Anesthesiology 2004; 101:1077– 83 © 2004 American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc.

Dexmedetomidine Pharmacodynamics: Part II

Crossover Comparison of the Analgesic Effect of Dexmedetomidine and
Remifentanil in Healthy Volunteers
Luis I. Cortinez, M.D.,* Yung-Wei Hsu, M.D.,† Sam T. Sum-Ping, M.B., Ch.B.,‡ Christopher Young, M.D.,§
John C. Keifer, M.D.,§ David MacLeod, F.R.C.A.,§ Kerri M. Robertson, M.D., F.R.C.P(C),§ David R. Wright, F.R.C.A.,储
Eugene W. Moretti, M.D.,储 Jacques Somma, M.D., F.R.C.P.(C)#

Background: Dexmedetomidine is a highly selective ␣2-adre- DEXMEDETOMIDINE is a highly selective ␣2-adrenocep-

noceptor agonist used for short-term sedation of mechanically tor agonist used for short-term sedation of mechanically
ventilated patients. The analgesic profile of dexmedetomidine

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

has not been fully characterized in humans. ventilated patients in intensive care units. The combina-
Methods: This study was designed to compare the analgesic tion of its analgesic, sedative/hypnotic, and anxiolytic
responses of six healthy male volunteers during stepwise tar- properties added to its minimal effect on ventilation
get-controlled infusions of remifentanil and dexmedetomidine. make dexmedetomidine suitable for use in the perioper-
A computer-controlled thermode was used to deliver painful
heat stimuli to the volar side of the forearms of the subjects. Six
ative period.1
sequential 5-s stimuli (ranging from 41° to 50°C) were delivered The analgesic profile of dexmedetomidine has not
in random order. The recorded visual analog scale was used to been fully characterized in humans.2– 4 This study was
fit an Emax model. designed to further quantify the analgesic effect of
Results: Compared to baseline, remifentanil infusions re-
dexmedetomidine over a wide plasma concentration
sulted in a right shift of the sigmoid curve (increased T50, the
temperature producing a visual analog scale score of 50% of the range during intravenous infusion. The analgesic effect
maximal effect, from 46.1°C at baseline to 48.4° and 49.1°C of opioids is well characterized.5 Therefore, to validate
during remifentanil infusions) without a change of the steep- our methods and provide a clinical point of reference to
ness of the curve (identical Hill coefficients ␥ during baseline the effects measured with dexmedetomidine, we com-
and remifentanil). Compared to baseline, dexmedetomidine in-
fusions resulted in both a right shift of the sigmoid curve (in-
pared the pharmacodynamic effects of dexmedetomi-
creased T50 to 47.2°C) and a decrease in the steepness of the dine to remifentanil, a very short-acting opioid. We mea-
curve (decreased ␥ from 3.24 during baseline and remifentanil sured and compared respiratory, analgesic, and sedative
infusions to 2.45 during dexmedetomidine infusions). There responses of healthy male volunteers during (1) a step-
was no difference in the pain responses between baseline and
wise target-controlled infusion (TCI) of remifentanil,
after recovery from remifentanil infusions (identical T50 and ␥).
Conclusion: As expected, dexmedetomidine is not as effective (2) a stepwise TCI of dexmedetomidine, and (3) a
an analgesic as the opioid remifentanil. The difference in the pseudonatural sleep session. This article focuses on the
quality of the analgesia with remifentanil may be a reflection of analgesic effects of dexmedetomidine, whereas a com-
a different mechanism of action or a consequence of the seda- panion article examines the respiratory properties of
tive effect of dexmedetomidine.
dexmedetomidine.

This article is accompanied by an Editorial View. Please see:

䉬 Maze M, Angst MS: Dexmedetomidine and opioid interactions: Materials and Methods
Defining the role of dexmedetomidine for intensive care unit Institutional Review Board and
sedation. ANESTHESIOLOGY 2004; 101:1059 – 61.
Inclusion/Exclusion Criteria
After this study was approved by the Institutional Re-
* Fellow, Human Pharmacology Laboratory, Visiting Associate, Department of
view Board (Duke University Medical Center, Durham,
Anesthesiology, Duke University Medical Center. Current position: Attending North Carolina), signed informed consent was obtained
Anesthesiologist, Department of Anesthesiology, Catholic University School of
Medicine, Santiago, Chile. † Fellow, Human Pharmacology Laboratory, Depart-
from each study subject. Eight male subjects, aged
ment of Anesthesiology, Duke University Medical Center. Current position: 21– 40 yr, with American Society of Anesthesiologists
Attending Anesthesiologist, Department of Anesthesiology, Mackay Memorial
Hospital, Taiwan. ‡ Professor of Anesthesiology and Pain Management, Univer- physical status I, were enrolled. Subjects with a history
sity of Texas Southwestern Medical Center, Dallas, Texas. Chief of Anesthesiol- of drug, tobacco, or alcohol abuse; chronic use of med-
ogy, VA North Texas Health Care System, Dallas, Texas. § Associate Clinical
Professor, 储 Assistant Clinical Professor, Department of Anesthesiology, # Assis- ications; gastroesophageal reflux; anticipated difficult
tant Professor and Director, Human Pharmacology Laboratory, Department of airway; body mass index of less than 18 or greater than
Anesthesiology, Duke University Medical Center.
28 kg/m2; or the presence of a beard or physiognomies
Received from Duke University Medical Center, Department of Anesthesiol-
ogy, Human Pharmacology Laboratory, Durham, North Carolina. Submitted for precluding a good fit of a facemask were excluded. The
publication May 19, 2003. Accepted for publication June 25, 2004. Supported by subjects underwent a screening session during which a
Abbott Laboratories, Abbott Park, Illinois.
Address reprint requests to Dr. Somma: Human Pharmacology Laboratory, De-
physical examination, medical history, electrocardio-
partment of Anesthesiology, Duke University Medical Center, Box 3094, Durham, gram, and laboratory tests were performed. During the
North Carolina 27710. Address electronic mail to: [email protected].
Individual article reprints may be purchased through the Journal Web site,
screening session, the subjects were familiarized exten-
www.anesthesiology.org. sively with the study procedures. Pain assessments after

Anesthesiology, V 101, No 5, Nov 2004 1077

1078 CORTINEZ ET AL.

heat stimuli delivered by a computer-controlled ther-

mode were also performed during the screening.

Conduct of the Study

The protocol consisted of three parts, over 24 h. Dur-
ing parts 1 and 2, the subjects received remifentanil or
dexmedetomidine, respectively, via TCI. During part 3,
no drugs were infused, and no pain data were collected.
During parts 1 and 2, the stepwise infusions were de-
signed to target and maintain remifentanil or dexmedeto-
midine plasma concentrations at four sequentially in-

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

creasing steps. Each step lasted 40 min. Steps 1– 4
targeted remifentanil (part 1) plasma concentrations of
1, 2, 3, and 4 ng/ml and dexmedetomidine (part 2)
plasma concentrations of 0.6, 1.2, 1.8, and 2.4 ng/ml.
Remifentanil plasma concentrations were chosen (from
clinical experience) to be high enough to produce re-
spiratory depression without apnea requiring assisted
ventilation, whereas dexmedetomidine plasma concen-
trations were chosen to range from a therapeutic level to
supratherapeutic levels. The first step of dexmedetomi-
dine (0.6 ng/ml) is a typical level used for sedating
patients in the surgical intensive care unit. During part 1,
at the end of step 4, the remifentanil infusion was
stopped, and a 90-min recovery period was allowed
before beginning part 2 to ensure that the subject re-
turned to baseline.6 Similarly, the dexmedetomidine in-
fusion was stopped, during part 2 at the end of step 4,
and, because the pharmacokinetics of dexmedetomidine
Fig. 1. Sequence of data collection during each 40-min step. ABG ⴝ
is slower than that of remifentanil, a 240-min recovery arterial blood gas; AEP ⴝ auditory evoked potentials; CO2 ⴝ car-
period was allowed before beginning part 3.7,8 bon dioxide; EEG ⴝ electroencephalogram; OAA/S ⴝ Observer’s
One of the eight subjects received a placebo instead of Assessment of Alertness/Sedation; VAS ⴝ visual analog scale.
remifentanil, and another subject received a placebo
instead of dexmedetomidine. This was randomly as- Measurements
signed, and the investigators were blinded to it. The two At each step, pharmacodynamic data were collected
placebo subjects were excluded from analysis to per- according to figure 1.9 Blood samples were collected for
form a crossover comparison. pharmacokinetic analysis at 15– 40 min of each step to
The subjects fasted for 8 h before the study and were ensure that a quasi–steady state was achieved.
asked to abstain from caffeine and alcohol consumption
for the preceding 24 h. On arrival in the morning, an Heat-pain Stimulation
18-gauge intravenous cannula was inserted, and lactated A 1-cm2 computer-controlled Peltier-type thermal stim-
Ringer’s solution was infused at 100 ml/h. A 20-gauge ulator (TSA-II; Medoc Advanced Medical Systems, Min-
catheter was inserted into the radial artery of the non- neapolis, MN) was used to deliver painful heat stimuli to
dominant hand. The subjects received 30 ml oral sodium the volar side of the forearms of the subjects. Before any
citrate, 10 mg intravenous metoclopramide, and 50 mg stimulus was applied, the subjects were familiarized
intravenous ranitidine to minimize the risk of pulmonary with the test procedure and were instructed on how to
aspiration. Electroencephalographic electrodes, a three- evaluate their pain using both a standard 100-mm-long
lead electrocardiogram, a noninvasive blood pressure paper visual analog scale (VAS) that ranged from 0 (no
cuff, pulse oximeter probes, and Respitrace (Non-Inva- pain) to 100 (worst possible pain) and a computerized
sive Monitoring Systems Inc., North Bay Village, FL) VAS box (COVAS; Medoc Advanced Medical Systems)
bands were also applied. connected to the computer controlling the TSA-II. The
After all the monitors were placed and the facemask subjects then slid a cursor on the control box, indicating
was carefully adjusted, the lights were dimmed, and the the level of pain from 0 (no pain) to 100 (worst possible
subject was requested to rest quietly in the bed in the pain). The position of the cursor was read by the com-
decubitus position for 30 – 60 min. The protocol was puter. The volar side of both forearms was divided into
then initiated with baseline measurements. six zones (fig. 2) and marked with an indelible marker.

Anesthesiology, V 101, No 5, Nov 2004

ANALGESIC EFFECTS OF REMIFENTANIL/DEXMEDETOMIDINE 1079

the models were based on the classic sigmoidal Emax

model:

E ⫽ 100 ⫻ 冉 T␥
T50␥ ⫹ T␥ 冊
, (1)

where E is the predicted effect (VAS from 0 to 100) for

a given temperature (T), T50 is the temperature produc-
ing 50% of the maximal effect, and ␥, the Hill coefficient,
Fig. 2. Volar side of the forearm divided into six zones.
is a measure of the steepness of the response.
The model was constrained to temperatures of 37°C or
greater with the assumption that the baseline tempera-

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

The computer-controlled thermode system was pro-
ture T ⫽ 37°C would result in zero effect E ⫽ 0. The
grammed to sequentially deliver one of six different 5-s
models were built using the data set collected from six
stimuli of predefined temperatures (41.0°, 42.8°, 44.6°,
individuals (i ⫽ 1– 6) during eight steps, where steps j ⫽
46.4°, 48.2°, and 50.0°C). A single probe was used and
1– 8 respectively represent baseline, remifentanil steps
moved from one zone to another between the stimuli. At
1– 4, recovery from remifentanil infusions, and dexme-
each step, the six stimuli were presented in a double-
detomidine steps 1 and 2. During dexmedetomidine
blind and pseudorandom fashion, such that each of the
steps 3 and 4, the subjects were too sedated to be
six zones of one forearm received a different stimulus.
included in the analysis. Equation 1 was thus rewritten as
The right and left forearms were alternatively used from
one step to another. During the 25 s between stimuli,
the VAS assessments were obtained by one investigator
while the probe was repositioned to the following zone
Eij ⫽ 100 ⫻ 冋 共T ⫺ 37兲␥j
册
共T50, j ⫺ 37兲␥j ⫹ 共T ⫺ 37兲␥j
⫹ ␧ij , (2)

by a second investigator. A third investigator (unblinded where ⑀ij represents the six residual errors of the ith
to the stimuli) operated the computer. The thermode individual during the jth steps for the six temperatures
was maintained at 37°C between stimuli. If the painful (T ⫽ 41.0°, 42.8°, 44.6°, 46.4°, 48.2°, and 50.0°C).
stimulus was not tolerated, it was stopped, and pain The initial model (equation 2) included eight ␥s and
scores were assessed with both methods. If the volun- eight T50s (one for each step). The process of model
teer was too sedated for pain score assessment, heat building consisted, in part, of finding which of these 16
stimuli were applied, the VAS and computerized VAS parameters were not needed. For example, all of the ␥s
were recorded as “unable to assess,” and withdrawal of during remifentanil steps (␥2, ␥3, ␥4, and ␥5) may be
the tested forearm was noted as indicated. identical and replaced by a single ␥Remi.
The quality of the fit was assessed by the values of the
Monitoring and Equipment Akaike Information Criteria, the magnitude of the stan-
A customized data acquisition system based on a Lab- dard errors on the parameters estimates, visual examina-
View platform (version 6.0; National Instruments, Aus- tion of the model fit to the raw data, and visual exami-
tin, TX) was used to collect vitals signs, extensive respi- nation of the residual plot.11
ratory variables, and the electrocardiogram. The
materials and methods of that data collection have been Results
described in detail by Hsu et al.9
The TCI pump devices consisted of two laptop com- Pharmacokinetics and Sedation
puters connected to infusion pumps (Harvard Pump 22; While remifentanil target plasma concentrations were
Harvard Apparatus, South Natick, MA). STANPUMP** 1, 2, 3, and 4 ng/ml, measured plasma concentrations
was used to run the TCI pumps. The pharmacokinetic were 0.78 ⫾ 0.19, 1.70 ⫾ 0.45, 2.25 ⫾ 0.52, and 3.12 ⫾
parameters used for the infusions of remifentanil and 1.28 ng/ml. Dexmedetomidine target plasma concentra-
dexmedetomidine were those published by Minto et tions were 0.6, 1.2, 1.8, and 2.4 ng/ml, and measured
al.10 and Dyck et al.,7 respectively. corresponding plasma concentrations were 0.67 ⫾ 0.07,
1.72 ⫾ 0.18, 2.81 ⫾ 0.20, and 3.78 ⫾ 0.36 ng/ml. For
both remifentanil and dexmedetomidine, there were no
Data Analysis
statistical differences between the first and second sam-
Nonlinear mixed effects models were used to analyze
ples drawn within each step.
the analgesic effect of dexmedetomidine and remifen-
Figure 3 shows the sedation assessments measured
tanil with S-PLUS (Insightful Corp, Seattle, WA).11 All of
with the Observer’s Assessment of Alertness/Sedation
sum.12 The scale ranges from 9 (completely unrespon-
** STANPUMP copyright S. L. Shafer, Palo Alto Department of Veterans
Affairs Medical Center, Palo Alto, California; software available at
sive) to 20 (awake and not sedated). The subjects were
http://anesthesia.stanford.edu/pkpd/. Accessed September 19, 2004. minimally sedated during remifentanil infusions. During

Anesthesiology, V 101, No 5, Nov 2004

1080 CORTINEZ ET AL.

dexmedetomidine infusion steps 1 and 2, the subjects

were sedated and arousable. In contrast, during steps 3
and 4, deeper levels of hypnosis were attained, and most
subjects were completely unarousable (four of the six
subjects having received dexmedetomidine).

Pain Model
Because most subjects were completely unarousable
during steps 3 and 4 of dexmedetomidine infusions, a
VAS score could not be obtained, and no modeling was
performed during these steps. However, of the four

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

subjects that were unarousable, three of them consis-
tently withdrew their arm when the higher heat stimuli
(T ⫽ 46.4°–50.0°C) were applied. The model building
steps are summarized in table 1, and the parameters of
the final model are summarized in table 2. The two
pivotal models, model 2 and the final one, model 7, are
Fig. 3. Observer’s Assessment of Alertness/Sedation (OAA/S) sum
scores (mean ⴞ SD) at baseline and at different remifentanil illustrated in figure 4. There was no difference in the
(REMI) and dexmedetomidine (DEX) steps. The scale ranges from pain response between baseline and recovery from
9 to 20 (maximum sedation to maximum alertness). remifentanil infusions (identical T50 and ␥). As seen in
figure 4, remifentanil infusions resulted in a right shift of
the sigmoid curve (increased T50 when compared with

Table 1. Nonlinear Mixed-effect Models

Number of Fixed
Model Fixed Effects Effects AIC Next Action

1 One ␥ and one T50% for each Eight ␥s and eight 2349.296 Considering similarities in their ␥s, one ␥ was
stage T50%s used for baseline and recov1, one ␥ was used
for all REMI stages, and one ␥ was used for the
two DEX stages.
2 One ␥ for baseline and Three ␥s and eight 2344.955 Considering that the ␥ values for baseline-
recovery, one ␥ for REMI, T50%s recovery and REMI were similar, only one ␥
and one ␥ for DEX; one was used for them.
T50% for each stage
3 One ␥ for baseline, recov1, and Two ␥s and eight 2342.965 Although different ␥ values were observed
REMI and one ␥ for DEX; T50%s between the two ␥s, a model with only one ␥
one T50% for each stage was further tested.
4 One ␥ for baseline, recov1 One ␥ and eight 2343.934 We kept two ␥s in our model based on the AIC
REMI, and DEX; one T50% T50%s criteria and the statistical differences between
for each stage them. The next step was to use the same T50%
for baseline and recov1 based on their
similarities.
5 One ␥ for baseline, recov1, and Two ␥s and seven 2341.151 Considering the similarities of the T50% values for
REMI and one ␥ for DEX; T50%s REMI2, 3, and 4, they were modeled with only
one common T50% for one T50%.
baseline and recov1
6 One ␥ for baseline, recov1, and Two ␥s and five 2337.283 Considering the similarities of the T50% values for
REMI, and one ␥ for DEX; T50%s DEX1 and 2, they were modeled with only one
one common T50% for T50%.
baseline and recov1 and one
common T50% for REMI2, 3,
and 4
7 ␥ for baseline, recov1, and Two ␥s and four 2333.294 We kept this model as our final model.
REMI and one ␥ for DEX; T50%s
one common T50% for
baseline and recov1; one
common T50% for REMI2, 3,
and 4 and one common
T50% for DEX1 and DEX2

AIC ⫽ Akaike Information Criteria; DEX ⫽ dexmedetomidine; ␥ ⫽ Hill coefficient; recov1 ⫽ recovery of remifentanil infusions; REMI ⫽ remifentanil; T50% ⫽
temperature producing a visual analog scale score of 50%.

Anesthesiology, V 101, No 5, Nov 2004

ANALGESIC EFFECTS OF REMIFENTANIL/DEXMEDETOMIDINE 1081

Table 2. Final Values of the Fixed Effects with Their Statistical Eisenach et al.17 used a heat pain experimental method
Significance in the Model plotting the pain–response as a function of temperature
Mean SE P Value (graph similar to figs. 4 and 5). Although their plot is similar
to ours, they only used it at baseline and did not model the
␥Base-Recov-Remi 3.24 0.23 ⬍ 0.0001 stimulus–response.
␥DEX 2.45 0.37 0.03
T50 Base-Recov 46.1 0.49 ⬍ 0.0001 In contrast, our approach, in human subjects, using an
T50 Remi1 48.4 0.53 ⬍ 0.0001 Emax model, analyzed the pain response as a function of
T50 Remi234 49.1 0.40 ⬍ 0.0001 variable–intensity stimuli. Although the Emax model has
T50 Dex 47.2 0.45 0.04
previously been used for pain analysis,18 –23 these studies
␥Base-Recov-Remi ⫽ Hill coefficient for baseline, recovery, and remifentanil modeled dose–responses where the pain–responses are
infusions; ␥DEX ⫽ Hill coefficient for dexmedetomidine infusions;
T50 Base-Recov ⫽ temperature producing a visual analog scale (VAS) score of

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

50% during baseline and during recovery from remifentanil infusions; T50 Dex
⫽ temperature producing a VAS score of 50% during dexmedetomidine
infusions; T50 Remi1 ⫽ temperature producing a VAS score of 50% during
remifentanil infusion step 1; T50 Remi 234 ⫽ temperature producing a VAS score
of 50% during remifentanil infusion steps 2, 3, and 4.

baseline), without a change of the steepness of the curve

(identical ␥s during baseline and remifentanil). Dexme-
detomidine infusions resulted in both a right shift of the
sigmoid curve (increased T50 when compared to base-
line) and a decrease in the steepness of the curve (sig-
nificantly smaller ␥ during dexmedetomidine when com-
pared with baseline).
Finally, an example of the raw data are found in figure 5.

Discussion
This study, investigating the analgesic effects of
dexmedetomidine and remifentanil, adds the following
three pieces of information to the current literature.
First, a new approach for experimental pain analysis is
presented by modeling VAS responses resulting from a
wide range (41°–50°C) of heat painful stimuli. Second,
the analgesic response of dexmedetomidine is character-
ized and compared with remifentanil using the experi-
mental heat pain model. Third, we demonstrated a sig-
nificantly different shape of the pain response during
dexmedetomidine infusions when compared with base-
line, recovery, and remifentanil infusions.

Approach to Pain Study Analysis

In most experimental pain studies, analgesic effects are Fig. 4. Population predictions according to model 2 for all the
steps (A). This model includes eight T50s, the temperature pro-
modeled using pain threshold, pain tolerance, or both.13,14 ducing a visual analog scale (VAS) score of 50% of the maximal
Hence, only one or two data points of the dose–response effect (one for each step), and three Hill coefficients (␥s) (one
curve (intensity of the pain stimuli vs. VAS response) are for baseline and recovery [Baseline–Recovery], one for all
remifentanil [REMI] steps, and one for all dexmedetomidine
considered. However, in certain conditions, there is a dif- [DEX] steps). The responses are clustered in three groups: Base-
ferential drug effect between pain threshold and pain tol- line–Recovery, REMI, and DEX infusions. Although not con-
erance.14 In addition, pain threshold has been shown to be strained to be identical both, ␥s for Baseline–Recovery and
REMI turned out to be identical. However, ␥ for DEX is smaller,
increased by pure hypnotic, and pain tolerance is thought resulting in a flatter response during DEX infusions when com-
to be more reliable in detecting true analgesic effects.13 pared with REMI or Baseline–Recovery. On this model, it can be
Few investigators have modeled the pain response in seen that Baseline and Recovery are identical, as well as DEX1
and DEX2. Although REMI1– 4 are clustered and almost identi-
function of variable–intensity of the stimulation. Morin and cal, it can be seen that REMI1 is apart from REMI2– 4. These
Bushnell15 modeled the entire stimuli–response but used subjective findings are statistically confirmed, and model 7 (B)
linear regression. Neugebauer and Li16 used an approach was selected as the final model. This model includes four T50s
(one for Baseline–Recovery, one for REMI1, one for REMI2– 4,
similar to ours (logistic equation) to model the sigmoid and one for DEX1 and 2) and two ␥s (one for Baseline–Recovery
stimulus–response curves in anesthetized rats. Finally, and REMI, and one for all DEX steps).

Anesthesiology, V 101, No 5, Nov 2004

1082 CORTINEZ ET AL.

imately equivalent to 2 ng/ml. From our clinical perspec-

tive, a ceiling effect of an opioid is surprising. This
apparent ceiling effect may have resulted from three
factors. First, the doses used were relatively small, con-
strained by the need to maintain spontaneous breathing.
Significantly higher doses (resulting in apnea) would
likely cause more profound analgesic effects. Second,
the current study is likely underpowered to detect subtle
changes in analgesia. Finally, although it is controver-
sial,26 the possibility of acute tolerance should be
mentioned.27

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

The observed analgesic effects of dexmedetomidine in
this study correlate well with the findings of both animal
and human studies. Animal studies have shown signifi-
cant analgesic effect after systemic administration of
clonidine or dexmedetomidine using thermal pain mod-
Fig. 5. Population predictions (line) and observed pain response
(circles) during baseline and recovery (after remifentanil, before
els.28,29 Jaakola et al.30 evaluated the analgesic effect of
dexmedetomidine). Both steps are modeled together (same fixed systemic administration of dexmedetomidine (0.25,
effects) according to our final model. VAS ⴝ visual analog scale. 0.50, and 1 ␮g/kg) and fentanyl (2 ␮g/kg) in healthy
volunteers, and they demonstrated a moderate analgesic
measured for a single-intensity stimulus in function of effect of dexmedetomidine with a ceiling effect at
variable drug concentrations or doses. Rather than ex- 0.5 ␮g/kg. This is equivalent to our first step of dexme-
amining a single point (e.g., pain threshold or pain tol- detomidine infusion (0.6 ng/ml). In contrast, Ebert et al.3
erance), we have modeled two parameters (T50 and ␥) used a cold pressor pain model to demonstrate a strong
that provide not only a measure of the shift of the curve dose-dependent analgesic effect of dexmedetomidine,
but also its shape. Two different drugs that shift the with no ceiling effect up to plasma concentrations of 8.4
curve equivalent amount (T50) but with different curve ng/ml. The absence of a ceiling effect in this study may be
shapes (␥) could suggest different mechanisms of action. explained by the use of a different pain model. Another
possible explanation is the fact that the painful stimulus of
Remifentanil and Dexmedetomidine Analgesic the cold pressor test could be related to peripheral vaso-
Effect constriction, whereas dexmedetomidine significantly mod-
In the current study, analgesic effects were docu- ulates peripheral vasoconstriction.1,31 However, the study
mented during both remifentanil and dexmedetomidine of Fuchs et al.32 showed that intradermal injection of ad-
infusions. A drug dose effect was identified with remifen- renergic agonists (norepinephrine and phenylephrine) re-
tanil, with an apparent ceiling effect at 2 ng/ml. A ceiling sulted in heat hyperalgesia, whereas injection of nonadren-
effect (or absence of drug dose effect) was possibly ergic vasoconstrictors (angiotensin II and vasopressin) did
observed with dexmedetomidine since no increase of not result in heat hyperalgesia, which suggests that adren-
analgesia was observed by increasing the dose of dexme- ergic-mediated mechanisms may play a role in the sensiti-
detomidine from step 1 to step 2. This impression of a zation of heat nociceptors.
ceiling effect is also supported by the fact that three of
the four unarousable subjects consistently withdrew Difference in the Shape of the Stimuli–Response
their arm when the higher heat stimuli (T ⫽ 46.4°– Curves
50.0°C) were applied. In addition, the magnitude of the Another interesting finding of this study is the effect of
analgesic effect of dexmedetomidine is smaller than that dexmedetomidine on the shape of the stimuli– dose re-
observed with remifentanil, which is consistent with the sponse. Remifentanil infusions resulted in an expected
clinical notion that the analgesic property of ␣2 agonists increase of T50 and an absence of change in the Hill
is not as effective as that of opioids. coefficient ␥, whereas dexmedetomidine infusions re-
There are little data on the effect of remifentanil on sulted in both an increase in T50 and a decrease in the
experimental pain.24,25 Gustorff et al.25 used the quanti- Hill coefficient ␥. Potential explanations of the flattened
tative sensory testing method on the heat pain threshold response include the sedative effect of dexmedetomi-
in volunteers. They derived an E50 of 0.05 ␮g · kg⫺1 · dine and a different mechanism of analgesic action of
min⫺1 for remifentanil, which is approximately equiva- dexmedetomidine.
lent to a plasma concentration of 1.2 ng/ml. Although It is known that pain perception is decreased when
our methodology is different, our results are in accord subjects are distracted from the painful stimulus.33,34 It has
with their study, because they also observed an apparent also been proposed that the adrenergic system may be
ceiling effect at 0.09 ␮g · kg⫺1 · min⫺1, which is approx- involved in cognitive modulation of pain.33 The sedative

Anesthesiology, V 101, No 5, Nov 2004

ANALGESIC EFFECTS OF REMIFENTANIL/DEXMEDETOMIDINE 1083

effect of dexmedetomidine may have resulted in a similar 10. Minto CF, Schnider TW, Shafer SL: Pharmacokinetics and pharmacody-
namics of remifentanil: II. Model application. ANESTHESIOLOGY 1997; 86:24 –33
phenomenon. 11. Pinheiro J, Bates D: Mixed-Effects Models in S and S-PLUS. New York,
Springer-Verlag, 2000
12. Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson AB,
Limitation of the Study Schwam EM, Siegel JL: Validity and reliability of the Observer’s Assessment of
The pharmacokinetic profile of dexmedetomidine pre- Alertness/Sedation Scale: Study with intravenous midazolam. J Clin Psychophar-
macol 1990; 10:244 –51
vented randomization of the sequence of administration 13. Luginbuhl M, Schnider TW, Petersen-Felix S, Arendt-Nielsen L, Zbinden
of the two drugs, and this is a limitation of the study. AM: Comparison of five experimental pain tests to measure analgesic effects of
alfentanil. ANESTHESIOLOGY 2001; 95:22–9
Although the development of acute opioid tolerance 14. Enggaard TP, Poulsen L, Arendt-Nielsen L, Hansen SH, Bjornsdottir I, Gram
after remifentanil infusion is still controversial,26,27,35–38 LF, Sindrup SH: The analgesic effect of codeine as compared to imipramine in
different human experimental pain models. Pain 2001; 92:277– 82
it must be carefully considered because cross-tolerance 15. Morin C, Bushnell MC: Temporal and qualitative properties of cold pain
between opioids and ␣2 agonists has been reported.39,40 and heat pain: A psychophysical study. Pain 1998; 74:67–73
16. Neugebauer V, Li W: Processing of nociceptive mechanical and thermal

Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/101/5/1077/356944/0000542-200411000-00006.pdf by guest on 15 February 2024

The return of the pain response curve to baseline after information in central amygdala neurons with knee-joint input. J Neurophysiol
recovery from remifentanil infusions strongly suggests 2002; 87:103–12
17. Eisenach JC, Hood DD, Curry R: Relative potency of epidural to intrathecal
an absence of acute tolerance in our study. clonidine differs between acute thermal pain and capsaicin-induced allodynia.
Pain 2000; 84:57– 64
18. Rodriguez M, Carlos MA, Ortega I, Suarez E, Calvo R, Lukas JC: Sex
Conclusion specificity in methadone analgesia in the rat: A population pharmacokinetic and
pharmacodynamic approach. Pharm Res 2002; 19:858 – 67
The analgesic effects of both remifentanil and dexme- 19. Gupta S, Modi NB, Sathyan G, Ho Pl PL, Aarons L: Pharmacokinetics of
controlled-release verapamil in healthy volunteers and patients with hyperten-
detomidine infusions were demonstrated. As expected, sion or angina. Biopharm Drug Dispos 2002; 23:17–31
the analgesic property of dexmedetomidine was less 20. Chen Y, Chan SY, Ho PC: Isobolographic analysis of the analgesic inter-
actions between ketamine and tramadol. J Pharm Pharmacol 2002; 54:623–31
effective than that of remifentanil, with the clinical im- 21. Carlos MA, Du Souich P, Carlos R, Suarez E, Lukas JC, Calvo R: Effect of
plication that dexmedetomidine could not replace the omeprazole on oral and intravenous RS-methadone pharmacokinetics and phar-
macodynamics in the rat. J Pharm Sci 2002; 91:1627–38
use of opioids. However, dexmedetomidine exhibited a 22. Kietzmann D, Bouillon T, Hamm C, Schwabe K, Schenk H, Gundert-Remy
qualitatively different analgesic effect as shown by a de- U, Kettler D: Pharmacodynamic modelling of the analgesic effects of piritramide
in postoperative patients. Acta Anaesthesiol Scand 1997; 41:888 –94
crease in the pain response slope. The effect of dexmedeto- 23. Hasselstrom J, Svensson JO, Sawe J, Wiesenfeld-Hallin Z, Yue QY, Xu XJ:
midine on the shape of the analgesic response should be Disposition and analgesic effects of systemic morphine, morphine-6-glucuronide
and normorphine in rat. Pharmacol Toxicol 1996; 79:40 – 6
further investigated because it may help to identify the 24. Petersen KL, Jones B, Segredo V, Dahl JB, Rowbotham MC: Effect of remifen-
clinical settings in which the analgesic properties of tanil on pain and secondary hyperalgesia associated with the heat– capsaicin sensi-
tization model in healthy volunteers. ANESTHESIOLOGY 2001; 94:15–20
dexmedetomidine can be used most effectively. 25. Gustorff B, Felleiter P, Nahlik G, Brannath W, Hoerauf KH, Spacek A, Kress
HG: The effect of remifentanil on the heat pain threshold in volunteers. Anesth
The authors thank D. Ryan Cook, M.D. (Professor of Anesthesiology, Duke University Analg 2001; 92:369 –74
Medical Center, Durham, North Carolina), and Richard Moon (Medical Director, Hyper- 26. Cortinez LI, Brandes V, Munoz HR, Guerrero ME, Mur M: No clinical
baric Center, and Professor of Anesthesiology, Duke University Medical Center, Durham, evidence of acute opioid tolerance after remifentanil-based anaesthesia. Br J
North Carolina), as well as Charles H. McLeskey, M.D. (Global Medical Director, Anes- Anaesth 2001; 87:866 –9
thesia/Sedation, Abbott Laboratories, Inc., Abbott Park, Illinois), and Victor S. B. Jorden, 27. Vinik HR, Kissin I: Rapid development of tolerance to analgesia during
M.D. (Associate Medical Director, Abbott Laboratories, Inc., Abbott Park, Illinois, and remifentanil infusion in humans. Anesth Analg 1998; 86:1307–11
Clinical Assistant Professor, Department of Anesthesiology, The Chicago Medical School, 28. Willenbring S, Stevens CW: Thermal, mechanical and chemical peripheral
Chicago, Illinois), for their support and critical review of the manuscript. sensation in amphibians: Opioid and adrenergic effects. Life Sci 1996; 58:125–33
29. Kayser V, Desmeules J, Guilbaud G: Systemic clonidine differentially mod-
ulates the abnormal reactions to mechanical and thermal stimuli in rats with
References peripheral mononeuropathy. Pain 1995; 60:275– 85
30. Jaakola ML, Salonen M, Lehtinen R, Scheinin H: The analgesic action of
1. Maze M, Tranquilli W: ␣2 Adrenoceptor agonists: Defining the role in dexmedetomidine—a novel alpha 2-adrenoceptor agonist—in healthy volun-
clinical anesthesia. ANESTHESIOLOGY 1991; 74:581– 605 teers. Pain 1991; 46:281–5
2. Eisenach JC, Lavand’homme P, Tong C, Cheng JK, Pan HL, Virtanen R, 31. Maze M, Scarfini C, Cavaliere F: New agents for sedation in the intensive
Nikkanen H, James R: Antinociceptive and hemodynamic effects of a novel care unit. Crit Care Clin 2001; 17:881–97
␣2-adrenergic agonist, MPV-2426, in sheep. ANESTHESIOLOGY 1999; 91:1425–36 32. Fuchs PN, Meyer RA, Raja SN: Heat, but not mechanical hyperalgesia,
3. Ebert TJ, Hall JE, Barney JA, Uhrich TD, Colinco MD: The effects of following adrenergic injections in normal human skin. Pain 2001; 90:15–23
increasing plasma concentrations of dexmedetomidine in humans. ANESTHESIOL- 33. Villemure C, Bushnell MC: Cognitive modulation of pain: How do atten-
OGY 2000; 93:382–94 tion and emotion influence pain processing? Pain 2002; 95:195–9
4. Hall JE, Uhrich TD, Barney JA, Arain SR, Ebert TJ: Sedative, amnestic, and 34. Miron D, Duncan GH, Bushnell MC: Effects of attention on the intensity
analgesic properties of small-dose dexmedetomidine infusions. Anesth Analg and unpleasantness of thermal pain. Pain 1989; 39:345–52
2000; 90:699 –705 35. Luginbuhl M, Gerber A, Schnider TW, Petersen-Felix S, Arendt-Nielsen L,
5. Leino K, Mildh L, Lertola K, Seppala T, Kirvela O: Time course of changes Curatolo M: Modulation of remifentanil-induced analgesia, hyperalgesia, and
in breathing pattern in morphine- and oxycodone-induced respiratory depres- tolerance by small-dose ketamine in humans. Anesth Analg 2003; 96:726 –32
sion. Anaesthesia 1999; 54:835– 40 36. Gustorff B, Nahlik G, Hoerauf KH, Kress HG: The absence of acute tolerance
6. Egan TD, Lemmens HJ, Fiset P, Hermann DJ, Muir KT, Stanski DR, Shafer SL: during remifentanil infusion in volunteers. Anesth Analg 2002; 94:1223– 8
The pharmacokinetics of the new short-acting opioid remifentanil (GI87084B) in 37. Guignard B, Bossard AE, Coste C, Sessler DI, Lebrault C, Alfonsi P, Fletcher
healthy adult male volunteers. ANESTHESIOLOGY 1993; 79:881–92 D, Chauvin M: Acute opioid tolerance: Intraoperative remifentanil increases
7. Dyck JB, Maze M, Haack C, Azarnoff DL, Vuorilehto L, Shafer SL: Computer- postoperative pain and morphine requirement. ANESTHESIOLOGY 2000; 93:409 –17
controlled infusion of intravenous dexmedetomidine hydrochloride in adult 38. Schraag S, Checketts MR, Kenny GN: Lack of rapid development of opioid
human volunteers. ANESTHESIOLOGY 1993; 78:821– 8 tolerance during alfentanil and remifentanil infusions for postoperative pain.
8. Scheinin H, Karhuvaara S, Olkkola KT, Kallio A, Anttila M, Vuorilehto L, Anesth Analg 1999; 89:753–7
Scheinin M: Pharmacodynamics and pharmacokinetics of intramuscular dexme- 39. Post C, Archer T, Minor BG: Evidence for crosstolerance to the analgesic
detomidine. Clin Pharmacol Ther 1992; 52:537– 46 effects between morphine and selective alpha 2-adrenoceptor agonists. J Neural
9. Hsu WY, Cortinez LI, Robertson KM, Keifer JC, Sum-Ping ST, Moretti EW, Transm 1988; 72:1–9
Young CC, Wright DR, McLeod DB, Somma J: Dexmedetomidine pharmacody- 40. Paalzow G: Development of tolerance to the analgesic effect of clonidine
namics: I. Crossover comparison of the respiratory effects of dexmedetomidine in rats: Cross-tolerance to morphine. Naunyn Schmiedebergs Arch Pharmacol
and remifentanil in healthy volunteers. ANESTHESIOLOGY 2004; 101:1066 –76 1978; 304:1– 4

Anesthesiology, V 101, No 5, Nov 2004