Clinical Hub,UW Health Clinical Tool Search,UW Health Clinical Tool Search,Questionnaires,Related

Assessment of Postoperative Nausea Using a Visual Analogue Scale

Assessment of Postoperative Nausea Using a Visual Analogue Scale - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related

Acta Anaesthesiol Scand 2000; 44: 470–474 Copyright C Acta Anaesthesiol Scand 2000
Printed in Denmark. All rights reserved
ISSN 0001-5172
Assessment of postoperative nausea using a visual
analogue scale
Departments of Anaesthesiology, University Hospital Centre, Charleroi and
Biostatistics, University of Liege, Belgium
Background: Assessment of postoperative nausea intensity is
difficult because nausea is a subjective and unpleasant sen-
sation. We propose using the Visual Analogue Scale (VAS) de-
vice to increase the efficiency and precision in the assessment of
nausea. We carried out a pilot study on postoperative patients
suffering from nausea to measure the degree of agreement be-
tween the VAS scores and those given on a 4-point verbal de-
scriptive scale (VDS).
Methods: Postoperative nausea was evaluated by means of a
classical VAS (0–10 cm) device and a 4-point VDS (0Ωno nausea,
1Ωmild, 2Ωmoderate, 3Ωsevere) in 128 surgical spontaneously
complaining patients. Evaluation was repeated 45 min after res-
cue medication given if nausea was intractable, lasted more than
10 min or at the request of the patient. Ordinal logistic re-
gression was used to measure the association between VAS and
VDS and to determine cut-off points on the VAS.
Results: The VAS device was easily understood and used by
patients. VAS scores decreased significantly from 5.5∫2.3 to
S RECENTLY stated by Apfel et al., there is no stan-
dard to measure postoperative nausea intensity
(1). Lack of universal agreement about what degree of
symptomatology constitutes clinically significant
nausea leads to lack of comparability among studies
Assessment of postoperative nausea is usually
based on verbal communication with the patient. The
rating scale most widely used is the verbal descriptive
scale (VDS) which comprises 4–5 word categories, i.e.,
‘none/mild/moderate/severe/unbearable nausea’
(5). The least intense descriptor is usually given a
score 0, the next a score of 1, and so on until each
adjective has a number score associated with it. The
patient’s intensity score is that number associated
with the word he/she chooses as most descriptive of
his/her nausea level. It forces the patient to translate
a feeling into words, which do not express exactly
what a person is experiencing. The same word does
not necessarily mean the same thing to each patient.
The intervals between the word categories do not rep-
1.4∫1.8 after rescue medication (PΩ0.002). Application of ordi-
nal logistic regression to pre- and post-medication data com-
bined yielded an agreement of 86% between VAS and VDS and
the cut-off points on the VAS were estimated as follows :
0–1 (no nausea), 1
–4 (mild), 4
–7 (moderate) and 7
–10 (se-
Conclusion: The VAS method proved to be useful for assessing
quantitative nausea intensity and for testing the efficacy of res-
cue medication. It was found that a cut-off value of 4 on the
VAS may be considered as a critical threshold triggering anaes-
thesiologists or nurses to administer rescue medication.
Received 10 June, accepted for publication 23 September 1999
Key words: Assessment; visual analog scale (VAS); postopera-
tive nausea; PONV.
c Acta Anaesthesiologica Scandinavica 44 (2000)
resent identical steps in nausea intensity. A continu-
ous sensation is artificially transferred into a categori-
cal system. A given change in nausea intensity leads
one individual to rate the nausea in the next category,
whereas another subject still uses the prior category.
Visual Analogue Scales (VAS) have been used to
measure a variety of subjective responses in both psy-
chology and education (6, 7). The VAS represents an-
other method of assessment of a ‘feeling’. Basically, it
consists of a 10-cm-long line. The left end corresponds
to ‘no symptom’ and the right to ‘unbearable symp-
tom’. Patients are asked to make a mark on the line
which represents their level of perceived symptom in-
tensity and the scale is scored by measuring the dis-
tance from the ‘no symptom’ end to the patient’s
mark. It also requests the transfer of a sensation into
another dimension. This is one drawback it has in
common with the VDS, while all others have been
substantially reduced. The reliability and sensitivity
of analogue scales have long been established in pain
studies (8).

Nausea intensity and visual analogue scale
The present study was designed to test the useful-
ness of a VAS in the assessment of postoperative
nausea intensity. In particular, it focuses on the degree
of the relationship and agreement with a 4-point VDS.
The simultaneous use of both scales enabled us to
compare and correlate the results obtained by each
method and to investigate whether or not the VDS
could be substituted by the VAS. Furthermore, the ef-
fect of antiemetic drugs on nausea intensity was also
assessed by using both rating scales simultaneously.
One hundred and twenty-eight consecutive inpatients
complaining spontaneously of postoperative nausea
and vomiting (PONV) were studied. Patients were in-
cluded if they were more than 15 years old, able to
read and understand French, with normal mental
health and hospitalised for elective surgery. Nausea
intensity was assessed using both a visual analogue
scale and a verbal descriptive scale. All patients were
questioned by the same investigator, a trained nurse
on call when a patient complained about postopera-
tive nausea. The assessment was performed at arrival
of the study nurse in the surgical ward. The training
of the research nurse included participation at educa-
tional seminars outlining the pathophysiology of
PONV, methods of nausea assessment focusing on
VAS, patients at highest risk, the pertinent aetiologic
factors, and the prevalence of PONV in hospitalised
The VAS device consisted of a 10-cm horizontal line
with no nausea’ at the left side and ‘worst imaginable
nausea’ at the right. The patient indicated nausea in-
tensity, thus yielding a score obtained by moving a
pointer on that device, by measuring the distance (cm)
from the left end, which represents no nausea, to the
pointer location. At the time of the preoperative visit,
patients were familiarised with a 10-cm VAS device
routinely used for pain assessment. The 4-point VDS
included the following word categories: 0Ωno nausea,
1Ωmild, 2Ωmoderate, and 3Ωsevere nausea (3).
Rescue medications for nausea were given if nausea
was intractable and lasted for at least 10 min, or any
time that the patient requested them. The choice of
the drug was left to the discretion of the attending
anaesthesiologist. Forty-five minutes after rescue
medication (i.v. metoclopramide, alizapride, tropise-
tron or neuroleptics), patients were re-evaluated using
the two scales.
The design of the present study was chosen to
create a situation as close as possible to the usual clin-
ical situation, in which decisions have to be made on
the basis of nausea assessment in each individual
Vomiting was recorded as either present or absent
by direct observation during the two assessment
phases. Retching, defined as a vomit not producing
any liquid, was assessed separately.
Results were expressed as mean∫standard deviation
(SD) or as frequencies. Comparison of VAS measure-
ments before and after treatment was done by Student
paired t-test, whereas the symmetry test was used for
VDS. To assess the degree of agreement between the
VAS and VDS scores, an ordinal logistic regression
analysis was applied (9). This model assumes that the
probability of having no nausea given a VAS measure-
ment can be expressed as P(no nausea)Ω
ªb.VAS), where L(t)Ωexp(t)/{1piexp(t)} is the
logistic function, q
is a threshold parameter and b
a regression coefficient. Likewise, the probabilities of
having mild, moderate or severe nausea, respectively,
can be expressed as:
P(mild nausea)ΩL(q
P(moderate nausea)ΩL(q
P(severe nausea)Ω1ªL(q
It is seen that 3 threshold parameters are needed to
define the 4 VDS categories, as expected. Correspond-
ing cut-off points on the VAS are simply obtained by
calculating the ratios c
/b, and c
b. Estimates of the unknown parameters q and b are
obtained by the maximum likelihood method together
with their standard errors (SE). Significance tests can
be performed on the qs and b. In particular, bΩ0
would indicate no relationship between VDS and
VAS. Likewise, if two consecutive threshold par-
ameters are equal (e.g. q
), then the correspond-
ing neighbouring VDS categories can not be dis-
tinguished with respect to VAS.
All calculations were carried out using the SAS stat-
istical package (SAS Institute, version 6.12, 1996). Re-
sults were considered to be significant at the 5% criti-
cal level (P∞0.05).
Among the 128 surgical patients suffering postopera-
tive nausea, there was a greater proportion of females
(nΩ83, 65%) than of males (nΩ45, 35%). Men, how-
ever, were significantly older than women (58∫17 vs.
47∫19 yr, PΩ0.001). Mean weight was 75.7∫14.4 kg
for men and 65.7∫14.4 kg for women (PΩ0.0003). Dis-
tribution of nauseated patients according to surgical

J. G. Boogaerts et al.
procedures and type of anaesthesia is displayed in
Table 1. There was an increased risk of nausea in pa-
tients undergoing thoracic, gynaecologic, maxillo-
facial, abdominal and neurological procedures
(P∞0.003). In the latter group, most of the patients
underwent disc surgery. As far as type of anaesthesia
Table 1
Distribution of the number of patients according to type of surgery
and type of anaesthesia (GAΩgeneral anaesthesia; LRAΩloco-re-
gional anaesthesia).
Type of anaesthesia
Type of surgery GA LRA Total
Orthopaedics 2 5 7
Neurosurgery 9 0 0
Maxillofacial 16 0 16
Gynaecologic 20 1 21
Urologic 8 9 17
Vascular 4 3 7
Abdominal 19 0 19
Thoracic 24 0 24
ENT 8 0 8
Total 110 (85.9%)18(14.1%) 128 (100%)
Table 2
Distribution of patients according to nausea Visual Descriptive Scale
(VDS) in the pre- and post-treatment phases. For each category,
mean∫SD of nausea Visual Analogue Scale (VAS) score is reported.
Pre-treatment Post-treatment
VDS n (%) VAS (cm) n (%) VAS (cm)
None 0 (0) – 66 (51.6) 0.13∫1.06
Mild 37 (29) 3.06∫1.06
43 (33.6) 2.21∫1.05*
Moderate 51 (40) 5.15∫1.18 10 (7.8) 4.39∫1.30*
Severe 40 (31) 8.00∫1.72 4 (3.1) 6.50∫2.20*
Total 128 (100) 5.46∫2.35 123 (96.1)
* P∞0.001.
one VAS value missing.
VAS and VDS values were missing for five patients due to refusal
of further participation in the study.
Table 3
Visual Analogue Scale (VAS) scores before and after rescue medi-
cations according to different antiemetics given to 121 patients.
Seven patients did not receive any drug.
VAS (cm) VAS (cm)
Drug N Pre-treatment Post-treatment P-value
Alizapride 76 5.41∫2.34 1.45∫1.84 ∞0.0001
Metoclopramide 22 5.52∫2.14 0.97∫1.92 ∞0.0001
Tropisetron 18 5.76∫2.68 1.48∫1.57 ∞0.0001
Neuroleptics 5 6.74∫1.94 2.35∫1.30 0.064
was concerned, we found 14.1% of patients operated
under loco-regional anaesthesia alone (Table 1).
The VAS device was easily understood and used by
patients. All patients completed the VDS rating in the
pre-treatment phase but one patient with mild nausea
did not supply the VAS score. In the post-treatment
phase, VDS and/or VAS values were missing for five
patients due to refusal of further participation in the
The distribution of both VDS and VAS in the pre-
and post-treatment phases are given in Table 2. For
each VDS category, the mean and standard deviation
of the VAS are also given. VAS mean scores increased
with nausea VDS categories in both phases of the
treatment. There was a significant improvement in the
distribution of VDS after rescue medication (PΩ0.002).
More than 50% of the patients recovered from nausea
after treatment, while only 3.3% declared that they
were still experiencing severe nausea as compared to
31% prior to rescue medication. By comparison, the
mean VAS decrease following rescue medication was
slightly more significant (5.46∫2.35 to 1.41∫1.83;
Drugs used for the treatment of nausea available in
121 patients were alizapride (nΩ76), metoclopramide
(nΩ22), tropisetron (nΩ18) or neuroleptics (nΩ5). The
efficacy of the different antiemetics assessed with the
VAS is presented in Table 3. The same significant dif-
ferences were observed when using the VDS.
To measure the degree of agreement between the
ordinal VDS and the continuous VAS in the assess-
ment of nausea in postoperative patients, we pooled
data of the pre- and post-treatment phases. As illus-
trated in Fig. 1, there was an excellent linear relation-
Fig. 1. Relationship between Visual Analogue Scale (VAS) and Verbal
Descriptive Scale (VDS) scores. Data from pre- and post-treatment

Nausea intensity and visual analogue scale
Table 4
Number of patients suffering retching (Retch) and vomiting (Vomit)
before (1) and after (2) rescue medications according to different anti-
emetic drugs.
Drug Retch 1 Retch 2 Vomit 1 Vomit 2
Alizapride 30 1* 14 2*
Metoclopramide 4 2 3 1
Tropisetron 10 1* 3 1
Neuroleptics 2 4 1 4
* P∞0.01.
ship between VDS categories and VAS scores (corre-
lation coefficient rΩ0.92). Mean values∫SDsofVAS
in each VDS category were 0.13∫0.29 (no nausea; nΩ
66), 2.60∫1.13 (mild; nΩ79), 5.02∫1.22 (moderate; nΩ
61), and 7.87∫1.80 (severe nausea; nΩ44), respectively.
By fitting an ordinal logistic regression model to the
VAS data, we found a highly significant regression co-
efficient between VAS and VDS of bΩ2.02 (SE: 0.186).
Moreover, all three thresholds were significantly dif-
ferent from each other, q
Ω2.06 (∫0.308), q
(∫0.768) and q
Ω13.1 (∫1.20), respectively, emphasis-
ing the discriminant ability of the VAS. The degree of
concordance between predicted probabilities of VDS
categories from VAS and observed frequencies
amounted to 86% (2% of discordant and 12% of ties).
From the results above, cut-off limits on the VAS
device were computed to match the VDS categories
(see statistics section). These estimated cut-off points
were respectively c
Ω1, c
Ω3.9 and c
Ω6.5. It follows
that, after some rounding off, the visual analogue
scale can be categorised as 0–1 for no nausea, 1
–4 for
mild nausea, 4
–7 for moderate nausea, and 7
–10 for
severe nausea. To make this point clear, if a patient
scores 5.1 on the VAS, this would almost certainly cor-
respond to ‘moderate nausea’ on the VDS.
Forty-six patients suffered retching during the first
assessment phase and 8 during the second. The same
profile was observed for vomiting with 21 emetic pa-
tients before treatment and 8 after. Results according
to the different antiemetic medications are displayed
in Table 4. The incidence of retching and vomiting in
patients who had received alizapride or tropisetron
was significantly lower than in those who had re-
ceived metoclopramide or neuroleptics (P∞0.01).
The present study validates the use of a VAS device for
nausea assessment in the postoperative period. The
distribution of nauseated patients in regard to the type
of surgery is in line with previous surveys for this clin-
ical setting (1–3, 5). Patients experienced nausea more
frequently after thoracic, gynaecologic, maxillofacial
and abdominal surgeries. Retching and/or vomiting
were also assessed as two different end-points. About
half of the patients with nausea also presented emetic
episodes at the pre-treatment phase.
Regarding the ease of administration and scoring,
from the description of the two scales given above,
both appeared to be extremely simple to use and
score. The estimation of nausea intensity with VAS re-
quires an ability to transform a complex subjective ex-
perience to a visuo-spatial display, which involves
perceptual judgement and accuracy. At this time, VAS
is routinely used by nurses in the PACU and on the
surgical wards for nausea assessment every four
hours during the 72 h postoperative period without
any special problem.
An advantage of VAS over other methods of nausea
intensity measurement is that continuous data have
ratio properties and are well suited for statistical
analysis, unlike categorical scales (10). In the conduct
of the present study, a database was elaborated, which
can serve as a source of information for further stat-
istical analysis and research, and ensures that trends
of postoperative nausea intensity can be addressed.
Difficulty in understanding the VAS scale could oc-
cur, especially among older patients, as shown by a
significant correlation between age and incorrect re-
sponse to the VAS in previous experiments (11). Re-
sults of studies on pain scales indicate that: (a) re-
sponse problems can occur with all scales when sub-
jects are asked to complete them in unsupervised
settings; (b) more response problems may occur if the
measurement task involves more effort on the part of
the patient; and (c) no scale is consistently associated
with more response problems than any other scales,
although older patients may have more difficulty
using the VAS as compared to younger patients (11–
Our study showed a strong relationship (rΩ0.92)
between the VAS and the VDS scores. Ordinal logistic
regression demonstrated a high degree of association
between the two scales (86% of agreement), indicating
that both may be considered as a useful measure of
subjective nausea intensity.
Should VDS categories be predicted from VAS
measures, then the following classes should be used:
0–1 for no nausea, 1
–4 for mild, 4
–7 for moderate,
and 7
–10 for severe nausea. A nausea VAS score of
4 appears to trigger rescue medication. This threshold
can thus be used for the establishment of treatment
guidelines in our hospital, according to the studied

J. G. Boogaerts et al.
Another criterion used to judge nausea intensity
measures is the relative ability to detect treatment ef-
fects. The two scales envisaged here showed a high
sensitivity to treatment effects. VAS and VDS were
able to detect significantly rescue medications effects,
with a slight superiority for VAS. Concerning effec-
tiveness of the different antiemetics, it appeared that
alizapride, tropisetron and metoclopramide were
more efficacious on nausea than neuroleptics but that
only the two former permitted to reduce significantly
retching and vomiting episodes.
A limitation of the study rests upon the choice of
the 4-point verbal descriptive scale. A weaker associ-
ation may have occurred if other scales had been em-
ployed. Also, further studies are needed and should
focus on the imprecision, i.e. repeatability of the
measuring device, and on its use in other settings,
e.g., nausea assessment after chemotherapy.
In conclusion, we have demonstrated that a linear
visual analogue scale is a suitable method for record-
ing the patient’s opinion about nausea intensity for
further research purposes or clinical practice.
The authors would like to thank Novartis Belgium for manufac-
turing the VAS device.
This material has been presented as an abstract at the 7th
European Society of Anaesthesiologists Annual Meeting which
was held in Amsterdam, May 29–June 1, 1999.
1. Apfel CC, Greim CA, Haubitz I, Goepfert C, Usadel J, Sefrin
P, Roewer N. A risk score to predict the probability of post-
operative vomiting in adults. Acta Anaesthesiol Scand 1998:
42: 495–501.
2. Korttila K. The study of postoperative nausea and vomiting.
Br J Anaesth 1992: 69: 20S–24S.
3. Korttila K. Can we predict who will vomit after surgery?
Acta Anaesthesiol Scand 1998: 42: 493–494.
4. Herrstedt J. We still need common criteria for the assess-
ment of nausea and vomiting. Eur J Cancer 1994: 30A: 1217.
5. Watcha MF, White PF. Postoperative nausea and vomiting.
Anesthesiology 1992: 77: 162–184.
6. Freyd M. The graphic rating scale. J Educ Psychol 1923: 43:
7. Aitken RCB. Measurement of feelings using visual analogue
scales. JRoySocMed1969: 62: 989–993.
8. Huskisson EC. Measurement of pain. Lancet 1974: 2: 1127–
9. Albert A, Harris EK. Multivariate interpretation of clinical
laboratory data. New York: Marcel Dekker Inc., 1987: 143–
10. Dexter F, Chestnut DH. Analysis of statistical tests to com-
pare visual analog scale measurements among groups. Anes-
thesiology 1995: 82: 896–902.
11. Kremer E, Atkinson JH, Ignelzi RJ. Measurement of pain:
patient preference does not confound pain measurement.
Pain 1981: 10: 241–248.
12. Ahles TA, Ruckdeschel JC, Blanchard EB. Cancer-related
pain. II. Assessment with visual analogue scales. J Psychosom
Res 1984: 28: 121–124.
13. Joyce CRB, Zutshi DW, Hrubes V, Mason RM. Comparison
of fixed interval and visual analogue scales for rating
chronic pain. Eur J Clin Pharmacol 1975: 8: 415–420.
14. Linton SJ, Gotestam KG. A clinical comparison of two pain
scales: correlation, remembering chronic pain and a measure
of compliance. Pain 1983: 17: 57–65.
15. Seymour RA. The use of pain scales in assessing the efficacy
of analgesics in postoperative dental pain. Eur J Clin Pharma-
col 1982: 23: 441–444.
J. G. Boogaerts, MD, PhD
Department of Anaesthesiology
University Hospital Centre
92 Boulevard Paul Janson
6000 Charleroi