Beruflich Dokumente
Kultur Dokumente
Original Contribution
Keywords:
Pain, postoperative;
Pupillometry;
Analgesia;
General anesthesia
Abstract
Study Objective: To investigate whether pupil diameter (PD) measured during scotopic conditions is
influenced by pain in conscious patients in the early postoperative period.
Design: Prospective, observational, cohort study.
Setting: Single-center, postanesthesia care unit (PACU).
Patients: Patients scheduled for a surgery during general anesthesia.
Interventions: Baseline PD was measured the day before surgery. Patients were observed on admission
to the PACU, immediately after extubation, during the different steps of analgesic intervention (demand,
relief, plus intermediate measures when relevant), and either at discharge or 3 hours after admission.
Measurements: PD, pain (numerical rating scale), and alertness (Observer's Assessment of Alertness/
Sedation scale).
Main Results: Of 103 patients enrolled, 80 required analgesia in the PACU and completed follow-up. Pain
intensity evolved in line with expectations (temporary increase then relief), and alertness increased with time.
PD increased from low mean values at admission to the PACU (40% of baseline) to a plateau throughout the
http://dx.doi.org/10.1016/j.jclinane.2014.09.006
0952-8180/ 2014 Elsevier Inc. All rights reserved.
24
C. Dual et al.
rest of the study period (80% of baseline) and was not related to pain intensity. Multivariate analyses
suggested that the factors influencing PD (or its value related to baseline) were time since extubation and the
type of opioid (remifentanil, sufentanil, or sufentanil at high doses) administered during surgery.
Conclusions: Because of a residual effect of intraoperative opioids and a level of nociceptive stimulation
lower than in surgical conditions, PD is not significantly influenced by early postoperative pain or pain relief.
2014 Elsevier Inc. All rights reserved.
1. Introduction
The current pharmacological options to counteract nociception are the intraoperative administration of intravenous opioids,
followed by postoperative analgesic regimens, guided by pain
assessment tools based mostly on the patient's subjective
sensation [1]. However, during surgery receiving general
anesthesia, the challenge is to provide sufficient analgesia to
avoid central sensitization of pain, without excessive amount of
opioids and the resultant hyperalgesia [2,3].
The measurement of scotopic (ie, under low light) pupil
diameter (PD) appears as an interesting approach to estimate
nociception and analgesia in unconscious patients. Indeed,
although nociceptive stimulation increases PD in an
intensity-dependent manner [4], opioids induce a dosedependent constriction (miosis) [5]. Pupil dilation induced by a
transient nociceptive stimulus, also known as the pupil
dilation reflex (PDR), is regularly used in experimental
research protocols [5-8], whereas infrared pupillometers
dedicated to the practice of anesthesia are now available on
the European market [9,10]. To build a mathematical model
able to predict excessive nociception during surgery, the
relationship between pain and PD was studied in conscious
patients. The available data in conscious subjects had been
obtained only with mild nociceptive stimulation [4,11,12].
Feeling that such conditions were far from the clinical
context, the immediate postoperative period was studied.
Two similar observational studies failed to find a significant
relation between early postoperative pain and PD [13,14].
25
Log transformations were done if relevant. Categorical data
were expressed as number of patients/observations and
percentage of the total. For comparison between subgroups,
the Student t test (numerical Gaussian), the Mann-Whitney U
test (other numerical or ordinal), and the 2 or Fisher exact test
(categorical) were used. Multivariable analysesin which the
dependent outcome was either PD or rPD measured repeatedly
were conducted using linear mixed models, considering the
first measurement as the reference for time. In these models
with random subject effects (random intercept and slope) and
fixed effects, parameters were estimated by restricted
maximum likelihood. Multicollinearity was avoided by
reducing the number of either modalities or covariates, with
the help of descriptive and factorial analysis. Statistical
analyses were performed using Stata V10 (StataCorp, College
Station, TX, USA) and XLStat (Addinsoft, Paris, France).
Significance level (type I error) was set at 0.05 for all analyses.
Figures were generated using Microsoft Office Excel 2003.
Sample size estimation was set to evidence a correlation
between PD and pain intensity, based on data collected in
similar conditions. First, we measured PD in 28 sedated
intensive care patients (78 observations) in current conditions
of a moderately painful current nursing (turning the patient
onto the side), sedation and analgesia being slightly increased
before nursing. Mean PD went from 2.9 0.5 mm before
nursing to 3.8 1.2 mm during nursing (P b .0001). We also
used data from studies conducted in healthy volunteers during
general anesthesia, in which a relevant noxious stimulation
was applied; mean PD was decreased with opioid analgesia by
1.7-3.4 mm [5,6]. We considered that 51 observations would
be necessary to show a difference of 1 mm with the greatest SD
noted in previous data (ie, 1.4 mm [6]) with = 5%, 1 =
95%, and = 0 (cross-over design, bilateral hypothesis).
Considering that only the patients who would require analgesic
intervention would be analyzed, recruitment was extended
to 100 patients.
3. Results
The flow chart of the participants is displayed in Fig. 1.
Eighty patients (77.7% of the included) needed analgesic
intervention in the PACU and were considered for primary
analysis (main sample). Tables 1-3 describe the whole
sample and subgroups. Only the highest level of previous
surgeries was significantly greater in the main sample than in
the patients who did not need additional analgesia.
A nominal variable was built to represent the type of
intraoperative opioid analgesia. As distribution for doses of
remifentanil was Gaussian, patients who received this drug
constituted 1 modality remifentanil. As distribution for
doses of sufentanil was bimodal, patients who received
sufentanil were classified as high sufentanil or normal
sufentanil, according to a cutoff value of 0.75 g/kg. As
some patients who received remifentanil or sufentanil also
26
C. Dual et al.
Fig. 1 Flow chart of the participants. The diagram at the bottom displays the number of patients within the main sample, that is, those who
needed an analgesic intervention in PACU and were considered for the primary analysis (n = 80), depending on (i) the number of observations
undertaken in the PACU after the 2 primary observations (T0, admission to the PACU; Text, early observation after extubation) and (ii) stay
of the patient over the third hour after admission (T0 + H3).
27
Patients' characteristics
Female gender
Age (y)
Age N 65 y
Weight (kg)
ASA score
I
II
III
Previous medical history
Hypertension
Diabetes
Stroke
Coronary disease
Dyslipidemia
Hepatic insufficiency
Renal insufficiency
Neuropsychiatric disorder
Cancer
Highest level of previously performed surgeries
None or minor
Intermediate
Major
Current medications
-Blocker
Calcium-channel blocker
Antidepressant
Neuroleptic
Benzodiazepine or equivalent
Opioid
Whole sample
(n = 103)
Main sample
(n = 80)
Between subgroups
(P value)
39 (37.9)
62.3 14.9
48 (46.6)
73.0 15.2
7 (30.4)
62.5 15.9
11 (47.8)
73.5 17.1
32 (40.0)
62.2 14.6
37 (46.3)
72.9 14.7
.555
.943
.894
.876
.983
21 (20.4)
55 (53.4)
27 (26.2)
5 (21.7)
12 (52.2)
6 (26.1)
16 (20.0)
43 (53.8)
21 (26.3)
37 (35.9)
13 (12.6)
15 (14.6)
17 (16.5)
13 (12.6)
2 (1.9)
2 (1.9)
10 (9.7)
4 (3.9)
6 (26.1)
3 (13.0)
4 (17.4)
3 (13.0)
2 (8.7)
0 (0)
0 (0)
2 (8.7)
1 (4.3)
31 (38.8)
10 (12.5)
11 (13.8)
14 (17.5)
11 (13.8)
2 (2.5)
2 (2.5)
8 (10.0)
3 (3.8)
46 (44.7)
35 (34.0)
22 (21.4)
16 (69.6)
4 (17.4)
3 (13.0)
30 (37.5)
31 (38.8)
19 (23.8)
30 (29.1)
0 (0.0)
14 (13.6)
9 (8.7)
16 (15.5)
9 (8.7)
6 (26.1)
0 (0)
1 (4.3)
1 (4.3)
4 (17.4)
3 (13.0)
24 (30.0)
0 (0.0)
13 (16.3)
8 (10.0)
12 (15.0)
6 (7.5)
NC
.024
NC
NC = not compared.
Description of the whole sample and the subgroups, as defined according to the need for analgesic intervention (see also Fig. 1). Numerical data are
expressed as mean SD or median (range). Categorical data are expressed as number of patients and percentages. Between-subgroups comparisons were
performed only if they were relevant and if variables were independent of each other (P values are given in the right column).
variables but the ASA score, plus the interactions between age
and each other variable. The model was predictive (P = .001),
diabetes being the only independent factor (mean PD, 4.2 vs 4.9
in diabetic and nondiabetic, respectively). Same analyses were
conducted with PD and rPD measured at T0 or Text as dependent
outcomes and the type of intraoperative opioid analgesia
replacing the level of the surgery in the model, as these 2
variables were collinear. None of them reached significance.
Table 5 shows the results of the linear mixed model
analyses. In analysis 1, the observations undertaken at T0 and
Text were excluded to keep conditions of good consciousness.
No significant effect was found, while the correlation between
PD/rPD and pain was poor (R2 b 0.1%). Analysis 2 checked if
adding factors related to the use of opioids (intraoperative or
postoperative) could influence the model. This was found for
the type of intraoperative opioid, only on rPD. When this factor
was kept as the sole fixed effect, the model was predictive for
PD and rPD (analysis 3). Fig. 3 shows the time course of PD
during the period of interest of these analyses, depending of the
4. Discussion
PD assessment in conscious postoperative patients had
already been reported, although with different aims and
conditions. In 14 patients observed after abdominal surgery,
who had analgesia by meperidine at request on the first
postoperative day, then an incremental administration of IV
alfentanil to achieve analgesia (6-8 g/kg- ), alfentanil
decreased PD from 4 to 3 mm, and a nonsignificant correlation
28
Table 2
C. Dual et al.
Intraoperative outcomes (surgery and anesthesia)
Surgical procedure
Parotidectomy
Thyroidectomy
Lumbar stenosis
Lumbar discal herniation
Total hip replacement
Carotid endarterectomy
Abdominal aortic surgery
Thoracotomy
Cystectomy
Nephrectomy
Complete prostatectomy
Level of surgery
I
II
III
Duration of surgery (min)
Premedication
Hydroxyzine
Gabapentin/pregabalin
Benzodiazepine or equivalent
Induction of general anesthesia a
Propofol
Thiopentone
Etomidate
Midazolam
Ketamine
Halogenated agent
Intravenous lidocaine
Propofol: dose for induction (mg/kg)
Maintenance of general anesthesia
Desflurane
Sevoflurane
Propofol
Propofol: mean infusion rate if target
controlled (mg/kg/h)
Intraoperative analgesia (opioids)
Remifentanil
Remifentanil + tramadol b
Sufentanil b 0.75 g/kg c
Sufentanil b 0.75 g/kg c + tramadol b
Sufentanil N 0.75 g/kg c
Sufentanil: total dose (g/kg)
Remifentanil: mean infusion
rate (g/kg/min)
Intraoperative coanalgesia (nonopioids)
Paracetamol
Nefopam
Ketoprofen
Steroid
Single-dose local anesthesia
Atropine (0.5-1 mg as single dose)
Ephedrine
Ephedrine: total dose (mg)
Droperidol (1.25 mg as single dose)
Whole sample
(n = 103)
1
10
11
21
26
17
1
6
4
2
3
0
2
3
5
5
3
1
1
0
0
2
11 (10.7)
76 (73.8)
16 (15.5)
100 [70-150] (25-430)
2 (8.7)
17 (73.9)
4 (17.4)
95 [65-143] (35-270)
75 (72.8)
24 (23.3)
6 (5.8)
17 (73.9)
7 (13.4)
2 (8.7)
101 (98.1)
1 (1.0)
2 (1.9)
32 (31.1)
11 (10.7)
0 (0)
5 (4.9)
2.5 1.1
23 (100)
0 (0)
0 (0)
7 (30.4)
2 (8.7)
0 (0)
2 (8.7)
2.5 0.8
12 (11.7)
65 (63.1)
28 (27.2)
0.16 0.07
3 (13.0)
15 (65.2)
5 (21.7)
0.18 0.11
Between subgroups
(P value)
NC
1
8
8
16
21
14
0
5
4
2
1
.916
9 (11.3)
59 (73.8)
12 (15.0)
105 [70-150] (25-430) .430
NC
58 (72.5)
17 (21.3)
4 (5.0)
NC
78 (97.5)
1 (1.3)
2 (2.5)
25 (31.3)
9 (11.3)
0 (0)
3 (3.8)
2.5 1.2
NC
.898
9 (11.3)
50 (62.5)
23 (28.8)
0.15 0.06
NC
.875 d
23 (22.3)
1 (1.0)
41 (39.8)
23 (22.3)
15 (14.6)
0.66 0.44
0.11 0.05
4 (17.4)
0 (0)
9 (39.1)
6 (26.1)
4 (17.4)
0.58 0.27
0.14 0.08
19 (23.8)
1 (1.3)
32 (40)
17 (21.3)
11 (13.8)
0.69 0.48
0.10 0.04
102 (99.0)
70 (68.0)
28 (27.2)
8 (7.8)
1 (1.0)
4 (3.9)
50 (48.5)
13.0 9.0
8 (7.8)
23 (100)
16 (69.6)
5 (21.7)
2 (8.7)
0 (0)
NA
13 (56.5)
15.5 8.4
2 (8.7)
79 (98.8)
54 (67.5)
23 (28.8)
6 (7.5)
1 (1.3)
4 (5)
37 (46.3)
12.2 9.2
6 (7.5)
NC
NC
NC
29
Whole sample
(n = 103)
Early observations
only (n = 23)
Main sample
(n = 80)
Between subgroups
(P value)
25 [15-44] (0-105)
25 [18-40] (0-90)
25 [15-45] (0-105)
.668
10 [5-10] (5-20)
10 [10-10] (5-20)
10 [5-10] (5-20)
.283
50 [30-70] (5-305)
55 [43-75] (15-130)
50 [30-66] (5-305)
.224
29 (28.2)
5 (21.7)
24 (30.0)
1.000
NC
Description of the whole sample and the subgroups, as defined according to the need for analgesic intervention (see also Fig. 1). Numerical data are
expressed as mean SD or median [interquartile range] (range). Categorical data are expressed as number of patients and percentages. Between-subgroups
comparisons were performed only if they were relevant and if variables were independent of each other (P values are given in the right column).
a
See Materials and methods for definition.
b
Only for the patients discharged before T0 + 3 hours.
Table 4
Whole sample
(n = 103)
Early observations
only (n = 23)
Main sample
(n = 80)
Between subgroups
(P value)
4.8 0.9
3.3 1.1
4.7 1.0
3.4 1.2
4.8 0.8
3.2 1.1
.544
.611
1.9 (1.6-2.4)
0.40 (0.36-0.50)
1.9 (1.5-2.3)
0.39 (0.36-0.44)
1.9 (1.6-2.6)
0.40 (0.36-0.54)
.140 a
.195 a
3.2 (2.2-4.0)
0.68 (0.53-0.85)
2.8 (2.5-3.4)
0.58 (0.52-0.79)
3.3 (2.6-4.1)
0.72 (0.55-0.82)
.140 a
.255 a
3.8 1.1
0.79 0.19
3.6 0.9
0.80 0.20
3.8 1.1
0.79 0.19
.492
.868
Scotopic PD measured at the right eye at the times of observation that were standardized for all patients, within the whole sample and the subgroups, as
defined according to the need for analgesic intervention (see also Fig. 1). Numerical data are expressed as mean SD or median (interquartile range). The
rPD is the ratio to the baseline preoperative value.
a
The Student t test performed on log-transformed values.
b
Either at discharge from PACU or at T0 + 3 hours.
Description of the whole sample and the subgroups, as defined according to the need for analgesic intervention (see also Fig. 1). Numerical data are
expressed as mean SD or median [interquartile range] (range). Categorical data are expressed as number of patients and percentages. Between-subgroups
comparisons were performed only if they were relevant and if variables were independent of each other (P values are given in the right column).
a
Possible combination of 2 drugs for induction, that is, the sum of rows may exceed 100%.
b
Fifty to one hundred milligrams given as a single dose at the end of surgery.
c
The cutoff value (0.75 g/kg) has been defined according to the distribution, see Results section for details.
d
For the test, the variable tramadol was nested within the 3 main modalities.
30
C. Dual et al.
As the mean PD throughout the stay in the PACU did not
exceed 80% of the preoperative values, we suspected a residual
effect of effects of general anesthesia. This was supported by
the significant effect of the type of intraoperative opioid, in
accordance with the elimination half-life of the drugs. Most of
our patients who received intraoperative remifentanil did not
receive a systematic compensating bolus of morphineas
often practicedbecause we currently use this drug for carotid
or thyroid surgery (here considered as level I). In nonpainful
conditions, the -agonist opioids have a direct miotic effect,
even at subanesthetic doses [5,21-23]. This direct action is
attributed to a blockade of inhibition of the Edinger-Westphal
nucleus neurons [8]. In nociceptive conditions, this direct
action is concomitant to the reduction of pain-induced pupil
dilation (PDR) through analgesia, a phenomenon evidenced in
subjects under general anesthesia [5-8]. In addition, the
opioid-induced myosis is dose dependentwhatever the
nociceptive stimulationwith a floor effect slightly inferior
to 2 mm [5].
As in the 2 previous studies conducted in PACU, we
intentionally included a heterogeneous population and had a
liberal attitude toward the practice of anesthesia and analgesia,
aiming to reflect routine clinical practice [13,14]. In such
observational design, it is difficult to distinguish the effects of
pain (and analgesia in general) and those of opioids
administered postoperatively. Nevertheless, our multivariable
analysis did not find any effect of the dose of opioid
administered before observation, whereas Kantor et al noted
in 27 patients who received morphine titration a mild but
significant reduction of PD by morphine (from 3.8 to 3.4 mm)
[14]. Lower doses of morphine were administered in our study
(7 vs 9 mg). The main information is the observed discrepancy
between the clinical effectiveness of postoperative opioids and
their weak effect on PD (when measured without additional
stimulation). Besides a masking effect of intraoperative opioids
on PD, it cannot be excluded that a part of postoperative
analgesia was a placebo effect, as already evidenced in a similar
setting [24]. In addition, the doses administered postoperatively
are low for safety reasons and then may not block nociceptive
pathways. For example, 1 mg/kg of morphine given in rats acts
on supraspinal aspects of pain perception [25], whereas a 6-fold
superior dose is needed to impair the response of spinal
nociceptive neurons [26].
The current study has some methodological limitations,
mostly due to its observational design. First, the accuracy of pain
scales may be impaired in elderly patients, leading to
underestimation in this population [27]. Second, pupil dilation
can also be elicited by nonpainful conditions, such as mental
tasks [28], anxiety [29], or even spontaneous awakeness [30].
However, PD changes in such conditions are in a much smaller
range than those observed in nociceptive PDR; furthermore,
mental aspects may also influence pain intensity [1]. Third,
another bias may be due to heterogeneity of the sample, in which
small subgroups of patients may have an impaired pupil function.
This is illustrated by the effect of diabetes that we observed and
which had already been reported [31], but impairment may also
31
Main sample
DSE
Tn
Tested
Not tested
Tested
Analysis 3
Analysis 4
DSE
Whole sample
Condition of
examination
Not tested
Not tested
Tested
PD rPD PD rPD
PD
rPD
PD
rPD
.293 .216 .183 .003
.017
.0002
b .0001 b .0001
ND
ND Rf nS and hS DSE (P = .017); Rf hS Rf nS and hS Baseline all
postoperative
conditions
Tn = time of observation in PACU; DSE = delay since extubation; ND = not done; Rf = intraoperative remifentanil; nS = intraoperative sufentanil, normal
doses; hS = intraoperative sufentanil, high doses.
Multivariable analyses using a linear mixed model. Post hoc analyses (Tukey-Kramer's test) were done only when the model was found predictive (ie, P
value for the model b .05). The rPD is the ratio to the baseline preoperative value. The variables Tn, DSE, and Condition of examination are defined in
the Materials and methods section. The modalities Rf, nS, and hS (type of intraoperative opioid) are defined in the Results section.
Acknowledgments
Assistance with the article:
Fig. 3 Intraoperative opioids and postoperative PD. Time course of
the scotopic PD measured at the right eye of the patient, depending on
the type of opioid analgesia administered during surgery. The diagram
displays the grand mean value for all observations and the confidence
interval limits (95%) limits. The 4 points of measurement were on
arrival at the PACU (T0); just after extubation (Text); during the
stay in PACU (PACU, pooled data from all the measured values);
end of the observation study (Tend), that is, either at discharge for
patients who stayed less than 3 hours in the PACU or at T0 + 3 hours
otherwise. The subgroup of patients who received remifentanil is
represented by white circles and a dotted line; those who received
sufentanil at normal doses are represented by light gray circles and a
dashed line; those who received sufentanil at high doses are
represented by dark gray circles and a full line (see Results section
for definition of normal and high sufentanil subgroups).
32
C. Dual et al.
References
[1] Breivik H, Borchgrevink PC, Allen SM, et al. Assessment of pain. Br J
Anaesth 2008;101:17-24.
[2] Wilder-Smith OH, Arendt-Nielsen L. Postoperative hyperalgesia: its
clinical importance and relevance. Anesthesiology 2006;104:601-7.
[3] Chu LF, Angst MS, Clark D. Opioid-induced hyperalgesia in humans:
molecular mechanisms and clinical considerations. Clin J Pain 2008;
24:479-96.
[4] Chapman CR, Oka S, Bradshaw DH, Jacobson RC, Donaldson GW.
Phasic pupil dilation response to noxious stimulation in normal
volunteers: relationship to brain evoked potentials and pain report.
Psychophysiology 1999;36:44-52.
[5] Barvais L, Engelman E, Eba JM, Coussaert E, Cantraine F, Kenny GN.
Effect site concentrations of remifentanil and pupil response to noxious
stimulation. Br J Anaesth 2003;91:347-52.
[6] Larson MD, Kurz A, Sessler DI, Dechert M, Bjorksten AR, Tayefeh F.
Alfentanil blocks reflex pupillary dilation in response to noxious
stimulation but does not diminish the light reflex. Anesthesiology
1997;87:849-55.
[7] Larson MD, Berry PD, May J, Bjorksten A, Sessler DI. Latency of
pupillary reflex dilation during general anesthesia. J Appl Physiol
2004;97:725-30.
[8] Larson MD. Mechanism of opioid-induced pupillary effects. Clin
Neurophysiol 2008;119:1358-64.
[9] IdMed. Analgesia. http://www.idmed.fr/pupillometrie-analgesie-105Analgesia.html. [Last access 27 June 2014].
[10] Synapsis. Video-AlgesiGraph. http://www.synapsys.fr/en/p-videoalgesigraph-vag-62.htm. [Last access 27 June 2014].
[11] Oka S, Chapman CR, Kim B, Nakajima I, Shimizu O, Oi Y. Pupil
dilation response to noxious stimulation: effect of varying nitrous
oxide concentration. Clin Neurophysiol 2007;118:2016-24.
[12] Yang LL, Niemann CU, Larson MD. Mechanism of pupillary reflex
dilation in awake volunteers and in organ donors. Anesthesiology
2003;99:1281-6.
[13] Aissou M, Snauwaert A, Dupuis C, Atchabahian A, Aubrun F,
Beaussier M. Objective assessment of the immediate postoperative
analgesia using pupillary reflex measurement: a prospective and
observational study. Anesthesiology 2012;116:1006-12.
[14] Kantor E, Montravers P, Longrois D, Guglielminotti J. Pain
assessment in the postanaesthesia care unit using pupillometry: a
cross-sectional study after standard anaesthetic care. Eur J Anaesthesiol 2014;31:91-7.
[15] Chanques G, Viel E, Constantin JM, et al. The measurement of pain in
intensive care unit: comparison of 5 self-report intensity scales. Pain
2010;151:711-21.
[16] Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the
Observer's Assessment of Alertness/Sedation Scale: study with
intravenous midazolam. J Clin Psychopharmacol 1990;10:244-51.
[17] Agence Nationale de Scurit du Mdicament et des produits de sant.
OXYNORM, Rsum des caractristiques du produit. http://agence-
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]