YIJOM-3516; No of Pages 8

Int. J. Oral Maxillofac. Surg. 2016; xxx: xxx–xxx

http://dx.doi.org/10.1016/j.ijom.2016.09.016, available online at http://www.sciencedirect.com

Research Paper
Clinical Pathology

Can progression of odontogenic K. Zemplenyi1, B. Lopez2,

M. Sardesai3, J. K. Dillon1
1
Department of Oral and Maxillofacial

infections to cervical necrotizing Surgery, University of Washington, Seattle,

WA, USA; 2School of Dentistry, University of
Washington, Seattle, WA, USA; 3Department
of Otolaryngology Head and Neck Surgery,

soft tissue infections be University of Washington, Seattle, WA, USA

§
predicted?
K. Zemplenyi, B. Lopez, M. Sardesai, J.K. Dillon: Can progression of odontogenic
infections to cervical necrotizing soft tissue infections be predicted?. Int. J. Oral
Maxillofac. Surg. 2016; xxx: xxx–xxx. # 2016 International Association of Oral and
Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.

Abstract. The progression of odontogenic infections to necrotizing soft tissue

infections (NSTIs) is unknown. The Laboratory Risk Indicator for Necrotizing
Fasciitis (LRINEC) score is used to predict risk of NSTI. This study aimed to (1)
estimate the frequency at which odontogenic infections progress to NSTIs, (2)
measure the value of LRINEC in predicting progression to NSTI, and (3) estimate
the charges associated with managing NSTIs. This retrospective cohort study
enrolled all subjects admitted for the management of odontogenic infections from
2001 to 2013. The primary predictor was the LRINEC score. The primary outcome
was NSTI. The secondary outcome was billing charges. Descriptive and bivariate
statistical analyses were performed, with significance set at a P-value of <0.05. Of
479 odontogenic infections, (1.0%) progressed to NSTI. The mean LRINEC for
NSTI was 5.8 and for odontogenic infection was 3.4 (P = 0.043). LRINEC
parameters for the prediction of NSTIs had 60% sensitivity, 68.4% specificity, 20%
positive predictive value, and 92.9% negative predictive value. The mean charge for Key words: LRINEC; necrotizing soft tissue
NSTI was $319,337 and for odontogenic infections was $19,291 (P = 0.051). One infections; odontogenic infections.
percent of odontogenic infections progressed to NSTIs. The LRINEC score was not
able to identify all NSTIs. NSTIs are 16 times more costly. Accepted for publication 26 September 2016

Necrotizing soft tissue infections (NSTIs)
and descending necrotizing mediastinitis
§
Presented at the 97th Annual Meeting of
the American Association of Oral and Maxil-
lofacial Surgeons, Oral Abstract Session IV,
September 2015.
are rare but rapidly progressive, usually
polymicrobial, infections with high limb
and life mortality. Mortality rates of be-
tween 30% and 50% have been reported
for descending necrotizing mediastinitis.1
With advancements made in the medical
field, other sources have reported a slight
decrease in the mortality rate to 20–40%.2
Early diagnosis and aggressive treatment
is critical to limit the associated morbidity
and mortality.
The progression of odontogenic
infections to NSTIs is well described,
but the frequency of this progression

0901-5027/000001+08 # 2016 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
YIJOM-3516; No of Pages 8

2 Zemplenyi et al.

and predictive factors are unclear.3 To inpatient costs associated with managing
better discriminate patients with NSTIs patients with NSTIs.
from those with other soft tissue infec-
tions, Wong et al. proposed the Labora-
Materials and methods
tory Risk Indicator for Necrotizing
Fasciitis (LRINEC) score4 . The LRI- Study design/sample
NEC score is a numeric score that ranges
After obtaining institutional review board
from 0 to 13 and is computed using six
approval, the investigators implemented a
laboratory indices: C-reactive protein
retrospective cohort study and enrolled a
(CRP), white blood cell (WBC) count,
sample derived from the population of
haemoglobin (Hb), sodium, creatinine,
subjects who presented to a medical center
and blood glucose. Individual point
for the evaluation and management of
values are summed to give the total
odontogenic infections between January
LRINEC score (Table 1). With this sys-
1, 2001 and December 31, 2013. To be
tem, a score 5 indicates <50% risk of
included in the study sample, the subjects
NSTI, a score of 6–7 indicates 50–75%
had to have one of the following discharge
risk of NSTI, and a score 8 indicates a
diagnoses, according to the International
>75% risk of NSTI.
Classification of Diseases, Ninth Revision
In the original study, which used patient
(ICD-9 codes): cellulitis (528.3), cellulitis
records from two tertiary hospitals over a
and abscess of face/neck (682, 682.1),
5-year period, all patients diagnosed with
mediastinitis (519.2), neck swelling/mass
necrotizing fasciitis were compared with a
(784.2), dental caries (521.09), and necro-
random cohort of patients diagnosed with
tizing fasciitis (728.86). Alternatively they
cellulitis or abscess. The aetiology and
could have a Current Procedural Termi-
location of the infections was not
nology (CPT) code for debridement
recorded.4 Thus, it is unclear whether this
(11,040–11,044, 41,000–41,008, 41,015–
diagnostic tool is useful in the early iden-
41,018, 42,725, 97,597), intraoral/extra-
tification of NSTI in its progression from
oral incision and drainage (21,501), ster-
odontogenic infection. It was hypothe-
nal debridement (21,627), open treatment
sized that the LRINEC score would be
of sternum fracture with or without skele-
positively associated with the risk of NSTI
tal fixation (21,825), thoracoscopy
in subjects with odontogenic infection.
(32,651–32,652), creation of pericardial
The specific aims of this study were (1)
window (33,025), mediastinotomy with
to estimate the frequency at which odon-
exploration (39,000, 39,010), removal of
togenic infection progresses to NSTI in an
devitalized tissue from wound(s) (97,602),
inpatient cohort, (2) to measure the value
and negative pressure therapy (97,605). A
of the LRINEC score in predicting
computer search of the electronic medical
this progression, and (3) to estimate the
records was used to identify potential
subjects for inclusion. Exclusion criteria
Table 1. Laboratory Risk Indicator of Necro-
were patient age <18 years, any non-
tizing Fasciitis (LRINEC).
odontogenic infection, and patient not
LRINEC pursuing treatment or for whom data were
Laboratory parameter, units score
not complete.
CRP, mg/l
<150 0
150 4 Variables
Total WBC count 109/l
<15 0 The primary predictor variable was the
15–25 1 LRINEC score. The LRINEC score was
>25 2 computed using the parameters described
Hb, g/dl by Wong et al. using admission laboratory
>13.5 0 values for CRP, WBC count, Hb, sodium,
11–13.5 1 creatinine, and blood glucose.4 For each
<11 2 patient, the LRINEC score was calculated
Sodium, mmol/l as the sum of these six laboratory values as
135 0
<135 2
per Table 1.
Creatinine, mg/dl The primary outcome variable was
1.6 0 whether or not the patient developed a
>1.6 2 NSTI and/or descending necrotizing med-
Glucose, mg/dl iastinitis. This was determined by search-
180 0 ing for ICD codes 728.86 (necrotizing
>180 1 fasciitis) and 519.2 (mediastinitis), respec-
CRP, C-reactive protein; WBC, white blood tively. These diagnoses were confirmed
cell; Hb, haemoglobin. clinically, operatively, and by biopsy of
necrotic tissue. The secondary outcome
variable was the total dollar amount billed
to each subject derived from billing data.
The data were collected in a Microsoft
Excel spreadsheet (Microsoft, Redmond,
WA, USA).
The other study variables were grouped
into the following sets: demographic data,
patient history, values on admission, pre-
sentation, evaluation, treatment, and out-
come. The demographic variables were
age (years), race (Caucasian, African
American, or other), and sex (male or
female). Weight (kg), height (cm), and
body mass index (BMI, kg/m2) were also
recorded. The patient history variables
were history of diabetes (insulin-depen-
dent and non-insulin-dependent), history
of dental disease, and other medical/social
history (cardiac disease, liver disease, hy-
pertension, chronic obstructive pulmonary
disease, asthma, HIV/AIDS, cancer, psy-
chiatric disorder, tobacco use, intravenous
drug use, alcohol abuse, and homeless-
ness).
Age was determined by subtracting the
date of birth in the demographic informa-
tion from the date of admission. Sex and
race were gathered from the patient demo-
graphic information. Weight, height, and
BMI were collected from the nursing notes
recorded on the day of admission. Medical
and social histories were determined from
the emergency department notes. Any his-
tory of dental disease specifically was
determined from the emergency depart-
ment notes or preoperative notes indicat-
ing decay of one or more teeth on
admission.
The values recorded on admission were
temperature (8C), CRP (mg/l), total WBC
count (109/l), Hb (g/dl), sodium (mmol/
l), creatinine (mg/dl), and blood glucose
(mg/dl). Temperature on admission was
that recorded in the emergency depart-
ment notes. Admission laboratory data
for CRP, WBC, Hb, sodium, creatinine,
and blood glucose were obtained from the
laboratory results section of the patient’s
chart.
The presentation variables were the
presence of a draining wound (neck or
oral), indication of gas on initial imaging,
and whether there was airway compro-
mise. Information on the use of imaging
was obtained from the radiology section
of the patient’s chart, and the assessment
of gas on imaging was obtained from the
radiology report and a review of the im-
aging. Airway compromise was deter-
mined by ‘airway involvement’ or
‘airway deviation’ on the radiology re-
port. The evaluation variables were the
number and type of imaging modalities

Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
YIJOM-3516; No of Pages 8
used: computed tomography (CT) and
magnetic resonance imaging (MRI).
The treatment variables were the treat-
ment rendered, use of steroid treatment,
type and route of antibiotics given, total
number of operating room procedures
within 48 h following imaging (including
all washouts, which were done exclusively
in the operating room), and whether sur-
gery involved the mediastinum. Steroid
and antibiotic use was determined from
the inpatient medication list. The number
and type of surgeries, including mediasti-
nal involvement, was determined from the
operative notes and discharge summaries.
Comparison of imaging dates and proce-
dure notes allowed the determination of
which surgeries occurred within 48 h of
imaging as per the imaging recommenda-
tions of Freeman et al.5
All of the NSTI cases were treated with
the same NSTI algorithm, as per the med-
ical center protocol. This included quad
antibiotic treatment (penicillin G, vanco-
mycin, clindamycin, and levofloxacin or
gentamicin ) until culture results were
obtained, emergent debridement, serial
washouts, and 48-h CT imaging.
The outcome variables were total days
spent in the intensive care unit (ICU), total
length of hospital stay, whether imaging
was obtained within 48 h of discharge, any
complications from the infection or treat-
ment, and whether the patient died during
the 90-day postoperative period. ICU use
was determined from the discharge sum-
mary and inpatient notes. The discharge
summary also indicated the total length of
hospital stay. If the patient was not admit-
ted overnight, the length of stay was
recorded as 0. Imaging dates were com-
pared to the discharge summary to deter-
mine whether or not imaging was obtained
within 48 h of discharge. Additional infor-
mation received from the billing depart-
ment included type of medical coverage,
actual payment, and remaining balance.
Data analyses
All data were abstracted from the subjects’
medical records and recorded using a
spreadsheet (Microsoft Excel, Redmond,
WA, USA). Descriptive statistics includ-
ing the mean, frequency, range, and stan-
dard deviations were computed for each
study variable. A Mann–Whitney U-test
was used to compare the NSTI and odon-
togenic infection groups with respect to
continuous variables. Fisher’s exact test
was used to evaluate differences between
groups with respect to categorical vari-
ables such as sex and race. To assess
the prognostic value of the LRINEC score,
Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
Progression of odontogenic infection to NSTI
diagnostic statistics were computed: sen-
sitivity, specificity, positive predictive
value (PPV), and negative predictive val-
ue (NPV); this was done for the subset of
patients for whom complete LRINEC lab-
oratory parameters were obtained. To ac-
count for the small sample size, binomial
proportion confidence intervals (CI) were
constructed using Clopper–Pearson inter-
vals rather than relying on the normal
approximation to the binomial distribu-
tion. Statistical calculations were carried
out in R version 3.2.2 (R Development
Core Team, Vienna, Austria; www.
R-project.org). Statistical significance
was set at P = 0.05.
Results
The study population was composed of
988 subjects. Subjects were excluded for
the following reasons: non-odontogenic
origin of infection, peritonsillar infection,
osteomyelitis, and cases involving
infected hardware. After applying the ex-
clusion criteria, the final sample was com-
posed of 479 subjects. Table 2 summarizes
the results, with a comparison of NSTI to
odontogenic infection for all variables
analyzed. Five (1.0%) cases of NSTI were
identified among the 479 subjects, two of
which progressed to descending necrotiz-
ing mediastinitis. One of the five NSTI
patients died within the 90-day follow-up
period from chronic kidney disease for
which dialysis was refused, while none
of the odontogenic infection patients died
within the same follow-up window.
Of the 479 patients, 297 (62.0% ) of the
odontogenic infection and four (80%) of
the NSTI patients were male. The mean
age of odontogenic infection patients was
37.7 years (range 18–93 years) and of
NSTI patients was 48.8 years (range 28–
66 years) (P = 0.094). The mean BMI of
the NSTI patients was 26.2 kg/m2 (range
20.5–42.5 kg/m2), while the mean BMI
for odontogenic infection patients was
27.2 kg/m2 (range 15.1–57 kg/m2)
(P = 0.344). Fifty-five of the 474 odonto-
genic infection patients (11.6%) had a
history of diabetes, either insulin-depen-
dent or non-insulin-dependent. Of the five
NSTI patients, one had non-insulin-depen-
dent diabetes and one had insulin-depen-
dent diabetes mellitus. Two hundred and
seventy-five of the 474 odontogenic infec-
tion patients (58.0%) reported a history of
tobacco use in any form. Of the five NSTI
patients, two reported tobacco use. Thirty-
four of the 474 (7.2%) admitted to intra-
venous drug use (IVDU) currently or in
the past. None of the five NSTI patients
3
admitted to IVDU, chronic alcohol abuse,
or homelessness.
Three of the five NSTI patients received
intravenous antibiotics prior to arrival at
the medical center, one of whom received
quad antibiotic therapy. All bacterial cul-
tures for the five NSTI patients were poly-
microbial oral flora. Only one NSTI
patient presented with bacterial resistance;
culture grew Bacteroides fragilis with
clindamycin resistance. Table 3 shows a
comparison of the bacterial isolates from
the NSTI cases and odontogenic infection
cases, as per Dillon et al.7
NSTI cases required a greater total
number of surgeries (7.6 vs. 1.3;
P < 0.001) and had a longer total length
of hospital stay (24.2 vs. 4.18 ;
P < 0.001), and total ICU stay (14.6 vs.
0.56; P < 0.001) as compared to odonto-
genic infection cases. All three of these
categories indicated a significant differ-
ence (Fig. 1).
LRINEC
The predictive power of the LRINEC
score in identifying NSTI was evaluated.
Only 38 of the odontogenic infection
patients had all six laboratory values for
calculating the LRINEC score (Fig. 2). For
the majority of patients (92.1%), CRP was
not ordered. The five NSTI cases had all
six necessary laboratory values, which
produced LRINEC scores of 4, 6, 7, 8,
and 4 based on laboratory values at ad-
mission (Table 4).
The average LRINEC score of the five
patients in the NSTI group was 5.8 (range
4–8). This was elevated compared to the
LRINEC score for the 38 patients with all
laboratory values in the odontogenic in-
fection group, which was 3.4 (range 0–9)
(P = 0.043). As described by Wong et al.
in 2004, a LRINEC score of 6 was used
as a predictive threshold for progression to
NSTI.4 Using this LRINEC value cut-off
for the 43 patients with all six LRINEC
laboratory values, the data indicated a
sensitivity of 60% (95% CI 15–95%),
specificity of 68.4% (95% CI 51–82%),
PPV of 20% (95% CI 4–48%), and NPV of
92.9% (95% CI 76–99%). Although the
average LRINEC score for the NSTI pool
was significantly greater than that for the
odontogenic infection pool, LRINEC was
unable to diagnose two of the five NSTIs,
leading to the sensitivity of 60%. In addi-
tion, 12 of the 38 odontogenic infection
patients had LRINEC scores greater than
6, leading to an elevated false-positive rate
and thus a low PPV. The limited sample
size of five NSTIs in this study is chal-
YIJOM-3516; No of Pages 8

4 Zemplenyi et al.

Table 2. Study variables grouped by necrotizing soft tissue infection status (odontogenic Discussion
infection vs. necrotizing soft tissue infection).a
A rare but documented complication of
Odontogenic Necrotizing soft
b odontogenic infections is progression to a
Variables infection tissue infection P-value
NSTI, for which predictive factors are
Demographic variables unknown. The LRINEC score has been
Age, years 37.7  14.4 48.8  15.7 0.094 used as a tool to predict the risk of pro-
Race 0.647
gression to NSTI. It was hypothesized that
African American 13.9% 20%
Caucasian 67.7% 80% the LRINEC score would be positively
Other 18.4% 0% associated with the risk of progression
Sex, male 62.0% 80% 0.656 to NSTI in subjects with odontogenic
Weight, kg 80.7  24.6 84.8  19.7 0.572 infection. Specifically, it was sought to
Height, cm 172.4  10.4 181.8  8.3 0.067 (1) estimate the frequency at which odon-
BMI, kg/m 2 27.2  7.8 26.2  9.14 0.344 togenic infection progresses to NSTI in an
Temperature, 8C 37.22  0.99 37.42  0.82 0.612 inpatient cohort, (2) measure the value of
Co-morbidities the LRINEC score in predicting this pro-
Hypertension 14.4% 40% 0.026 gression, and (3) estimate the inpatient
Tobacco 58.0% 40% 0.722 costs associated with managing patients
Alcohol 11.8% 0% 1.0 with NSTIs.
HIV 1.48% 0% 1.0 In the patient pool seen at the study
Hepatitis C 2.74% 0% 1.0 medical center between 2001 and 2013,
Intravenous drug use 7.2% 0% 1.0 odontogenic infection progressed to NSTI
Cancer 1.69% 0% 1.0
Psychiatric illness 9.28% 0% 1.0
in five of 479 cases (1.0%). Analysis of the
Renal disease 0.63% 20% 0.023 data indicated that there was no significant
Diabetes 11.6% 40% 0.110 difference between patients who pro-
Dental disease 93.80% 100% 1.0 gressed to NSTIs and those who did not
Homelessness 0.6% 0% 1.0 with regard to the risk factors, specifically
Laboratory variables
tobacco use, IVDU, alcohol abuse, diabe-
CRP, mg/l 116.46  96.6 101.52  96.69 0.919 tes, and homelessness. This corroborates
Total WBC count, 109/l 13.68  4.63 14.59  5.68 0.555 the findings of the study by Ylijoki et al.,
Hb, g/dl 13.36  2.75 10.9  1.12 0.004 in which patients with severe odontogenic
Sodium, g/dl 135.79  2.98 135  2.65 0.395 infections who required ICU care were
Creatinine, mg/dl 0.90  0.63 1.92  1.8 0.387 compared to those who did not require
Glucose, mg/dl 126.24  62.4 127  71.1 0.947 ICU care; they found no particular anam-
Hospital course nestic background variable to be associat-
Gas on imaging 12.60% 80% 0.001 ed with the need for intensive care.6 In
Draining wound 96.83% 100% 1.00 addition, the present results did not indi-
Airway compromise 22.40% 60% 0.081 cate a statistically significant association
Number of CTs 1.14  1.01 2.4  3.71 0.988 between demographic variables and pro-
Number of MRIs 0.006  0.08 0 0.042 gression to NSTI, although the total num-
Mediastinal involvement 1.47% 40% 0.003 ber of surgeries, total ICU stay, and total
Total number of surgeries 1.3  0.757 7.6  6.43 <0.001
length of hospital stay were significantly
ICU, days 0.56  2.26 14.6  11.8 <0.001
LOS, days 4.18  4.54 24.2  10.4 <0.001 increased within the NSTI pool as com-
Imaging within 48 h of discharge 58.50% 20% 0.166 pared to the odontogenic infection pool
(Table 2).
Financials All five NSTI infections were polymi-
Amount billed, US$ 19,291  25,915 319,337  271,506 0.051
crobial in nature and stemmed from the
Amount collected, US$ 8975  14,663 177,329  248,915 0.2
odontogenic flora. NSTIs and odontogenic
BMI, body mass index; HIV, human immunodeficiency virus; CRP, C-reactive protein; WBC, infections may be either polymicrobial or
white blood cell; Hb, haemoglobin; CT, computed tomography; MRI, magnetic resonance monomicrobial, and the present results
imaging; ICU, intensive care unit; LOS, length of stay.
a
The results are presented as the number and percentage, or the mean  standard deviation.
indicate no bacterial species distinguish-
b
P < 0.05 significant. ing the NSTI cases. The bacterial species
of odontogenic infection given in Table 3
were drawn from the study of Dillon
et al.,7 which investigated odontogenic
infection and causative organisms during
lenging and leads to large confidence $660,426). By comparison, the average
a time period similar to that of this study.
intervals. cost of odontogenic infection was
The authors of the present study operated
$19,291 ($476–$335,953). This is a 16-
under the assumption that the microorgan-
fold increase in cost (P = 0.051) (Fig. 3,
Financials isms of odontogenic infection have
Table 5). Comparison of the average col-
remained unchanged since the time of
An additional aim of the study was to lection for NSTI vs. non-NSTI odonto-
the aforementioned study and thus remain
evaluate the cost of NSTIs to hospitals. genic infection was $177,329 ($0–
applicable. Of note, only one of the NSTI
The average cost in the case of NSTI $607,342) vs. $8975 ($0–$163,647)
cases showed antimicrobial resistance to
patients was $319,337 (range $114,842– (P = 0.2) (Table 2).

Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
YIJOM-3516; No of Pages 8
Table 3. Frequency of pathogens isolated, in descending order of occurrence.
Necrotizing soft tissue infection
(of five total)
Streptococcus milleri (n = 5)
Coagulase-negative Staphylococcus species
(n = 3)
Anaerobic Gram-negative rods (n = 3)
a-Hemolytic Streptococcus species (n = 2)
Propionibacterium (n = 1)
Eikenella species (n = 1)
Pseudomonas species (n = 1)
clindamycin, and this patient had been on
several rounds of antibiotic therapy with-
out definitive treatment prior to admission.
The second aim of the study was to
determine the value of the LRINEC score
in predicting progression to NSTI. The
Fig. 1. Comparison of average number of surgeries, average number of CT scans, average
hospital stay (days), and average ICU stay (days) for necrotizing soft tissue infection (NSTI) vs.
odontogenic infection (OI) patients.
Fig. 2. Inclusion criteria.
Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
Progression of odontogenic infection to NSTI
Odontogenic infection
(from Dillon et al., 2012)7
Anaerobic Gram-negative rods
a-Hemolytic Streptococcus species
Streptococcus milleri
Anaerobic non-spore-forming
Gram-positive rods
Coagulase-negative Staphylococcus species
Anaerobic Gram-positive cocci
Neisseria species
average LRINEC score of the five patients
with an NSTI was 5.8, which was signifi-
cantly higher than the LRINEC score of
the 38 patients with all laboratory values
in the odontogenic infection group, which
was 3.4 (P = 0.043). These data, specifi-
5
cally for the pool of 43 patients with full
LRINEC laboratory values, indicated a
sensitivity of 60% (95% CI 15–95%),
specificity of 68.4% (CI 51–82%), PPV
of 20% (CI 4–48%), and NPV of 92.9%
(CI 76–99%).
The large confidence intervals involved
in a pool of only five subjects should be
noted and caution should be maintained
when interpreting the results. It is pro-
posed that the results of this study are
combined with those for NSTIs resulting
from odontogenic infection at other insti-
tutions in order to increase the power of
the study and look more closely at the role
of the LRINEC. However, since NSTI
originating from odontogenic infection
is such a rare entity, the data presented
do give some guiding information. Impor-
tantly, the LRINEC was unable to predict
the progression to NSTI in two of the five
NSTI cases. In a study performed in 2012,
Thomas and Meyer sought to assess the
prognostic value of the LRINEC score in
predicting 17 cases of cervical necrotizing
fasciitis and found that a cut-off level of
6 had a sensitivity of 56%, specificity of
60%, PPV of 25%, and NPV of 85%. The
conclusion from that study was that the
LRINEC was not useful for distinguishing
cervical necrotizing fasciitis from non-
cervical necrotizing fasciitis.8 Similar de-
scriptive statistics were described by Hol-
land in 2009 who looked at 10 NSTIs (not
specific to body region) and found a sen-
sitivity of 80%, specificity of 67%, PPV of
57%, and NPV of 86%.9 Holland reported
a sensitivity higher than that found in the
present study, 80% vs. 60%, however the
difference may be because the former
cohort had a larger sample size, with 10
positive NSTIs compared to five positive
cases in the present study. This study
focused on NSTIs of odontogenic origin,
while the study by Holland did not dis-
criminate by origin of infection.
A recent publication by Sandner et al.
set out to determine the value of the
LRINEC score as a predictor of cervical
necrotizing fasciitis from all head and
neck sources (trauma, odontogenic, para-
tonsillar, parapharyngeal, retropharyn-
geal, and any other acute lymphadenitis,
cellulitis, or abscess of the neck). Using a
LRINEC score of 6 as the cut-off for
elevated risk, these researchers found a
sensitivity of 0.94 (95% CI 0.92–0.96), a
specificity of 0.94 (95% CI 0.70–1.00), a
PPV of 0.29 (95% CI 0.17–0.44), and a
NPV of 0.99 (95% CI 0.99–1.00).10 They
concluded that the LRINEC has value in
the early identification of cervical necro-
tizing fasciitis and should be calculated
during the evaluation of such conditions.
YIJOM-3516; No of Pages 8

6 Zemplenyi et al.

Table 4. Necrotizing soft tissue infection: patient demographics. The LRINEC score itself weights CRP
Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 levels heavily, in an all-or-none fashion: 4
points for a CRP 150 mg/l and 0 points
Age, years 28 45 63 42 66
BMI, kg/m 2 20.5 22.6 22.8 22.8 42.5 for CRP <150 mg/l. By contrast, the other
LOS, days 34 34 27 15 12 laboratory values contribute a maximum 2
Days in ICU 27 28 5 5 8 points. Therefore, a CRP laboratory value
Number of surgeries 11 17 6 2 2 is essential for the LRINEC score to be of
LRINEC score 4 6 7 8 4 any value as a diagnostic tool, and if this
BMI, body mass index; LOS, length of stay; ICU, intensive care unit; LRINEC, Laboratory Risk variable is missing, it will significantly
Indicator of Necrotizing Fasciitis. alter the LRINEC. The weight given to
the CRP value could be reconsidered ei-
ther by lowering the threshold CRP cut-off
or by creating a gradient rather than a
binary 0 or 4 points within the LRINEC
algorithm. Borschitz et al. tested this no-
tion in their recent 2015 study. The group
compared 138 clinical and laboratory fea-
tures of 29 NSTI cases to 59 cases of
cellulitis (without discrimination for the
origin of infection). After logistic regres-
sion analysis, multiple modified LRINEC
score calculations were trialled, most of
which employed an added intermediate
CRP point value of 2 points for a CRP
>100 mg/l . This modification served to
increase the sensitivity and NPV of the
LRINEC.11 Given a larger sample size, as
in a multicenter study, similar metrics
Fig. 3. Total amount billed and collected for necrotizing soft tissue infection (NSTI) vs. would be valuable in calculating NSTI
odontogenic infection (OI) patients. from the narrowed pool of odontogenic
infection. The authors of the present study
intend to apply a similar analysis and
The authors did not specifically look at et al. reported six LRINEC predictor vari- modified LRINEC score in prospective
odontogenic infections, nor was it clear ables for all patients, in both the control studies, as this study is limited by its
whether they were responsible for all and experimental groups. In the present sample size.
odontogenic infections at their institution. study pool, while six LRINEC predictor CRP is not ordered routinely in odonto-
At the medical center at which the present variables were reported for all five NSTI genic infection as it is a non-specific
study was performed, the majority of head patients, these were reported for only 8.0% marker of inflammation and a raised
and neck infections (irrespective of sever- (38/474) of the control patients. Five pre- CRP value in itself does not necessarily
ity) presenting to the emergency depart- dictive variables, besides CRP, were indicate infection. That all five of the
ment are odontogenic in nature. Patients reported for 83.5% (396/474) of the con- NSTI patients had a CRP value recorded
with odontogenic infections are seen by trol group. This discrepancy may be due to may or may not speak to a clinical suspi-
the oral and maxillofacial service, and a fundamental difference in diagnostic cion of NSTI early in the hospital course,
patients with NSTIs are seen by either protocols between the two sites (different since the reasons for ordering a CRP are
the oral and maxillofacial service or the demographics, social/health care policies, not clear. CRP is not a standard laboratory
department of otolaryngology head and and available resources). The exclusion of test for infections and may be more reflec-
neck surgery, depending upon the infec- CRP values from a large percentage of tive of the preferences and/or training of
tion source. This study focused specifical- odontogenic infection patients, although a the ordering emergency department phy-
ly on NSTIs from an odontogenic origin, major weakness of this study, still leads us sician as opposed to infection severity per
strictly excluding all other head and neck to question the relevance of the CRP value se. Since the CRP and preoperative labo-
sources. Given the rarity of NSTI and the within the LRINEC scoring system given ratory data were ordered by the emergency
exclusion criteria applied, the NSTI sam- the additional cost of the test and the ad department upon patient presentation, the
ple size was reduced. Of note, Sandner hoc frequency with which it is obtained. oral and maxillofacial service did not have
control over when or for which cases CRP
was ordered. As this was a retrospective
Table 5. Finances for five necrotizing soft tissue infection patients.
study, it can only be hypothesized as to
Patient 1 Patient 2 Patient 3 Patient 4 Patient 5 why certain laboratory tests were ordered
Method of payment Montana medicaid Group health Self pay Regence Medicare for some patients and not for others. Fur-
LOS, days 34 34 27 15 12 thermore, the value from CRP comes from
Total bill $487,193 $660,426 $209,388 $114,842 $124,833 trends and not solitary readings. Sharma
Amount collected $166,036 $607,342 0 $89,930 $23,339 et al. analyzed the value of CRP in cases of
% Reimbursement 34.1% 92% 0% 78.3% 18.7% odontogenic infection, recording serial
LOS, length of stay. CRP readings on admission, day 4, and
day 8 of the patient’s hospital course.12

Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
YIJOM-3516; No of Pages 8
This group concluded that CRP is a signif-
icant predictor of the severity of infection
and effectiveness of the treatment regime,
with upwards trending CRP values indicat-
ing a higher risk of complicated neck
infections and longer hospital stays.12
The LRINEC score captures only a snap-
shot CRP value. The question then follows
whether obtaining serial CRP values for all
severe odontogenic infections makes finan-
cial sense. At an average cost of $50 per
CRP test, with three CRP tests per odonto-
genic infection, this would amount to an
additional $70,950 total, or $5400/year,
over the time period of the study (2001–
2013). While this may seem insignificant,
if serial CRPs were additionally applied to
all non-odontogenic infections and non-
odontogenic NSTIs, the total cost could
present a significant burden.
Table 5 shows the hospital bills incurred
by the five NSTI patients and the range of
reimbursement rates, with an average cost
of $319,337 for NSTI compared to
$19,291 for odontogenic infection. The
data indicate that although NSTIs repre-
sent only 1% of odontogenic infections,
they disproportionately account for over
10% of the cost. This is not surprising
given the increase in hospital stay, in
particular the ICU, increased surgical
interventions, and routine tests done, such
as complete blood counts, basic metabolic
panels, and serial imaging to assess
progress/progression of the infection. Al-
though odontogenic infection progressing
to NSTI appeared to have occurred at a
frequency of one every 2 years on average,
the hundreds of thousands of dollars spent
and left unreimbursed on their care are
notable. The hospital was reimbursed 0%
for patient 3, who was uninsured. Accord-
ing to Dillon et al., uninsured patients
represented 61.9% of the odontogenic in-
fection patient pool in King County prior
to the Affordable Care Act.7,13 This per-
centage of uninsured has decreased with
the Affordable Care Act, which covers the
cost of these infections. However, for any
insurance to be of benefit, patients must
enrol and not everyone will do so.
Of interest is the overall lack of NSTI
cases in the head and neck region. The
study medical centre is the only tertiary
level I trauma center in the Pacific North-
west and serves the US states of Washing-
ton, Wyoming, Alaska, Montana, and
Idaho, with a total population of 10.5
million. Additionally, it is the primary
referral center for NSTI for this popula-
tion. In 2015, there were a total of 98 NSTI
cases with the following sites of origin:
45% extremity, 33% Fournier, 18% trunk,
4% head and neck. In postulating as to
Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016
Progression of odontogenic infection to NSTI
why the majority of NSTI cases occur
outside the head and neck region, the
etiological factors need to be reviewed.
Well-documented risk factors for the de-
velopment of NSTI, regardless of the ori-
gin of infection, include diabetes, vascular
disease, and immune compromise. Both
diabetes and peripheral vascular disease
affect the blood supply. It is hypothesized
that there is an increased rate of injury to
the genitalia and extremities due to their
inherent anatomical vulnerability to vas-
cular compromise. Compartment syn-
drome renders ischaemia and necrosis to
the injured area, in turn making the tissue
more defenseless to microbiological at-
tack. This tissue damage triggers a con-
siderable inflammatory response, resulting
in large increases in CRP and the other
biochemical changes seen. Compared to
these sites, the head neck region has a rich
blood supply, is normally bathed in a large
variety of oral flora to which it has
adapted, and is enveloped in fascial
planes. This rich blood supply and anato-
my renders compartment syndrome much
less likely in head and neck NSTI. How-
ever, the fascial planes afford limited
boundaries to the spread of infection, with
an increased potential for rapid descend-
ing infection to the mediastinum and more
commonly, airway compromise.
In conclusion, the progression of odon-
togenic infection to NSTI is small-only
1.0% in the present cohort over a 12-year
period. However the financial burden
these infections place on an institution is
substantial. Although the LRINEC was
statistically significant in detecting NSTIs,
it cannot be relied upon as the sole diag-
nostic tool for progression to NSTI as it
failed to correctly diagnose two of the five
NSTI cases. It behoves us to more accu-
rately predict progression to NSTI, as they
are costly infections-over 16 times the
amount for odontogenic infections. Given
the rarity of NSTIs of odontogenic aetiol-
ogy, there would be great benefit in a
multi-institutional study for improved di-
agnosis and management.
Funding
This study was supported in part by the
University of Washington Department of
Oral and Maxillofacial Surgery Research
and Education Fund.
Competing interests
No competing interests.
Ethical approval
Institutional Review Board of the Univer-
sity of Washington: #44316.
7
Patient consent
Not applicable.
Acknowledgement. Dr. Thomas B. Dodson
DMD, MPH for statistical advice and
input on the organization of the manu-
script.
M. Zemplenyi (PhD Candidate, Depart-
ment of Biostatistics, Harvard University,
Boston, MA, USA) for statistical support.
References
1. Pinto A, Scaglione M, Scuderi MG, Tortora
G, Daniele S, Romano L. Infections of the
neck leading to descending necrotizing med-
iastinitis: role of multi-detector row comput-
er tomography. Eur J Radiol 2007;65:389–
94.
2. Sarna T, Sengupta T, Miloro M, Kolokythas
A. Cervical necrotizing fasciitis with des-
cending mediastinitis: literature review and
case report. J Oral Maxillofac Surg
2012;70:1342–50.
3. Sancho L, Minamoto H, Fernandez A, Sennes
L, Jatene F. Descending necrotizing medias-
tinitis: a retrospective surgical experience.
Eur J Cardiothorac Surg 1999;16:200–5.
4. Wong CH, Khin LW, Heng KS, Tan KC, Low
CO, The LRINEC. (Laboratory Risk Indica-
tor for Necrotizing Fasciitis) score: a tool for
distinguishing necrotizing fasciitis from oth-
er soft tissue infections. Crit Care Med
2004;32:1535–41.
5. Freeman RK, Vallières E, Verrier ED,
Karmy-Jones R, Wood DE. Descending nec-
rotizing mediastinitis: an analysis of the
effects of serial surgical debridement on
patient mortality. J Thorac Cardiovasc Surg
2000;119:260–7.
6. Ylijoki S, Suuronen R, Jousimies-Somer H,
Meurman JH, Lindqvist C. Differences be-
tween patients with or without the need for
intensive care due to severe odontogenic infec-
tions. J Oral Maxillofac Surg 2001;59:867–72.
7. Dillon JK, Christensen B, Han M. The cause
of cost in the management of odontogenic
infections 1: a demographic survey and mul-
tivariate analysis. J Oral Maxillofac Surg
2013;71:2058–67.
8. Thomas AJ, Meyer TK. Retrospective eval-
uation of laboratory-based diagnostic tools
for cervical necrotizing fasciitis. Laryngo-
scope 2012;122:2683–7.
9. Holland MJ. Application of the Laboratory
Risk Indicator in Necrotising Fasciitis (LRI-
NEC) score to patients in a tropical tertiary
referral centre. Anaesth Intensive Care
2009;37:588–92.
10. Sandner A, Moritz S, Unverzagt S, Plontke SK,
Metz D. Cervical necrotizing fasciitis—the
value of the Laboratory Risk Indicator for
Necrotizing Fasciitis score as an indicative
parameter. J Oral Maxillofac Surg
2015;73:2319–33.
YIJOM-3516; No of Pages 8

8 Zemplenyi et al.

11. Borschitz T, Schlicht S, Siegel E, Hanke E, 2012;3:148–51. http://dx.doi.org/10.4103/ University of Washington

von Stebut E. Improvement of a clinical 0975-5950.111369. 4 West Clinic
score for necrotizing fasciitis: ‘pain out of 13. Dillon JK, Christensen B, Han M. The cause 325 Ninth Ave.
proportion’ and high CRP levels aid the of cost in the management of odontogenic Box 359893
diagnosis. PLOS ONE 2015;10:e0132775. infections 2: multivariate outcome analyses. Seattle
http://dx.doi.org/10.1371/journal.- J Oral Maxillofac Surg 2013;71: 2068–76. WA 98104
pone.0132775.2015. USA
12. Sharma A, Gokkulakrishnan S, Shahi AK, Address: Tel: +1 206 744 4124;
Jasjit Dillon Fax: +1 206 744 2810
Kumar V. Efficacy of serum CRP levels as
Department of Oral and Maxillofacial E-mail: dillonj5@uw.edu
monitoring tools for patients with fascial space
infections of odontogenic origin: a clinicobio- Surgery
chemical study. Natl J Maxillofac Surg Harborview Medical Center

Please cite this article in press as: Zemplenyi K, et al. Can progression of odontogenic infections to cervical necrotizing soft tissue
infections be predicted?, Int J Oral Maxillofac Surg (2016), http://dx.doi.org/10.1016/j.ijom.2016.09.016