DENTOALVEOLAR SURGERY

Increasing Use of Intensive Care Unit

for Odontogenic Infection Over One
Decade: Incidence and Predictors
Benjamin Fu, BDSc (Hons), MD,* Kelly McGowan, BOH, Grad Dip Dent,y
Hansen Sun, BPH, MPH, PhD,z and
Martin Batstone, BDSc (Hons), MBBS, MPhil(Surg), FRACDS (OMS), FRCS (OMFS)x
Purpose: To determine whether intensive care unit (ICU) admissions for odontogenic infections have
increased during the past decade and whether certain clinical features are associated with a greater rate
of ICU admission.
Materials and Methods: The data from patients undergoing surgery for odontogenic infections at the
Royal Brisbane and Women’s Hospital in the 24 month from January 2003 to December 2004 were
compared with those from patients treated from January 2013 to December 2014. A c2 analysis was
used to compare the demographic, admission, and clinical patient data in each cohort. A multiple logistic
regression model was used to determine which clinical features were associated with greater rates of ICU
admission.
Results: The rate of ICU admission increased significantly from 7 to 24% during the decade (c2 = 12.74;
P = .000), although the clinical presentation of patients admitted to the ICU was similar in both cohorts.
The mean number of days spent in the ICU increased significantly from 1.7 ! 0.5 to 3.24 ! 2.5 days
(t = "3.63; P = .001), and the overall length of stay increased from 1.7 ! 0.5 to 3.5 ! 4.1 days
(t = 2.99; P = .004). The use of preoperative computed tomography (CT) increased significantly from
42.9 to 93.3% (c2 = 13.25; P = .000). The most significant predictors of ICU admission were lower third
molar involvement (P = .026), dysphagia (P = .020), and C-reactive protein (CRP) levels exceeding
150 mg/L (P = .039).
Conclusions: The use of the ICU in the management of odontogenic infection has increased significantly
at the Royal Brisbane and Women’s Hospital over 1 decade. The demographic data and clinical presenta-
tion of the patients admitted to the ICU did not change significantly. However, the length of ICU stay and
the total length of stay have both increased. A significant increase in CT usage for odontogenic infections
also occurred. Third molar infections, dysphagia, and elevated CRP might be relevant clinical predictors of
a more complicated course of care requiring ICU admission. More judicious use of CT scanning, combined
with prompt surgical consultation and intervention, might reduce the rate of ICU admissions for odonto-
genic infections.
! 2018 Published by Elsevier Inc. on behalf of the American Association of Oral and Maxillofacial
Surgeons
J Oral Maxillofac Surg -:1-8, 2018

*Registrar, Department of Oral and Maxillofacial Surgery, Royal
Melbourne Hospital, Melbourne, VIC, Australia and Masters
Candidate, School of Medicine, University of Queensland, QLD,
Australia.
yAssociate Lecturer, School of Dentistry, University of
Queensland, QLD, Australia and PhD Candidate, School of
Dentistry and Oral Health, Griffith University, QLD, Australia.
zAdjunct Associate Professor, School of Public Health and Social
Work, Queensland University of Technology, Brisbane, QLD, Australia.
xDirector, Department of Oral and Maxillofacial Surgery, Royal
Brisbane and Women’s Hospital, Herston, QLD, Australia.
Conflict of Interest Disclosures: None of the authors have any
relevant financial relationship(s) with a commercial interest.
Address correspondence and reprint requests to Dr Fu: Depart-
ment of Oral and Maxillofacial Surgery, Royal Melbourne Hospital,
Victoria 3050, Australia; e-mail: s4027318@gmail.com
Received February 19 2018
Accepted May 16 2018
! 2018 Published by Elsevier Inc. on behalf of the American Association of Oral
and Maxillofacial Surgeons
0278-2391/18/30496-8
https://doi.org/10.1016/j.joms.2018.05.021

1
Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
2 INCREASING USE OF ICU FOR ODONTOGENIC INFECTIONS
Odontogenic infections have the potential to result in
significant morbidity and mortality.1,2 The early stages
of the disease are generally well-localized to the
affected tooth and respond well to surgical interven-
tion, such as elimination of the source of infection
by extraction or root canal therapy.3 However, without
early surgical intervention, the inflammatory process
can erode through the bony cortex of the maxilla or
mandible and spread by direct extension along the
fascial planes.4 Antibiotics alone without surgical
drainage will be ineffective in altering the course of
the disease process.5 As the infection progresses, the
soft tissue in the face and neck region will be displaced
to accommodate the accumulating pus and inflamma-
tory exudate.3
The path of spread of odontogenic infections has
been well documented.6 In the maxilla, extension
into the canine space, buccal, and masticator space
is common.7 Although they rarely pose a threat to
the airway,6 they do have the potential to cause serious
complications such as orbital abscess,8 cavernous
sinus thrombosis,9 and cerebral abscess.10 In the
mandible, the thin lingual cortex in the posterior
molar region allows for rapid spread of an odontogenic
infection into the sublingual, submandibular, and para-
pharyngeal spaces.11 Acute upper airway obstruction
can occur owing to the mass effect of the deep neck
infection,12 and further spread of the infection by
way of the fascial planes can lead to necrotizing fasci-
itis13 and mediastinitis.14 In the most severe cases,
odontogenic infection can be fatal, usually due to up-
per airway obstruction or multiorgan failure.15 The
intensive care unit (ICU), therefore, plays a critical
role in airway protection and perioperative support
for severe cases of odontogenic infections.
The number of patients presenting to the emer-
gency department (ED) with odontogenic infections
has been increasing,16-20 likely due to a number of
factors, including cost, fear, mental illness, substance
abuse, health literacy, and perceptions that oral
disease is of low importance.21,22 Although some of
these patients can be discharged safely from the ED
for treatment in outpatient clinics, others will
require inpatient admission for more urgent incision
and drainage of their abscess. Severity scoring
systems have been reported to aid in clinical
decision-making in the ED, with a focus on ‘‘red flag’’
symptoms such as trismus, dysphagia, sepsis, and
airway compromise.23 A proportion of patients with
severe odontogenic infections will also require ICU
support in the immediate postoperative period after
surgical drainage.
The rate of ICU admission for odontogenic infection
differs widely in reported studies. Jundt and Gutta24 re-
ported an ICU admission rate of 14%, and Ylijoki et al25
reported a rate of 18%. A recent retrospective review
Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
by Gams et al26 reported that 45% of the cohort
required ICU admission after surgery. The difference
in reported rates likely resulted from the different pa-
tient populations and local hospital policies. Also,
because the ICU is not standardized across different
hospitals, it would be useful to investigate the pattern
of ICU use in 1 institution over a period of time. We
investigated the pattern of ICU use for odontogenic in-
fections during 1 decade in a major Australian tertiary
hospital to determine 1) whether the use of the ICU
for patients surgically treated for odontogenic infec-
tion had changed during the course of the decade;
2) whether the demographic data and clinical presen-
tation of the patients admitted to the ICU had changed;
and 3) whether certain clinical features could be used
as predictors for ICU admission.
Materials and Methods
The Royal Brisbane and Women’s Hospital (RBWH)
human research ethics committee provided ethical
approval for the project (ID no. EC00172, HREC/14/
QRBW/351). A retrospective review was conducted
at the maxillofacial surgery department of the RBWH
to compare the 24-month period from January 2003
to 2004 with the admission data from January 2013
to December 2014. The patients were identified
from the hospital operation record database. Patients
admitted as an inpatient who had undergone a surgical
procedure for incision and drainage of an odontogenic
infection were eligible for inclusion. The RBWH is an
adult hospital; therefore, the study population did
not include any pediatric patients. The medical re-
cords were individually reviewed, and the patients
admitted to the ICU for postoperative support were
identified manually.
The following data were de-identified and recorded
into an Excel spreadsheet (Microsoft, Redmond, WA)
for analysis:
1. Patient outcomes: number of ICU admissions for
odontogenic infections, number of days in the
ICU, and the overall length of stay
2. Demographic information: gender, age, aborig-
inal and Torres Strait Islander (ATSI) status, and
smoking status
3. Comorbidities: diabetes, chronic renal disease,
immunosuppression (human immunodeficiency
virus, leukemia, lymphoma, neutropenia, organ
transplantation, current malignancy, corticoste-
roid dosage >20 mg for $4 weeks), and Charlson
comorbidity index score27
4. Clinical presentation: site of infection, third
molar involvement, trismus, dysphagia, dyspnea,
tongue elevation, submandibular swelling, pre-
scription of antibiotics before admission, and
FU ET AL
whether the patient had sought dental treatment
for their infection
5. Vital signs and investigations: fever (>38.0# C),
tachycardia (heart rate >90 bpm), altered respira-
tory rate (>20 breaths per minute or partial pres-
sure of carbon dioxide in arterial blood <4.3 kPa),
peak white blood cell (WBC) count, and C-reac-
tive protein (CRP) level
6. Medical imaging: use of preoperative orthopan-
tomography (OPG) and computed tomography
(CT) scans, and patterns of imaging requests by
clinicians
The statistical analyses were conducted using IBM
SPSS Statistics, version 24.0.0 (IBM Corp, Armonk,
NY). A c2 analysis was conducted to compare the
demographic, admission, and clinical data from the
patients in the earlier cohort (2003 to 2004) with the
data from those in the later cohort (2013 to 2014).
An independent samples t test was used to determine
whether differences in patient age, days spent in the
ICU, and overall length of stay were statisti-
cally significant.
A statistical analysis of the clinical features associ-
ated with ICU admission was also conducted. A c2
analysis was used to compare the recorded demo-
graphic and clinical characteristics of the patients
who had required admission to the ICU with those
of the patients who had not required ICU admission.
The factors that were significant were considered as
potential variables for multiple logistic regression.
Third molar involvement, dysphagia, dyspnea, tongue
elevation, CRP, and systemic inflammatory response
syndrome (SIRS) status were selected as the most clin-
ically relevant predictor variables for the regression
model. Trismus and submandibular swelling were
omitted, because they were both constant in all pa-
tients admitted to the ICU. Based on previous research,
CRP was selected instead of the WBC count because it
reacts faster to acute infection,25 and the 2 parameters
could not both be used owing to collinearity. SIRS
status was used because it incorporates body temper-
ature, heart rate, respiratory rate, and WBC count
while maintaining the most parsimonious model. No
significant differences in patient demographic data
were detected in the c2 analysis and were therefore
not included as potential confounders. First, univariate
analysis was performed to determine the association
between each main effect and the outcome. Next, a
multivariable model was constructed using the pro-
cess of backward elimination. Collinearity diagnostics
were run to confirm tolerance and that the variance
inflation factors were within acceptable limits. Next,
all candidate main effects were entered into the
model, and nonsignificant variables were removed
Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
3
sequentially. At each step, the variable removed was
the least significant. The final model was selected
when all remaining variables were significant
at P < .05.
Results
In the 24 months from January 2003 to December
2004, 101 patients were admitted to the RBWH for
incision and drainage of an odontogenic infection; of
these patients, 7 were admitted to the ICU. From
January 2013 to December 2014, 191 patients were
admitted for surgical management of odontogenic
infections, of whom, 45 required ICU admission. The
admission details for the 2 cohorts are listed in
Table 1. This increase in ICU use from 7 to 24% was sta-
tistically significant (c2 = 12.74; P = .000). No deaths
were recorded in either group. The mean number of
days spent in the ICU increased significantly from
1.7 ! 0.5 days in 2003 to 2004 to 3.24 ! 2.5 days in
2013 to 2014 (t = "3.63; P = .001). The patients in
the later cohort also required significantly more days
in the hospital overall, increasing from a mean of
1.7 ! 0.5 days in 2003 to 2004 to 3.5 ! 4.1 days in
2013 to 2014 (t = 2.99; P = .004).
The demographic data of the patients admitted to
the ICU in both cohorts were similar with regard to
gender, ATSI status, smoking habits, diabetes, and
immunosuppression (Table 2). The mean age of the
patients in the 2003 to 2004 cohort was
34.5 ! 13.8 years compared with 37.2 ! 14.4 years
in the 2013 to 2014 cohort. However, the difference
was not statistically significant (t = 1.60; P = .158).
The clinical presentation of the patients admitted to
ICU in both cohorts was also similar, with no signifi-
cant differences detected in any of the signs and symp-
toms (Table 3). In both cohorts, all patients admitted
to the ICU had odontogenic infections in the lower
jaw, palpable submandibular swelling, and trismus.
Of the 52 ICU admissions recorded in total, 67%
were related to lower third molars.
Apart from trismus and submandibular swelling,
which were found in all patients, the most common
clinical sign on presentation was dysphagia (79%), fol-
lowed by tongue elevation (58%), fever (33%), and
tachycardia (29%). Dyspnea (12%), hypotension
(7.7%), and stridor (4%) were uncommon; 37% of
the patients met the criteria for SIRS. Only 1 patient
had a positive blood culture result, and no patient
recorded a reduced Glasgow coma scale score.
The proportion of patients with CT scans increased
significantly from 42.9% in 2003 to 2004 to 93.3% in
2013 to 2014 (c2 = 13.25; P = .000). Because this in-
crease seemed substantial, the data set was re-
examined to determine whether this trend was limited
4 INCREASING USE OF ICU FOR ODONTOGENIC INFECTIONS

Table 1. TREATMENT OUTCOMES

Variable 2003-04 (n = 101) 2013-14 (n = 191) c2 Test P Value

Interval from ED presentation to surgery (hr) 8.74 .013

0-12 68 (68.0) 102 (53.7)
13-24 25 (25.0) 52 (27.4)
$25 7 (7.0) 36 (18.9)
ICU admission 12.74 .000
Yes 7 (6.9) 45 (23.8)
No 94 (93.1) 144 (76.2)
ICU duration (days) 5.48 .019
1-2 7 (100.0) 24 (53.3)
$3 0 (0.0) 21 (46.7)
Patient required return to surgery 1.16 .281
Yes 3 (3.0) 11 (5.8)
No 98 (97.0) 178 (94.2)
Overall length of stay (days) 9.89 .020
1-2 28 (27.7) 36 (18.8)
3-4 60 (59.4) 118 (61.8)
5-6 12 (11.9) 19 (9.9)
$7 1 (1.0) 18 (9.4)
Data presented as n (%).
Patients admitted from 2013 to 2014 were more often admitted to the ICU and tended to have a longer overall stay than
patients admitted in the earlier cohort; percentages were calculated using the number of valid cases; therefore, because of
missing data, the percentages might not correlate with the cohort total.
Abbreviations: ED, emergency department; ICU, intensive care unit.
Fu et al. Increasing Use of ICU for Odontogenic Infections. J Oral Maxillofac Surg 2018.

Table 2. DEMOGRAPHIC DATA OF PATIENTS ADMITTED TO ICU

Variable 2003-2004 (n = 7) 2013-2014 (n = 45) c2 Test P Value

Gender 1.47 .224

Male 6 (85.7) 28 (62.2)
Female 1 (14.3) 17 (37.8)
ATSI 0.50 .482
No 6 (85.7) 42 (93.3)
Yes 1 (14.3) 3 (6.7)
Smoker 0.76 .384
No 1 (20.0) 16 (40.0)
Yes 4 (80.0) 24 (60.0)
Diabetes 0.50 .482
No 7 (100.0) 42 (93.3)
Yes 0 (0.0) 3 (6.7)
Immunosuppression* 0.32 .569
No 7 (100.0) 43 (95.6)
Yes 0 (0.0) 2 (4.4)

Data presented as n (%).

No statistically significant differences in patient demographic data were found between the 2 cohorts; percentages were
calculated using the number of valid cases; therefore, because of missing data, percentages might not correlate with the cohort
total.
Abbreviations: ATSI, aboriginal and Torres Strait Islander; ICU, intensive care unit.
* Defined as the presence of human immunodeficiency virus, leukemia, lymphoma, neutropenia, organ transplantation, cur-
rent malignancy, prolonged corticosteroid use ($4 weeks at >20 mg), or other significant immunosuppression.
Fu et al. Increasing Use of ICU for Odontogenic Infections. J Oral Maxillofac Surg 2018.

Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
FU ET AL 5

Table 3. CLINICAL PRESENTATION OF PATIENTS ADMITTED TO THE ICU

Variable 2003-2004 (n = 7) 2013-2014 (n = 45) c2 Test P Value

Location of infection NA NA
Maxilla 0 (0.0) 0 (0.0)
Mandible 7 (100.0) 45 (100.0)
Lower third molar infection 1.25 .264
No 1 (14.3) 16 (35.6)
Yes 6 (85.7) 29 (64.4)
Postoperative complication 0.139 .709
No 4 (57.1) 29 (64.4)
Yes 3 (42.9) 16 (35.6)
Submandibular swelling* NA NA
No 0 (0.0) 0 (0.0)
Yes 7 (100.0) 43 (100.0)
Trismus* NA NA
No 0 (0.0) 0 (0.0)
Yes 7 (100.0) 43 (100.0)
Dysphagia 0.27 .605
No 2 (28.6) 9 (20.0)
Yes 5 (71.4) 36 (80.0)
Dyspnea 2.30 .129
No 5 (71.4) 41 (91.1)
Yes 2 (28.6) 4 (8.9)
Stridor 2.38 .123
No 6 (85.7) 44 (97.8)
Yes 1 (14.3) 1 (2.2)
Tongue elevation 0.66 .719
No 3 (42.9) 18 (40.0)
Yes 4 (57.1) 27 (60.0)
Fever (>38.0# C) 0.06 .803
No 5 (71.4) 30 (66.7)
Yes 2 (28.6) 15 (33.3)
Tachycardia (HR >90 bpm) 0.84 .361
No 6 (85.7) 31 (68.9)
Yes 1 (14.3) 14 (31.1)
Hypotensive (SBP <100 mm Hg) 0.81 .368
No 7 (100.0) 34 (89.5)
Yes 0 (0.0) 4 (10.5)
RR >22 breaths/min or PaCO2 <32 1.06 .304
No 7 (100.0) 39 (86.7)
Yes 0 (0.0) 6 (13.3)

Data presented as n (%).

No statistically significant differences in clinical presentation were found for patients admitted to the ICU from 2003 to 2004
and those admitted from 2013 to 2014; percentages were calculated using the number of valid cases; therefore, because of
missing data, percentages might not correlate with the cohort total.
Abbreviations: bpm, beats per minute; HR, heart rate; ICU, intensive care unit; NA, not applicable; PaCO2, partial pressure of
carbon dioxide in arterial blood; RR, respiratory rate; SBP, systolic blood pressure.
* No statistical analysis could be performed for location, trismus, or submandibular swelling because these clinical features
were constants for all patients admitted to the ICU.
Fu et al. Increasing Use of ICU for Odontogenic Infections. J Oral Maxillofac Surg 2018.

to patients admitted to the ICU or whether it was
observed in all admissions for odontogenic infections
during the study period. The increased uptake of CT
scanning was more pronounced when comparing all
admissions, with only 6% of patients in 2003 to 2004
undergoing CT scanning preoperatively compared
with 79% in 2013 to 2014 (c2 = 142.08; P = .000).
A multiple regression model was used to determine
which factors were associated with ICU admission
(Table 4). The demographic data of the patients

Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
6 INCREASING USE OF ICU FOR ODONTOGENIC INFECTIONS

Table 4. MULTIPLE LOGISTIC REGRESSION MODELING OF PREDICTORS OF ICU ADMISSION*

Parameter Crude OR (95% CI) P Value Adjusted OR (95% CI) P Value

Lower third molar 4.15 (2.19-7.87) .000 6.41 (1.25-33.33) .026

Dysphagia 8.26 (4.02-16.97) .000 6.49 (1.33-31.25) .020
Dyspnea 7.63 (2.07-27.77) .002 Removed NA
Tongue elevation 5.15 (2.74-9.71) .000 Removed NA
SIRS 3.03 (1.56-5.88) .001 Removed NA
CRP $150 mg/L 4.50 (1.21-16.67) .024 5.00 (1.08-22.72) .039
CT scan 7.35 (3.18-16.95) .000 Removed NA

Abbreviations: CI, confidence interval; CRP, C-reactive protein; CT, computed tomography; ICU, intensive care unit; NA, not
applicable; OR, odds ratio; SIRS, systemic inflammatory response syndrome.
* c2 = 16.43; P = .001; R2 = 40.4%.
Fu et al. Increasing Use of ICU for Odontogenic Infections. J Oral Maxillofac Surg 2018.

admitted to the ICU were compared with the data
from those who did not require ICU admission. The re-
sults of the c2 analysis are presented in Supplemental
Table 1. The results of the preliminary analysis showed
that third molar involvement, dysphagia, dyspnea,
tongue elevation, CRP level, SIRS status, and the use
of CT scans were the most clinically relevant predictor
variables for the regression model. Although all the
variables were statistically significant on univariate
analysis, the results of the multivariable analysis
suggested that lower third molar involvement
(P = .026), dysphagia (P = .020), and CRP level
exceeding 150 mg/L (P = .039) are the most significant
clinical predictors of ICU admission (c2 = 16.43;
P = .001), with the final model accounting for 40.4%
of the variation in ICU admissions. Patients admitted
to the ICU were approximately 6.5 times more likely
to have lower third molar infections or dysphagia
and 5 times more likely to present with CRP levels
greater than 150 mg/L.
Discussion
At the RBWH, the proportion of patients admitted to
the ICU after surgical management of odontogenic
infection has increased by 3.4-fold during the course
of 1 decade. Patients admitted to ICU in the later
cohort had a longer average stay in the ICU and a
longer overall hospitalization. Despite the increase in
incidence, no statistically significant difference was
found in terms of patient demographic data or clinical
presentation between the 2 cohorts. Dysphagia, lower
third molar involvement, and CRP were found to be
significant predictors for ICU admission.
The rates of ICU admission have varied widely in the
reported data, ranging from 14 to 45%.24-26 Our most
recent ICU admission rate of 24% was in the middle
of this range; however, we found a huge increase
from only 7% in the previous decade. A similar study
conducted in Helsinki comparing odontogenic
infections 1 decade apart also found a significant
increase in the ICU admission rate from 11 to 32%.28
Their length of stay in the ICU was 2 days in 1994
to 1995 and 3 days in 2004 to 2005. Although this
increase was not statistically significant, it appears
that this general trend of increased ICU use is not iso-
lated to our institution.
Population growth could be a potential contributor
to the increased ICU admission rate observed. Queens-
land’s population increased by 22% from 3.88 million to
4.72 million in the 10-year period of our study.29 The
clinical severity of the patients presenting with odonto-
genic infections to the RBWH during the study period
had also worsened slightly, which might have also
contributed to the greater number of ICU admissions.
The proportion of patients presenting with mandibular
infections increased by 20%, and the proportion of
patients presenting with a lower third molar infection
doubled. This was associated with an increased propor-
tion of patients presenting with trismus and submandib-
ular swelling. Because mandibular odontogenic
infections are more likely to threaten the airway,12 this
increased incidence could account for some of the
increased ICU admissions. However, no other clinical
parameters were significantly different among the 2
cohorts, with the rates of dysphagia, dyspnea, stridor,
and tongue elevation similar between the 2 groups.
The demographic data of the population in terms of
age, gender, and medical comorbidities between the
2 decades was also not significantly different; therefore,
the 3.4-fold increase in ICU admissions could not be
adequately explained by the population growth and
worsened clinical severity alone.
Fifty patients were treated for maxillary infections
during the study period; however, none required
admission to the ICU postoperatively. The path of
spread of odontogenic infections originating from
the maxilla is less likely to threaten the airway and
require ICU admission for airway protection. Certainly,
severe odontogenic maxillary infections have been

Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
FU ET AL
treated at the RBWH; however, none were captured
during the study period.
We hypothesized that 1 potential contributor to the
increased rate of ICU use was the significant increase
in CT scanning during the decade. CT has gained wide-
spread use as the imaging modality of choice for deep
neck space infections in the emergency setting,
because it overcomes the field of view limitations of
ultrasonography and is less time-consuming and
more accessible than magnetic resonance imaging.30
The overall rate of preoperative CT scanning for odon-
togenic infections at the RBWH increased from 6% in
the earlier cohort to 79% in the later cohort. Also, for
58% of the patients admitted from 2013 to 2014, CT
scanning was the initial imaging modality requested,
in preference to OPG. Of these CT scans, 91% were
ordered by ED or referring hospitals and 89% were or-
dered before clinical review by a maxillofacial surgery
registrar. Of the patients with maxillary infections,
55% had a preoperative CT scan ordered in addition
to OPG. These data show a significant change in the
way CT scans are being ordered at the RBWH.
Just as for any imaging modality, CT scans have
certain limitations. Multiple studies have examined
the positive predictive value of CT scans with intrao-
perative findings in deep neck space infections. Lazor
et al31 reported that CT scan results have a positive
predictive value of 76% for deep neck space abscesses,
with a false-positive rate of 13.2%. Freling et al32
reported a similar positive predictive value of 82%;
however, when rim enhancement is lacking or the
collection is smaller, the positive predictive value di-
minishes. Smith et al33 reported a negative exploration
rate of 25% and stressed that although CT can add valu-
able information, the decision to surgically drain a
neck space abscess should be made clinically. The
importance of a thorough clinical history and examina-
tion has been validated in reported studies. Rosenthal
et al34 compared the preoperative CT and surgical find-
ings and found that the diagnostic accuracy of maxillo-
facial surgeons reviewing the CT scan was greater than
that of the radiologist, likely because the surgeons
were able to correlate the radiologic evidence with
the clinical examination findings but the radiologist
had to rely on the imaging studies alone.
We also hypothesized that the significant increase in
preoperative CT scans could have a flow-on effect in
how clinicians subsequently treat these patients.
Deep tissue changes that were previously not visual-
ized can now be readily seen on CT scans, and subtle
changes in the upper airway anatomy seen on imaging
studies have the potential to make clinicians warier of
the possibility of potential airway difficulties and
resort to more invasive intubation techniques. These
patients are then more likely to remain intubated
and require ICU admission for airway protection.
Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
7
Radiographic demonstration of swelling of the upper
airway anatomy in the axial plane can also influence
the length of intubation in the ICU and can lead
to an increased length of stay in the ICU. We
hypothesized that in the earlier cohort, owing to the
very low rate of CT use, clinicians relied more on the
physical examination findings and clinical judgment
to establish whether a patient can be safely extubated
at the end of surgery. It is possible that more patients
would have required ICU admission in the earlier
cohort if the rate of CT scan use had been similar.
CT scans undoubtedly will continue to be an invalu-
able imaging modality for assessing undifferentiated
swelling in the neck. However, we would argue that
when a patient presents to the ED with clear history
of dental pain suggestive of odontogenic infection, a
clinician with a good understanding of the disease pro-
cess and fascial spaces should be able to accurately
diagnose and triage most cases. OPGs allow for rapid
identification of large carious lesions and periapical
abscesses12 and should be the first-line imaging modal-
ity for these patients. Source control should be per-
formed, regardless of whether the causative dental
pathology is associated cellulitis or a drainable collec-
tion, and early surgical consultation should be sought
to guide patient treatment. An efficient pathway in the
management of odontogenic infections will prevent
deterioration of the condition to the point that ICU
admission is required. The cost of 1 night of ICU stay
at the RBWH has been estimated to be $4500 (Lipman
J, RBWH ICU, personal communication, December
2017). The estimated cost of an ICU stay alone in our
study amounted to $765,000. Given that odontogenic
infection is a preventable disease, optimization of pri-
mary prevention, prehospital care, and efficient triage
and management of the disease remains a priority.
In conclusion, the use of the ICU in the management
of odontogenic infection has increased significantly at
the RBWH during 1 decade. The demographic data
and clinical presentation of the patients admitted to
ICU remained the same. The ICU length of stay and
the total length of stay have both increased. Third molar
infections, dysphagia, and elevated CRP levels are strong
predictors of ICU admission. We found a significant in-
crease in CT usage for odontogenic infections over the
decade, with most of these CT scans ordered by refer-
ring clinicians before the surgical review. More judicious
use of CT scanning, combined with prompt surgical
consultation and intervention, might reduce the rate
of ICU admissions for odontogenic infections.
References
1. Green A, Flower E, New N: Case report: Mortality associated
with odontogenic infection! Br Dent J 190:529, 2001
2. Carter L, Lowis E: Death from overwhelming odontogenic
sepsis: A case report. Br Dent J 203:241, 2007
8 INCREASING USE OF ICU FOR ODONTOGENIC INFECTIONS
3. Robertson D, Smith A: The microbiology of the acute dental
abscess. J Med Microbiol 58:155, 2009
4. Bridgeman A, Wiesenfeld D, Newland S: Anatomical consider-
ations in the diagnosis and management of acute maxillofacial
bacterial infections. Aust Dent J 41:238, 1996
5. Brennan MT, Runyon MS, Batts JJ, et al: Odontogenic signs and
symptoms as predictors of odontogenic infection: A clinical trial.
J Am Dent Assoc 137:62, 2006
6. Yonetsu K, Izumi M, Nakamura T: Deep facial infections of odon-
togenic origin: CT assessment of pathways of space involve-
ment. Am J Neuroradiol 19:123, 1998
7. Obayashi N, Ariji Y, Goto M, et al: Spread of odontogenic infec-
tion originating in the maxillary teeth: Computerized tomo-
graphic assessment. Oral Surg Oral Med Oral Pathol Oral
Radiol Endod 98:223, 2002
8. Kim I-K, Kim J-R, Jang K-S, et al: Orbital abscess from an odonto-
genic infection. Oral Surg Oral Med Oral Pathol Oral Radiol En-
dod 103:e1, 2007
9. Ogundiya DA, Keith DA, Mirowski J: Cavernous sinus throm-
bosis and blindness as complications of an odontogenic infec-
tion: Report of a case and review of literature. J Oral
Maxillofac Surg 47:1317, 1989
10. Mylonas AI, Tzerbos FH, Mihalaki M, et al: Cerebral abscess of
odontogenic origin. J Craniomaxillofac Surg 35:63, 2007
11. Ariji Y, Gotoh M, Kimura Y, et al: Odontogenic infection pathway
to the submandibular space: Imaging assessment. Int J Oral
Maxillofac Surg 31:165, 2002
12. Flynn TR: The swollen face: Severe odontogenic infections.
Emerg Med Clin North Am 18:481, 2000
13. Tung-Yiu W, Jehn-Shyun H, Ching-Hung C, Hung-An C: Cervical
necrotizing fasciitis of odontogenic origin: A report of 11 cases. J
Oral Maxillofac Surg 58:1347, 2000
14. Kinzer S, Pfeiffer J, Becker S, Ridder GJ: Severe deep neck space
infections and mediastinitis of odontogenic origin: Clinical rele-
vance and implications for diagnosis and treatment. Acta Otolar-
yngol 129:62, 2009
15. Hwang T, Antoun JS, Lee KH: Features of odontogenic infections
in hospitalised and non-hospitalised settings. Emerg Med J 28:
766, 2011
16. Salomon D, Heidel RE, Kolokythas A, et al: Does restriction of
public health care dental benefits affect the volume, severity,
or cost of dental-related hospital visits? J Oral Maxillofac Surg
75:467, 2017
17. Sepp€anen L, Lauhio A, Lindqvist C, et al: Analysis of systemic and
local odontogenic infection complications requiring hospital
care. J Infect 57:116, 2008
18. Sundararajan K, Gopaldas J, Somehsa H, et al: Morbidity and
mortality in patients admitted with submandibular space in-
fections to the intensive care unit. Anaesth Intensive Care
43:420, 2015
Downloaded for Anonymous User (n/a) at Universitas Indonesia from ClinicalKey.com by Elsevier on October 02, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
19. Thomas SJ, Atkinson C, Hughes C, et al: Is there an epidemic of
admissions for surgical treatment of dental abscesses in the UK?
BMJ 336:1219, 2008
20. Uluibau I, Jaunay T, Goss A: Severe odontogenic infections. Aust
Dent J 50:s2, 2005
21. Kegeles SS: Why people seek dental care: A review of present
knowledge. Am J Pub Health Nations Health 51:1306, 1961
22. Cohen LA, Bonito AJ, Akin DR, et al: Toothache pain: A compar-
ison of visits to physicians, emergency departments and den-
tists. J Am Dent Assoc 139:1205, 2008
23. Sainuddin S, Hague R, Howson K, Clark S: New admission
scoring criteria for patients with odontogenic infections: A pilot
study. Br J Oral Maxillofac Surg 55:86, 2017
24. Jundt JS, Gutta R: Characteristics and cost impact of severe
odontogenic infections. Oral Surg Oral Med Oral Pathol Oral Ra-
diol 114:558, 2012
25. Ylijoki S, Suuronen R, Jousimies-Somer H, et al: Differences be-
tween patients with or without the need for intensive care
due to severe odontogenic infections. J Oral Maxillofac Surg
59:867, 2001
26. Gams K, Shewale J, Demian N, et al: Characteristics, length of
stay, and hospital bills associated with severe odontogenic infec-
tions in Houston, TX. J Am Dent Assoc 148:221, 2017
27. Charlson M, Szatrowski TP, Peterson J, Gold J: Validation
of a combined comorbidity index. J Clin Epidemiol 47:
1245, 1994
28. Sepp€anen L, Rautemaa R, Lindqvist C, Lauhio A: Changing clin-
ical features of odontogenic maxillofacial infections. Clin Oral
Investig 14:459, 2010
29. Australian Bureau of Statistics: Regional Population Growth,
Australia, 2013-14: Queensland State Summary. Available at:
http://www.abs.gov.au/ausstats/abs@.nsf/Previousproducts/
3218.0Main Features302013-14?opendocument&tabname=
Summary&prodno=3218.0&issue=2013-14&num=&view=’’.
Accessed January 2018
30. Maroldi R, Farina D, Ravanelli M, et al: Emergency imaging
assessment of deep neck space infections. Semin Ultrasound
CT MR 33:432, 2012
31. Lazor JB, Cunningham MJ, Eavey RD, Weber AL: Comparison of
computed tomography and surgical findings in deep neck infec-
tions. Otolaryngol Head Neck Surg 111:746, 1994
32. Freling N, Roele E, Schaefer-Prokop C, Fokkens W: Prediction of
deep neck abscesses by contrast-enhanced computerized to-
mography in 76 clinically suspect consecutive patients. Laryngo-
scope 119:1745, 2009
33. Smith JL, Hsu JM, Chang J: Predicting deep neck space abscess
using computed tomography. Am J Otolaryngol 27:244, 2006
34. Rosenthal M, Oreadi D, Kraus J, et al: Comparison of preopera-
tive computed tomography and surgical findings in maxillofacial
infections. J Oral Maxillofac Surg 69:1651, 2011