University of Southern California Dissertations and Theses

Propofol Effects On Eeg And Levels Of Sedation

(USC Thesis Other)

Propofol Effects On Eeg And Levels Of Sedation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand corner and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
fonn at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zecb Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600

PROPOFOL EFFECTS
ON EEG AND LEVELS OF SEDATION
by
Monica G. Doss
A Thesis Presented to the
FACULTY OF THE SCHOOL OF ENGINEERING
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(Biomedical Engineering)
December 1995
UMI Numbers 1379578
Copyright 1996 by
Doss, Monica 6.
All rights reserved.
UMI Microform 1379578
Copyright 1996, by UMI Company. All rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
UNIVERSITY O F SOUTHERN CALIFORNIA
THE GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, CALIFORNIA 0 0 0 0 7
Thts thesis, written by
Monica 6. Doss
under the direction of h&X. Thesis Committee,
and approved by all its members, has been pre
sented to and accepted by the Dean of The
Graduate School, in partial fulfillment of the
requirements for the degree of
Dean
jy afe December 20, 1995
THESIS COMMITTEE
* • * * • * ■ ■ ■ * * « * ■ • • • » * « r ~ i | p ---------- n i r -------------1 1 ■ t " t
_
ACKNOWLEDGEMENTS
I would like to thank Dr. David D'Argenio for his guidance, support and
encouragement in the avenue of research taken. His confidence in allowing me to conduct
the research in a different city was greatly appreciated. I would also like to thank Dr.
Juliana Barr for her supervision, assistance and willingness to answer my many questions
in the conduct of this research project. Her insight and advice not only in the research but
also in career matters was extremely helpful. Lastly, I would like to thank Charles Minto
for allowing me to view the intricate workings of his software and taking the time to
explain its various aspects.
CONTENTS
Acknowledgements................................................................ ii
List of Tables....................................................................................................................... iv
List of Figures.................................................................................................................... v
Introduction........................................................................................... ,........................... 1
EEG........................................................................................................................ I
Propofol.................,............................................................................................... 3
Specific Aim......................................... 4
Methods ....................................................................... 5
Pli arma cokenitics/Ph amia co dynamics................. 8
Results ................................................................................. 12
Individual Pliarmacokenitic Analysis..................................................................... 12
Population Pharmacokenitic Analysis................................................................... 12
Pharmacodynamic Analysis.................................................................. 14
Discussion and Conclusion............................................................................................... 31
Reference List............................. 35
* * *
in
LIST OF TABLES
1. Patient Demographics................................................... .....................................................6
2. Sedation Scale for Assessing the Level of Sedation in ICU Patients............................. 7
3. Two Stage Propofol PK Analysis.................................................................................... 13
4 . Pooled Propofol PK Analysis......................................................................................... 15
5. Pooled Analysis: Performance Measures....................................................................... 16
6. Convergence for Parameters of Interest............................................ 17
7. Bispectral Index Pharmacodynamic Summary. ..................................................... 18
8. Comparison of PD Modeling for Bispectral Index Using PredCe vs. PrcdCp............ 28
9. Convergence for Individual Sedation Probability Curves..............................................29
LIST OF FIGURES
1 . Link Model ......................................................................................................... 9
2. Heteroscedastic NONMEM 2 Compartment................................................................19
3. Heteroscedastic NONMEM 3 Compartment......................................................... 20
4. Homoscedastic NONMEM 2 Compartment................................................................ 21
5. Homoscedastic NONMEM 3 Compartment................................................................ 22
6. Heteroscedastic NPD 2 Compartment................................ 23
7. Heteroscedastic NPD 3 Compartment................ 24
8. Homoscedastic NPD 2 Compartment........................................................................... 25
9. Homoscedastic NPD 3 Compartment...........................................................................26
10. Pooled Probability Curves for Sedation vs. Concentration...........................................30
v
INTRODUCTION
»
Critically ill patients in the intensive care unit (ICU) who are intubated and
mechanically ventilated often experience pain, anxiety, agitation and discomfort [1,2,3,4].
In order to minimize these physical and psychological discomforts, sedation is typically
required for such patients. However, this raises concerns for the anesthesiologist.
Excessive dosing of a sedative may lead to respiratory depression, systemic hypotension,
and persistent sedation following its discontinuation [5,6,7]. Conversely, inadequate
dosing will not alleviate the patient's discomfort and may subsequently lead to
cardiopulmonary and myocardial complications [8,9]. An objective means of assessing the
level of sedation in unconscious patients while they are intubated would be extremely
helpful to prevent the hazards of inadequate or excessive sedation.
Such a means would consequently take advantage of the pharmacodynamic
modeling principles which correlates dose to response. In pharmacodynamic modeling a
measurement of the pharmacological response intensity is required [10], For drugs acting
on the central nervous system, this requirement may possibly be fulfilled through
electroencephalogram (EEG) analysis [11], Hie continuous signal intrinsic to the EEG
parameters are consistent with ideal pharmacological measures.
EEG
Interpretation of traditional EEG tracings takes years of training and typically
requires hours. Consequently, their use and applications in events requiring real time
decision such as that in the ICU has been less than desirable in the past. The current
1
availability of fast microprocessors has led to the development of numerous devices and
techniques which have been utilized for real time interpreting of a patient's EEG. In an
attempt to obtain an objective measurement to aid in therapeutic decisions, such new
technology has the possible potential to be employed in critically ill ICU patients receiving
anesthesia [12],
Like any periodic signal, the EEG can be broken down into its component
frequencies or bands. The delta, theta, alpha, and beta bands are composed of frequencies
less than 4Hz, 4-7.5Hz, 8.0-13.5Hz, and 13.75-30 Hz, respectively. One common
method of displaying these parameters is through power spectral analysis (PSA). PSA
represents the amplitudes or power of each of the sine wave components as a function of
the frequency. The PSA bands may be reported as an absolute value (ABS) or a relative
value (REL) of the total power spectrum. Two additional parameters are sometimes
utilized to characterize EEGs. The median power frequency is that frequency that bisects
the power spectrum. Half the power is below and half the power is above the median
power frequency. .The spectral edge frequency reduces the power spectrum to a single
number that represents the highest frequency present in the EEG[14].
Power spectral analysis assumes a Gaussian, stationary first order, linear model of
the signal. Consequently, PSA cannot quantify relationships that may exist between the
components themselves. By utilizing this method of analysis, one assumes that each sine
wave component is independent of the others and that there is no relationship between the
components[l 1,15 16], Only frequency and power estimates are considered, while phase
information is typically ignored.
2
Biological systems, on the other hand, are typically nonlinear. Consequently, such
systems would not conform to the assumptions of the conventional power spectrum. A
method of analysis which detects and quantifies aspects of non-linear changes may
therefore better characterize the dynamic changes of the EEG.
Bispectral analysis is a technique that quantifies the interaction between the
components that make up the signal. This is accomplished by quantifying the non-linear
phase locking and quadratic energy coupling within various frequencies of the EEG [17,
18]. The bispectral index is a univariant parameter computed from the EEG bispectrum
[19].
The important issue is the choice of the EEG parameter and its relevance to the
clinical environment of the ICU. In order for the EEG parameter to be useful to clinicians
in an ICU setting it must correlate to some depth of sedation measurement. Currently,
studies are inconclusive and the correlation still needs to be established [12, 13]. The
difficulty arises when different anesthetics produce different EEG effects at differing doses
or when the sedatives are administered in conjunction with each other [13, 14],
PROPOFOL
Propofol is one possible sedative which may be given to patients in the ICU
setting. Although a significant number of studies cite EEG and propofol in their
methodology, there is a scarcity of studies which correlates specific EEG changes to
dosing amount and even fewer which correlate EEG changes to sedation levels. Some
studies have shown that propofol produces an increase in the delta and the beta bands
while other studies have shown that a significant decrease is produced
[20,21,22,23,25,26,27], Herrigod et. al. showed that a biphasic response occurs in the
delta band when propofol is administered [24], Studies investigating bispectral index,
spectral edge frequency, and median frequency are few. One study found that the
bispectral index significantly decreased [19], Inquiry into the effects on the spectral edge
by propofol revealed that a significant increase may occur [20,25], Results investigating
median frequency were mixed, showing significant increase, significant decrease and no
significant change [20,23,25,28], Although these studies attempted to characterize
significant changes, none of them attempted to develop pharmacodynamic models.
SPECIFIC AIM
An objective means of assessing the level of sedation in unconscious ICU patients
while they are intubated and receiving propofol is desirable. EEG parameter changes and
its possible correlation to dosing and to sedation level may provide such an objective
measurement. The following provides an outline of the objectives investigating this
pharmacokinetic/pharmacodynamic model:
1) Characterize a pharmacokinetic model for propofol
2) Investigate various EEG parameters and identify a parameter where convergence
occurs for all patients during individual pharmacodynamic modeling
3) Describe possible correlation between sedation levels and propofol plasma
concentrations via probability curves.
4) Describe possible correlation via probability curves between sedation levels and the
EEG parameter which convergence occurred for all patients.
4
METHODS
Subsequent to Institutional Review Board approval, informed consent was obtain
for 7 patients prior to their surgery. Patient demographics are depicted in Table 1. All
patients received general anesthesia during surgery. Furthermore, all patients required
postoperative intubation and mechanical ventilation for greater than 12 hours in the
intensive care unit. Patients received propofol for sedation via a computer-controlled
infusion pump (Harvard Apparatus 22 syringe-based infusion pump) which permitted
targeting plasma propofol concentration via Stanpump kinetics'. Concentrations were
adjusted for gender, height, and weight. Patients were assessed for sedation utilizing a
modified Ramsey Sedation Scale (Table 2) [29].
Prior to the initiation of the propofol, a SOug bolus iv followed by an infusion of
25 to 150ug/hr fentany 1 was administered for analgesia. A targeted propofol plasma
concentration of 0.25ug/ml was started. Each targeted propofol change was accompanied
with a Ramsey Sedation Score (RSS). The predicted concentration was increased
0.25ug/ml every 15 minutes until the patient acquired a RSS of 5, a mean arterial pressure
(MAP) decrease to <60 mmHg, or a MAP 30% below baseline. Once the patient attained
a RSS of 5, the targeted propofol plasma concentration was maintain and the patient
received a loading mivacurium (Mivacron® , Burroughs Wellcome) bolus up to 0.2mg/kg
iv followed by a maintenance infusion of 1 to 15ug/kg/min for 60 minutes. Subsequently,
1 STANPUMP software was written by Steven L. Shafer, MD, Anesthesiology Services (112A),
PAVAMC, 3801 Miranda Ave., Palo Alto, CA.
5
Table 1. PATIENT DEMOGRAPHICS
m ean std dev
age (yrs)
65.57 12.35
Gender
6M&1F na
Height (cm)
178.04 7.01
Weight(kg)
83.08 14.03
Lean body
mass (kg)
62.03 8.26
Fat body
mass (kg)
21.05 7.83
Tabic 2. Sedation Scale for Assessing the Level of Sedation in ICU Patients
SEDATION SCORE CLINICAL RESPONSE
0 patient paralyzed; unable to evaluate
1 patient awake
2 patient lightly sedated but oriented
3 patient deeply sedated, bur follows commands
4 patient deeply sedated, unable to follow commands, but
responds to non-pain stimuli
5 patient deeply sedated; responds only to pain stimuli
6 patient deeply sedated; unresponsive to painful stimuli
7
target plasma propofol concentrations were decreased by 0.25ug/ml every 15 minutes to
Oug/ml or until the patient attained a RSS of 2. Blood plasma samples were collected
from patients at baseline, prior to and 15 minutes subsequent to each titration of the
propofol infusion.
EEG monitoring was performed with an Aspect A -1000 monitor (Natick, MA,
USA) with a two channel referential montage. Sampling occurred every 10 seconds. In
accordance with the International 10-20 electrode system, 3 electrodes were placed at At],
At2, with reference at Fp z and ground at Fp J or Fp 2 . Electrode impedance was less than
50000 during the study. Segments EEG power spectrum and bispectral index were
chosen for analysis. Extracted segments comprised of EEG data collected in the controlled
environment of the ICU during time epochs that are artifact free and when there was no
non-constant concurrent procedure which may provide constant stimulus to the patient.
PHARMACOKINETICS/PHARMACODYNAMICS
Pharmacology is divided into two disciplines, pharmacokinetics and
pharmacodynamics. Pharmacokinetics describes the process which governs the drug
concentrations subsequent to drug dosing. Pharmacodynamics describes the equilibrium
relationship between concentration and the effects of interest.
The "link model" assess the pharmacodynamics by linking it to the
pharmacokinetics [39], Tlius the pharmacokinetics predicts the dose-concentration
relationship, thereby providing the foundation for assessing the concentration-effect
relationship of the pharmacodynamics. Figure 1 provides a schematic which describes this
linking relationship.
Figure 1. "LINK MODEL"
Dose — PK PD PKPD - c P — —Ce~ Teffect
Assuming a 3 compartment mamillary model, the pharmacokinetic curve is described by:
C = Ae ^ + Be 'pi+De "f it
The rate constants for each compartment are the roots of the quadratic equation. So if,
b=ct(D+BHBfA+DH8fA+BI
A+B+D
c= apD+ct8B+B8A
A+B+D
k» = fb -tb 2-4cl03l
2
k2i = tb + tb2- 4 c f 5 I
2
Tlien:
kio = a p S
kziksi
kn = IBS + aB + aSV bifa + p+SI-kmkn+fhil2
k jt- k ii
loi = a + P +5 — (kio+ku+kii+kjt)
Volumes for compartments 1 , 2, and 3 (VI, V2, and V3 respectively) may be calculated
subsequently;
VI = dose V2 =Vlk» V3 = Vlk»
A + B +D kii ksi
9
It is the linking of the plasma concentration (Cp) with the effect site concentration
(Ce) via linking of the pharmacokinetic and pharmacodynamic analysis which allows the
investigator to model the measured effect as a function of the drug.
One possible pharmacodynamic model is the sigmoid Emax model:
E = Emax CpY (Equation 1)
ECm + Cpr
where E is the measured effect, Cp is the plasma concentration, Emax is the maximum
effect attributed to the drug, y influences the slope of the curve, and EC» is the plasma
concentration producing 50% of Emax, Assuming that drug enters and leaves the effect
site via a first order process, then if the plasma concentrations change, the time course of
the drug's accumulation at the effect site can be predicted.
The KeO is the rate constant which characterizes the disequilibrium between
plasma concentration and the effect site concentration. Tlius, the equilibrium half-time
between the effect site and the plasma is In2/Ke0. This rate constant cannot be estimated
from the Cp but may be estimated from the time coarse of the drug effect if the
pharmacokinetic model is defined and the pharmacodynamic form is given. Hie predicted
effect site concentration utilizes the parameters of the pharmacokinetic model for the
plasma concentrations and the equilibrium constant. If Ce is utilized instead of Cp in
equation 1, then the KeO, Emax, EC» and y may be estimated through non-linear
regression analysis.
Pharmacokinetic and pharmacodynamic modeling was performed using PK/PD
Tools fo r Excel with XLMEM® (Version 1.02), created by Charles Minto, MB.ChB., and
10
Thomas Schnider, M.D. Individual pliarmacodynamic parameters were based upon the
best fit for the individual pharmacokinetic models (2 or 3 compartment). Population
pharmacokinetics (two staged pooled, naive pooled, and nonlinear mixed effects)
modeling was conducted. EEG parameters; absolute beta and delta, relative beta and
delta, median frequency, spectral edge frequency, and bispectral index were assessed for
pharmacodynamic modeling. Further analysis was performed on the parameter which
converged for all patients.
Probability analysis (between level of sedation and propofol plasma concentration)
was completed for each patient and as a pooled analysis. Probability analysis was also
completed between level of sedation and bispectral index.
11
RESULTS
INDIVIDUAL PHARMACOKENITIC ANALYSIS
All 7 patients were included in the pharmacokinetic analysis. Patient 6 and patient
7 both had a history of alcohol abuse. Patient 5 experienced renal failure during the study.
Patient 1 had additional plasma samplings at time 1, 2, and 5 minute post start of propofol
infusion. Individual two compartment and three compartment pharmacokinetic
heteroscedastic analysis was conducted for each patient. The log likelihood (LL) was
chosen as the objection function and was maximized such that the most probable
pharmacokinetic parameters (i.e. volumes and clearances) were estimated. The difference
of twice the log likelihood between the two and three compartment model was calculated.
The three compartment model was chosen as the optimal model for a given patient when
the difference exceeded 8 (P=0.05). Variance for each model and patient was calculated.
As depicted in Table 3, three patients fitted best to a two compartment model whereas
four patients fitted best to a three compartment model. Log average of parameter values
for two and three compartment model was calculated (Table 3). The standard deviation in
the log domain was determined to be approximate to the coefficient of variation in the
standard domain (Table 3).
POPULATION PHARMACOKENITIC ANALYSIS
In addition to the two staged pooled analysis, homoscedastic and heteroscedastic
naive pooled (NPD) and nonlinear mixed effects model (NONMEM) pharmacokinetic
analysis was conducted for the seven patients. Hie extended least squares (ELS) was
12
Table 3. TWO STAGE PROPOFOL PK ANALYSIS
P T 1 PT2 PT3 PT4
v l 23.70 152.96 66.99 124.04
v2 2365.29 3152.22 1458.18 5120.68
v3 0.00 0.00 0.00 0.00
ell 3.65 2.72 3.16 3.54
cl2 3.61 3.09 1.59 3.36
cl3 0.00 0.00 0.00 0.00
var 0.08 0.05 0.03 0.02
*log likelihood -8.41 21.44 21.63 45.60
v l 20.29 145.80 50.62 94.87
v2 143.71 1461.69 248.79 300.00
v3 2168.98 9965.78 7381.58 7491.55
e ll 3.64 1.98 2.33 3.18
cl2 9.04 2.90 1.09 1.50
cl3 2.14 1.55 1.75 3.04
var 0.06 0.05 0.02 0.02
*log likelihood 1.47 22.67 27.99 47.29
difference o f 2x LL 19.75564 2.474037 12.71249 3.382485
biff greater than 8 (P=0.05) 3 COMP 2 COMP 3 COMP 2 COMP
YLOG LIKELIHOOD SHOULD BE MAXIMIZED
log
PT5 PT6 PT7 AVG AVG CV%
154.30 100.55 76.74 99.90 86.37 26.31
4149.77 4665.16 2995.53 3415.26 3171.66 17.56
0.00 0.00 0.00 0.00 0.00 0
3.70 2.62 4.13 3.36 3.32 6.78
3.69 2.84 2.57 2.97 2.87 11.65
0.00 0.00 0.00 0.00 0.00 0
0.05 0.02 0.02 0.04
54.33 50.72 45.12 ■ 32.92
49.22 70.37 76.66 72.55 60.53 24.97
440.55 143.29 2978.66 816.67 315.68 46.66
10710.96 4772.83 5000.00 6784.52 6275.99 22.06
2.80 2.59 4.06 2.94 2.70 10.10
3.87 1.09 2.57 3.15 2.38 30.70
3.12 2.61 0.08 2.04 2.29 11.53
0.02 0.02 0.02 0.03
63.06 51.10 45.15 36.96
17.46029 0.759293 0.069185 8.087632
3 COMP 2 COMP 2 COMP 3 COMP
chosen as the objection function and was minimized such that the most probable
pharmacokinetic parameters (i.e., volumes and clearances) were estimated. The difference
of twice the log likelihood between the two and three compartment model was calculated.
The three compartment model was chosen as the optimal model for a given patient when
the difference exceeded 8 (P=0.05). The heteroscedastic model for NPD and for
NONMEM model fitted best in a three compartment model while the homoscedastic
model for NPD and NONMEM fitted best in a two compartment model (Table 4). Bias,
precision, divergence and wobble were calculated for both NPD and NONMEM and are
depicted in Table 5. Figures 2-9 shows the ratio of the observed over predicted verses
time for homoscedastic and heteroscedastic NPD and NONMEM models.
PHARMACODYNAMIC ANALYSIS
Pharmacodynamic modeling was conducted on five patients. Table 6 depicts
which EEG parameters converged for each patient. Only patient 1 showed convergence
for all parameters. The bispectral index was the only parameter which converged for all
five patients studied.
A closer look at the pharmacodynamic values obtained for the bispectral index
reflects an extreme range for the KeO. Patient 2 and patient 5 had extremely large values
whereas patient 5 and patient 7 had very small values. These KeO values translated into
equilibrium half times of 1.26, 70.26, and 31.04 minutes for patient 1, 6, and 7
respectively. Equilibrium half times for patients 2 and 5 were basically instantaneous.
Such large KeO values for patient 2 and 5 reflected possibly that steady state had already
been achieved for these patients prior to plasma sampling (Table 7).
14
Table 4. POOLED PROPOFOL PK ANALYSIS
7PT
2 STAGE
homoscedastic
NPD
vl 99.90 103.17
v2 3415.26 2854.64
v3 0 0
ell 3.36 3.14
cl2 197 189
cl3 0 0
var 0.04 0.04
Hog likelihood 3192 ELS -506.33
Vl 7155
v2 816.67
v3 6784.52
ell 194
cl2 3.15
cl3 104
var 0.03
Hog likelihood 36.96
8.09
ELS
liff greater than 8 (P=0.05) 3 COMP
ELS IS MINIMIZED AND IS ~ EQUIVALENT TO 2LL
k LOGLIKELIHOOD SHOULD BE MAXIMIZED
83.46
32.16
2965.03
3.11
1.64
2.84
0.04
-507.73
1.40
2 COMP
log
heteroscedastic
NPD
104.48
3896.18
0
3.19
2.73
0
0.14
-227.01
homoscedastic
NONMEM
109.45
2328.61
0
3.36
2.64
0
0.03
-547.25
log
heteroscedastic
NONMEM
93.14
4564.62
0
3.48
3.29
0
0.11
-278.71
58.70
249.19
5597.00
3.01
2.97
141
0.13
-255.05
100.00
15.76
2387.94
3.26
138
2.77
0.04
-549.22
18.02
484.02
724149
3.79
5.21
3.53
0.07
-328.89
28.03
3 COMP
1.96
2 COMP
50.18
3 COMP
Table 5. POOLED ANALYSIS; PERFORMANCE MEASURE
NPD NPD MEM MEM
1 compartment
homoscedastic heteroscedastic homoscedastic hctroscedastic
bias -0.036 0.048 0.004 0.163
precision 0.188 0.213 0.199 0.259
divergence 0.002 0.001 0.003 0.001
wobble 0.116 0.137 0.139 0.149
3 compartment
bias -0.031 0.055 -0.025 0.279
precision 0.192 0.187 0.201 0.338
divergence 0.002 0.000 0.003 -0.002
wobble 0.124 0.131 0.137 0.179
Table 6. CONVERGENCE FOR PARAMETERS OF INTEREST
PATIENT 1 PATIENT 2 PATIENT 5 PATIENT 6 PATIENT 7
ABS BETA X X X X
ABS DELTA X X X
REL BETA X X X X
REL DELTA X X
MED FREQ X X X
SEF X X
BIS INDEX X X X X X
17
T abic 7. BISPECTRAL INDEX PHARMACODYNAMIC SUMMARY
PT 1 PT2 PT5 PT6 PT7
COMPARTMENT MODEL 3 2 3 2 2
KeO 0.55 15736.25 8671.81 0.01 0.02
E O 85.80 88.84 87.24 96.32 85.95
Emax 70.38 49.02 54.53 59.36 48.58
EC50 1.62 1.06 0.23 0.50 0.17
gamma 10.45 14.77 15.00 7.09 4.64
SumSq 19956.36 44003.97 44572.51 204005.28 6945.52
tl/2 (min) 1.26 0.00 0.00 70.26 31.04
steady state 6.29 0.00 0.00 351.31 155.18
18
Figure 2.
HETEROSCEDASTIC NON MEM
2 COMPARTMENT
T T
6 -
5 -
*D
4 )
■ * - *
u
T3 4
V
•a
0 1
£ 3
0 ) 0
v >
- Q
o
2 -
1 -
1000 2000 3000 4000
Elapsed Time (min)
5000 6000 7000 8000
vs
Figure 3.
HETEROSCEDASTIC NONMEWI
3 COMPARTMENT
6 -
T 3
0 )
5
•o 4
T J
V
£ 3
u «
1 0
• Q
O
2 -
1000 2000 3000 4000
Elapsed Time (min)
5000 6000 7000 8000
Observed / Predicted
Figure 4.
HOMOSCEDASTIC NON MEM
2 COMPARTMENT
0 1000 2000 3000 4000 5000 6000 7000 8000
Elapsed Time (min)
Figure 5.
HOMOSCEDASTIC NONMEM
3 COMPARTMENT
6 -
5 -
T3
V
■ G
T J 4
< 0
■ u
v
2 3
v " 5
M
■ Q
o
2 -
1 -
1000 2000 3000 4000
Elapsed Time (min)
5000 6000 7000 8000
Figure 6.
HETEROSCEDASTIC NPD
2 COMPARTMENT
T >
ID
G
T3 A
a >
k .
CL
-a
ID
£ 3
ID
.8
o
2 -
1 -
1000 2000 3000 4000 5000
Elapsed Time (min)
6000 7000 8000
Figure 7.
HETEROSCEDASTIC NPD
3 COMPARTMENT
7
6
5
4
3
2
1
0
0 1000 2000 3000 4000 5000 6000 7000 8000
Elapsed Time (min)
Observed / Predicted
Figure 8.
HOMOSCEDASTIC NPD
2COMPARTMENT
T
0 1000 2000 3000 4000 5000 6000 7000 8000
Elapsed Time (min)
Figure 9.
HOMOSCEDASTIC NPD
3 COMPARTMENT
6 -
5 -
■ D
0 1
o
T J 4
«
'a
0 9
£ 3 A
st
O
1 -
1000 2000 3000 4000
Elapsed Time (min)
5000 6000 7000 8000
This was further investigated by repeating the pharmacological analysis,
substituting predicted plasma concentration values for predicted effect site concentration
values and not solving for the KeO. The difference of twice the log likelihood between the
5 parameter model (KeO, Eo, Emax, EC50 and gamma) and the 4 parameter model
(excluding KeO) was calculated. The inclusion of an additional pharmacodynamic model
of KeO made a significant difference for all the patients except patient 2 and patient 5
(Table 8).
Sedation levels at varying plasma concentrations were assessed. Table 9 depicted
for which patients and sedation levels probability curves were obtained. Plasma
concentrations and corresponding sedation scores were pooled for those individual
patients which curves were obtained. Pooled probability curves were subsequently
constructed (Figure 10).
Sedation levels at varying corresponding bispectral index values were also assessed
for individual patients as well as for a pooled analysis. Probability curves were
unattainable for either.
27
Table 8. COMPARISON OFPD MODELING FOR BISPECTR/IL INDEX USING PredCe VS. PredCp
P T 1 P T 2 P T 5 P T 6 P T 7
COM P# 3 2 3 2 2
WITH PredCe
KeO 0.551 175321.008 8671.813 0.010 0.022
EO 85.798 90.575 87.236 96.318 85.953
Emax 70,385 43.335 54.530 59.358 48.584
EC50 1.618 1.116 0.232 0.505 0.172
gam m a 10.448 6.598 15.000 7.088 4.644
Sum Sq 19956.356 44003.971 44572.506 204005.282 6945.517
log likelihood -1242.327 -2547.832 -2023.254 -5003.080 -946.465
variance 54.976 48.220 79.452 161.521 21.705
WITH PredCp
KeO
EO 85.646 90.575 87.232 66.803 82.370
Emax 71.614 43.335 54.532 32.434 48.648
EC50 1.587 1.116 0.232 0.951 0.533
gam m a 10.665 6.598 15.000 9.276 11.112
log likelihood -1251.336 -2547.832 -2023.254 -5501.609 -992.545
variance 57.774 48.220 79.452 181.419 28.949
difference of 2x LL 18.017 0.000 -0.001
not as good a fit
997.059 92.159
diff greater than 4 (P=0.05) Pred Ce p n i g i n m i
isw v lvv w tvwwwivi'is^v. s s-Aw. t-v-v
Pred Ce Pred Ce
28
Table 9. CONVERGENCE FOR INDIVIDUAL SEDATION PROBABILITY CURVE
PATIENT >2 >3 >4 >5
1 X X X X
2 X X X X
3 X X
4
X
5
6 X X X
7 X X
29
Figure 10.
POOLED SEDATION ANALYSIS
for Converged Individual Analysis
to -
/
• Sedation S care >2
■ Sedation Score >3
A Sedation Score >4
0 r
Propofol Plasma Concentration
u
o
DISCUSSION
Tabic 3 indicates that approximately half of the patients fitted into a three
compartment model better and half in a two compartment. Looking at the two
compartment model, the coefficient of variation is relatively low across all parameters.
Although the coefficient of variance is low in the parameters of a three compartment
model, the volume and the clearance for the second compartment in this model are
relatively higher than the rest. For V2, this is most likely due to patient 2 and 7 who had
values one order of magnitude greater than the rest. For C12, it appears that patient 1 may
have been the contributing factor to the relatively higher coefficient of variation. A further
examination of the individual patients1 pharmacokinetics reveals that in the two
compartment model, patients 1, 2, and 5 had the greatest variance at 0.08, 0,05,and 0.05,
respectively. In the three compartment model patient 1 at 0.06 and patient 2 at 0.05 had
the greatest variance. In both compartment model the variance is less than 10%.
Comparing the 2 compartment pharmacokinetic model for the various pooled
analysis techniques, the parameter values remain consistent in order of magnitude (Table
4). Those techniques which fitted better to a three compartment model showed parameter
values consistent in order of magnitude.
Defining precision as the criteria for best fit, the NPD heteroscedastic model for 3
compartment appeared to best describe the data. However, a closer look shows that the
NPD homoscedastic 2 compartment model is 0.001 point higher than the 3 compartment
heteroscedastic NPD model. In accordance with the convention to describe systems in the
31
simplest technique or theory that best describes the data, one could make the argument for
the homoscedastic 2 compartment model.
An additional look at the performance of the models supports however, the
heteroscedastic 3 compartment NPD model as the better fitted model. Figures 2 through
9 depicts the performance error ratio for NPD and NONMEM verses time. An
Observed/Predicted value of 1 indicates that the observed exactly matched the predicted.
Thus values above 1 indicate positive bias and values below I indicate negative bias.
Overall the 3 compartment models reigned in outlier values present in the corresponding 2
compartment models. Furthermore, comparing heteroscedastic 3 compartment NPD and
NONMEM (Figure 3 and 7), the NPD values are more tightly and evenly distributed
about the value of 1 and consequently a better fit. The NONMEM model (Figure 3)
appears to have a greater bias than the NPD values. This is supported by the actual
precision and bias calculated values for 3 compartment heteroscedastic NPD and
NONMEM (Table 5).
Examining the results in Table 7, it becomes evident that the bispectral index
pharmacodynamic parameters are rather consistent across patients except for the KeO.
With respect to the KeO, it appears that three possible populations prevail. Patient l's KeO
leads to an approximate time to reach steady state of 6 minutes. If we are to believe the
KeO values for patients 2 and 5, steady state is reached instantaneously. Lastly, patients 6
and 7 have KeO values which lead to a 2 to 6 hour time period for steady state to occur.
One possible explanation for these results relates specifically to the plasma
sampling times. Patient 1 had received additional samplings at 1,2, and 5 minutes. Thus
32
concentration values along the equilibrium curve were captured. Patients 2 and 5 had
their samplings at 15 minutes post step up dosing. If we are to believe the results from
patient 1, steady state had already occurred at this time. Consequently, analysis would
have erroneously lead to large KeO values. Patient 6 and 7 took much longer to reach
steady state. Consequently, the 15 minute post step up dosing strategy captured
concentrations along the equilibrium curve.
Patients 6 and 7 r s past history of alcohol abuse may have been a contributing
factor to the lengthening time for steady state to be achieved. A number of studies have
correlated higher lipid concentrations in the blood of alcoholics than for individuals
without such a history [30, 31]. Camps et. al. found that even in individuals who no
longer abuse alcohol, the lipid concentration remained higher than individuals without
history o f abuse [30]. If patient 6 and 7 were consistent with such studies, then we would
expect that patient 6 and 7 would have higher than normal blood lipid concentration.
Propofol is an extremely lipophilic compound and is prepared in a 10% fat emulsion
[32,33]. Consequently, it is possible that since alcoholics may have higher lipid
concentrations in their plasma, the propofol may be more proned to remain in the blood
stream rather than move into the effect site. Thus, a longer time is required to achieve
steady state. This explanation would be consistent with case reports that alcoholics
require unusually higher doses of propofol and adjunctive non-lipophilic sedatives in order
for sedation to be achieved [34, 35,36]. Furthermore, In one study, alcoholics required
significantly greater doses for induction to occur and subsequently had significantly
greater propofol plasma concentrations at induction than non-alcoholics. [37]. In another
33
study, greater doses of propofol were required to achieve sedation in alcoholics.
However, the propofol plasma level was the same as that of non-alcoholics once loss of
consciousness had occurred [38],
The completed probability curves implies that a corresponding relationship may
exist between the level of sedation and the propofol concentration. As the concentration
level increases there is a greater probability that the level of sedation will increase. When
the relationship between sedation level and bispectral index was examined via probability
curves, no relationship was discernible. Since steady state may not have been achieved at
the time of the sedation score for patients 2 and 5, this pooled analysis may not reflect the
relationship accurately. In addition, the possibility of three differing populations, as
reflected in the kinetics, may have contributed to the sedation level - bispect probability
results.
In conclusion, this study elucidated the importance of plasma sampling at 1,3, and
5 minute post step-up dosing to acquire values along the equilibrium curve. Furthermore,
future studies should carefully include either individual with or individuals without a
history of alcohol abuse in order to properly characterized the pharmacokinetics and
pharmacodynamics for that population. The probability curves reflected that a correlation
between the sedation level and plasma concentration exist. However, the relationship
between sedation level and bispectral index was not seen. Future studies should continue
to examine the possible relationship in the ICU setting. Sedation scores should be taken at
steady state levels in order to characterize this relationship.
34
REFERENCES
1 . Crippen, D.W. The Role of Sedation in the ICU Patient With Pain and Agitation.
Critical Care Clin 6: 369-392, 1990.
2. Hansell, H.N. The Behavioral Effects of Noise on Man: The Patient with "Intensive
Care Unit Psychosis". Heart Lung, 13: 59-65, 1984.
3. Margolis, G.J. Postoperative Psychosis in the Intensive Care Unit. Comparative
Psychiatry, 8: 227-232, 1967.
4. Wilson, L.M. Intensive Care Delirium. Arch Intern Med, 130:225-226, 1972.
5. Shelly, M.P., Mendel, L,, Park, G.R. Failure of Critically 1 1 1 Patients to Metabolize
Midazolam. Anaesthesia,42:619-626, 1987.
6. Bergman, I.., Steeves, M,, Burckart, G., Thompson, A., Reversible Neurological
Abnormalities Associated with Prolonged Intravenous Midazolam and Fentanyl
Administration. Journal o f Pediatrics, 119: 644-649, 1991.
7. Riker, R.R., Fraser, G.L., Cox, P.M., Continuous Infusion of Haloperidol Controls
Agitation in Critically H I Patients. Critical Care Medicine, 22: 433-440, 1994.
8. Goldman, L., Supraventricular Tachyarrhythmias in Hospitalized Adults After
Surgery. Clinical Correlates in patients Over 40 Years of Age After Major Noncardiac
Surgery. Chest, 73: 450-454, 1978.
s' Udelsman, R., Norton, J.A., Jelenich, S.E., et al. Responses of the Hypothalamic-
pituitary-adrenal and Renin-agiotensin Axes and the Sympathetic System During
Controlled Surgical And Anesthetic Stress. Journal o f Clinical Endocrinological
Metabolism,64: 986-994, 1987.
10. Dingemanse, J., Danhof, M., Breimer, D.D., Pharmacokinetic-Pharmacodynamic
Modellinng of CNS Drug Effects. An Overview. Pharmacology Therapy, 38: 1-52,
1988.
11. Mandema, J.M., Danhof, M,t Electroecephalogram Effect Measures and Relationships
Between Phramacokinetics and Pharmacodynamics of Psychotropic Drugs, Clinical
Pharmacokin 23: 191-215, 1992.
12. Pichlmayr, I., Lips, U., EEG Monitoring in Anesthesiology and Intensive Care,
Neuropsychobiology, 10: 239-248, 1983.
35
13. Black, S., Malila, M.E., Cucchiara, R.F., Neurologic Monitoring , in Miller's
Anesthesia, 3rd Edtion, 1990,
14. Rampil, I.J., Wliat Every Neuroanesthesiologist Should Know About
Electroencephalograms and Computerized Monitors, In Cerebral Protection,
Resuscitation and Monitoring: a Look Into The Future of Neuroanesthesia.
15. Levy, W.J., Intraoperative EEG Patterns: Implications for EEG Monitoring,
Anesthesiology, 60: 430-434, 1984.
16. Shwilden H., Stoeckel, H., The Derivation of EEG Parameters for Modelling and
Control o f Anesthetic Drug Effect. In Stoeckel H. (Ed.) Quantitative, Modelling and
Control in Anaesthesia, pp. 196-210, Thieme, Stuttgart, 1985.
17. Huber, P.J., Kleiner, B., Gasser, T., Dumermuth, G., Statistical Methods for
Investigating Phase Relations in Stationaty Stochastic Processing, IEEE Trans Audio
Electroacoust, 19: 78-86, 1971.
1 8 Sigl, J.C., Chamoun, N.G., An Introduction to Bispectral Analysis fortlie
Electroencephalogram, Journal o f Clin Monit, in press, 1995.
19. Kearse, L.A., Manberg, P., Chamoun, N., deBros., F.,Zaslavsky, A., Bispectral
Analysis of the Electroencephogram Correlates with Patient Movement to Skin
Incision During Propofol/Nitrous Oxide Anesthesia, Anesthesia, 81: 1365-1370,
1994.
20. Veselis, R.A., Reinsel, R.A., Marino, P., Wronski, M., EEG Power Spectrum Changes
During Propofol Sedation, Anesthesiology, 75 (3A),_A181, 1991.
21. Yates, M..P., Maynard, D.E., Major, E., Frank, M., Vemiquet, A.J.W., The Cerebral
Function Analyzing Monitor: A Study Using Bolus Doses on ICI 35,868, British
Journal O f Anaesthesia 56(11), 1298P-1299P, 1984.
22. Seifert, H.A., Blouin, R.T., Conard, P.F., Gross, J.B., Sedative Doses of Propofol
Increase Beta Activity o f Processed EEG, Anesthiology, 77(3A) , A219, 1992.
23. Scliwilden, H., Stoeckel, H., Schuttler, J. Closed-Loop Feedback Control of Propofol
Anesthesia by Quantitative EEG Analysis in Humans, British Journal O f
Anaesthesia,62, 290-296.
24. Herrigods, L., Roily, G., Mortier, E., Bogaert, M,, Mergaert, C., EEG and SEMG
Monitoring and Maintenance of Anesthesia With Propofol, Journal o f Clinical
Monitoring and Computing ,6, 67-73, 1989.
36
25. Verelis, KA., Reinsel, R.A., Wronski, M., Marino, P., Tong., & Bedford, R.F., HHG
and Memory Effects of Low-Dose Infusions of Propofol, British Journal O f
Anaesthesia, 69, 246-254, 1992.
26. Reddy, R.V., Moorthy, S.S., Mattice, T., Dierdorf, S.F., Deitch, R.D., An
Electroencephalographic Comparison of Effects of Propofol and Methohexital,
Electroencephalography and Clinical Neurophysiology, 83, 162-168, 1992.
27. Seifert, H.A., Blouin, R.T., Conard, P.P., Gross, J.B., Sedative Doses of Propofol
Increase Beta Activity of Processed EEG, Anesthiology and Analgesics,76, 976-978,
1993.
28. Forrest, F.C., Tooley, M.A., Saunders, P.R., Piys-Roberts,C., Propofol Infusion and
the Suppression of Consciousness : the EEG and Dose Requirements, British Journal
O f Anaesthesia, 72, 35-41. 1994.
29. Ramsey, M.A.E., Savage, T.M., Simpson, B.R.J., Goodwin, R., Controlled Sedation
With Alphaxalone-alphadalone, British Medical Journal, 2, 656-658, 1974.
30. Camps, J., Pizarro, I., Prats, E., La Ville, A., et al., Plasma Lipoprotein Alteration in
Patients With Chronic Hepatocellular Lever Disease Resulting From Alcohol Abuse:
Effects of Alcohol Intake Cessation, J Hepatol, 21(5) : 704-709, 1994.
31. Lin, R.C., Miller, B.A., Kelly, T.J., Concentration of Apolipoprotein AI, All, and E in
Plasm and Lipoprotein Fractions of Alcoholic Patients: Gender Differences in the
Effects of Alcohol, Hepatology, 21(4): 942-9, 1995.
32. Altmayer, P., Buech, U., Larson., R , Buech, H.P., Binding of Propofol to Native
Human Serum, Human Serum Albumen, and Human Hemoglobin, In tJ Clin
Pharmacol Ther Toxicol, 30: 269, 1992.
33. Barr, J., Propofol: A New Drug for Sedation in the Intensive Care Unit, International
Anesthesiology Clinics, 33(1): 131-154, 1995.
34. Brimacombe, J., An Extreme Case of Resistance to Anaesthetic Agents, Anaesthesia
and Intensive Care, 22(2): 236, 1994.
35. Du Calier, J., Dathis, F., Eledjam, J.J., Bonnet, M., C., Propofol et Ethylisme, Annales
Francoises d'Anesthesie et de Reanimation, 29: 332-333, 1987.
36. Warmington, A., Extreme Resistance to General Anasthetics, Anaesthesia and
Intensive Care, 22: 627, 1997.
37
37. Fassoulake, A, Farmotti, R,. Servin, F., Desmonts, J.M., Chronic Alcoholism Increases
the Inductiou Dose of Propofol in Human, Anesthesia and Analgesia, 77: 553-556,
1993.
38. Servin, F., Diprivan et foie, Annafes Francoises d'Anesthesie et de Reanimation 13:
477-479, 1994.
39. Holford, N.G.G., Sheiner, L., Understanding the Dose-Relationship: Clinical
Application of Pharmacokinetic-Pharmacodynamic Models, Clinical
Pharmacokinetics 6 :429-453, 1981.
38

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A physiologic model of granulopoiesis

PDF

Bayesian estimation using Markov chain Monte Carlo methods in pharmacokinetic system analysis

PDF

Three-dimensional functional mapping of the human visual cortex using magnetic resonance imaging

PDF

Auditory brainstem responses (ABR): variable effects of click polarity on auditory brainstem response, analyses of narrow-band ABR's, explanations

PDF

Effects of prenatal cocaine exposure in quantitative sleep measures in infants

PDF

Estimation of upper airway dynamics using neck inductive plethysmography

PDF

Cross-correlation methods for quantification of nonlinear input-output transformations of enural systems using a Poisson random test input

PDF

Auditory brainstem responses (ABR): quality estimation of auditory brainstem responsses by means of various techniques

PDF

Respiratory system impedance at the resting breathing frequency range

PDF

Head injury biomechanics: Quantification of head injury measures in rear-end motor vehicle collisions

PDF

A user interface for the ADAPT II pharmacokinetic/pharmacodynamic systems analysis software under Windows 2000

PDF

English phoneme and word recognition by nonnative English speakers as a function of spectral resolution and English experience

PDF

A preliminary investigation to determine the effects of a crosslinking reagent on the fatigue resistance of the posterior annulus of the intervertebral disc

PDF

Comparison of evacuation and compression for cough assist

PDF

Functional water MR spectroscopy of stimulated visual cortex using single voxel

PDF

A model of upper airway dynamics in obstructive sleep apnea syndrome

PDF

Comparisons of deconvolution algorithms in pharmacokinetic analysis

PDF

Cellular kinetic models of the antiviral agent (R)-9-(2-phosphonylmethoxypropyl)adenine (PMPA)

PDF

Processing And Visualization In Functional Magnetic Resonance Imaging (Fmri) Of The Human Brain

PDF

Pulse oximetry failure rates

Asset Metadata

Creator Doss, Monica G. (author)

Core Title Propofol Effects On Eeg And Levels Of Sedation

Degree Master of Science

Degree Program Biomedical Engineering

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag engineering, biomedical,Health Sciences, Pharmacology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor D'Argenio, David B. (committee chair), Kalaba, Robert (committee member), Khoo, Michael Chee-Kuan. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-13538

Unique identifier UC11356876

Identifier 1379578.pdf (filename),usctheses-c18-13538 (legacy record id)

Legacy Identifier 1379578-0.pdf

Dmrecord 13538

Document Type Thesis

Rights Doss, Monica G.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA