CLINICAL INVESTIGATION
Propofol Pharmacodynamics and Bispectral Index During
Key Moments of Awake Craniotomy
Martin Soehle, MD, PhD, MHBA,* Christina F. Wolf, MD,w Melanie J. Priston, PhD,z
Georg Neuloh, MD,y Christian G. Bien, MD,8z Andreas Hoeft, MD, PhD,* and
Downloaded from https://journals.lww.com/jnsa by +Y1bMDlExgYVjFR5GS+Q70ml9n5kHUB1DNh07x0GIPo2/XSOTDUY1cv/VGpjJSjihKf+zdBOF5jud9+NfEOQyy/QvgS7d6EaXr0iSjcYiu4Zl/8y2L7eTS4fIRgGXEcd7maeyrds8ds= on 06/10/2021
Richard K. Ellerkmann, MD, PhD*
Background: During awake craniotomy, the patient’s language
centers are identified by neurological testing requiring a fully
awake and cooperative patient. Hence, anesthesia aims for an
unconscious patient at the beginning and end of surgery but an
awake and responsive patient in between. We investigated the
plasma (Cplasma) and effect-site (Ceffect-site) propofol concen-
tration as well as the related Bispectral Index (BIS) required for
intraoperative return of consciousness and begin of neurological
testing.
Materials and Methods: In 13 patients, arterial Cplasma were
measured by high-pressure liquid chromatography and Ceffect-site
was estimated based on the Marsh and Schnider pharmacoki-
netic/dynamic (pk/pd) models. The BIS, Cplasma and Ceffect-site
were compared during the intraoperative awakening period at
designated time points such as return of consciousness and start
of the Boston Naming Test (neurological test).
Received for publication April 2, 2016; accepted August 30, 2016.
From the Departments of *Anesthesiology and Intensive Care Medicine;
8Epileptology, University of Bonn, Bonn; wDepartment of
Anesthesiology and Intensive Care Medicine, Sana Clinic Berlin-
Lichtenberg, Berlin; yDepartment of Neurosurgery, University of
Aachen, Aachen; zEpilepsy Centre Bethel, Mara Hospital, Bielefeld,
Germany; and zC3P Analysis, Pharmacy Department, Plymouth
Hospitals NHS Trust, Plymouth, UK.
M.S. and C.F.W. contributed equally.
Supported by departmental sources. Covidien plc supplied compli-
mentary electroencephalographic electrodes for study purposes.
Data presented here are part of the doctoral thesis of C.F.W.
M.S. and R.K.E. have received honoraria for lectures from Covidien
Germany (Neustadt/Donau, Germany). C.G.B. gave scientific advice
to Eisai (Frankfurt, Germany) and UCB (Monheim, Germany),
undertook industry-funded travel with support of Eisai (Frankfurt,
Germany), UCB (Monheim, Germany), Desitin (Hamburg,
Germany), and Grifols (Frankfurt, Germany), obtained honoraria
for speaking engagements from Eisai (Frankfurt, Germany), UCB
(Monheim, Germany), Desitin (Hamburg, Germany), diamed (Köln,
Germany), Fresenius Medical Care (Bad Homburg, Germany), and
Biogen (Ismaning, Germany) received research support from diamed
(Köln, Germany) and Fresenius Medical Care (Bad Homburg,
Germany). C.G.B. is a consultant to the Laboratory Krone, Bad
Salzuflen, Germany, regarding neural antibodies. The remaining
authors have no conflicts of interest to disclose.
Address correspondence to: Martin Soehle, MD, PhD, MHBA,
Department of Anesthesiology and Intensive Care Medicine, Uni-
versity of Bonn Hospital, Sigmund-Freud-Str. 25, D-53105 Bonn,
Germany (e-mail: martin.soehle@ukb.uni-bonn.de).
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
DOI: 10.1097/ANA.0000000000000378
32 | www.jnsa.com
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
Results: Return of consciousness occurred at a BIS of 77 ± 7
(mean ± SD) and a measured Cplasma of 1.2 ± 0.4 mg/mL. The
Marsh model predicted a significantly (P < 0.001) higher Cplasma
of 1.9 ± 0.4 mg/mL as compared with the Schnider model
(Cplasma = 1.4 ± 0.4 mg/mL) at return of consciousness. Neuro-
logical testing was possible as soon as the BIS had increased to
92 ± 6 and measured Cplasma had decreased to 0.8 ± 0.3 mg/mL.
This translated into a time delay of 23 ± 12 minutes between
return of consciousness and begin of neurological testing.
At begin of neurological testing, Cplasma according to Marsh
(Cplasma = 1.3 ± 0.5 mg/mL) was significantly (P = 0.002)
higher as compared with the Schnider model (Cplasma = 1.0 ±
0.4 mg/mL).
Conclusions: To perform intraoperative neurological testing,
patients are required to be fully awake with plasma propofol
concentrations as low as 0.8 mg/mL. Following our clinical set-
up, the Schnider pk/pd model estimates propofol concentrations
significantly more accurate as compared with the Marsh model
at this neurologically crucial time point.
Key Words: awake craniotomy, propofol, pharmacokinetics,
pharmacodynamics, target-controlled infusion, depth of anes-
thesia, bispectral index
(J Neurosurg Anesthesiol 2018;30:32–38)
R esections of brain tumors and pharmacotherapy re-
sistant epileptogenic areas are often performed using
the approach of an awake craniotomy.1–4 This means
that the patient is asleep during the beginning and the end
of the surgery but is fully conscious and alert during parts
of the operation to undergo neurological testing.5,6
Usually, propofol is used as the hypnotic agent and
is applied via continuous infusion. To manage the in-
fusion rates, anesthesiologists can either depend on
measures of anesthetic depths such as electroencephalo-
gram (EEG)-derived parameters (eg, Bispectral Index
[BIS], Narcotrend, Patient State Index) or use target-
controlled infusion (TCI) pumps which alter propofol
infusion rates to achieve certain plasma (Cplasma) or effect-
site (Ceffect-site) concentrations based on pharmacokinetic/
dynamic (pk/pd) models.7 At least 6 pk/pd models have
been published for propofol application in adults,8–14 of
which the pk/pd models proposed by Marsh et al9 and
Schnider et al10 are broadly used. As the different models
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
vary greatly in their estimations of propofol concen-
trations for the same patient, it is important to know
which model is implemented in the TCI pump at use.
Recently, we compared the predictive performance
of the Marsh and the Schnider pk/pd models during
awake craniotomy, and observed a significantly higher
accuracy and trend toward a lower bias when applying
the Schnider model.15
Intraoperative language center identification re-
quires an awake, alert, and responsive patient who is able
to perform a dedicated neurological test. As the patient is
anesthetized during the preceding period of surgery, it
takes time until propofol redistributes and propofol
effect-site concentration is low enough to permit neuro-
logical testing. The aim of this study was to determine the
time interval between propofol discontinuation, intra-
operative return of consciousness, and begin of neuro-
logical testing, respectively. In addition, we measured
propofol plasma concentrations as well as anesthetic
depth at the time points “propofol discontinuation,”
“return of consciousness,” and “begin of neurological
testing,” and compared it with propofol plasma concen-
trations and effect-site concentrations estimated accord-
ing to the 3 most frequently used pk/pd models
implemented in TCI pumps.
MATERIALS AND METHODS
Patients and Procedure
Ethical approval for this study (Ethical Committee
No. 226/08) was provided by the Ethical Committee of
the Bonn University Hospital, Bonn, Germany (Chair-
man: Prof. K. Racké). After obtaining written informed
consent from the study participants, patients undergoing
awake craniotomy were investigated in a prospective
observational study. Pregnancy or an age below 18 years
were exclusion criteria. Pharmacokinetic results related to
the prediction error of the Marsh and the Schnider model
were obtained from the same patients and published re-
cently,15 whereas pharmacodynamic data related to the
time points “propofol discontinuation,” “return of con-
sciousness,” and “begin of neurological testing” are pre-
sented here. All patients were anaesthetized using the
“asleep-awake-asleep” technique.16
Upon arrival in the induction room, standard
monitoring was applied and a peripheral venous cannula
as well as a radial artery catheter were inserted. The
electroencephalographic effect of propofol was quantified
using the BIS. The BIS ranges between 0 and 100, in-
dicating an isoelectric EEG and full wakefulness, re-
spectively.17 Therefore, decreasing the depth of
anesthesia, as performed for intraoperative awakening, is
associated with increasing BIS values. The BIS-electrode
(XP-Sensor) was attached to the right forehead and
connected to an A-2000 BIS monitor (Version XP, BIS
version 4.0; Covidien plc, Dublin, Ireland). The propofol
infusion was started at a rate of 2000 mg/h until the BIS
dropped to the desired range between 40 and 60. There-
after, the infusion rate was reduced stepwise to keep the
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
Propofol PD and BIS During Awake Craniotomy
patient in the recommended BIS range. A Portex tube was
inserted through the nose, with the tip above the vocal
cords and the glottis. Therefore, the patients’ ability to
speak intraoperatively was not impeded. However, the
trachea could be intubated by advancing the tube under
fiberoptic guidance if necessary. Oxygen (4 L/min) was
applied by the nasal tube and spontaneous breathing was
maintained throughout surgery.
The patients were transferred to the operation
room, where a scalp block was performed.18 In addition,
the skin was infiltrated with local anesthetics at the site of
the Mayfield pins and skin incision. The neurosurgeons
fixed the patient’s head in a Mayfield holder and per-
formed the craniotomy. After opening of the dura, the
propofol infusion was discontinued which defined the
time point “propofol discontinuation.” Patients were al-
lowed to wake up and “return of consciousness” was
defined as response to verbal command. To do so, pa-
tients were asked every 20 to 30 seconds to open their
eyes. In terms of neurological testing, the Boston naming
test was applied19 being a measure of confrontation
naming. This consists of 60 black and white drawings.
The test was started (time point “begin of neurological
testing”) as soon as the patients were alert enough to
watch the drawings on a screen and to name them. If a
patient developed naming difficulties during electric
stimulation of a brain region of interest, this area was
considered as eloquent and was not removed during
surgery. After completion of the test, propofol infusion
was resumed and tailored to keep the BIS in the men-
tioned range between 40 and 60. After the skin suture was
completed, propofol infusion was stopped again and after
emergence from anesthesia, the nasal tube was removed
while still in the operating room. Subsequently, patients
were transferred to the neurosurgical intensive care unit.
Patients received remifentanil only temporarily
during the painful periods of Mayfield fixation, skin in-
cision and skin suture using a TCI system. To do so, the
pk/pd model described by Minto et al20 was applied and a
low-target effect-site concentration of 0.8 ng/mL was set,
which permitted spontaneous breathing. To assess any
potential interaction between propofol and remifentanil,
we estimated the remifentanil effect-site concentration at
the 3 investigated time points. To do so, the above-
mentioned remifentanil pk/pd model was applied.20
Throughout the operation, BIS values and the
propofol infusion rates were recorded and stored on a
laptop computer every 5 seconds using the Win Log-
Software (version 1.0, Aspect Medical Systems) and TCI
Bonn, an in-house build software, respectively. In addi-
tion, crucial events during surgery such as propofol
discontinuation, return of consciousness or begin of
neurological testing were manually recorded and later
paired with BIS, propofol plasma concentrations and
propofol effect-site concentrations at the time of occur-
rence. If a patient had to undergo 2 testing periods re-
sulting in 2 values of BIS, propofol Cplasma and Ceffect-site
for the time points “intraoperative return of conscious-
ness” and “begin of neurological testing,” the mean value
www.jnsa.com | 33
Soehle et al
was calculated. Pharmacokinetic and dynamic analysis
was performed offline.
Pharmacokinetic and Dynamic Analysis
Propofol Plasma Concentration (Cplasma)
The propofol plasma concentration (Cplasma) was
measured by high-pressure liquid chromatography in a
certified laboratory (C3P analysis, Pharmacy Department,
Plymouth Hospitals NHS Trust, UK). To do so, arterial
blood samples were drawn intraoperatively at propofol
discontinuation, return of consciousness, and begin of
neurological testing amongst others. Plasma was sepa-
rated by centrifugation and stored at  201C until anal-
ysis. In addition, propofol plasma concentration was
simulated in 5-second intervals based on the pk/pd models
of Marsh et al9 and Schnider et al10 according to the re-
corded propofol infusion rates.21 Calculations were per-
formed in an Excel spreadsheet called “excelsior.xls”
developed by Bruhn and Bouillon22 and validated in
comparison with the Tivatrainer-software (version 9.1,
GuttaBV, available at http://www.eurosiva.eu/tivatrainer/
TTweb/TTdownload.html).
Propofol Effect-Site Concentration (Ceffect-site)
Propofol effect-site concentration was calculated by
simultaneous pk/pd modeling as described in detail else-
where23 based on the following differential equation:



dCeffect-site /dt ¼ Cplasma  Ceffect-site ke0 ;
ke0 denotes the equilibration constant and determines the
rate by which propofol is distributed between the plasma
and effect compartment. In this study, 2 different ap-
proaches of how to choose appropriate ke0 values were
applied: Firstly, predefined (published) ke0 values of 0.26/
minutes (Marsh et al9) and 0.456/min (Schnider et al24) as
implemented in the Diprifusor and Braun Space TCI,
respectively were used. Secondly, a fixed “time-to-peak”
approach was applied as described in detail by Minto
et al.25 Here, ke0 was calculated so that the maximum
propofol effect will occur at a predefined time-to-peak
(tpeak) of 1.6 minutes. This approach is implemented in
the Asena PK TCI.
Statistics
SigmaPlot for Windows version 12.3 (Systat Soft-
ware Inc., Germany) was used for the statistical analysis.
In case of normal distribution, the parameter sets were
compared applying a paired t test, otherwise Wilcoxon
rank sum test was performed. A statistical significance
was assumed when Pr0.05. All data are displayed as
mean value ± SD unless mentioned otherwise.
RESULTS
Thirteen patients were included, whose character-
istics are shown in Table 1. Patients received remifentanil
for a duration of 35 minutes (median, 25% quartile =
22 min, 75% quartile = 39 min) during the painful period
of the first asleep period. All patients regained con-
34 | www.jnsa.com
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
TABLE 1. Patient Characteristics
No. patients 13
Sex 7 women, 6 men
Age (y) 43 ± 15
Weight (kg) 75 ± 11
Height (cm) 174 ± 6
Body mass index 24.9 ± 3.2
Duration of surgery 253 ± 42
(min)
Duration of awake 107 ± 32
period (min)
Brain pathology Therapy resistant epilepsy (focal cortical
dysplasia, n = 3)
Brain tumor (glioma, n = 10):
Astrocytoma or oligodendroglioma
(WHO grade II, n = 3)
Astrocytoma or oligoastrocytoma
(WHO grade III, n = 5)
Glioblastoma (WHO grade IV, n = 2)
Data are shown as mean ± SD.
WHO indicates World Health Organization.
sciousness and were able to undergo neurological testing
during the awake period.
Three patients had to be tested twice and therefore
had 2 periods of intraoperative awakening. Table 2
and Figures 1–3 show BIS, propofol plasma concen-
tration and propofol effect-site concentration at the time
points “propofol discontinuation,” “return of conscious-
ness,” and “begin of neurological testing” for the pk/pd
models of the Diprifusor (Marsh), Braun Space (Schnider),
and Asena PK (Schnider tpeak) TCI.
Propofol Discontinuation
When the propofol infusion was stopped, patients
were asleep with a BIS of 40 ± 12 and a measured Cplasma
of 2.5 ± 0.8 mg/mL (Table 2). Cplasma was significantly
(P < 0.01) overestimated by both the Marsh (Cplasma =
3.8 ± 0.7 mg/mL) and the Schnider model (Cplasma = 3.5
± 1.0 mg/mL, Fig. 1). In contrast, Ceffect-site did not differ
between the models (Fig. 1).
The remifentanil infusion was stopped 26 minutes
(median, 25% quartile = 17 min, 75% quartile = 63 min)
before. A remaining median remifentanil effect-site con-
centration of 0.05 ng/mL (25% quartile = 0 ng/mL, 75%
quartile = 0.13 ng/mL) was calculated for the time point
of propofol discontinuation.
Intraoperative Return of Consciousness
Patients regained consciousness 10.5 ± 3.2 minutes
after stopping the propofol infusion. At return of con-
sciousness, a BIS of 77 ± 7 as well as a Cplasma of
1.2 ± 0.4 mg/mL were measured. The Schnider model
predicted a similar Cplasma of 1.4 ± 0.4 mg/mL, whereas
the Marsh model predicted a significantly (P < 0.001)
higher Cplasma of 1.9 ± 0.4 mg/mL at “return of con-
sciousness” (Table 2 and Fig. 2). The propofol effect-
site concentration according to the Marsh model
(Ceffect-site = 2.3 ± 0.6 mg/mL) was significantly (P < 0.001)
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018 Propofol PD and BIS During Awake Craniotomy

TABLE 2. Propofol Concentration and EEG Effect According to the Marsh and Schnider Model at 3 Time points
Discontinuation of Propofol Return of Begin of Neurologic
Parameters Infusion Consciousness Testing
BIS 39.5 ± 12.1 76.6 ± 7.1 92.4 ± 5.7
Cpl measured (mg/mL) 2.47 ± 0.76 1.15 ± 0.40 0.84 ± 0.27
Cpl Schnider (mg/mL) 3.47 ± 0.95 1.39 ± 0.41 1.01 ± 0.43
Cpl Marsh (implemented in Diprifusor) (mg/mL) 3.82 ± 0.69 1.88 ± 0.44 1.33 ± 0.52
Ce Schnider (implemented in. Braun Space) (mg/mL) 3.83 ± 1.07 1.52 ± 0.44 1.06 ± 0.48
Ce Schnider tpeak (implemented in Asena PK) (mg/mL) 3.84 ± 1.07 1.58 ± 0.49 1.08 ± 0.50
Ce Marsh (mg/mL) 4.11 ± 0.74 2.27 ± 0.60 1.51 ± 0.69
Data are shown as mean ± SD.
BIS indicates Bispectral Index; Ce, effect-site concentration of propofol; Cpl, plasma concentration of propofol; tpeak, modified pharmacodynamic model to obtain a
fixed time to peak effect of 1.6 minutes.

higher as compared with the Schnider model (Ceffect-site = DISCUSSION

1.5 ± 0.4 mg/mL) and the Schnider tpeak model (Ceffect-site =
1.6 ± 0.5 mg/mL).
Intraoperative Begin of Neurological Testing
Patients were able to perform the Boston Naming
Test 23 ± 12 minutes after cessation of the propofol in-
fusion. At begin of neurological testing, a BIS of 92 ± 6
and a Cplasma of 0.8 ± 0.3 mg/mL were measured. The
Marsh model predicted a significantly (P = 0.002) higher
Cplasma of 1.3 ± 0.5 mg/mL as well as a significantly
(P < 0.001) higher Ceffect-site of 1.5 ± 0.7 mg/mL as com-
pared with the Schnider model (Cplasma = 1.0 ± 0.4 mg/
mL, Ceffect-site = 1.1 ± 0.5 mg/mL, Ceffect-site tpeak = 1.1 ±
0.5 mg/mL, Table 2 and Fig. 3).
To our knowledge, this is the first study that ac-
tually measured Cplasma (1.2 ± 0.4 mg/mL) at intra-
operative return of consciousness during awake
craniotomy. The Schnider pk model predicted a similar
Cplasma (1.4 ± 0.4 mg/mL), whereas the Marsh pk model
significantly overestimated Cplasma (1.9 ± 0.4 mg/mL) at
the time point of “return of consciousness.” This differ-
ence can be explained based on pk/pd parameters: In
comparison with the Marsh model, Schnider and col-
leagues postulate an B3-fold smaller central compart-
ment volume V1 (Table 3), resulting in a faster increase in
Cplasma during the onset of anesthesia. However after a
while, this effect vanishes as propofol distributes to the
peripheral compartments V2 and V3, which is determined
by the micro rate constants k12 and k13.26 Both k12 and k13
are much higher for the Schnider than the Marsh model

FIGURE 1. Plasma (left side) and effect-site (right side) con-

centration of propofol at the time point when the propofol
infusion was stopped intraoperatively. The boxplots denote
the propofol concentration as actually measured (white), and
as estimated according to the Marsh (dark grey) or Schnider
(grey) model. The modified Schnider pd model, which applies
a modified ke0 to achieve a fixed time to peak effect is in-
dicated as “Schnider tpeak.”
FIGURE 2. Propofol concentrations as measured in plasma
(white), and as calculated based on the Marsh (dark grey) or
Schnider (grey) pk/pd model. Concentrations were de-
termined in plasma (left side) and effect-site (right side) at
return of consciousness. “Schnider tpeak” denotes the modified
Schnider pd model with a fixed time to peak of 1.6 minutes.

Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved. www.jnsa.com | 35

Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.

Soehle et al
FIGURE 3. Plasma (left side) and effect-site (right side) con-
centration of propofol at the time point when the Boston
Naming Test was started. The boxplots denote the propofol
concentration as actually measured (white), and as estimated
according to the Marsh (dark grey) or Schnider (grey) model.
The modified Schnider pd model, which applies a modified ke0
to achieve a fixed time to peak effect is indicated as “Schnider
tpeak.”
(Table 3), therefore a higher amount of propofol dis-
tributes to V2 and V3, resulting in a lower (yet non-
significant) Cplasma for the Schnider model at the time
point at which the propofol infusion stops (propofol
discontinuation). Subsequently, propofol is eliminated as
determined by k10 and redistributed according to k21 and
k31. Cplasma decreases faster according to the Schnider
model (Fig. 4), since it postulates an B4-fold higher
elimination constant k10 (Table 3), resulting in significant
lower Cplasma at return of consciousness and begin of
neurological testing in comparison with the Marsh model
(Figs. 2–4). In addition, the Schnider model is charac-
terized by an almost 2-fold higher equilibration constant
ke0, resulting in a faster equilibration between plasma and
effect-site as well as a faster decrease in Ceffect-site (Fig. 4)
during periods of decreasing plasma concentrations (and
vice versa). Hence a significantly lower Ceffect-site is esti-
mated at return of consciousness and begin of neuro-
logical testing by the Schnider versus the Marsh model.
TABLE 3. Central Compartment Volume and Microconstants
According to the Marsh and Schnider Model for the Example
of a 54-Year-Old Woman (Weight = 64 kg, Height = 176 cm)
Parameters Marsh Model Schnider Model
V1 (L) 14.6 4.3
k10 (L/min) 0.119 0.459
k12 (L/min) 0.114 0.297
k21 (L/min) 0.055 0.068
k13 (L/min) 0.0419 0.1958
k31 (L/min) 0.0033 0.0035
ke0 (L/min) 0.26 0.456
36 | www.jnsa.com
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
FIGURE 4. Example of the propofol concentrations obtained
from a 54-year-old woman (weight = 64 kg, height = 176 cm)
following dura opening during awake craniotomy. Time
points of note are marked as “propofol discontinuation,”
“return of consciousness,” and “begin of neurological
testing.” Propofol concentrations in plasma (solid line) or
effect-site (dashed line) as predicted by the Marsh and
Schnider model are shown in black and grey, respectively.
Time points at which plasma samples were drawn are in-
dicated by a black cross.
Sahinovic et al27 measured a lower median Cplasma
of 0.7 mg/mL (interquartile range [IQR], 0.5 to 0.8 mg/mL)
at postoperative return of consciousness in neurosurgical
tumor patients and predicted a lower Cplasma according to
the Schnider pk model of 0.9 mg/mL (IQR, 0.8 to 1.1 mg/
mL). Lobo and Beiras28 reported a higher estimated
Ceffect-site Schnider at return of consciousness of 2.0 ±
0.1 mg/mL and a higher BIS (85 ± 2) as compared with
our data (Ceffect-site Schnider = 1.5 ± 0.4 mg/mL, BIS = 77
± 7) despite the same surgery, endpoint (eye opening and
following orders) and pk/pd model (Schnider) used. In
our study, patients regained consciousness intra-
operatively 11 ± 3 minutes after propofol discontinuation
which is similar to the median of 9 minutes (IQR range, 6
to 13 min, n = 98) reported by Keifer et al,29 and in be-
tween the 4.7 and 15 minutes observed by Lobo and
Beiras28 and Olsen,30 respectively.
Obviously, regaining consciousness is not enough to
successfully perform a dedicated neurological test such as
the Boston naming test. In fact, patients need to be alert
enough to watch the screen, to concentrate on the
drawing, to recognize and to name the presented item.
This explains, why further time subsequent to return of
consciousness is required for the propofol concentration
to decrease to values as low as measured Cplasma of
0.8 ± 0.3 mg/mL, and an estimated Ceffect-site of
1.5 ± 0.7 mg/mL (Marsh model) or 1.1 ± 0.5 mg/mL
(Schnider model). Consequently, the BIS increased from
BIS = 77 ± 7 (return of consciousness) to BIS = 92 ± 6
(begin of neurological testing).
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
As Table 2 demonstrates, calculated plasma and
effect-site concentrations at return of consciousness and
begin of neurological testing varied significantly accord-
ing to the pk/pd model applied. When using certain es-
timated values to manage anesthesia in awake
craniotomies, it is therefore essential to know which pk/
pd model the TCI device is applying to correctly interpret
estimated Cplasma and Ceffect-site values displayed by the
TCI pump. The Diprifusor incorporates the Marsh pk
model with a ke0 of 0.26/minutes. For both the Asena PK
and the Braun Space TCI pump, the user can choose
between the Marsh and the Schnider pk model as well as
between plasma and effect-site targeting. However, for the
Schnider model in effect-site mode, the Braun Space ap-
plies a fixed ke0 of 0.456/min, whereas the Asena PK
calculates an individual ke0 to achieve a fixed time-to-
peak effect of 1.6 minutes. Therefore, we chose appro-
priate pk/pd models to simulate the performance of the
TCI pumps Diprifusor (Marsh model), Braun Space
(Schnider model), and Asena PK (Schnider tpeak model).
Accordingly, we expected that intraoperative return of
consciousness occurs at different propofol concentrations
during awake craniotomy when using these TCI pumps.
For the Diprifusor, this would translate to a Cplasma of
B1.9 mg/mL, for the Braun Space to a Ceffect-site of 1.5 mg/
mL, and for the Asena PK to a Ceffect-site of 1.6 mg/mL
(Asena PK). Considering the far lower interindividual
variability of BIS (CoVar = coefficient of variation =
SD/mean value = 9.3% and 6.2% at return of con-
sciousness and begin of neurological testing, respectively)
compared with Cplasma (CoVar between 23.4% and
42.6%) and Ceffect-site (CoVar between 26.4% and 46.3%),
it seems recommendable to use a processed EEG device to
account for the interindividual variability.
Limitations
Both TCI and BIS monitoring are intended to sup-
port and guide the anesthesiologist, but limitations must
be kept in mind: For instance, the pk/pd models im-
plemented in TCI devices were obtained in relatively small
study populations from which a given patient might de-
viate to some extent. However, Schnider et al31 “found no
evidence that TCI is not at least as safe as anesthetic ad-
ministration using constant rate infusions.” This is based
on experience with TCI which has been used for >20 years
in millions of patients. Therefore, TCI is considered a
“mature technology” and is approved or available in at
least 96 countries.32 The BIS monitor has been shown to
correlate with the estimated effect-site concentration of
anesthetics,33,34 which is an accepted indicator of anes-
thetic depth.34,35 Even though this correlation is close
(coefficient of determination r2 between 0.8335 and 0.92,21
it is not perfect (r2 = 1.0). These limitations have to be
taken into account when using these technologies. How-
ever when doing so, both TCI and BIS monitoring are very
helpful in guiding anesthesia in our opinion.
Arterial blood sampling could produce higher pro-
pofol concentrations than venous sampling. Therefore
our study might have favored the Schnider model, since
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
Propofol PD and BIS During Awake Craniotomy
both were based on the same (arterial) sampling method.
In contrast, the Marsh model, which is based on venous
sampling might underestimate the arterial propofol con-
centration. However, this holds true only during running
propofol infusions. At least 60 seconds after stopping the
propofol infusion, it makes no difference, as to whether ar-
terial or venous samples are drawn (personal written com-
munication with Gavin N.C. Kenny 2016, coauthor of the
Marsh model). Since the data presented here were obtained
following cessation of the propofol infusion, the method of
blood sampling would not have affected our results.
Our results were obtained during awake craniotomy
in neurosurgical patients and are not necessary trans-
ferable to other patients and other anesthetic techniques.
For instance, Barakat et al36 reported that changes in BIS
and sedation scores correlated better with the Marsh than
with the Schnider effect-site concentrations in patients
undergoing orthopedic surgery under spinal anesthesia
with propofol sedation.
Our study is limited by the number of 13 patients
included, which is attributable to the fact, that awake
craniotomies are restricted to specific and rare indications
such as tumors or epileptic foci in the vicinity of the
language center when extraoperative mapping (the
standard procedure in Bonn)37 using subdural electrodes
is not feasible. Patients received remifentanil during short
and painful periods such as Mayfield fixation, skin in-
cision and suture which might have confounded our
study. Any remaining remifentanil effect during the 3
investigated time points is unlikely, given the short re-
mifentanil half-time of 3 minutes38 and the fact that the
remifentanil was stopped 26 minutes earlier than the
propofol infusion. The remifentanil effect-site concen-
tration of 0.05 ng/mL, which was calculated at the time
point of propofol discontinuation, is at least a magnitude
below the concentration, where a clinical relevant inter-
action with propofol would be expected.39 Therefore, we
exclude a confounding effect of remifentanil at the first
time point (ie, propofol discontinuation) and even more
at the subsequent time points of return of consciousness
and intraoperative testing.
Patients were evaluated not constantly but rather
every 20 to 30 seconds for the occurrence of return of
consciousness and the ability to begin with neurological
testing adding an error of a few seconds to the time points
measured.
CONCLUSIONS
To perform the Boston Naming Test during awake
craniotomies, patients are required to be fully awake and
cooperative with plasma propofol concentrations as low
as 0.8 mg/mL. At this time point the Schnider model es-
timates similar propofol concentrations, whereas the
Marsh pk/pd model significantly overestimates propofol
concentrations. Moreover, the Marsh model significantly
overestimates the actual propofol plasma concentration
during neurologically crucial time points such as return of
consciousness. Therefore, in a clinical setting of awake
www.jnsa.com | 37
Soehle et al
craniotomy in spontaneous breathing patients, we advo-
cate to use the Schnider model instead. To adjust for in-
terindividual variations in pk/pd, it seems advisable to
monitor depth of anesthesia using dedicated EEG devices.
REFERENCES
1. Bilotta F, Rosa G. ‘Anesthesia’ for awake neurosurgery. Curr Opin
Anaesthesiol. 2009;22:560–565.
2. Brown T, Shah AH, Bregy A, et al. Awake craniotomy for brain
tumor resection: the rule rather than the exception? J Neurosurg
Anesthesiol. 2013;25:240–247.
3. Frost EA, Booij LH. Anesthesia in the patient for awake
craniotomy. Curr Opin Anaesthesiol. 2007;20:331–335.
4. Sarang A, Dinsmore J. Anaesthesia for awake craniotomy—
evolution of a technique that facilitates awake neurological testing.
Br J Anaesth. 2003;90:161–165.
5. Huncke K, Van de Wiele B, Fried I, et al. The asleep-awake-asleep
anesthetic technique for intraoperative language mapping. Neuro-
surgery. 1998;42:1312–1316.
6. Skucas AP, Artru AA. Anesthetic complications of awake
craniotomies for epilepsy surgery. Anesth Analg. 2006;102:882–887.
7. Gepts E. Pharmacokinetic concepts for TCI anaesthesia. Anaes-
thesia. 1998;53(suppl 1):4–12.
8. Gepts E, Camu F, Cockshott ID, et al. Disposition of propofol
administered as constant rate intravenous infusions in humans.
Anesth Analg. 1987;66:1256–1263.
9. Marsh B, White M, Morton N, et al. Pharmacokinetic model driven
infusion of propofol in children. Br J Anaesth. 1991;67:41–48.
10. Schnider TW, Minto CF, Gambus PL, et al. The influence of
method of administration and covariates on the pharmacokinetics of
propofol in adult volunteers. Anesthesiology. 1998;88:1170–1182.
11. Schuttler J, Ihmsen H. Population pharmacokinetics of propofol: a
multicenter study. Anesthesiology. 2000;92:727–738.
12. Shafer A, Doze VA, Shafer SL, et al. Pharmacokinetics and
pharmacodynamics of propofol infusions during general anesthesia.
Anesthesiology. 1988;69:348–356.
13. Tackley RM, Lewis GT, Prys-Roberts C, et al. Computer controlled
infusion of propofol. Br J Anaesth. 1989;62:46–53.
14. Upton RN, Ludbrook G. A physiologically based, recirculatory
model of the kinetics and dynamics of propofol in man.
Anesthesiology. 2005;103:344–352.
15. Soehle M, Wolf CF, Priston MJ, et al. Comparison of propofol
pharmacokinetic and pharmacodynamic models for awake craniot-
omy: a prospective observational study. Eur J Anaesthesiol. 2015;32:
527–534.
16. Schulz U, Keh D, Fritz G, et al. “Asleep-awake-asleep”-anaesthetic
technique for awake craniotomy. Anaesthesist. 2006;55:585–598.
17. Soehle M, Ellerkmann RK, Grube M, et al. Comparison between
bispectral index and patient state index as measures of the
electroencephalographic effects of sevoflurane. Anesthesiology. 2008;
109:799–805.
18. Costello TG, Cormack JR. Anaesthesia for awake craniotomy: a
modern approach. J Clin Neurosci. 2004;11:16–19.
19. Kaplan E, Goodglass H, Weintraub S. Boston Naming Test.
Philadelphia, PA: Lea & Febiger; 1983.
20. Minto CF, Schnider TW, Egan TD, et al. Influence of age and
gender on the pharmacokinetics and pharmacodynamics of remi-
fentanil. I. Model development. Anesthesiology. 1997;86:10–23.
38 | www.jnsa.com
Copyright r 2018 Wolters Kluwer Health, Inc. All rights reserved.
J Neurosurg Anesthesiol  Volume 30, Number 1, January 2018
21. Soehle M, Kuech M, Grube M, et al. Patient state index vs
bispectral index as measures of the electroencephalographic effects
of propofol. Br J Anaesth. 2010;105:172–178.
22. Bruhn J, Bouillon T. A simple Excel spreadsheet approach to predict
i.v. anesthetic drug concentrations [in German]. Anaesth Inten-
sivmed. 2002;43:708–710.
23. Sheiner LB, Stanski DR, Vozeh S, et al. Simultaneous modeling
of pharmacokinetics and pharmacodynamics: application to
d-tubocurarine. Clin Pharmacol Ther. 1979;25:358–371.
24. Schnider TW, Minto CF, Shafer SL, et al. The influence of age on
propofol pharmacodynamics. Anesthesiology. 1999;90:1502–1516.
25. Minto CF, Schnider TW, Gregg KM, et al. Using the time of
maximum effect site concentration to combine pharmacokinetics
and pharmacodynamics. Anesthesiology. 2003;99:324–333.
26. Schwinn DA, Shafer SL. Basic principles of pharmacology related to
anesthesia. In: Miller RD, ed. Anesthesia Fifth Edition. Philadelphia,
PA: Churchill Livingstone; 2000;1:15–47.
27. Sahinovic MM, Beese U, Heeremans EH, et al. Bispectral index
values and propofol concentrations at loss and return of conscious-
ness in patients with frontal brain tumours and control patients. Br J
Anaesth. 2014;112:110–117.
28. Lobo F, Beiras A. Propofol and remifentanil effect-site concen-
trations estimated by pharmacokinetic simulation and bispectral
index monitoring during craniotomy with intraoperative awakening
for brain tumor resection. J Neurosurg Anesthesiol. 2007;19:
183–189.
29. Keifer JC, Dentchev D, Little K, et al. A retrospective analysis of a
remifentanil/propofol general anesthetic for craniotomy before
awake functional brain mapping. Anesth Analg. 2005;101:502–508.
30. Olsen KS. The asleep-awake technique using propofol-remifentanil
anaesthesia for awake craniotomy for cerebral tumours. Eur J
Anaesthesiol. 2008;25:662–669.
31. Schnider TW, Minto CF, Struys MM, et al. The safety of target-
controlled infusions. Anesth Analg. 2016;122:79–85.
32. Absalom AR, Glen JI, Zwart GJ, et al. Target-controlled infusion: a
mature technology. Anesth Analg. 2016;122:70–78.
33. Kreuer S, Bruhn J, Larsen R, et al. Application of Bispectral Index
and Narcotrend Index to the measurement of the electroencephalo-
graphic effects of isoflurane with and without burst suppression.
Anesthesiology. 2004;101:847–854.
34. Struys MM, Jensen EW, Smith W, et al. Performance of the ARX-
derived auditory evoked potential index as an indicator of anesthetic
depth: a comparison with bispectral index and hemodynamic
measures during propofol administration. Anesthesiology. 2002;96:
803–816.
35. Billard V, Gambus PL, Chamoun N, et al. A comparison of spectral
edge, delta power, and bispectral index as EEG measures of
alfentanil, propofol, and midazolam drug effect. Clin Pharmacol
Ther. 1997;61:45–58.
36. Barakat AR, Sutcliffe N, Schwab M. Effect site concentration
during propofol TCI sedation: a comparison of sedation score with
two pharmacokinetic models. Anaesthesia. 2007;62:661–666.
37. Wellmer J, Weber C, Mende M, et al. Multitask electrical
stimulation for cortical language mapping: hints for necessity and
economic mode of application. Epilepsia. 2009;50:2267–2275.
38. Kapila A, Glass PS, Jacobs JR, et al. Measured context-sensitive
half-times of remifentanil and alfentanil. Anesthesiology. 1995;83:
968–975.
39. Nieuwenhuijs DJ, Olofsen E, Romberg RR, et al. Response
surface modeling of remifentanil-propofol interaction on cardiores-
piratory control and bispectral index. Anesthesiology. 2003;98:
312–322.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.