Int. J. Oral Maxillofac. Surg.

2020; 49: 22–27

https://doi.org/10.1016/j.ijom.2019.06.016, available online at https://www.sciencedirect.com

Research Paper
Clinical Pathology

Predicting postoperative D. M. Laskin1, C. K. Carrico2,

J. Wood3
1
Department of Oral and Maxillofacial

nausea and vomiting in patients Surgery, Virginia Commonwealth University

School of Dentistry, Richmond, VA, USA;
2
Department of Oral Health & Community
Outreach, Virginia Commonwealth University

undergoing oral and School of Dentistry, Richmond, VA, USA;

3
Private Practice, Spartanburg, SC, USA

maxillofacial surgery
D. M. Laskin, C. K. Carrico, J. Wood: Predicting postoperative nausea and vomiting
in patients undergoing oral and maxillofacial surgery. Int. J. Oral Maxillofac. Surg.
2020; 49: 22–27. ã 2019 International Association of Oral and Maxillofacial
Surgeons. Published by Elsevier Ltd. All rights reserved.

Abstract. A common predictive measure of postoperative nausea and vomiting

(PONV) is the Apfel score. Although tested in many different operations, it has not
been tested extensively in oral and maxillofacial surgery (OMFS). This study was
designed to determine whether it applied to OMFS and whether there were other
factors in this population that would improve its accuracy. A retrospective chart
review was carried out on a randomly selected group of patients who had OMFS
during a 10-month period. In addition to the Apfel score risk factors, PONV data
were collected in relation to type of anesthetic induction and maintenance, type of
surgery, use of maxillomandibular fixation (MMF), use of opioids, and anesthesia
and surgery times. One-hundred and sixty-seven patients were included in the
analysis; 24% had nausea and 11% had nausea and vomiting. Patients who had
Key words: Postoperative nausea; postopera-
orthognathic or temporomandibular joint surgery had the highest rate of PONV. tive vomiting; Apfel score.
Young age, anesthesia and operation time, and use of MMF were also associated
with increased PONV. Adding age, MMF or limited postoperative mouth opening, Accepted for publication
and surgery type to the Apfel score should make it more predictive in OMFS. Available online 21 June 2019

Postoperative nausea and vomiting
(PONV) is always a concern after general
anesthesia. Without prophylaxis it can
occur in approximately 30% of children
and adults1–3, often resulting in delayed
hospital discharge, increased cost, and
patient dissatisfaction.
The Apfel score was developed to
assess patients’ risk for PONV so that
appropriate preventive measures could
be initiated4. Predictive factors that in-
crease the risk for PONV included in
the Apfel score are female gender, a
history of motion sickness or PONV,
being a non-smoker and postoperative
opioids. The predictive risk of PONV is
21% if one factor is present, 39% if
there are two, 61% when three are pres-
ent, and 79% when there are four. Al-
though the accuracy of the Apfel score
in predicting PONV has been tested in
patients undergoing a variety of differ-
ent operations5–7, it has not been stud-
ied extensively in patients undergoing
oral and maxillofacial surgery. The few
studies that have been carried out have
focused on orthognathic surgery8–10.

0901-5027/01022 + 06 ã 2019 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.
Table 1. Area under the curve for Apfel score and modified Apfel scores when predicting nausea and vomiting.
Area under curve (P*)
Vomiting Nausea or vomiting
Apfel score
Female, previous PONV, postoperative opioids, non-smoker 0.53 (4) 0.58 (4)
Modified Apfel 1
Female, previous PONV, postoperative opioids, non-smoker, 0.74 (0.0006) 0.63 (0.2797)
age 40 years, orthognathic surgery, any MMF
Modified Apfel 2
Age 40 years, orthognathic surgery, any MMF 0.75 (0.0033) 0.60 (0.7904)
MMF, maxillomandibular fixation; PONV, postoperative nausea and vomiting.
*
P test of difference in area under curve for modified scores compared to original Apfel score using nonparametric test for comparing the area
under the curve of correlated receiver operating characteristic curves.

The aim of this study was to determine orthognathic surgery, as well as to ana- Materials and Methods
whether the Apfel score risk factors lyse for additional risk factors specific
This study was approved by the Institu-
were applicable to other types of oral to this population that may make it more
tional Review Board at Virginia Common-
and maxillofacial surgery besides accurate.
wealth University. A retrospective chart
review was completed for a group of
Table 2. Patient demographics and surgical factors. randomly selected patients who under-
went oral and maxillofacial surgery pro-
Personal factors Mean SD
cedures under general anesthesia in the
Age 33.15 19.13 operating room between December 2016
n % and September 2017. Inclusion criteria
included patients who remained in the
Age (categorical) hospital for at least 1 day postoperatively
40 years 109 65%
>40 years 58 35%
and had undergone a surgical procedure
Gender for a dentofacial anomaly, maxillofacial
pathology, facial trauma, temporomandib-
Male 92 55% ular joint pathology, incision and drainage
Female 75 45% for an odontogenic infection, alveolar cleft
Smoking status repair, or dentoalveolar surgery. Possible
Smoker 43 26% risk factors that were recorded included
Non-smoker 124 74% age, sex, non-smoker, previous PONV,
Surgical factors n % inhalation vs. intravenous anesthesia in-
Surgery type duction, maintenance with intravenous vs.
Cleft 21 13% inhaled anesthetic, neuromuscular block-
Fracture 28 17% ade, endotracheal tube vs. face mask, in-
Hardware removal 2 1% traoperative administration of an
Incision and drainage 31 19% antiemetic, intraoperative opioids, anes-
Orthognathic 36 22% thesia time, type of surgery, surgery time,
Pathology 10 6%
TMJ 23 14%
maxillomandibular fixation, and postoper-
Extractions 15 9% ative opioids. Charts were also reviewed
Laceration 1 1% to assess for nausea and vomiting postop-
eratively.
Antiemetic Data were summarized using descrip-
Yes 161 96%
No 6 4%
tive statistics. Individual predictors [age,
gender, smoking status, surgery time, an-
MMF esthesia time, surgery type, antiemetic,
None 100 60% maxillomandibular fixation (MMF), and
Light elastics 34 20% previous PONV] were tested against each
Wire fixation 23 14%
Heavy elastics 10 6%
of the two outcome variables: nausea (yes
or no), vomiting (yes or no), using t-tests
Postoperative opioids and Fisher’s exact test. Any variables that
Yes 140 84% were significant at 0.10 level in bivariate
No 27 16% models were considered for overall logis-
Previous PONV tic regression models. Backwards elimi-
Yes 10 6% nation was used to reach an overall logistic
No 157 94% regression model with significance set at
MMF, maxillomandibular fixation; PONV, postoperative nausea and vomiting; SD, standard 0.05. Receiver operating characteristic
deviation; TMJ, temporomandibular joint. (ROC) curves and the corresponding area
24 Laskin et al.

under the curve (AUC) were calculated for Results patient characteristics, contributing surgi-
the Apfel score and two modified Apfel cal factors and the incidence of PONV are
A total of 167 patients were included in the
scores based on results from the overall given in Tables 3 and 4.
analysis. Demographics of the patients and
logistic regression model also were devel-
descriptors of their surgical procedures are
oped (Table 1). One consisted of all the
given in Table 2. The average age of the
original Apfel measures (female, previous Postoperative Nausea
patients was 33.15 years [standard devia-
PONV, postoperative opioids, and non-
tion (SD) = 19.13], with 65% (n = 109) 40 Twenty-four percent of the 167 patients
smoker) along with the significant factors
years or younger and 35% (n = 58) over 40 (n = 40) experienced nausea and 11% of
from the final overall model (age: 40,
years. There was a roughly equal number of this group (n = 18) also experienced
orthognathic surgery, and MMF). The sec-
males and females (55% and 45%, respec- vomiting (none of the patients experienced
ond modified Apfel score only included
tively). Seventy-four percent of patients vomiting without nausea). There was a
the significant factors from the overall
could be confirmed as current smokers significant relationship between surgery
model (age: 40, orthognathic surgery,
through the chart notes (n = 124). The most type [orthognathic, fracture, temporoman-
and MMF). AUC was compared across
common procedure was orthognathic sur- dibular joint (TMJ), other] and nausea
the three scores using the methods de-
gery (n = 36, 22%) followed by incision and (P = 0.0084). For both fractures and other
scribed by DeLong et al.11. The age cutoff
drainage (n = 31, 19%) and treatment of surgery types (incision and drainage, cleft,
of 40 years was selected because the ma-
fractures (n = 28, 17%). Ninety-six percent etc.) the rate of reported nausea was <20%
jority of vomiting occurred in patients
of patients (n = 161) received an intraoper- compared to >40% for those with orthog-
under the age of 40 (17/18 with reported
ative antiemetic. Only 6% (n = 10) had a nathic surgery. TMJ surgery was in the
vomiting were 40; 16% of those 40 and
confirmed previous PONV experience. middle at 35%. Therefore, fractures were
under compared to 2% of those over 40).
Eighty-four percent of patients received combined into the ‘Other’ category due to
SAS EG v.6.1 software was used for all
postoperative opioids. A summary of the the similar rate of PONV. After combining
analyses.

Table 3. Personal and surgical factors related to nausea.

No nausea (n = 121, 60%) Nausea (n = 40, 24%) P*
BMI (mean, SD) 25.03, 7.33 26.26, 7.47 0.3590
Age 0.8495
40 years 75% 25%
>40 years 78% 22%
Gender 0.4608
Male 80% 20%
Female 71% 29%
Smoking status 0.1645
Smoker 81% 19%
Non-smoker 74% 26%
Surgical factors No nausea (n = 121, 60%) Nausea (n = 40, 24%) P*
Surgical time (mean, SD) 169.9, 125.0 176.2, 126.9 0.7810
Anesthesia time (mean, SD) 235.3, 132.4 244.0, 133.4 0.7178
Surgery type 0.0094
Orthognathic 58% 42%
TMJ 65% 35%
Fracture 86% 14%
Other 81% 19%
Antiemetic 0.1495
Yes 77% 23%
No 50% 50%
MMF 0.0277
None 82% 18%
Any 67% 33%
Post-op opioids 0.8075
Yes 76% 24%
No 74% 26%
Previous PONV 0.369
Yes 70% 30%
No 76% 24%
BMI, body mass index; MMF, maxillomandibular fixation; PONV, postoperative nausea and vomiting; SD, standard deviation; TMJ,
temporomandibular joint.
*
P from t-test for continuous variables and Fisher’s exact test for categorical.
 Predicting postoperative nausea and vomiting 25

Table 4. Personal and surgical factors related to vomiting.

No vomiting (n = 143, 89%) Vomiting (n = 18, 11%) P*
BMI (mean, SD) 24.0, 7.71 25.5, 7.32 0.4361
Age 0.0070
40 years 84% 16%
>40 years 98% 2%
Gender 0.2848
Male 87% 13%
Female 92% 8%
Smoking status 0.5681
Smoker 93% 7%
Non-smoker 88% 12%
Surgical factors No vomiting (n = 143, 89%) Vomiting (n = 18, 11%) P*
Surgical time (mean, SD) 166.4, 127.9 213.0, 91.6 0.1355
Anesthesia time (mean, SD) 231.1, 134.8 289.4, 97.7 0.0773
Surgery type 0.0003
Orthognathic 69% 31%
TMJ 87% 13%
Fracture 96% 4%
Other 95% 5%
Antiemetic 1
Yes 89% 11%
No 100% 0%
MMF – any 0.0210
None 94% 6%
MMF 82% 18%
Post-op opioids 0.7406
Yes 89% 11%
No 93% 7%
Previous PONV 0.2937
Yes 80% 20%
No 90% 10%
BMI, body mass index; MMF, maxillomandibular fixation; PONV, postoperative nausea and vomiting; SD, standard deviation; TMJ,
temporomandibular joint.
*
P from t-test for continuous variables and Fisher’s exact test for categorical.

categories, surgery type was significantly
associated with increased nausea
(P = 0.0029), with orthognathic surgery
having the highest rate (42%), followed
by TMJ surgery (35%), and all others with
16%. MMF was also significantly associ-
ated with nausea (P = 0.0277). Because all
the patients who had orthognathic surgery
also had some form of MMF, an overall
logistic regression model could not be
constructed with these two variables due
to the collinearity. Postoperative opioid
use was not associated with nausea
(P = 0.7930).
Postoperative Vomiting
Younger age (P = 0.0070), MMF
(P = 0.0210), and surgery type
(P = 0.0001) were associated with a great-
er risk for vomiting. Specifically, younger
patients (40 years) had a higher rate of
vomiting than older patients (16% vs. 2%).
Those undergoing orthognathic surgery
had the highest rate of vomiting (31%),
followed by TMJ surgery (13%), and all
others with 7% (P = 0.0001). Again, MMF
was also associated with increased vomit-
ing (18% vs 6%, P = 0.0210). Anesthesia
time was marginally associated with
vomiting (P = 0.0773). Patients who ex-
perienced vomiting had longer anesthesia
times on average (289.4 min vs.
231.1 min). Postoperative opioid prescrip-
tion was not associated with vomiting
(P = 0.5373). Twenty percent of females
undergoing orthognathic surgery experi-
enced vomiting compared to 38% of
males, but this difference was not statisti-
cally significant (P = 0.2951).
An overall logistic regression model
was constructed with the variables of in-
terest, except for MMF due its collinearity
with surgery type. After removing anes-
thesia time (P = 0.8633), surgery type
(P = 0.0062) was significantly associated
with increased risk for vomiting and age
(P = 0.1172) was marginally associated
with increased risk for vomiting. Patients
aged 40 years or younger were 5.7 times as
likely to experience vomiting than those
older than 40 years (95% CI: 0.65–49.95).
Patients undergoing orthognathic surgery
were 6.75 times as likely to experience
vomiting than those who had other types
of surgery (95% CI: 1.66–27.41). Those
with TMJ surgery were 4.1 times as likely
to experience vomiting than those with
other surgery types (95% CI: 0.73–22.47).
Apfel scores
The area under the curve (a measure of
the ability of a composite score to predict
an outcome) for the original Apfel score
predicting vomiting was 0.53, which was
significantly lower than both of the new
modified Apfel scores: (original factors
+ significant factors: 0.74, P = 0.0006;
significant factors only: 0.75,
P = 0.0033) (Table 1, Fig. 1). The ROC
curves and corresponding AUC were near-
ly identical for the modified Apfel 1 and
Apfel 2 scores indicating that for these
patients, age, surgery type, and MMF may
be more important regarding PONV than
the factors normally considered based on
26 Laskin et al.
Fig. 1. Receiver operating characteristic (ROC) curves for Apfel score and modified Apfel scores for predicting postoperative vomiting.
the original Apfel score. When predicting
nausea alone, there were no significant
differences in the area under the curve
for any of the scores (Table 1).
Discussion
The overall incidence of PONV in this
study was 24%, which is consistent with
previous studies1–3. This indicates that
patients undergoing oral and maxillofacial
surgery are not at a higher risk of PONV
than those individuals undergoing other
types of surgical procedures. However,
some of the predictive factors involved
in the Apfel score did not serve the same
purpose in this study. This was at least
partially due to the young age of a large
percentage of the included patients likely
limiting those with a prior history of
PONV and the fact that although 84%
received postoperative opioids, 97% had
been given an antiemetic as part of the
standard anesthesia protocol. Although
young age (average 20.2 years), which
is not a risk factor in the Apfel score,
appeared to be a risk factor in this study,
it may have been related to the relatively
large group of young orthognathic surgery
patients and may not be generalizable.
Within this study, the factor that greatly
increased the risk of PONV was the surgery
type. This appeared to be related to the
anesthesia time and the possible restriction
of mouth opening. With both orthognathic
surgery and TMJ surgery the anesthesia
times were the longest (average orthog-
nathic 310 min; average TMJ 333 min; all
others average 188 min). Moreover, re-
stricted mouth opening was a factor in both,
due to placement of orthognathic surgery
patients in light elastics or maxillomandib-
ular fixation postoperatively and the post-
operative limitation in mouth opening that
generally occurs after TMJ surgery. The
latter did not seem to be a factor in fracture
cases because most were treated with rigid
fixation and not placed in postoperative
MMF. The amount of bleeding associated
with orthognathic surgery, and the potential
for blood being swallowed, may also be a
factor in such patients (10).
Because of the various factors that
could lead to PONV, any brief predictive
risk score has its limitations. However, the
results of this study show that adding the
use of maxillomandibular fixation, age,
and surgery type to the Apfel score should
improve its value in oral and maxillofacial
surgery when predicting postoperative
vomiting.
Funding
None.
Competing interests
None.
Ethical approval
Ethical approval was given by the VCVU
IRB No. HM20011502.
Patient consent
Not Required.

References
1. Pierre S, Whelan R. Nausea and vomiting
7.
after surgery. Contin Educ Anesth Crit Care
Pain 2013;13:28–32.
2. Conway B. Prevention and management of
postoperative nausea and vomiting in adults.
AORN J 2009;90:391–413.
3. Watcha MF. Postoperative nausea and emesis.
8.
Anesthesiol Clin North Am 2002;20:709–22.
4. Apfel CC, Laara E, Koivuranta M, Greim
CA, Roewer N. A simplified risk score for
predicting postoperative nausea and vomit-
ing: conclusions from cross-validation be-
9.
tween two centers. Anesthesiology
1999;91:693–700.
5. Sherif L, Hegde R, Manswami M, Ollapally
A. Validation of the Apfel scoring system for
10.
identification of high-risk patients for PONV.
Karnatka Anaesth J 2015;1:115–7.
6. Pierre S, Benais H, Pouymayou J. Apfel’s
simplified score may favourably predict the
Predicting postoperative nausea and vomiting 27
risk of postoperative nausea and vomiting. 11. DeLong ER, DeLong DM, Clarke-Pearson
Can J Anaesth 2002;49:237–42. DL. Comparing the areas under two or more
Weilbach C, Rahe-Mayer N, Rayondos K, correlated receiver operating characteristic
Weissig A, Scheinichen D, Piepenbrock S. curves: a nonparametric approach. Bio-
Postoperative nausea and vomiting (PONV): metrics 1988;44(3):837–45.
usefulness of the Apfel-Score for identifica-
Address:
tion of high risk patients for PONV. Acta
Daniel M. Laskin
Anaesth Belg 2006;57:361–3.
Department of Oral and Maxillofacial
Silva AC, O’Ryan F, Poor DB. Postoperative
Surgery
nausea and vomiting (PONV) after orthog-
Virginia Commonwealth University
nathic surgery: a retrospective study and
521 N. 11th Street
literature review. J Oral Maxillofac Surg
2006;64:1385–94. Richmond
VA 23298-0566
Phillips C, Brookes CD, Rich J, Arbon J,
USA
Turvey TA. Postoperative nausea and vomit-
Tel.: +1 804 828 3547
ing following orthognathic surgery. Int J
E-mail: dmlaskinW@vcu.edu
Oral Maxillofac Surg 2015;44:745–51.
Dobbeleir M, De Coster J, Coucke W, Politis
C. Postoperative nausea and vomiting after
oral and maxillofacial surgery. Int J Oral
Maxillofac Surg 2018;47:721–5.