Journal of Critical Care (2009) 24, 628.e7628.

e17

Behavioral and physiologic indicators during a nociceptive

procedure in conscious and unconscious mechanically
ventilated adults: Similar or different?
Cline Glinas RN, PhD , Caroline Arbour RN, BSc, MScA(c)
School of Nursing, McGill University, Montreal, Quebec, Canada H3A 2A7
Center for Nursing Research, Jewish General Hospital, Montreal, Quebec, Canada

Keywords:
Behaviors;
Physiologic indicators;
Mechanically ventilated
patients;
Adult population;
Critical care

Abstract
Purpose: The purpose of this study was to describe behavioral and physiologic indicators to a
nociceptive procedure in conscious and unconscious mechanically ventilated adults and to examine their
association with the patients' self-reports of pain.
Materials and methods: A total of 257 intensive care unit patients (144 conscious and 113 unconscious)
from 4 clinical settings in Canada participated. Patients were observed at rest, during a nociceptive
procedure, and 20 minutes postprocedure. Behavioral indicators were measured with the Critical Care Pain
Observation Tool (CPOT), and physiologic indicators were documented from the available monitoring.
Results: Conscious patients showed more intense behaviors and a more elevated mean arterial pressure
compared with unconscious patients at all measurements. Conscious patients who self-reported being in pain
during the procedure showed more intense behaviors compared with patients without pain. Only the CPOT
score could predict the presence or the absence of pain on the basis of the patient's self-report during the
procedure.
Conclusions: The use of behaviors is strongly recommended for pain assessment in unconscious patients,
and results from this study support this clinical guideline. Vital signs should be used with caution for the
detection of pain as they can be influenced by other factors besides pain.
2009 Elsevier Inc. All rights reserved.

1. Introduction
Moderate to severe pain is experienced by critically ill
patients in the intensive care unit (ICU) [1]. Although the
patient's self-report remains the most valid measure for pain,
many patients are unable to provide it because of the use of
Corresponding author. School of Nursing, McGill University,
Montreal, Quebec, Canada H3A 2A7. Tel.: +1 514 398 6157.
E-mail address: celine.gelinas@mcgill.ca (C. Glinas).
0883-9441/$ see front matter 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcrc.2009.01.013

sedative agents, mechanical ventilation, and change in the

level of consciousness [2]. For this reason, pain assessment
methods often need to be adapted to the patient communication capabilities of the patient. In nonverbal patients unable to
self-report, the use of behavioral or physiologic indicators are
strongly recommended for the detection of pain [3-5].
As many ICU patients are unable to self-report their pain,
it is relevant to document behaviors and physiologic
responses to pain. Although some studies have been focused
on the validation of behavioral scale total scores in critically
ill nonverbal adults [6-10], little research explored the

628.e8
specific behavioral reactions to pain in this population.
Indeed, pain behaviors during nociceptive procedures in
critically ill patients able to self-report were explored in one
international study [11]. According to physiologic indicators,
some vital signs (eg, heart rate [HR], mean arterial pressure
[MAP]) were previously studied in critically ill verbal
nonverbal patients [8-10]. The aim of this study was to
document both behavioral and physiologic indicators during
a nociceptive procedure in conscious and unconscious
mechanically ventilated adults. Besides being innovative,
this study is important as it could contribute to validate
nonverbal indicators of the presence of pain that could be
used to assess pain and to evaluate the effectiveness of pain
management protocols in this vulnerable adult population.

2. Background
Although behavioral indicators are strongly recommended for pain assessment in nonverbal patients [3,5],
few tools have been developed and tested in critically ill
adults as follows: Behavioral Pain Scale (BPS) [9], Critical
Care Pain Observation Tool (CPOT) [7], PACU-BPRS [12],
and PAIN [13]. From those tools, both the BPS and the
CPOT have been suggested by experts for use in
uncommunicative critically ill adults [14,15] and in the
clinical practice recommendations of a Task Force of the
American Society for Pain Management Nursing [5].
The Behavioral Pain Scale [9] and the CPOT [7] are 2
instruments that were validated in critically ill nonverbal
adults. The Behavioral Pain Scale includes 3 behaviors: (1)
facial expression, (2) movements of upper limbs, and (3)
compliance with the ventilator. Each behavior is rated on a
scale from 1 to 4 for a possible total score from 3 to 12. The
CPOT includes 4 behaviors: (1) facial expression, (2) body
movements, (3) compliance with the ventilator, and (4) muscle
tension. Each behavior is rated from 0 to 2 for a possible total
score ranging from 0 to 8. The process for item selection and
content validity of the CPOT were previously described [16].
The Behavioral Pain Scale was tested in 3 different samples
of unconscious ICU patients with different diagnosis (trauma,
surgical, or medical) for a total of 104 patients [6,9,10]. In all 3
studies, higher BPS scores were obtained with nociceptive
procedures (eg, turning, endotracheal suctioning) compared
with rest or nonnociceptive procedures (eg, compression
stocking application, central venous catheter dressing change,
eye care). So, it can be concluded that patients express more
intense behaviors during procedures known to cause pain.
However, because results from those studies focused on the
total score of the scale, the specific behaviors expressed by the
patients during the nociceptive procedures were not described.
The Critical Care Pain Observation Tool was validated
with 2 different samples of ICU patients, able to self-report
or not. A total of 160 patients with various diagnosis (trauma,
surgical, or medical) participated [7,8]. Again, in these 2

C. Glinas, C. Arbour
studies, higher CPOT scores were obtained with a nociceptive procedure (turning with or without suctioning) compared with rest or a nonnociceptive procedure (noninvasive
blood pressure). Interestingly, among mechanically ventilated patients, those who were conscious obtained higher
CPOT scores compared with those who were unconscious.
Moreover, for patients who were able to self-report, higher
CPOT scores were observed for those who had pain
compared with those who did not experience pain during
the nociceptive procedure. Once again, only CPOT total
scores were reported in these studies not allowing to describe
the specific behaviors expressed by the patients.
Besides the validation of those 2 behavioral scales, some
physiologic indicators were documented. In the study of
Payen and colleagues [9], MAP and HR were found to
increase with nociceptive procedures compared with rest.
Similar results were found in the study of Glinas and
Johnston [8]. Mean arterial pressure, HR, and respiratory rate
(RR) significantly increased, whereas SpO2 decreased with
the nociceptive procedure compared with the nonnociceptive
procedure. On the opposite, findings from the study of Young
and colleagues [10] showed that the hemodynamic parameters (blood pressure, HR) increased both during nociceptive and nonnociceptive procedures.
Results from the study of the Thunder Project II published
by Puntillo and colleagues [11] allowed to describe the
specific behavioral reactions of 5957 critically ill adults able
to self-report during different nociceptive procedures (eg,
turning, wound drain removal, wound care, tracheal suctioning). Grimacing, rigidity, wincing, shutting of eyes, and
verbal complaints were exhibited by at least 20% of patients
who reported procedural pain. On average, there were
significantly more behaviors exhibited by patients with vs
without procedural pain. Patients with pain exhibited an
average of 3.5 behaviors, whereas patient without pain
exhibited an average of 1.8 behaviors. Patients with pain were
2.8 to 10.3 times more likely to have increased facial, body
movement, or verbal responses during the procedure than
patients without pain. Although this large international study
was the first to identify pain-related behaviors in patients able
to self-report, future studies are warranted to explore these
behaviors in critically ill patients unable to self-report.

3. Study aim and objectives

The aim of this study was to examine the behavioral
indicators, using the CPOT [7], and physiologic signs (ie,
MAP, HR, RR, SpO2, and end-tidal carbon dioxide [CO2]) in
conscious and unconscious mechanically ventilated adults
with different diagnosis during a nociceptive procedure. The
following specific objectives were achieved:
1. Describe and compare the behavioral and physiologic
indicators in mechanically ventilated adults according

Behavioral and physiologic indicators in mechanically ventilated adults

to their level of consciousness (conscious vs unconscious) and to their diagnosis at admission in the ICU.
2. Describe and compare the behavioral and physiologic
indicators in conscious patients who had pain and those
who had no pain during the nociceptive procedure.
3. Examine the relationship between behavioral/physiologic behaviors and the conscious patient's self-report
of pain during the nociceptive procedure.
4. Explore which indicators best predict the presence of
pain in mechanically ventilated conscious adults able to
self-report.

628.e9

Quebec in Canada agreed to participate in this study. It must

be noted that complementary data collected from 154
patients who participated in the validation of the CPOT
previously published [7,8] were used for this study. The
principal investigator (CG) was the coordinator of all sites.
The patients recruited met the following inclusion criteria:
(1) age of 18 years and older, (2) admitted at the ICU and
mechanically ventilated, and (3) conscious or unconscious
(Glasgow Coma Scale 8). Quadriplegic patients, those
receiving neuromuscular blocking agents, or being investigated for brain death were excluded.

4.2. Ethical considerations

4. Method
4.1. Study design and sample
A descriptive-correlational design was chosen for this
study. A convenience sample of 144 conscious patients and
113 unconscious patients (see Table 1 for demographics)
from 4 different university health centers in the province of
Table 1

Ethics approval was provided by the Human Research

Ethics Board of each hospital. Whenever possible, consent
was obtained from the patient either the day before surgery
(for elective surgical patients) or in the ICU. In the case
where patients were not able to give their own consent, the
consent form was given to the decision maker (family
member) to complete on their behalf. The study was
explained in detail to the patient or the decision maker, and

Descriptive statistics of the conscious and the unconscious patient groups (n = 257)

Variable

Sex (n)
Male
Female
Age
Mean (SD)
Diagnostic group (n)
Trauma with head injury
Other trauma
Cardiac surgery
Other surgery (eg, thoracic, abdominal)
Medical
APACHE
Mean (SD)
Analgesia and sedation (n)
No infusions
Fentanyl infusion (g/h)
Mean (SD)
Propofol infusion (mg/h)
Mean (SD)
Midazolam infusion (mg/h)
Mean (SD)
Procedure at P2 (n)
Turning
Turning + endotracheal (ET) suctioning
Turning + other a
Turning + ET suctioning + other a
ET suctioning

Conscious patients
(n = 144)

Unconscious patients
(n = 113)

Total
(n = 257)

104
40

84
29

188
69

Statistical test
2

0.14

61.38 (10.76)
4
4
107
7
22

54.65 (18.70)
39
5
35
8
26

58.42 (15.12)

22.04 (4.57)

22.21 (4.50)

21
110
68.01 (36.91)
11
110.77 (74.90)
2
3.50 (3.54)

28
51
78.41 (44.28)
65
98.11 (81.56)
3
3.33 (2.08)

49
161
71.30 (39.55)
76
99.95 (80.27)
5
3.40 (2.30)

64
26
17
5
1

116
95
36
9
1

APACHE indicates Acute Physiology and Chronic Health Evaluation.

a
Dressing change, mouth suctioning, and abdominal palpation were the procedures referred as other.
P .001.

13.09

43
9
142
15
48

22.35 (4.46)

52
69
19
4
0

62.68

0.30

2.43
0.23
0.01
18.46

628.e10
a copy of the consent signed was given to them. During
data collection, ICU nurses were told to provide the usual
care to their patients that means that they could medicate
their patients for pain relief according to their pain
management protocol.

4.3. Instruments
Behavioral indicators (ie, facial expressions, body
movements, compliance with the ventilator, and muscle
tension) were measured using the CPOT [7]. Both the
French version and the English version of the CPOT
showed good psychometric qualities. Weighted from 0.52
to 0.88 and interclass correlation coefficients from 0.80 to
0.93 were obtained for interrater reliability. Discriminant
validity was supported with higher CPOT scores with a
nociceptive procedure (turning) compared with rest or a
nonnociceptive procedure (paired t tests or RM-MANOVAs, P .05). For criterion validity, the CPOT scores were
related with the patient's self-report of pain (analyses of
variance with P .001 for presence or absence of pain,
correlation coefficients from 0.40 to 0.71 with pain intensity
from 0 to 10) [7,8].
Physiologic indicators (ie, MAP, HR, RR, SpO2, and endtidal CO2) were examined with the monitoring equipment
available in the ICUs. Mechanically ventilated patients able
to self-report were asked if they had pain or not (presence or
absence of pain) immediately after being scored with the
CPOT. This method was specified so that raters (principal
investigator, research staff, and ICU nurses) were not biased
by the patient's self-report of pain when they scored with
the CPOT.

4.4. Procedure
Patients participating in the study were assessed by
members of the research team and ICU nurses during a
nociceptive procedure (ie, turning [1] with or without
endotracheal suctioning or other care). All evaluators were
blinded to each other's scores. Assessments were completed
at rest preprocedure (P1), during the nociceptive procedure
(P2), and 20 minutes postprocedure (P3) for a total of 3
assessments. Twenty minutes was selected as a postprocedure
rest assessment because that amount of time is required for the
liberation, the reaction, and the elimination of stress
hormones (epinephrine and norepinephrine), response activated by a stressor (turning). The epinephrine and norepinephrine half-life is short, 1 to 3 minutes, and these hormones
are completely eliminated after 15 to 20 minutes [17].
To complete the CPOT, patients were observed during 1
minute at rest preprocedure and postprocedure. During the
nociceptive procedure, patients were observed for the
duration of the procedure, which could be a few minutes to
detect any behavior of the patient. Physiologic indicators
from the monitoring were documented as well.

C. Glinas, C. Arbour
More than 80 ICU nurses were trained to use the CPOT in
the different sites. The 1-hour training session provided by
the principal investigator (CG) included the following: (1)
objectives, sample criteria and procedure of the study, (2)
description of the CPOT indicators and the scoring, and (3)
completion of data collection sheet.

4.5. Data analysis

Descriptive statistics were calculated for all variables. For
ordinal variables (individual behavioral indicators), 2 tests
were used to compare the different patient groups. For
continuous variables (physiologic indicators, CPOT total
score), analyses of variance were used to compare the different
patient groups. To identify the best indicators that predict
the presence of pain in mechanically ventilated conscious
adults able to self-report, logistic regression was performed.

5. Results
5.1. Sample
A total of 257 patients participated in this study (see
Table 1 for descriptive statistics). More than half of them
(56%) were conscious, whereas 44% were unconscious
during the time of this data collection. Most were males
(73%). The mean age of conscious patients was slightly
higher compared with unconscious patients. Most patients of
the conscious group were in the postoperative phase,
whereas the unconscious patient group was an even mix of
trauma, postoperative cardiac, and medical patients. Both
patient groups had a similar Acute Physiology and Chronic
Health Evaluation mean score. In analgesia and sedation,
15% and 25% of conscious and unconscious patients,
respectively, were not receiving any analgesic or sedative
agents. According to analgesia, 76% of the conscious
patients and 45% of the unconscious patients were on
fentanyl infusions. Regarding sedation, 58% of the unconscious patients and only 8% of the conscious patients were
on propofol infusions. Rarely, patients received midazolam
infusions. Moreover, only a few conscious (n = 16/144) and
unconscious (n = 21/113) ventilated patients received a
subcutaneous dose or an intravenous bolus of a sedative or an
analgesic agent before the nociceptive procedure.

5.2. Nociceptive procedure (P2)

Except for one patient who was only suctioned via the
endotracheal tube, every other patient was turned with or
without other care (see Table 1). Almost half of conscious
(48%) also underwent endotracheal suctioning compared
with only 23% of unconscious patients. Few patients
received other care including dressing change, mouth
suctioning, or abdominal palpation.

Behavioral and physiologic indicators in mechanically ventilated adults

5.3. Behavioral and physiologic indicators

in conscious and unconscious mechanically
ventilated adults
According to behavioral indicators measured with the
CPOT, conscious patients obtained significant higher scores
on the CPOT compared with unconscious patients at all times
(P1 to P3; see Table 2). At rest preprocedure (P1), it can be
observed that a higher proportion of conscious patients
showed a tense face, protection movements, and muscle
tension. During the nociceptive procedure (P2), 88.9% of
conscious patients showed a tense or a grimacing face
compared with 69% of unconscious patients. For body
movements, most conscious patients (61.1%) showed protection movements, whereas most unconscious patients (54%)
remained immobile. Similar observations were obtained for
muscle tension. Most conscious patients (72.2%) were tense
or very tense, whereas unconscious patients remained relaxed
(65.5%). Regarding compliance with the ventilator, more than
half of the conscious patients (55.6%) were coughing but
tolerating, and some were fighting their ventilator, whereas
most unconscious patients (52.2%) were tolerating their
ventilator. At rest postprocedure (P3), behavioral reactions
were comparable with those observed preprocedure (P1).
In physiologic indicators, some variations were observed
for all of them in both patient groups (see Table 3). Most
physiologic indicators (MAP, HR, RR, end-tidal CO2)
increased during the nociceptive procedure (from P1 to P2),
whereas the SpO2 slightly decreased. A significant difference
was obtained only for MAP between the 2 groups for both the
reactivity to the nociceptive procedure (P1 to P2) and the
recovery (P2 to P3).
When the 5 diagnostic patient groups were compared,
CPOT scores were statistically different during the nociceptive procedure (P2) (F = 5.11; df = 4; P .001). Postoperative
patients (cardiac surgery and other surgery) obtained the
highest CPOT score (3.20 [SD, 1.40] and 3.40 [SD, 1.24],
respectively), whereas the head trauma patients showed the
lowest CPOT score (2.05 [SD, 1.93]). In the individual
behavioral indicators, only facial expressions (2 = 61.70; df
= 8; P .001) and muscle tension (2 = 54.89; df = 8; P
.001) were statistically significant. More than half of the head
trauma patients (51.2%) showed a relaxed face, whereas most
other patient groups (N75%) expressed a tense or a grimacing
face. According to muscle tension, most trauma patients
(72.1%) were relaxed during the procedure compared with
postoperative patients (46.7%-75.4%) who were more tensed.
Interestingly, specific facial reactions were observed for
head trauma patients. Most (69.8%) did not show frowning
or brow lowering and had closed eyes. A higher proportion
of head trauma patients presented weeping eyes (14%) and
opened their eyes (16%) during the nociceptive procedure. In
all groups, 6% to 15% of the patients bit their endotracheal
tube when being turned.
Again, when the 5 diagnostic groups were compared
regarding physiologic indicators during the nociceptive

628.e11

procedure (P2), the MAP, HR, and RR increased, and the

SpO2 decreased. However, only the change in SpO2 was
statistically significant (F = 4.10; df = 4; P .001). The
trauma patients as well as medical patients showed the highest
variability in SpO2 (1.42 to 1.78 [SD, 1.72-3.34])
compared with postoperative patients (0.20 to 0.53 [SD,
1.70-1.96]). The variability of physiologic indicators was
quite different at recovery (P2 to P3). Indeed, most changes in
vital signs were statistically significant including the MAP
(F = 2.70; df = 4; P .05), HR (F = 4.02; df = 4; P .01), RR
(F = 2.43; df = 4; P .05), and SpO2 (F = 6.07; df = 4; P
.001). Trauma patients showed the highest variability in these
physiologic indicators compared with other groups.

5.4. Relationship between behavioral/physiologic

indicators and conscious patient's self-report of pain
Conscious patients who reported the presence of pain
during the nociceptive procedure obtained a significant
higher CPOT mean score compared with patients who did
not experience pain (see Table 4). Facial expression,
compliance with the ventilator, and muscle tension were
statistically significant between the 2 groups (presence vs
absence of pain). In fact, 94.3% of patients with pain showed
a tense or a grimacing face compared with 71.4% of patients
without pain. A proportion of 68.9% of patients with pain
coughed or fought their ventilator compared with 45.7% of
the patients without pain. The difference between these 2
groups according to muscle tension was wider with 83% of
patients with pain who were tense or very tense compared
with only 37.1% of the patients without pain.
According to physiologic indicators, significant results
were obtained only at recovery (from P2 to P3) (see Table 5). It
can be observed that the decreases in MAP, HR, and end-tidal
CO2 were more important in patients without pain compared
with those with pain. During the nociceptive procedure (from
P1 to P2), the descriptive statistics showed an increase in most
of the vital signs, except for SpO2 for which a slight decrease
was observed. However, no statistical difference was observed
between patients with pain from those without pain when
exposed to the nociceptive procedure.

5.5. Predictors of the conscious patient's

self-report of pain
To determine the best indicators that predict the conscious
patient's self-report of pain (presence or absence), a logistic
regression was performed (see Table 6). Only the CPOT score
could predict the presence or the absence of pain on the basis of
the patient's self-report during the nociceptive procedure.

6. Discussion
Accurate detection of the critically ill patient's pain is
not an easy task for ICU nurses especially when the

628.e12

C. Glinas, C. Arbour

Table 2 Frequency of each behavioral indicator of the CPOT, mean, and SD of the scale total score at rest preprocedure (P1), during the
nociceptive procedure (P2), and 20 minutes postprocedure (P3) in the conscious and the unconscious patient groups
Behavioral indicators

P1
Facial expression
Relaxed, neutral
Tense
Grimacing
Body movements
Absence of movements
Protection
Restlessness
Compliance with the ventilator
Tolerating ventilator
Coughing but tolerating
Fighting ventilator
Muscle tension
Relaxed
Tense, rigid
Very tense or rigid
P1 total
Mean (SD)
P2
Facial expression
Relaxed, neutral
Tense
Grimacing
Body movements
Absence of movements
Protection
Restlessness
Compliance with the ventilator
Tolerating ventilator
Coughing but tolerating
Fighting ventilator
Muscle tension
Relaxed
Tense, rigid
Very tense or rigid
P2 total
Mean (SD)
P3
Facial expression
Relaxed, neutral
Tense
Grimacing
Body movements
Absence of movements
Protection
Restlessness
Compliance with the ventilator
Tolerating ventilator
Coughing but tolerating
Fighting ventilator
Muscle tension
Relaxed
Tense, rigid
Very tense or rigid

Conscious patients
(n = 144)

Unconscious patients
(n = 113)

113 (78.47%)
30 (20.83%)
1 (0.69%)

106 (93.81%)
7 (6.19%)
0 (0%)

112 (77.78%)
31 (21.53%)
1 (0.69%)

102 (90.27%)
10 (8.85%)
1 (0.88%)

133 (92.36%)
10 (6.94%)
1 (0.69%)

110 (97.35%)
2 (1.77%)
1 (0.88%)

40 (27.78%)
64 (44.44%)
0 (0%)

94 (83.19%)
17 (15.04%)
2 (1.77%)

0.98 (1.17)

0.39 (0.77)

16 (11.11%)
90 (62.50%)
38 (26.39%)

35 (30.97%)
65 (57.52%)
13 (11.50%)

42 (29.17%)
88 (61.11%)
14 (9.72%)

61 (53.98%)
48 (42.48%)
4 (3.54%)

53 (36.81%)
80 (55.56%)
11 (7.64%)

59 (52.21%)
48 (42.48%)
6 (5.31%)

40 (27.78%)
92 (63.89%)
12 (8.33%)

74 (65.49%)
33 (29.20%)
6 (5.31%)

3.47 (1.51)

2.23 (1.48)

109 (75.69%)
35 (24.31%)
0 (0%)

105 (92.92%)
8 (7.08%)
0 (0%)

110 (76.39%)
31 (21.53%)
3 (2.08%)

103 (91.15%)
7 (6.19%)
3 (2.65%)

129 (89.58%)
15 (10.42%)
0 (0%)

110 (97.35%)
3 (2.65%)
0 (0%)

88 (61.11%)
54 (37.50%)
2 (1.39%)

88 (77.88%)
23 (20.35%)
2 (1.77%)

Statistical test
F

df

21.45

43.56

Behavioral and physiologic indicators in mechanically ventilated adults

628.e13

Table 2 (continued)
Behavioral indicators

P3 total
Mean (SD)

Conscious patients
(n = 144)

Unconscious patients
(n = 113)

Statistical test
F

df

1.01 (1.34)

0.45 (0.80)

15.24

P b .001.

patient is unable to self-report or cannot verbalize pain

because of mechanical ventilation. Although today's
guidelines strongly suggest the use of a standardized
behavioral pain scale to nurses who care for uncommunicative patients [5], further research is still needed to
fully understand the behavioral and the physiologic
responses of critically ill patients who are experiencing
pain. This study allowed to describe the behavioral and the
physiologic-related indicators of pain at rest (preprocedure
and postprocedure), and during a nociceptive procedure in
conscious and unconscious mechanically ventilated
patients with different type of diagnosis, and to examine
the relationship of these indicators with the conscious
patient's self-report of pain.
At all times and especially during the nociceptive
procedure (P2), a higher proportion of conscious patients
showed tense facial expressions, protection movements,
coughing on their endotracheal tube, and muscle rigidity
as opposed to unconscious patients. For this reason,
higher CPOT scores were obtained with conscious
patients. These results may be explained in different
ways. First, a lower proportion of conscious patients (9%)
compared with unconscious patients (60.2%) received an
infusion of a sedative agent (midazolam or propofol).
Therefore, conscious patients may have been more aware
of the pain sensation and the presence of the endotracheal

tube. On the other hand, we know that the use of

sedative agents may blur behavioral reactions to nociceptive procedures. In previous studies, the higher the
dosage of midazolam, the lower were the behavioral
scores on the BPS [9,10]. Second, 64% of conscious
patients were exposed to turning with other care
compared with 43% of the unconscious patients. Being
exposed to more than one noxious stimulus may
contribute to higher pain levels that may be expressed
by more intense behavioral reactions.
When comparing the behaviors in the different
diagnostic groups during the nociceptive procedure,
significant results were obtained for facial expressions
and muscle tension. Regarding facial expressions, similar
reactions (including frowning and brow lowering) were
observed in patients with trauma, in the postoperative
phase or with a medical diagnosis. Such observations are
similar to findings from previous studies conducted with
critically ill patients able to self-report [11,13]. On the
other hand, the head injury patients reacted very
differently. They were less likely to express frowning
and eyelid contractions but were more likely to have
weeping eyes and to open their eyes during the nociceptive
procedure. Moreover, they were relaxed during the
nociceptive procedure. It must be noted that most of
head injury patients were unconscious and more heavily

Table 3 Description of the changes in vital signs during the nociceptive procedure (P1 to P2) and at recovery (P2 to P3) in the conscious
and unconscious patient groups
Vital signs

P1 to P2
MAP
HR
RR
SpO2
End-tidal CO2
P2 to P3
MAP
HR
RR
SpO2
End-tidal CO2

Conscious patients (n = 144)

Unconscious patients (n = 113)

Total (n = 257)

ANOVA

Mean (SD)

Mean (SD)

Mean (SD)

df

7.88 (10.47)
4.78 (8.20)
4.01 (6.26)
0.77 (2.79)
2.11 (5.04)

5.28
4.86
3.50
0.64
1.28

(10.20)
(9.10)
(6.12)
(2.03)
(2.69)

6.76 (10.42)
4.82 (8.59)
3.79 (6.19)
0.71 (2.48)
1.74 (4.16)+

3.92
0.01
0.42
0.17
1.76

1
1
1
1
1

6.62 (12.18)
2.80 (8.25)
3.94 (7.29)
0.56 (3.15)
0.39 (5.19)

3.42
4.36
3.08
0.65
0.22

(11.42)
(7.90)
(6.35)
(2.29)
(5.29)

5.22 (11.94)
3.49 (8.12)
3.57 (6.90)
0.60 (2.80)
0.31 (5.22)+

4.54
2.34
0.99
0.07
0.05

1
1
1
1
1

ANOVA indicates analysis of variance.

+A total of 175 patients had end-tidal CO2 monitoring.
P .05.

628.e14

C. Glinas, C. Arbour

Table 4 Frequency of each behavioral indicator of the CPOT, mean, and SD of the scale total score in association with the conscious
patient's self-report of pain during the nociceptive procedure (P2)
Behavioral indicators

Patient's self-report

Patient's self-report

Yes, presence of pain (n = 106)

No, absence of pain (n = 35)

P2
Facial expression
Relaxed, neutral
Tense
Grimacing
Body movements
Absence of movements
Protection
Restlessness
Compliance with the ventilator
Tolerating ventilator
Coughing but tolerating
Fighting ventilator
Muscle tension
Relaxed
Tense, rigid
Very tense or rigid

P2 total
Mean (SD)

Total
(n = 141)

6 (5.66%)
69 (65.09%)
31 (29.25%)

10 (28.57%)
19 (54.29%)
6 (17.14%)

16 (11.35%)
88 (62.41%)
37 (26.24%)

29 (27.36%)
67 (63.21%)
10 (10.60%)

13 (37.14%)
19 (54.29%)
3 (8.57%)

42 (29.79%)
86 (60.99%)
13 (9.22%)

33 (31.13%)
64 (60.38%)
9 (8.49%)

19 (54.29%)
14 (40.00%)
2 (5.71%)

52 (36.88%)
78 (55.32%)
11 (7.80%)

18 (16.98%)
77 (72.64%)
11 (10.38%)

22 (62.86%)
13 (37.14%)
0 (0%)

40 (28.37%)
90 (63.83%)
11 (7.80%)

3.76 (1.42)

2.49 (1.42)

Statistical
test

3.45 (1.52)

df

14.13

1.21

6.06

28.35

df

21.21

P .05.
P .001.

sedated. For those reasons, the need for further research

remains to examine the specificities of behavioral reactions
of head injury patients to nociceptive procedures. In fact,
lower CPOT scores obtained during nociceptive stimulation in head injury patients may not necessarily reflect
decreased pain intensity in this group.

The use of physiologic indicators for the detection of pain

is easily accessible with the monitoring in the ICU. It was
also noted in the study of Puntillo and colleagues [13] that
increased blood pressure and HR where the vital signs most
frequently associated with acute pain by ICU nurses.
However, contradictory results have been found in previous

Table 5 Description of the changes in vital signs in conscious patients during the nociceptive procedure (P1 to P2) and at recovery (P2 to
P3) in association with the patient's self-report of pain (yes or no)
Vital signs

P1 to P2
MAO
HR
RR
SpO2
End-tidal CO2
P2 to P3
MAP
HR
RR
SpO2
End-tidal CO2

Patient's self-report

Patient's self-report

Yes, presence of pain (n = 106)

No, absence of pain (n = 35)

8.09 (10.58)
5.22 (8.51)
4.52 (6.56)
0.99 (2.97)
2.47 (5.26)
3.65 (11.27)
0.48 (6.12)
2.66 (7.68)
0.13 (2.31)
1.07 (5.13)

ANOVA indicates analysis of variance.

+A total of 95 patients had end-tidal CO2 monitoring.
P .05.
P .01.

Total
(n = 141)

ANOVA
F

df

7.71 (9.55)
3.43 (7.05)
2.57 (5.29)
0.11 (2.15)
0.67 (3.94)

8.00
4.77
4.04
0.77
2.13

(10.30)
(8.18)
(6.30)
(2.81)
(5.07)+

0.04
1.26
2.54
2.59
1.86

1
1
1
1
1

8.89 (12.75)
4.12 (9.32)
5.16 (7.20)
1.06 (3.81)
1.83 (4.51)

6.37
2.36
3.95
0.49
-0.08

(12.30)
(8.11)
(7.51)
(3.22)
(5.08)+

6.26
6.92
3.76
4.60
7.66

1
1
1
1
1

Behavioral and physiologic indicators in mechanically ventilated adults

Table 6 Pain (presence or absence) logistic regression analysis
during the nociceptive procedure (P2) (n = 95)
Variable

R2

Wald

CPOT score at P2
Constant
Variables not in the equation
MAP
HR
RR
SpO2
End-tidal CO2

0.21

1.17
1.93

14.59
5.23
0.47
0.01
0.35
1.20
0.001

P .05.
P .001.

studies that cannot support their sensitivity to nociceptive

procedures in critically ill adults [8,10].
Regarding the variations in physiologic indicators
during the nociceptive procedure (from P1 to P2) and at
recovery (from P2 to P3), only the MAP significantly
varied between the conscious and the unconscious patients.
In their study with critically ill unconscious patients (n =
30), Payen and colleagues [9] observed that both MAP and
HR were significantly increasing during nociceptive
procedures. When comparing the diagnostic groups during
the nociceptive procedure (from P1 to P2), SpO2 was the
only vital sign to fluctuate in a significant way. The
patients who had the more pronounced decrease of SpO2
during procedural pain were the trauma and medical
patients. A high proportion of these patients were
unconscious and were receiving either propofol or fentanyl
infusions. Interestingly, all physiologic signs, except for
end-tidal CO2, varied significantly at recovery (from P2 to
P3). The patients with head injury or other trauma showed
the highest variations in vital signs. Such results highlight
their unstable hemodynamic condition when exposed to
nociceptive stimuli.
The CPOT score was associated with the conscious
patients' self-reports of pain and was the only predictor of the
presence or the absence of pain. More important, this study
allowed to explore the relationship of specific behaviors
included in the CPOT with the patient's self-report of pain
during the nociceptive procedure. Facial expressions, muscle
tension, and compliance with the ventilator were the 3
behaviors that were significantly related with the patients'
self-report of pain. To our knowledge, facial expressions are
part of all existing behavioral pain scales. Also, facial
expressions in pediatric and adult populations were extensively studied with the FACS [18] and the NFCS [19].
Therefore, patients' facial expressions represent an essential
component to take into account in pain assessment in
nonverbal patients. Further analysis exploring the sensitivity
and specificity of individual behaviors included in the CPOT
has to be done.
It must be emphasized that even if observation of
behaviors is recognized as a valid approach to pain

628.e15

assessment, they are not always accurate reflections of

pain intensity. Experts also strengthen that behavioral
assessment tool scores should not be interpreted as a pain
intensity score nor compared with standard categories of
pain levels. A behavioral score may be helpful in
identifying the presence of pain and in evaluating
interventions aimed to relieve pain [5]. Results from a
previous study support this recommendation. In fact,
behavioral displays of pain were more often observed in
critically ill adults with severe pain than those with mild or
moderate pain [20].
Interestingly, besides most conscious patients were able to
self-report experienced pain (n = 106/141) during the
nociceptive procedure, only 11.1% of them received a dose
of an analgesic or a sedative agent before the procedure. This
is consistent with the results of the Thunder Project II in
which less than 20% of the patients (n = 5957) were
administered an analgesic before nociceptive procedures
known to be painful [21]. Those observations highlight the
emergency to apply clinical recommendations into ICU
practice that support that pathologic conditions and common
problems or procedures known to cause pain should trigger
an intervention [5].
Surprisingly, changes of MAP, HR, SPO2, and end-tidal
CO2 were associated with the conscious patients' selfreports of pain only at recovery (from P2 to P3). Indeed,
marked decrease of MAP, HR, and end-tidal CO2, and a
better recovery of SpO2 were found in pain-free patients.
However, such results are difficult to apply into practice.
Indeed, it makes more sense to look at reactivity during a
nociceptive procedure instead of recovery postprocedure to
make a decision about treatment of pain management.
Also, the logistic regression showed that the variations of
vital signs could not predict the presence or the absence of
pain during the nociceptive procedure. Experts recommend
that vital signs should not be considered as primary
indicators of pain as they can be attributed to other distress
conditions, homeostatic changes, and medications [5].
Even in healthy volunteers exposed to experimental painful
stimuli, variable results have been obtained and a sex
effect has been observed. Indeed, a positive correlation
between HR and pain was found in men but not in women
[22]. Such evidence reinforces the lack of validity for vital
signs and their use for pain assessment should be
considered with some cautious. This is even truer in the
context of critical care where the patient's condition is
unstable and where many drugs may influence the vital
sign variability.
All those findings support the guiding principles for the
assessment of pain in the critically ill nonverbal patient in
which the use of behavioral indicators represent the primary
indices to consider. Indeed, relying on changes in vital signs
as primary indicators of pain can be misleading because these
changes may also be attributed to other factors [5]. Nevertheless, because the absence of increased vital signs does not
indicate the absence of pain [13,23], changes in vital signs

628.e16
can be considered as a cue to begin further assessment of
pain or other stressors [5].

6.1. Limitations
This study was not without limitations. First, only 175
patients had end-tidal CO2 monitoring contributing to
missing data for this physiologic indicator and some
statistical analyses. Second, raters could not be blind to the
nociceptive procedure and were aware that turning was
painful to patients. Therefore, the raters may have perceived
more behaviors during turning. Third, neither nociceptive
procedure nor therapeutic regimen could be standardized.
Finally, as very few patients (11.1% of the conscious patients
and 18.6% of the unconscious patients) received a dose of an
analgesic or a sedative agent before the nociceptive
procedure, we could not explore the impact of receiving
such a dose on the behavioral and physiologic indicators.

6.2. Future directions

Further research is warranted to describe the pain-related
behaviors in head injury patients who seem to react
differently from the other patients. Also, the inconsistent
findings regarding physiologic indicators in pain assessment
in critically ill patients mark the time of looking at other
measures than vital signs. As a matter of fact, new
technologies have been developed and some implemented
in ICU settings. The near-infrared spectroscopy (NIRS) and
the bispectral index are available technologies that may be
used for the measurement of the cortical responses to pain.
The NIRS that allows for the measurement of changes in
cerebral oxygenation in a specific cortical region have been
examined in neonates exposed to a nociceptive stimulus in 3
previous studies [24-26]. Also, a pilot study has just been
completed in which the NIRS was explored in cardiac
surgery adults exposed to common nociceptive procedures in
the operating room [27]. In summary for the NIRS, results of
those recent studies suggest that the cerebral hemodynamic
indicators are sensitive to nociceptive procedures in both
infants and adults. Regarding the bispectral index that reflects
a signal-processed electroencephalogram and is primarily
used as an objective measure of sedation, it has been recently
studied in the context of pain assessment in critically ill
sedated adults [28]. The bispectral index has also been shown
to be sensitive to nociceptive procedures. Further studies on
these innovative measures could be conducted in critically ill
patients to examine their validity for the detection of pain.

Acknowledgments
This work was supported by a Research Development
Grant from the Jewish General Hospital Foundation
(Montreal, Quebec, Canada).

C. Glinas, C. Arbour

References
[1] Puntillo KA, White C, Morris AB, et al. Patients' perceptions and
responses to procedural pain: Results from thunder project II. Am J
Crit Care 2001;10:238-51.
[2] Shannon K, Bucknall T. Pain assessment in critical care: what have we
learnt from research. Intensive Crit Care Nurs 2003;19:154-62.
[3] Anand KJS, Craig KD. New perspectives on the definition of pain.
Pain 1996;67:3-6.
[4] Jacobi J, Fraser GL, Coursin DB, et al. Clinical practice guidelines for
the sustained use of sedatives and analgesics in the critically ill adult.
Crit Care Med 2002;30:119-41.
[5] Herr K, Coyne PJ, Key T, et al. Pain assessment in the nonverbal
patient: position statement with clinical practice recommendations.
Pain Manag Nurs 2006;7:44-52.
[6] Assaoui Y, Zeggwagh AA, Zekraoui A, et al. Validation of a
behavioral pain scale in critically ill, sedated, and mechanically
ventilated patients. Anesth Analg 2005;101:1470-6.
[7] Glinas C, Fillion L, Puntillo KA, et al. Validation of the Critical-Care
Pain Observation Tool in adult patients. Am J Crit Care 2006;15:420-7.
[8] Gelinas C, Johnston C. Pain assessment in the critically ill ventilated
adult: validation of the Critical-Care Pain Observation Tool and
physiologic indicators. Clin J Pain 2007;23:497-505.
[9] Payen JF, Bru O, Bosson JL, et al. Assessing pain in the critically ill
sedated patients by using a Behavioral Pain Scale. Crit Care Med 2001;
29:2258-63.
[10] Young J, Siffleet J, Nikoletti S, et al. Use of a Behavioural Pain Scale
to assess pain in ventilated, unconscious and/or sedated patients.
Intensive Crit Care Nurs 2006;22:32-9.
[11] Puntillo KA, Morris AB, Thompson CL, et al. Pain behaviors observed
during six common procedures: results from thunder project II. Crit
Care Med 2004;32:421-7.
[12] Mateo OM, Krenzischek DA. A pilot study to assess the relationship
between behavioral manifestations and self-report of pain in
postanesthesia care unit patients. J Post Anesth Nurs 1992;7:15-21.
[13] Puntillo KA, Miaskowski C, Kehrle K, et al. Relationship between
behavioral and physiologic indicators of pain, critical care self-reports
of pain, and opioid administration. Crit Care Med 1997;25:1159-66.
[14] Li D, Puntillo K, Miaskowski C. A review of objective pain measures
for use with critical care adult patients unable to self-report. J Pain
2008;9:2-10.
[15] Sessler CN, Grap MJ, Ramsay MAE. Evaluating and monitoring
analgesia and sedation in the intensive care unit. Crit Care 2008;12
Available online. http://ccforum.com/content/12/S3/S2.
[16] Glinas C, Fillion L, Puntillo KA. Item selection and content validity
of the Critical-Care Pain Observation Tool: an instrument to assess
pain in non-verbal critically ill patients. J Adv Nurs 2009;65:203-16.
[17] Berne RM, Levy MN. Physiology. St Louis: Mosby Inc; 1983.
p. 1033-68.
[18] Ekman P, Friesen WV. Manual for the Facial Action Coding System.
Palo Alto: Consulting Psychologists Press; 1992.
[19] Grunau RVE, Craig KD. Facial activity as a measure of neonatal pain
expression. In: Tyler DC, Krane EJ, editors. Advances in pain research
and therapy, vol. 15. New York (NY): Raven; 1990. p. 147-55.
[20] Glinas C, Harel F, Fillion L, et al. Sensitivity and specificity of the
Critical-Care Pain Observation Tool for the detection of pain in
intubated adults following cardiac surgery. J Pain Symptom Manage
2009;37:58-67.
[21] Puntillo KA, Rietman L, Morris AB, et al. Practices and predictors of
analgesic interventions for adults undergoing painful procedures. Am J
Crit Care 2002;11:415-29.
[22] Tousignant-Laflamme Y, Rainville P, Marchand S. Establishing a link
between heart rate and pain in healthy subjects: a gender effect. J Pain
2005;6:341-7.
[23] Glinas C, Viens C, Fortier M, et al. Les indicateurs de la douleur en
soins critiques. Perspect Infirm 2005;2:12-22.

Behavioral and physiologic indicators in mechanically ventilated adults

[24] Bartocci M, Bergqvist LL, Lagercrantz H, et al. Pain activates cortical
areas in the preterm newborn brain. Pain 2006;122:109-17.
[25] Slater R, Cantarella A, Gallella S, et al. Cortical pain responses in
human infants. J Neurosci 2006;26:3662-6.
[26] Slater R, Cantarella A, Franck L, et al. How well do clinical
pain assessment tools reflect pain in infants? Plos Med 2008;5:
928-33.

628.e17

[27] Glinas C, Choinire M, Johnston C, et al. Towards a new approach for

the detection of pain in adults undergoing cardiac surgery: the nearinfrared spectroscopy (NIRS)preliminary results. Anesth Anal (under
review).
[28] Li DTY, Miaskowski CA, Burkhardt DH, et al. Physiologic and
behavioral responses associated with noxious procedures in sedated
critically ill patients. J Crit Care [in press].