ORIGINAL ARTICLE

Skin conductance as a pain assessment tool during chest tube

removal: An observational study
J.O. Hansen1, H. Storm1, A. Boglino-Ho
€ rlin2,3, M. Le Guen2,3, E. Gayat4,5, M. Fischler2,3
1 The Simulation Center, Division of Emergencies and Critical Care, Institute of Clinical Medicine, University of Oslo, Norway
2 Department of Anesthesiology, Ho ^pital Foch, Suresnes, France
3 Universit
e Versailles Saint-Quentin en Yvelines, France
4 Department of Anesthesiology and Critical Care Medicine, Ho ^pital Saint Louis-Lariboisie
re-Fernand Widal, Paris, France
5 UMR-S 942, INSERM, University Paris 7, Diderot, France

Correspondence Abstract
Julian Okazaki Hansen
E-mail: [email protected] Background: Skin conductance variability to assess pain has shown
varying results. Skin conductance responses per second (SCR) during a
Funding sources standardized painful stimulus in awake adults may give further
Funding was supplied by the Department of understanding of the method’s validity. The purpose of this study was to
Anesthesiology, Foch University Hospital,
validate the SCR with the visual analogue scale (VAS) for pain (P-VAS)
Suresnes, France.
and anxiety (A-VAS) during chest tube removal (CTR).
Conflicts of interest Methods: Ninety-five patients receiving epidural or non-epidural
Hanne Storm, Professor, MD, PhD, The Sim- treatment, scheduled for CTR, were studied. Pain or anxiety was
ulation Center, Division of Emergencies and considered when VAS > 30 mm; the SCR cut-off value reflecting pain
Critical Care, Department of Anesthesiology, was ≥0.2 SCR.
Institute of Clinical Medicine, University of Results: SCR values could not be recorded in eight cases before CTR,
Oslo, Norway is founder and co-owner of
six cases during CTR and seven cases after CTR. CTR induced increases
Med-Storm Innovation AS which owns the
in SCR, P-VAS and A-VAS (p < 0.001). Seventy-seven percent of all
patents for the skin conductance technology
used to assess pain in this study. pairs of P-VAS and SCR values were well-classified; P-VAS ≤ 30 mm
and SCR < 0.2 or P-VAS > 30 mm and SCR ≥ 0.2. SCR obtained before
CTR differentiates between patients with and without pain during CTR
Accepted for publication in all patients (p = 0.04) and in the subgroup of non-anxious patients
29 November 2016 (p = 0.02), but not in the subgroup of anxious patients. SCR obtained
during CTR had similar values in patients with and without pain in all
doi:10.1002/ejp.999
patients and in the subgroup of anxious patients, but in the subgroup of
non-anxious patients SCR during CTR differentiates patients with and
without pain (p = 0.009).
Conclusions: SCR increases during painful procedures. Preprocedural
SCR may help predict reported pain in patients exposed to painful
procedures. SCR during CTR differentiates between patients with and
without pain only in non-anxious patients.
Significance: Preprocedural SCR may help predict reported pain in
patients exposed to painful procedures. Procedural SCR accuracy
improves in a subgroup of non-anxious patients. P-VAS is influenced by
anxiety different from SCR.

1. Introduction and tears, has a low specificity for pain and must
therefore be replaced by a more specific monitoring
Pain assessment in anaesthetized and non-communi-
technique. At least four methods are commercially
cative patients remains a problem. Use of clinical
available: pupillometry, analgesia/nociception index
signs such as tachycardia and hypertension, or sweat

© 2017 European Pain Federation - EFICâ Eur J Pain  (2017) – 1

Skin conductance and pain assessment J. O. Hansen et al.

(ANI), surgical pleth index and the skin conductance behavioural factors. Reported pain is influenced by
algesimeter (SCA). These techniques assess nocicep- factors such as previous pain experiences, gender,
tion and anti-nociception through the balance cultural background, sociodemographic factors, anxi-
between activity of the sympathetic and parasympa- ety and depression. The pain experience may also be
thetic nervous system. Pupillometry provides accu- intensified when accompanied by pain-related anxi-
rate measurement of the pupillary reflex to a ety (Kain et al., 2000; Lid
en et al., 2009).
nociceptive stimulus using an infrared pupilometer Chest tube removal (CTR) is a standardized,
(Aissou et al., 2012). ANI is a 0–100 index derived potentially painful stimulus which is commonly
from the heart rate variability reflecting the associated with pain, burning and pulling (Gift et al.,
parasympathetic tone (Boselli et al., 2014). SPI pro- 1991), and can be the worst hospital experience for
vides an index of the nociception–anti-nociception some patients (Casey et al., 2010). In a study, ‘fear-
balance obtained from pulse and photoplethysmog- ful’ was the most frequently used word to describe
raphy, both mirroring sympathetic activity (Wen- the affective dimension of this CTR-related pain
nervirta et al., 2008). (Puntillo, 1994). The purpose of this study was to
SCA is based on skin conductance variability compare the self-evaluation scores of patients using
assessment which is a non-invasive method for indi- the visual analogue scale (VAS) for pain and anxiety
rect estimation of skin sympathetic nerve activity. and the SCR during CTR in patients receiving two
Pain is associated with increased sympathetic ner- different modalities of postoperative analgesic treat-
vous activity leading to increased palmar and plantar ment during routine care (epidural analgesia and
secretion of sweat, which in turn causes increased multimodal parenteral analgesia) to validate the SCR
skin conductance responses per second (SCR); the to assess pain.
SCA index. An increase in the SCR may be inter-
preted as increased activity in the sympathetic ner-
2. Materials and methods
vous system and may thus serve as a surrogate
marker of pain. Measuring of SCR has the advan- This prospective study was conducted at the Respira-
tage, over ANI and SPI, of not being influenced by tory Intensive Care Unit of the Department of Tho-
haemodynamic changes, neuromuscular blockade, racic Surgery at Foch University Hospital (Suresnes,
adrenergic receptor active agents or environmental France). The study was approved by the Ile de
temperature within normal range, because the skin France VIII Ethical Committee (June 5, 2013; N°
sympathetic nerve releases acetylcholine that acts on 130529) and registered on the Clinicaltrials.gov web
muscarinic receptors (Hagbarth et al., 1972; Wallin site (NCT02015858).
et al., 1975; Bini et al., 1980; Macefield and Wallin, Patients were recruited for the study between
1996). Each time the skin sympathetic nervous sys- June 2013 and April 2014. After informed consent
tem is activated, the palmar and plantar sweat glands was obtained, patients aged between 18 and 75 years
are filled up, the skin resistance is reduced, and skin with a chest tube in place after surgical lung proce-
conductance increases before the sweat is reabsorbed dure were included. Patients were excluded if they
and skin conductance again decreases (Edelberg, suffered from a central or peripheral neurological
1967; Christie, 1981). This creates a skin conduc- disease or insulin-dependent diabetes with dysau-
tance response and the size of the response depends tonomia. Pregnant or lactating patients were also
on how forcefully the skin sympathetic nerve is fir- excluded along with patients with a previous history
ing. The SCR is specific for the stimulus, which of chest drainage.
induces the response and is evident within 1–2 s The choice of analgesic therapy for each patient
after stimulation. Therefore SCR may better correlate was not modified by the study. In patients with an
to pain responses than changes in blood pressure epidural catheter at the time of CTR, a combined
and heart rate, at least in circulatory unstable bolus of a local anaesthetic (levobupivacaine
patients (Gjerstad et al., 2008). 0.125%) and an opioid (sufentanil 0.025%), both
Validation of all these techniques is difficult and in varying doses was administered 10–30 min
the use of a standardized, potentially painful stimu- before the CTR. Patients with no epidural catheter
lus in awake patients permits study of the relation- at the time of CTR received peroral analgesics such
ship between pain reported by the patients and the as opioids, paracetamol and NSAIDs in different
monitor’s evaluation. Evaluation in awake patients combinations and in varying dosages; the timing of
must take into account that pain is a subjective these medications was not modified by the time
experience involving sensory, emotional and of CTR.

2 Eur J Pain  (2017) – © 2017 European Pain Federation - EFICâ

J. O. Hansen et al. Skin conductance and pain assessment

A skin conductance monitor, the SCA (Med-Storm 2.1 Statistics

Innovation AS, Oslo, Norway), connected to three
In this study, the primary endpoint was to study if
electrodes on the palmar skin, was used (Fig. 1).
the SCA index increased during the painful stimulus.
Two electrodes, C (current) and R (reference) were
According to other studies, 20 patients and even
used in a feedback configuration to apply an exact
fewer patients have been sufficient to find differ-
and constant alternating current between R and the
ences from before to during painful stimuli
third M (measurement) electrode. The recorded
(Ledowski et al., 2006, 2007; Gjerstad et al., 2007;
return current from M then provided direct informa-
Eriksson et al., 2008; Gunther et al., 2013), 95
tion on skin conductance (Eriksson et al., 2008). To
patients should therefore be sufficient. Results are
minimize movement artefacts, the electrodes were
expressed as medians and first to third quartiles or
attached to the hand least likely to move. The SCA
counts and percentages. The main outcome measure
was connected to a laptop computer whose software
was the pain score obtained before, during and after
recorded the skin conductance data in three inter-
CTR. Marginal associations between single variables
vals, a 30-s interval during drain removal and 60-s
and outcome were assessed by Wilcoxon’s rank-sum
intervals before and after drain removal. Skin con-
test for quantitative variables and Fisher’s exact test.
ductance is measured as SCR where the change in
Patients’ characteristics, including SCR before, during
conductance is defined as greater than a chosen
and after CTR were compared using Friedman’s test
threshold to define a response, which was set at
for multiple comparisons. The effect of anxiety when
0.005 lsiemens.
SCR was validated was assessed by repeating com-
VAS for anxiety (A-VAS) and for perceived pain
parison in the subgroup of patients with and without
(P-VAS), heart rate (HR) and respiratory rate (RR)
anxiety. Furthermore, to test the strength of the
were recorded 5 min before CTR ‘before’, during CTR
data, cut-off values for pain of 20 and 40 mm were
‘during’ and 5 min after CTR ‘after’. VAS consisted of
used as well. Moreover, to investigate whether high
a horizontal line, 100 mm in length anchored by two
levels of anxiety induced a stronger association
verbal descriptors, one for each symptom extreme.
between anxiety and pain, linear regression analysis
We chose a cut-off of 30 mm between none to mild
was used between pain and anxiety, before, during
and moderate to severe pain (Collins et al., 1997)
and after CTR. A 2-sided p-value of 0.05 was consid-
and 0.20 for SCR (Ledowski et al., 2006, 2007; Gun-
ered significant. All of the analyses were performed
ther et al., 2013) to determine a significant change in
with R statistical software, version 2.10.2 (R Founda-
SCR. Additionally, for preprocedural SCR we studied
tion for Statistical Computing, Vienna, Austria).
how these values predicted the procedural P-VAS,
and for procedural SCR how these values are associ-
ated with the procedural P-VAS with the different 3. Results
cut-off values for anxiety A-VAS: 30, 40 and 50 mm
Patients were consecutively studied owing they
(Facco et al., 2013). Also, P-VAS cut-off values of 20
agreed to participate and if one of the investigators
and 40 mm were used to test the strength of the
(J.H. or M.L.) was present. There were 100 patients
results (Jensen et al., 2003; Hirschfeld and Zernikow,
who were recruited, of these two patients had
2013a,b; Boonstra et al., 2014).

Figure 1 A skin conductance monitor, the Skin Conductance Algesimeter (SCA) (Med-Storm Innovation AS), connected to three electrodes on the
palmar skin was used.

© 2017 European Pain Federation - EFICâ Eur J Pain  (2017) – 3

Skin conductance and pain assessment J. O. Hansen et al.

technological problems with the skin conductance Table 2 Patient demographics including comparisons between the
equipment and therefore withdrawn and three patient groups receiving non-epidural and epidural analgesics. Mann–
Whitney U-test was used to compare the different patient groups.
patients had withdrawn their consent forms. A total
of 95 patients undergoing the CTR procedure were Non-epidural Epidural
therefore studied. SCR values could not be recorded analgesia analgesia
(n = 51) (n = 44) p-value
in eight cases before CTR, six cases during CTR and
seven cases after CTR due to technical reasons or to Age (years) 55 (35–65) 57 (47.8–66) 0.399
significant movements of the patients’ hands. Female gender 24 (47.1) 17 (38,6) 0.466
Demographics for all patients are presented in BMI (kg/m²) 22.8 (19.3–26.4) 22.5 (19–24.9) 0.823
Type of surgery
Table 1, and detailed information regarding the
Partial or total lung 21 (41) 32 (74)
patients who received epidural and non-epidural
resection
treatments is shown in Table 2. Patients received Lung transplant 13 (26) 4 (9)
24F or 28F chest tubes. Videothoracoscopy 14 (28) 6 (14)
Numbers of patients with anxiety and pain at any Surgery 15/36 7/37 0.122
time during the procedure are reported in Table 3. (thoracoscopy/
CTR induced immediate increases in SCR, A-VAS, thoracotomy)
Number of chest tubes
P-VAS, HR and RR as compared to the preprocedural
1 11 (21.6) 7 (15.9) 0.491
values (p < 0.001), and these variables decreased
2 28 (54.9) 33 (75)
within 5 min after drain removal (p < 0.001) 4 12 (23.5) 4 (9.1)
(Table 3). Postoperative day 5 (3–8) 3 (3–4) <0.0001
All pairs of SCR and P-VAS data collected before, VAS for anxiety
during and after CTR are presented in (Fig. 2 panel Before 20 (6–45) 5 (0–18.8) 0.002
A). ‘Well-Classified’ pairs correspond to two cases: During 50 (15–80) 13.5 (0–39.8) 0.000
5 min after 0 (0–30) 0 (0–1.5) 0.005
P-VAS ≤ 30 mm and SCR < 0.2, P-VAS > 30 mm
Heart rate
and SCR ≥ 0.2. Seventy-seven percent of all pairs
Before 84 (72–101) 87 (80.3–97.8) 0.514
were well-classified (Figure 2, panel B). Seventy- During 93 (80–105) 90.5 (84–99.5) 0.560
eight percent of pairs of SCR and P-VAS data were 5 min after 84 (72–102) 87 (76–94.8) 0.988
well-classified among non-anxious patients while Respiratory rate
this number is 76% among anxious ones (Figure 2, Before 20 (16–24) 20 (16.3–24) 0.834
panel B). During 24 (20–29) 20.5 (17–27.5) 0.250
5 min after 20 (16–26) 20 (16–24) 0.661
SCR obtained before CTR differentiated between
VAS for pain
patients with and without pain during CTR (Fig. 3,
Before 10 (0–20) 0 (0–18.75) 0.499
During 60 (35–80) 17.5 (0–40) <0.0001
Table 1 Patient demographics. 5 min after 10 (0–25) 0 (0–10) 0.021
Skin conductance
All patients
Before 0.05 (0–0.085) 0.05 (0.02–0.125) 0.337
(n = 95)
During 0.33 (0.27–0.4) 0.33 (0.2–0.435) 0.633
Age (years) 55.8 (42.5–65.8) 5 min after 0.05 (0–0.1225) 0.07 (0.015–0.12) 0.633
Female gender 41 (44%)
BMI (kg/m²) 22.5 (19.2–25.0)
Surgical procedures panel B; p = 0.04) and SCR were similar during CTR
Partial or total lung resection 53 (56%) in patients with pain (P-VAS > 30 mm) and in
Pleural procedure 21 (22%) patients with no pain (P-VAS ≤ 30 mm) (Fig. 3,
Lung transplant 17 (18%) panel A). The same predictive property of preproce-
Other 4 (4%) dural SCR is found among non-anxious patients
Surgical approach (thoracotomy/thoracoscopy) 74/21
(Table 4), but not among anxious patients (Table 4).
Number of chest tubes
1 17 (18%)
In the subgroup of non-anxious patients during CTR,
2 62 (65%) SCR differentiated between patients with and with-
4 16 (17%) out pain during CTR (Table 5), different from the
Postoperative day 4 (3–5) subgroup of anxious patients (Table 5). When testing
Epidural analgesia 44 (46%) the strength of these results, the cut-off value for
Results are expressed as median (first quartile – third quartile) or pain was increased to 40 mm and decreased to
count (percentage) Abbreviations: BMI, body mass index; VAS, visual 20 mm. The predictive strength of the SCR for
analogue scale. patients with A-VAS ≤ 30 mm before CTR

4 Eur J Pain  (2017) – © 2017 European Pain Federation - EFICâ

J. O. Hansen et al. Skin conductance and pain assessment

Table 3 Anxiety-VAS (A-VAS), pain-VAS (P-VAS), skin conductance responses per sec (SCR), heart rate (HR) and respiratory rate (RR) before, during
and after chest tube removal (CTR).

Before CTR During CTR After CTR p-value

A-VAS 15 (0–35) 30 (10–60) 0 (0–10) <0.001

% of anxious patients 26% 48% 9% <0.001
P-VAS 5 (0–20) 40 (10–70) 5 (0–18) <0.001
% of patients with pain 16% 62% 14% <0.001
SCR 0.05 (0.00–0.10) 0.33 (0.20–0.40) 0.05 (0.00–0.12) <0.001
% of patients with SCR ≥ 0.2 10% 84% 6% <0.001
HR 86 (78–98) 91 (82–101) 87 (76–100) <0.001
RR 20 (16–23) 22 (17–26) 20 (16–24) <0.001

Results are expressed as median [first quartile – third quartile] or count (percentage). A patient is qualified as anxious when Visual Analogue Scale
for anxiety (A-VAS) is >30 on a 0–100 mm scale. A patient is qualified as painful when Visual Analogue Scale for pain (P-VAS) is >30 on a 0–
100 mm scale. Friedman’s test for multiple comparisons was used to test for equalities.

Figure 2 Pairs of skin conductance responses per second (SCR) and Pain-Visual Analogue Scale (P-VAS) data points. Panel A: Scatterplot of all
pairs collected before, during and after chest tube removal (CTR). Panel B: Percentages of ‘well-classified’ patients. A ‘Well-Classified’ pair corre-
sponds to two cases: P-VAS ≤ 30 mm and SCR <0.2, P-VAS > 30 mm and SCR ≥ 0.2. Percentages are provided for all patients and for before, dur-
ing and after CTR, as well as in non-anxious and anxious patients.

disappeared when the cut-off for P-VAS was high pain, with 0.33 (0.20–0.47) (n = 37) and 0.40
≤40 mm during CTR (see amendment Table 6). (0.33–0.60) (n = 9) respectively, p = 0.02 (see
Interestingly, the predictive value of SCR was sus- amendment Table 7). During CTR, for patients with
tained for patients with A-VAS ≤ 30 mm before CTR A-VAS ≤ 30 mm and P-VAS ≤ 20 mm, the differ-
when the cut-off value for P-VAS was ≤20 mm dur- ence between no or mild pain and higher levels of
ing CTR, with SCR before CTR of 0.02 (0–0.06) pain was also sustained with 0.27 (0.13–0.40)
(n = 28) for no pain or mild pain during CTR and (n = 27) and 0.40 (0.33–0.53) (n = 19), p = 0.003,
SCR before CTR of 0.05 (0.03–0.1) (n = 36) for the respectively. Interestingly, when anxiety level was
higher level of pain during CTR, p = 0.02. For correlated with the pain level before CTR, during
patients with A-VAS ≤ 30 mm during CTR, cut-off CTR, and after CTR, the correlation level increased
value of P-VAS < 40 mm during CTR, sustained the with R = 0.39 (p < 0.001), R = 0.50 (p < 0.001) and
difference between no or mild pain and moderate or R = 0.60 (p < 0.001), respectively.

© 2017 European Pain Federation - EFICâ Eur J Pain  (2017) – 5

Skin conductance and pain assessment J. O. Hansen et al.

Figure 3 Skin conductance responses per second (SCR) measured before (Panel A) and during (Panel B) in all patients without and with pain dur-
ing chest tube removal (CTR).

Table 4 This table shows how Anxiety-VAS (A-VAS) before chest tube removal (CTR) influences the SCR before CTR to predict the Pain-VAS (P-
VAS) during CTR. Skin conductance responses per sec (SCR) is given in median and 25 and 75 quartiles as well as number (n) of patients in each
group. P-VAS cut-off 30 mm.

A-VAS cut-off values and P-VAS ≤ 30 mm P-VAS > 30 mm

SCR before CTR during CTR during CTR p-value

A-VAS ≤ 30 mm 0.02 (0–0.05) (n = 29) 0.05 (0.02–0.1) (n = 35) 0.018

A-VAS ≤ 40 mm 0.02 (0–0.05) (n = 29) 0.05 (0.03–0.08) (n = 40) 0.016
A-VAS ≤ 50 mm 0.02 (0–0.06) (n = 30) 0.05 (0–0.1) (n = 46) 0.046
A-VAS > 30 mm 0.1 (0.02–0.18) (n = 5) 0.07 (0.03–0.12) (n = 18) 0.57
A-VAS > 40 mm 0.1 (0.02–0.18) (n = 5) 0.07 (0.02–0.15) (n = 13) 0.77
A-VAS > 50 mm 0.06 (0.02–0.12) (n = 4) 0.1 (0.07–0.14) (n = 7) 0.57
All patients 0.02 (0–0.07) (n = 34) 0.07 (0.02–0.1) (n = 53) 0.036

Table 5 This table shows how Anxiety-VAS (A-VAS) during chest tube removal (CTR) influences the SCR during CTR to correlate with Pain-VAS (P-
VAS) during CTR. Skin conductance responses per sec (SCR) is given in median and 25 and 75 quartiles as well as number (n) of patients in each
group. P-VAS cut-off 30 mm.

A-VAS cut-off values and P-VAS ≤ 30 mm P-VAS > 30 mm

SCR during CTR during CTR during CTR p-value

A-VAS ≤ 30 mm 0.27 (0.18–0.42) (n = 28) 0.4 (0.33–0.52) (n = 18) 0.0088

A-VAS ≤ 40 mm 0.27 (0.2–0.4) (n = 30) 0.36 (0.29–0.47) (n = 26) 0.078
A-VAS ≤ 50 mm 0.27 (0.16–0.4) (n = 31) 0.33 (0.25–0.47) (n = 32) 0.12
A-VAS > 30 mm 0.4 (0.3–0.44) (n = 7) 0.33 (0.2–0.4) (n = 36) 0.33
A-VAS > 40 mm 0.33 (0.27–0.47) (n = 5) 0.33 (0.27–0.4) (n = 28) 0.74
A-VAS > 50 mm 0.4 (0.32–0.48) (n = 4) 0.36 (0.27–0.4) (n = 22) 0.36
All patients 0.27 (0.2–0.44) (n = 35) 0.33 (0.27–0.4) (n = 54) 0.18

4. Discussion
greater in patients who will report pain during CTR
During CTR, patients showed an increase in A-VAS, in the whole group of patients and in the subgroup
P-VAS, SCR, HR and RR. Preprocedural measure- of patients without anxiety. This result seems to
ment of SCR brings an important message since it is depend upon the chosen threshold to define pain.

6 Eur J Pain  (2017) – © 2017 European Pain Federation - EFICâ

J. O. Hansen et al. Skin conductance and pain assessment

Table 6 This table shows how Anxiety-VAS (A-VAS) before chest tube removal (CTR) influences the SCR during CTR to correlate with Pain-VAS (P-
VAS) during CTR. Skin conductance responses per sec (SCR) is given in median and 25 and 75 quartiles as well as number (n) of patients in each
group. P-VAS cut-off 40 mm.

A-VAS cut-off values and P-VAS ≤ 40 mm P-VAS > 40 mm

SCR before CTR during CTR during CTR p-value

A-VAS ≤ 30 mm 0.03 (0–0.07) (n = 42) 0.05 (0–0.13) (n = 22) 0.32

A-VAS ≤ 40 mm 0.03 (0–0.07) (n = 43) 0.05 (0–0.08) (n = 26) 0.34
A-VAS ≤ 50 mm 0.04 (0–0.75) (n = 44) 0.05 (0–0.11) (n = 32) 0.48
A-VAS > 30 mm 0.08 (0.02–0.18) (n = 6) 0.07 (0.02–0.12) (n = 17) 0.75
A-VAS > 40 mm 0.08 (0.02–0.18) (n = 6) 0.07 (0.02–0.12) (n = 17) 0.73
A-VAS > 50 mm 0.06 (0.02–0.15) (n = 4) 0.1 (0.07–0.14) (n = 7) 0.50
All patients 0.04 (0–0.08) (n = 48) 0.07 (0.01–0.13) (n = 39) 0.24

Table 7 This table shows how Anxiety-VAS (A-VAS) during chest tube removal (CTR) influences the SCR during CTR to correlate with Pain-VAS (P-
VAS) during CTR. Skin conductance responses per sec (SCR) is given in median and 25 and 75 quartiles as well as number (n) of patients in each
group. P-VAS cut-off 40 mm.

A-VAS cut-off values and P-VAS ≤ 40 mm P-VAS > 40 mm

SCR during CTR during CTR during CTR p-value

A-VAS ≤ 30 mm 0.33 (0.20–0.47) (n = 37) 0.40 (0.33–0.60) (n = 9) 0.02

A-VAS ≤ 40 mm 0.33 (0.20–0.47) (n = 41) 0.33 (0.24–0.50) (n = 15) 0.54
A-VAS ≤ 50 mm 0.33 (0.20–0.47) (n = 42) 0.33 (0.20–0.40) (n = 21) 0.80
A-VAS > 30 mm 0.40 (0.30–0.44) (n = 12) 0.33 (0.20–0.40) (n = 31) 0.18
A-VAS > 40 mm 0.33 (0.20–0.44) (n = 8) 0.33 (0.27–0.40) (n = 25) 0.97
A-VAS > 50 mm 0.33 (0.30–0.44) (n = 7) 0.40 (0.27–0.40) (n = 19) 0.84
All patients 0.33 (0.20–0.47) (n = 49) 0.33 (0.27–0.40) (n = 40) 0.69

Higher cut-off values for pain do not sustain the pre- Several methods have been used to evaluate the
dictive demeanour of SCR different from lower cut- usefulness of skin conductance as a monitor of pain.
off values for pain. The reason may be that a high SCR has been shown to change in the same way
cut-off value for pain during CTR may include more as stress measures such as blood pressure, heart rate,
patients with anxiety and the accuracy between SCR norepinephrine and epinephrine levels during gen-
and pain decreases. Interestingly, the procedural eral anaesthesia (Storm et al., 2002), but less con-
measurement of SCR was greater in patients with vincing results have been shown in a similar study
pain when compared to patients without pain only (Ledowski et al., 2010).
in the subgroup of patients without anxiety. This Some of the studies were performed during gen-
result did not depend upon the chosen cut-off values eral anaesthesia using a tetanic stimulation as a cali-
to define pain between 20 and 40 mm. Moreover, brated painful stimulus. Thus, Gjerstad et al. (2007)
the modality of analgesia also influenced the level of and Storm et al. (2013) showed that SCR increased
reported pain and anxiety. The patients receiving after the stimulus. On the opposite, Sabourdin et al.
epidural analgesic treatment reported less pain and (Sabourdin et al., 2013) reported that it was not
anxiety compared to the non-epidural patient group. modified in a similar experiment performed in chil-
The difference in reported anxiety and pain between dren, but they used a threshold to analyse SCR
the treatment groups may in part be explained by which was not in accordance with the recommenda-
the fact that the patient with non-epidural pain tions from the manufacturer (Storm, 2013).
analgesia treatment had their chest tubes in place for More interesting for SCR are the situations where
a longer time period compared to the patients who pain can be evaluated using validated pain scales.
had epidural analgesia treatment. Patients with Use of SCR to detect postoperative pain during the
epidural analgesia may have had a postoperative early postoperative period in the postanaesthesia
period with less complications and therefore less care unit was studied by Ledowski et al. (2009) and
pain and anxiety when compared to the non- by Czaplik et al. (2012) in adults. Ledowski et al.
epidural patient group. The surgery methods did (2006, 2007) showed that the sensitivity to discover
not influence the reported pain, anxiety and SCR moderate and severe pain was about 90% and the
during CTR. specificity was about 70%. Moreover, Ledowski

© 2017 European Pain Federation - EFICâ Eur J Pain  (2017) – 7

Skin conductance and pain assessment J. O. Hansen et al.

et al. (2009) showed that the SCR and surgical stress evaluation of a pain monitor in a conscious patient
index differed significantly when the numeric rating has its own limitation since it is difficult or even
scale was lower or >5 on a 0–10 scale while these impossible to distinguish between reported pain per
indices identified time points with moderate to sev- se and anxiety. It could be interesting to repeat simi-
ere pain with only moderate sensitivity and speci- lar studies with anxious patients receiving anxiolytic
ficity, different from Ledowski’s previous studies treatment before the painful procedure.
(Ledowski et al., 2006, 2007). Czaplik et al. (2012) Our study has some limitations
used skin conductance to direct pain therapy and We did not calculate a number of patients to be
concluded that it failed to distinguish pain from included, but our sample of patients appears large
other stressors in postoperative adult patients. Other enough, and in accordance with similar studies
studies have been performed during the same period (Ledowski et al., 2006, 2007, 2009; Hullett et al.,
but in children. Hullett et al. (2009) concluded that 2009; Choo et al., 2010; Czaplik et al., 2012), which
measurement of SCR may provide an additional tool included patients with various levels of pain and of
for pain assessment while Choo et al. (2010) stated anxiety.
that SCR measurement is not specific for postopera- The patients were given different pain medication
tive pain intensity being unable to identify analgesia and it is therefore difficult to develop a scale saying
requirements when compared with self-report mea- which patient received more or less pain killers
sures. Furthermore, Ledowski et al. (2007) showed when local anaesthetic, opioids, paracetamol and
that SCR decreases after analgesia treatment, and NSAIDS were used separately or in combination. We
reported that both pain and SCR decrease after anal- have therefore not done quantification of analgesics
gesia treatment (Ledowski et al., 2006). used and correlated these levels of pain medication
All these methods of evaluation of a pain monitor with the reported pain, the reported anxiety or the
device have their pitfalls. No study was performed SCR. Moreover, opioids have anxiolytic properties
on surgical patients to improve our knowledge. This that can be a confounding factor. It has been shown
constitutes a simple model since patients can self- in the postoperative setting that anxiolytics does not
evaluate their pain. We used CTR as the painful pro- attenuate SCR (Hullett et al., 2009; Choo et al.,
cedure and did not change the patients’ care, some 2010).
of them receiving epidural analgesia, the others mul- We did not record temperature in this study. It is
timodal parenteral analgesia. In this setting, our found that skin resistance is not influenced by the
results showed the importance of asking the patients sudomotor activity, palmary and plantary, between
about anxiety when validating the reported pain. 22 and 40 °C (Bini et al., 1980), indicating that SCR
Different from the reported pain, SCR has not been is not influenced by changes in temperature within
found to be influenced by anxiety in the postopera- normal range.
tive setting (Hullett et al., 2009; Choo et al., 2010). In the postoperative setting, Jensen et al. (2003)
However, the skin conductance level, in microsie- suggest a P-VAS > 44 mm to define moderate pain,
mens, has been influenced by anxiety outside the while a recent study considering chronic muscu-
perioperative setting (McDonnel and Carpenter, loskeletal pain suggests P-VAS > 34 mm (Boonstra
1959; Kissel and Littig, 1962; Lader, 1967; Siepmann et al., 2014). Other authors have highlighted that
et al., 2007). The fact that this P-VAS-predictive and optimal cut-off points are strongly influenced by ran-
P-VAS-procedural property of SCR improve in sub- dom fluctuations within a sample (Hirschfeld and
groups with low reported anxiety corroborates with Zernikow, 2013a,b). According to the protocol, we
earlier findings that a pain experience may be inten- dichotomized patients to groups of perceived pain,
sified when accompanied by pain-related anxiety no/mild and moderate/high with a P-VAS cut-off
(Kain et al., 2000; Ozalp et al., 2003) and the fact value of 30 mm (Collins et al., 1997). To test the
that anxiety has been shown to predict reported pain strength of the results, we also tested with P-VAS
(Kain et al., 2000). The reported pain, to a larger cut-off values of 20 and 40 mm. We also used differ-
extent, mirrors the perceived anxiety. Interestingly, ent A-VAS cut-off values to test how anxiety influ-
anxiety and pain levels were correlated more enced the correlation between pain and SCR. The
strongly after CTR than during, and more strongly latter is disputable, but more extensive anxiety
during than before CTR. This association did there- assessment tools, as the Spielberger’s State Trait
fore not seem to be related to the level of pain and Anxiety Inventory, would be difficult to use in our
anxiety, because both pain and anxiety were highest setting. The A-VAS was introduced in 1976 (Horn-
during CTR. Consequently, our method of blow and Kidson, 1976) and some studies have

8 Eur J Pain  (2017) – © 2017 European Pain Federation - EFICâ

J. O. Hansen et al. Skin conductance and pain assessment

assessed its reliability mainly comparing it to STAI immediate postoperative pain using the analgesia/nociception index:
A prospective observational study. Br J Anaesth 112, 715–721.
(Millar et al., 1995; Palmer-Bouva et al., 1998; Kind- Casey, E., Lane, A., Kuriakose, D., McGeary, S., Hayes, N., Phelan, D.,
ler et al., 2000; Boker et al., 2002). It has also been Buggy, D. (2010). Bolus remifentanil for chest drain removal in
used for the evaluation of premedication or anaes- ICU: A randomized double-blind comparison of three modes of
analgesia in post-cardiac surgical patients. Intensive Care Med 36,
thetic drug requirements (Chen et al., 2008; Tirault 1380–1385.
et al., 2010). We used the same cut-off of 30 mm to Chen, C.C., Lin, C.S., Ko, Y.P., Hung, Y.C., Lao, H.C., Hsu, Y.W.
separate non-anxious and anxious patients while (2008). Premedication with mirtazapine reduces preoperative anxiety
and postoperative nausea and vomiting. Anesth Analg 106, 109–113.
Facco et al. (2013) reported that A-VAS around Choo, E.K., Magruder, W., Montgomery, C.J., Lim, J., Brant, R.,
50 mm is a reliable threshold for a clinically mean- Ansermino, J.M. (2010). Skin conductance fluctuations correlate
ingful level of preoperative anxiety. Interestingly, poorly with postoperative self-report pain measures in school-aged
children. Anesthesiology 113, 175–182.
when we included patients with anxiety level higher Christie, M.J. (1981). Electrodermal activity in the 1980s: A review. J R
than 30 mm in the statistical analyses, the statistical Soc Med 74, 616–622.
correlation between SCR and P-VAS decreased. Collins, S.L., Moore, R.A., McQuay, H.J. (1997). The visual analogue
pain intensity scale: What is moderate pain in millimetres? Pain 72,
Finally, we used a cut-off of 0.2 for SCR since 95–97.
Ledowski et al. (2007) have shown that this level Czaplik, M., Hubner, C., Kony, M., Kaliciak, J., Kezze, F., Leonhardt,
was associated with a numerical rating scale of four S., Rossaint, R. (2012). Acute pain therapy in postanesthesia care
unit directed by skin conductance: A randomized controlled trial.
and above. PLoS ONE 7, e41758.
In conclusion, our results suggest that skin con- Edelberg, R. (1967). Electrical properties of the skin. In Methods in
ductance could be a method that may guide care- Psychophysiology, C.C. Brown, ed. (Baltimore: Williams Wilkins) pp. 1–59.
Eriksson, M., Storm, H., Fremming, A., Schollin, J. (2008). Skin
givers to proactively identify patients at risk for conductance compared to a combined behavioural and physiological
insufficient analgesia before and during painful pro- pain measure in newborn infants. Acta Paediatr 97, 27–30.
cedures in patients without anxiety. These findings Facco, E., Stellini, E., Bacci, C., Manani, G., Pavan, C., Cavallin, F.,
Zanette, G. (2013). Validation of visual analogue scale for anxiety
offer new therapeutic objectives such as pretreat- (VAS-A) in preanesthesia evaluation. Minerva Anestesiol 79, 1389–
ment with anxiolytics of anxious patients ahead of 1395.
evaluation of analgesia needed prior to painful pro- Gift, A.G., Bolgiano, C.S., Cunningham, J. (1991). Sensations during
chest tube removal. Heart Lung 20, 131–137.
cedures. Gjerstad, A.C., Storm, H., Hagen, R., Huiku, M., Qvigstad, E., Raeder, J.
(2007). Comparison of skin conductance with entropy during
intubation, tetanic stimulation and emergence from general
Author contributions anaesthesia. Acta Anaesthesiol Scand 51, 8–15.
Gjerstad, A.C., Wagner, K., Henrichsen, T., Storm, H. (2008). Skin
J.H., H.S., M.L. and M.F. contributed to the conception conductance versus the modified COMFORT sedation score as a
and design of the study. J.H. and M.L. collected data. J.H., measure of discomfort in artificially ventilated children. Pediatrics 122,
H.S., A.B., E.G. and M.F. participated in data analysis and e848–e853.
Gunther, A.C., Bottai, M., Schandl, A.R., Storm, H., Rossi, P., Sackey,
interpretation. J.H. and H.S. drafted the manuscript and all
P.V. (2013). Palmar skin conductance variability and the relation to
authors were involved in the revision and final approval of stimulation, pain and the motor activity assessment scale in intensive
the submitted manuscript. H.S. has ownership to the skin care unit patients. Crit Care 17, R51.
conductance device used, see conflict of interest declara- Hagbarth, K.E., Hallin, R.G., Hongell, A., Torebjork, H.E., Wallin, B.G.
(1972). General characteristics of sympathetic activity in human skin
tion. The other authors report no competing interests.
nerves. Acta Physiol Scand 84, 164–176.
Hirschfeld, G., Zernikow, B. (2013a). Cut points for mild, moderate,
and severe pain on the VAS for children and adolescents: What can
References be learned from 10 million ANOVAs? Pain 154, 2626–2632.
Hirschfeld, G., Zernikow, B. (2013b). Variability of “optimal” cut points
Aissou, M., Snauwaert, A., Dupuis, C., Atchabahian, A., Aubrun, F., for mild, moderate, and severe pain: Neglected problems when
Beaussier, M. (2012). Objective assessment of the immediate comparing groups. Pain 154, 154–159.
postoperative analgesia using pupillary reflex measurement: A Hornblow, A.R., Kidson, M.A. (1976). The visual analogue scale for
prospective and observational study. Anesthesiology 116, 1006–1012. anxiety: A validation study. Aust N Z J Psychiatry 10, 339–341.
Bini, G., Hagbarth, K.E., Hynninen, P., Wallin, B.G. (1980). Hullett, B., Chambers, N., Preuss, J., Zamudio, I., Lange, J., Pascoe, E.,
Thermoregulatory and rhythm-generating mechanisms governing the Ledowski, T. (2009). Monitoring electrical skin conductance: A tool
sudomotor and vasoconstrictor outflow in human cutaneous nerves. for the assessment of postoperative pain in children? Anesthesiology
J Physiol 306, 537–552. 111, 513–517.
Boker, A., Brownell, L., Donen, N. (2002). The Amsterdam Jensen, M.P., Chen, C., Brugger, A.M. (2003). Interpretation of visual
preoperative anxiety and information scale provides a simple and analog scale ratings and change scores: A reanalysis of two clinical
reliable measure of preoperative anxiety. Can J Anaesth 49, 792–798. trials of postoperative pain. J Pain 4, 407–414.
Boonstra, A.M., Schiphorst Preuper, H.R., Balk, G.A., Stewart, R.E. Kain, Z.N., Sevarino, F., Alexander, G.M., Pincus, S., Mayes, L.C.
(2014). Cut-off points for mild, moderate, and severe pain on the (2000). Preoperative anxiety and postoperative pain in women
visual analogue scale for pain in patients with chronic undergoing hysterectomy. A repeated-measures design. J Psychosom
musculoskeletal pain. Pain 155, 2545–2550. Res 49, 417–422.
Boselli, E., Bouvet, L., Begou, G., Dabouz, R., Davidson, J., Deloste, Kindler, C.H., Harms, C., Amsler, F., Ihde-Scholl, T., Scheidegger, D.
J.Y., Rahali, N., Zadam, A., Allaouchiche, B. (2014). Prediction of (2000). The visual analog scale allows effective measurement of

© 2017 European Pain Federation - EFICâ Eur J Pain  (2017) – 9

Skin conductance and pain assessment J. O. Hansen et al.

preoperative anxiety and detection of patients’ anesthetic concerns. norepinephrine, the VAS, and CDAS in four different procedures. Gen
Anesth Analg 90, 706–712. Dent 46, 356–360.
Kissel, S., Littig, L.W. (1962). Test anxiety and skin conductance. J Puntillo, K.A. (1994). Dimensions of procedural pain and its analgesic
Abnorm Soc Psychol 65, 276–278. management in critically ill surgical patients. Am J Crit Care 3, 116–122.
Lader, M.H. (1967). Palmar skin conductance measures in anxiety and Sabourdin, N., Arnaout, M., Louvet, N., Guye, M.L., Piana, F.,
phobic states. J Psychosom Res 11, 271–281. Constant, I. (2013). Pain monitoring in anesthetized children: First
Ledowski, T., Bromilow, J., Paech, M.J., Storm, H., Hacking, R., Schug, assessment of skin conductance and analgesia-nociception index at
S.A. (2006). Monitoring of skin conductance to assess postoperative different infusion rates of remifentanil. Paediatr Anaesth 23, 149–155.
pain intensity. Br J Anaesth 97, 862–865. Siepmann, M., Heine, B., Kluge, A., Ziemssen, T., Muck-Weymann, M.,
Ledowski, T., Bromilow, J., Wu, J., Paech, M.J., Storm, H., Schug, S.A. Kirch, W. (2007). The effects of lorazepam on skin conductance
(2007). The assessment of postoperative pain by monitoring skin responses to aversive stimuli in healthy subjects. Clin Auton Res 17,
conductance: Results of a prospective study. Anaesthesia 62, 989–993. 160–164.
Ledowski, T., Ang, B., Schmarbeck, T., Rhodes, J. (2009). Monitoring of Storm, H. (2013). “Pain monitoring in anesthetized children: First
sympathetic tone to assess postoperative pain: Skin conductance vs assessment of skin conductance and analgesia-nociception index at
surgical stress index. Anaesthesia 64, 727–731. different infusion rates of remifentanil”, recommended preset values
Ledowski, T., Pascoe, E., Ang, B., Schmarbeck, T., Clarke, M.W., Fuller, for the skin conductance equipment was not used. Pediatr Anaesth 23,
C., Kapoor, V. (2010). Monitoring of intra-operative nociception: 760–763.
Skin conductance and surgical stress index versus stress hormone Storm, H., Myre, K., Rostrup, M., Stokland, O., Lien, M.D., Raeder, J.C.
plasma levels. Anaesthesia 65, 1001–1006. (2002). Skin conductance correlates with perioperative stress. Acta
Lid
en, Y., Landgren, O., Arner, S., Sj€
oLund, K.F., Johansson, E. (2009). Anaesthesiol Scand 46, 887–895.
Procedure-related pain among adult patients with hematologic Storm, H., Stoen, R., Klepstad, P., Skorpen, F., Qvigstad, E., Raeder, J.
malignancies. Acta Anaesthesiol Scand 53, 354–363. (2013). Nociceptive stimuli responses at different levels of
Macefield, V.G., Wallin, B.G. (1996). The discharge behaviour of single general anaesthesia and genetic variability. Acta Anaesthesiol Scand 57,
sympathetic neurones supplying human sweat glands. J Auton Nerv 89–99.
Syst 61, 277–286. Tirault, M., Foucan, L., Debaene, B., Frasca, D., Lebrun, T., Bernard,
McDonnel, G., Carpenter, J.A. (1959). Anxiety, skin conductance and J.C., Sandefo, I.Van., Elstraete, A.C. (2010). Gabapentin
alcohol; a study of the relation between anxiety and skin premedication: Assessment of preoperative anxiolysis and
conductance and the effect of alcohol on the conductance of subjects postoperative patient satisfaction. Acta Anaesthesiol Belg 61, 203–209.
in a group. Q J Stud Alcohol 20, 38–52. Wallin, B.G., Sundlof, G., Delius, W. (1975). The effect of carotid sinus
Millar, K., Jelicic, M., Bonke, B., Asbury, A.J. (1995). Assessment of nerve stimulation on muscle and skin nerve sympathetic activity in
preoperative anxiety: Comparison of measures in patients awaiting man. Pflugers Arch 358, 101–110.
surgery for breast cancer. Br J Anaesth 74, 180–183. Wennervirta, J., Hynynen, M., Koivusalo, A.M., Uutela, K., Huiku, M.,
Ozalp, G., Sarioglu, R., Tuncel, G., Aslan, K., Kadiogullari, N. (2003). Vakkuri, A. (2008). Surgical stress index as a measure of nociception/
Preoperative emotional states in patients with breast cancer and antinociception balance during general anesthesia. Acta Anaesthesiol
postoperative pain. Acta Anaesthesiol Scand 47, 26–29. Scand 52, 1038–1045.
Palmer-Bouva, C., Oosting, J., deVries, R., Abraham-Inpijn, L.
(1998). Stress in elective dental treatment: Epinephrine,

10 Eur J Pain  (2017) – © 2017 European Pain Federation - EFICâ