N %

Total agreements 250 60.2

Total underestimations 39 9.4

Total overestimations 126 30.4

Abbreviation: fVPA = free valproic acid serum concentration.

a
Therapeutic levels of total VPA serum levels according to Cook et al.46
b
Therapeutic levels of free fraction of VPA according to Parent et al.14 and Hermida and Tutor.13

calculated fVPA underestimated low therapeutic measured fVPA
as subtherapeutic. In the other 11 measurements, calculated and
measured fVPA were both categorized as subtherapeutic.
Discussion
In this retrospective study, we analyzed the accuracy of a pre-
viously published equation to calculate fVPA in adult patients
with status epilepticus. By using this equation, we detected a
considerable discordance between calculated and actual mea-
sured fVPA in these patients. Serum albumin levels and tVPA
independently influenced the accuracy of calculated fVPA with
increasing inaccuracy with higher levels of albumin and tVPA.
The calculation of fVPA based on this equation resulted in a
misinterpretation in 40% of measurements with about 75% of
them overestimating the actual measured fVPA. In our cohort,
treatment decisions based on the calculated fVPA would have
led to a reduction of the administered dosages in up to a third of
patients and supratherapeutic levels would have been missed in
up to 8% as compared to the measured fVPA.
Drug monitoring is advised for VPA due to its narrow thera-
peutic range.12 Dose-dependent toxicities, including CNS
dysfunction, hepatotoxicity, pancreatitis, and thrombocytopenia,
have mainly been described with supratherapeutic VPA serum
levels.8,9 For patients with epilepsy, tVPA from 50 to 100 mg/L
corresponds to effective treatment while minimizing toxicity.8,9
However, seizure control may be achieved at drug concentra-
tions above or below this reference range.9 Current guidelines for
status epilepticus treatment do not specify the therapeutic range

Neurology.org/N Neurology | Volume 96, Number 1 | January 5, 2021 e107

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
 for VPA.4–6 Drug monitoring based on tVPA may be misleading
in many situations, especially in patients with status epilepticus
with hypoalbuminemia.14,30 fVPA is the biologically active form
that crosses the blood–brain barrier and correlates best with
neurologic adverse symptoms.30,31 Therefore, it is the most
suitable candidate for therapeutic VPA monitoring.11,12,32 Be-
cause its measurement is not readily available in many institu-
tions, an equation, based on tVPA and serum albumin levels
between 18 and 42 g/L, had been proposed that predicts fVPA
with an error of ≤2.5 mg/L, if tVPA is less than 75 mg/L and
there is absence of uremia or jaundice.13,14 Our results indicate
that this equation is inaccurate in patients with status epilepticus.
Our findings are in line with recent reports of ICU patients who
demonstrated poor prediction of fVPA by the equation. One
recent study analyzed 174 patients, of whom almost 90% were
hospitalized.33 This study showed a discordance between calcu-
lated and measured fVPA in approximately 43% of measurements
when stratified to subtherapeutic, supratherapeutic, or thera-
peutic levels. The highest proportion of discordance was found in
ICU patients (58%), followed by non-ICU hospitalized patients
(37%) and outpatients (35%). Of all discordant comparisons,
97% of calculated fVPA measurements underestimated the actual
measured fVPA. As a limiting factor, the indication for VPA in this
study remains unknown, subgroup analyses regarding patients
with status epilepticus were lacking, and albumin levels were
drawn up to 7 days of VPA measurements. Another study
compared calculated with measured fVPA in 15 ICU patients, of
whom 6 received VPA for seizures.34 The results demonstrated a
poor correlation of calculated and measured fVPA with an un-
derestimation of mean 11.9 mg/L of calculated fVPA.
fVPA has been reported to be influenced by several endogenous
compounds, including blood urea nitrogen (BUN) or bilirubin.
Remarkably, the inaccuracies of calculated free fractions were not
strikingly different in patients with and without renal or liver
insufficiencies and the analyzed equation was inaccurate re-
gardless of the level of serum albumin.
This is the first study that investigated the accuracy of a proposed
equation for calculating fVPA in a cohort of adult patients with
status epilepticus. However, we also report some limitations. The
single-center observational design limits the generalizability of this
study. The fact, however, that the clinical characteristics in our
population are similar to those in other status epilepticus studies
including age,35–39 outcome,37,40 etiologies,37–40 status epilepticus
severity,37,38 and types of status epilepticus35,40,41 indicate that our
cohort has several characteristics typical of other large status epi-
lepticus cohorts. Our study included all patients regardless of lab-
oratory abnormalities to reflect a more general patient population,
in contrast to Hermida and Tutor,13 where patients with elevated
urea, creatinine, and bilirubin were excluded. Their rationale for
excluding patients with hyperbilirubinemia was its displacement of
VPA from albumin, resulting in free fraction elevation, but this
assumption is contradicted by 2 evaluations indicating that VPA
displaces bilirubin.42,43 In addition, renal insufficiency is described
as potentially influencing free VPA serum levels in only 2 patients
e108 Neurology | Volume 96, Number 1 | January 5, 2021
Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
from the original study and a casual interaction is not proven.14
However, to enhance accuracy, subgroup analyses were performed
in patients with and without renal or liver insufficiency without
significant changes of results. Multiple drug–drug interactions with
VPA are known, including certain ASDs and anti-infectives.12 As in
neurocritical care, special attention is needed for carbapenems,
which lower VPA serum levels.12,29 We identified only a few pa-
tients who received carbapenems during VPA measurements that
did not change the main results of the study. Many endogenous
and exogenous compounds, such as serum free fatty lipids, BUN,
bilirubin, or salicylates, and nonsteroidal anti-inflammatory drugs
were reported to alter the unbound fraction of VPA.12,32,44 Al-
though our study did not aim at identifying all potential causes of
the observed differences between calculated and measured fVPA,
these aspects underscore the complexity of VPA metabolism and
may at least to a certain degree explain the inaccuracy of the
equation examined. We could not analyze the effects regarding
treatment decisions, as this was not the aim of our study and data
assessment. Lastly, the optimal therapeutic range for VPA for status
epilepticus treatment needs to be determined. Despite these lim-
itations, our these results are worrisome, as they uncover a potential
high risk of status epilepticus being undertreated or being treated
with potentially toxic VPA levels.
Our study reveals that in patients with status epilepticus, the
calculated fVPA is inaccurate when using the proposed
equation. Because of the potential critical effect of over-
estimation and the considerable variability of fVPA, we ad-
vocate therapeutic drug monitoring in critically ill patients
with status epilepticus receiving VPA as a second-line ASD to
be based on measured free fractions.
Acknowledgment
The authors thank Sarah Tschudin-Sutter, MD, MSc (University
Hospital Basel), for technical help with the statistical analysis.
Study Funding
From the University Hospital Basel, Switzerland. The funder
(University Hospital Basel) had no role in the design of the study;
in the collection, analyses, or interpretation of data; in the writing
of the manuscript; or in the decision to publish the results. This
study was performed and designed without the input or support of
any pharmaceutical company or other commercial interest.
Disclosure
Urs Fisch, Sira M. Baumann, and Saskia Semmlack report no
disclosures. Stephan Rüegg received unconditional research
grants from UCB Pharma; honoraria from serving on the sci-
entific advisory boards of Arvelle, Desitin, Eisai, GSK, and UCB
Pharma; travel grants from GSK, Janssen-Cilag, and UCB
Pharma; and speaker fees from UCB Pharma and from serving
as a consultant for Eisai, GlaxoSmithKline, Janssen-Cilag,
Pfizer, Novartis, and UCB Pharma. He does not hold any
stocks of any pharmaceutical industries or manufacturers of
medical devices. He received funding from UCB Pharma,
Novartis, and Swiss National Science Foundation grants, grant
320030_169,379/1, and coapplicant for grants 33CM30_
Neurology.org/N
125,115/1 and 33CM30_140,338/1; is the president of the 3. Sutter R, Marsch S, Fuhr P, Rüegg S. Mortality and recovery from refractory
status epilepticus in the ICU: a 7-year observational study. Epilepsia 2013;54:
Swiss League against Epilepsy (no payments); and is Editor of 502–511.
EPILEPTOLOGIE (Journal of the Swiss League against Epi- 4. Brophy GM, Bell R, Claassen J, et al. Guidelines for the evaluation and management of
lepsy) (no payments) and Editor of the Swiss EEG Bulletin 5.
status epilepticus. Neurocrit Care 2012;17:3–23.
Glauser T, Shinnar S, Gloss D, et al. Evidence-based guideline: treatment of con-
(payments from UCB). Stephan Marsch reports no disclosures. vulsive status epilepticus in children and adults: report of the guideline committee of
the American Epilepsy Society. Epilepsy Curr 2016;16:48–61.
Raoul Sutter received research grants from the Swiss National 6. Meierkord H, Boon P, Engelsen B, et al. EFNS guideline on the management of status
Foundation (320030_169,379), the Research Fund of the epilepticus in adults. Eur J Neurol 2010;17:348–355.
7. Kapur J, Elm J, Chamberlain JM, et al. Randomized trial of three anticonvulsant
University Basel, the Scientific Society Basel, and the Gottfried medications for status epilepticus. N Engl J Med 2019;381:2103–2113.
Julia Bangerter-Rhyner Foundation and personal grants from 8. Valproic acid [package insert]. Maple Grove: Upsher-Smith Laboratories; 2016.
UCB Pharma and holds stocks in Novartis and Roche. Go to 9. Pharmacotherapy of the epilepsies. In: Hilal-Dandan R, Brunton LL, eds. Goodman
and Gilman’s Manual of Pharmacology and Therapeutics, 2nd ed. New York:
Neurology.org/N for full disclosures. McGraw-Hill Education; 2016.
10. Levy RH, Schmidt D. Utility of free level monitoring of antiepileptic drugs. Epilepsia
1985;26:199–205.
Publication History 11. Nasreddine W, Dirani M, Atweh S, Makki A, Beydoun A. Determinants of free serum
Received by Neurology April 27, 2020. Accepted in final form valproate concentration: a prospective study in patients on divalproex sodium
monotherapy. Seizure 2018;59:24–27.
August 20, 2020. 12. Patsalos PN, Spencer EP, Berry DJ. Therapeutic drug monitoring of antiepileptic
drugs in epilepsy: a 2018 update. Ther Drug Monit 2018;40:526–548.
13. Hermida J, Tutor JC. A theoretical method for normalizing total serum valproic
acid concentration in hypoalbuminemic patients. J Pharmacol Sci 2005;97:
Appendix 1 Authors 489–493.
14. Parent X, Marzullo C, Gutbub AM. [Valproic acid: a simple method for the estimation
Name Location Contribution of free serum concentration.] Ann Biol Clin 1993;51:649–650.
15. Sutter R, Grize L, Fuhr P, Rüegg S, Marsch S. Acute phase proteins and mortality in
Urs Fisch, University Hospital Basel, Major role in the acquisition status epilepticus: a five-year observational cohort study. Crit Care Med 2013;41:
MD, PhD Switzerland of data, designed and 1526–1533.
conceptualized study, 16. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Ob-
interpreted the data, drafted servational Studies in Epidemiology (STROBE) statement: guidelines for reporting
parts of the manuscript observational studies. Lancet 2007;370:1453–1457.
17. Sutter R, Semmlack S, Opic P, et al. Untangling operational failures of the Status
Sira M. University Hospital Basel, Major role in the acquisition of Epilepticus Severity Score (STESS). Neurology 2019;92:e1948–e1956.
Baumann Switzerland data, interpreted the data, 18. Sutter R, Semmlack S, Kaplan PW, Opic P, Marsch S, Ruegg S. Prolonged status
revised parts of the epilepticus: early recognition and prediction of full recovery in a 12-year cohort.
manuscript Epilepsia 2019;60:42–52.
19. Semmlack S, Kaplan PW, Spiegel R, et al. Illness severity scoring in status epi-
Saskia University Hospital Basel, Major role in the acquisition of lepticus: when STESS meets APACHE II, SAPS II, and SOFA. Epilepsia 2019;60:
Semmlack, Switzerland data, interpreted the data, 189–200.
MD revised parts of the 20. Rybitschka A, Semmlack S, Kaplan PW, et al. Calorie intake during status epilepticus
manuscript for intellectual and outcome: a 5-year cohort study. Crit Care Med 2019;47:1106–1115.
content 21. Trinka E, Cock H, Hesdorffer D, et al. A definition and classification of status epi-
lepticus: report of the ILAE task force on classification of status epilepticus. Epilepsia
Stephan University Hospital Basel Interpreted the data, 2015;56:1515–1523.
Marsch, and medical faculty of the revised the manuscript for 22. Sutter R, Marsch S, Fuhr P, Kaplan PW, Rüegg S. Anesthetic drugs in status epi-
MD, DPhil University of Basel, intellectual content lepticus: risk or rescue? A six-year cohort study. Neurology 2014;82:656–664.
Switzerland 23. Rossetti AO, Logroscino G, Bromfield EB. A clinical score for prognosis of status
epilepticus in adults. Neurology 2006;66:1736–1738.
Stephan University Hospital Basel Interpreted the data, 24. Rossetti AO, Logroscino G, Milligan TA, Michaelides C, Ruffieux C,
Rüegg, MD and medical faculty of the revised the manuscript for Bromfield EB. Status Epilepticus Severity Score (STESS): a tool to orient
University of Basel, intellectual content early treatment strategy. J Neurol 2008;255:1561–1566.
Switzerland 25. Sutter R, Kaplan PW, Ruegg S. Independent external validation of the status epi-
lepticus severity score. Crit Care Med 2013;41:e475–479.
Raoul University Hospital Basel Major role in the acquisition 26. Rossetti AO, Hurwitz S, Logroscino G, Bromfield EB. Prognosis of status epilepticus:
Sutter, MD and medical faculty of the of data, designed and role of aetiology, age, and consciousness impairment at presentation. J Neurol
University of Basel, conceptualized study, Neurosurg Psychiatry 2006;77:611–615.
Switzerland analyzed the data, drafted 27. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic
the manuscript for comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–383.
intellectual content 28. Available at: kdigo.org/. Accessed July 12, 2020.
29. Al-Quteimat O, Laila A. Valproate interaction with carbapenems: review and rec-
ommendations. Hosp Pharm 2020;55:181–187.
30. Dore M, San Juan AE, Frenette AJ, Williamson D. Clinical importance of monitoring unbound
valproic acid concentration in patients with hypoalbuminemia. Pharmacotherapy 2017;37:
900–907.
Appendix 2 Coinvestigators 31. Rapeport WG, Mendelow AD, French G, et al. Plasma protein-binding and CSF concentrations
of valproic acid in man following acute oral dosing. Br J Clin Pharmacol 1983;16:365–369.
Sarah University Hospital Contributor Provided 32. Dasgupta A. Usefulness of monitoring free (unbound) concentrations of therapeutic
Tschudin- Basel and medical technical help drugs in patient management. Clin Chim Acta 2007;377:1–13.
Sutter, faculty of the University for the 33. Drisaldi A, Weeda E, Neyens R, et al. Accuracy of valproic acid concentration cor-
MD, MSc of Basel, Switzerland statistical rection based on serum albumin. Neurocrit Care 2019;30:301–306.
analysis 34. Riker RR, Gagnon DJ, Hatton C, et al. Valproate protein binding is highly variable in ICU patients
and not predicted by total serum concentrations: a case series and literature review. Pharmacotherapy
2017;37:500–508.
35. Novy J, Logroscino G, Rossetti AO. Refractory status epilepticus: a prospective ob-
servational study. Epilepsia 2010;51:251–256.
References 36. Agan K, Afsar N, Midi I, Us O, Aktan S, Aykut-Bingol C. Predictors of refractoriness in
1. Sutter R, Kaplan PW, Rüegg S. Outcome predictors for status epilepticus: what really a Turkish status epilepticus data bank. Epilepsy Behav 2009;14:651–654.
counts. Nat Rev Neurol 2013;9:525–534. 37. Fatuzzo D, Novy J, Rossetti AO. Use of newer antiepileptic drugs and prognosis in
2. Sutter R, Semmlack S, Kaplan PW. Nonconvulsive status epilepticus in adults: insights adults with status epilepticus: comparison between 2009 and 2017. Epilepsia 2018;59:
into the invisible. Nat Rev Neurol 2016;12:281–293. e98–e102.

Neurology.org/N Neurology | Volume 96, Number 1 | January 5, 2021 e109

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.
 38. Beuchat I, Novy J, Rossetti AO. Newer antiepileptic drugs in status epilepticus: prescription
trends and outcomes in comparison with traditional agents. CNS Drugs 2017;31:327–334.
39. Strzelczyk A, Ansorge S, Hapfelmeier J, Bonthapally V, Erder MH, Rosenow F. Costs,
length of stay, and mortality of super-refractory status epilepticus: a population-based
study from Germany. Epilepsia 2017;58:1533–1541.
40. Knake S, Rosenow F, Vescovi M, et al. Incidence of status epilepticus in adults in
Germany: a prospective, population-based study. Epilepsia 2001;42:714–718.
41. Agan K, Afsar N, Midi I, Us O, Balosso S, Aykut-Bingol C. Predictors of refractoriness
in a Turkish status epilepticus data bank. Epilepsy Behav 2009;14:651–654.
42. Gough H, Goggin T, Crowley M, Callaghan N. Serum bilirubin levels with antiepi-
leptic drugs. Epilepsia 1989;30:597–602.
43. Jankovic S. Increase in bilirubin plasma concentration in patients taking sodium
valproate. Pharmacology 2005;1:196–197.
44. Levy RH, Koch KM. Drug interactions with valproic acid. Drugs 1982;24:543–556.
45. Sutter R, Kaplan PW, Rüegg S. Independent external validation of the status epi-
lepticus severity score. Crit Care Med 2013;41:e475–479.
46. Cook AM, Zafar MS, Mathias S, et al. Pharmacokinetics and clinical utility of valproic
acid administered via continuous infusion. CNS Drugs 2016;30:71–77.

e110 Neurology | Volume 96, Number 1 | January 5, 2021 Neurology.org/N

Copyright © 2020 American Academy of Neurology. Unauthorized reproduction of this article is prohibited.