10.1007_s00405-011-1761-1

HEAD AND NECK

Deep neck infections: a study of 365 cases highlightingrecommendations for management and treatment

Paolo Boscolo-Rizzo Marco Stellin Enrico Muzzi

Monica Mantovani Roberto Fuson Valentina Lupato

Franco Trabalzini Maria Cristina Da Mosto

Received: 2 May 2011 / Accepted: 26 August 2011 / Published online: 14 September 2011

Springer-Verlag 2011

Abstract The aims of this investigation were to review

the clinical behavior of deep neck infections (DNIs) treated

in our institution in order to identify the predisposing

factors of life-threatening complications and propose

valuable recommendations for management and treatment.

A total of 365 adult patients with DNIs were retrospec-

tively identified. One-hundred and thirty-nine patients

(38.1%) underwent surgical drainage. Overall, 226 patients

(61.9%) responded effectively to intravenous antimicrobial

therapy only. There were 67 patients (18.4%) developing

life-threatening complications. Diabetes mellitus (odd ratio

5.43; P \ 0.001) and multiple deep neck spaces involve-ment (odd ratio 4.92; P \ 0.001) were the strongest inde-pendent predictors of complications. The mortality rate was

0.3%. Airway obstruction and descending mediastinitis are

the most troublesome complications of DNIs. In selected

patients, a trial of intravenous antibiotic therapy associated

with an intensive computed tomography-based wait-and-

watch policy may avoid an unnecessary surgical procedure.

However, about one-fourth of patients present significant

comorbidities, which may negatively affect the course of

the infection. In these cases and in patients with large or

multiple spaces infections, a more aggressive surgical

strategy is mandatory.

Keywords Abscess Complications Computedtomography Deep neck infections Diagnosis Treatment

Introduction

The deep neck spaces are regions of loose connective tissue

filling the areas between the three layers of deep cervical

fascia. Deep neck infections (DNIs) are suppurative

infections that develop within deep neck spaces.

Deep neck infections usually starts as cellulitis in the

soft tissues adjacent to the source of upper aero-digestive

tract infection: if left untreated and depending on the vir-

ulence of the causative pathogen, the infection will even-

tually lead to an abscess and spread along cervical into to

the mediastinum [1].

The insidious evolution of this pathology still represents

an open problem. An unsuspecting physician may under-

estimate an initially localized infection, which could

shortly present as airway collapse or descending

mediastinitis.

In most of cases, the source of the infection is a peri-

apical infection, involving the mandibular second or third

molar teeth, or an acute follicular tonsillitis [2, 3]. The

microbiology of DNIs reflects the normal endogenous

P. Boscolo-Rizzo M. Stellin R. Fuson V. Lupato M. C. Da Mosto

Department of Medical and Surgical Specialities,

University of Padua, School of Medicine, Padua, Italy

P. Boscolo-Rizzo M. Stellin M. Mantovani R. Fuson V. Lupato M. C. Da MostoRegional Center for Head and Neck Cancer,

University of Padua, School of Medicine,

Treviso Regional Hospital,

Treviso, Italy

P. Boscolo-Rizzo (&)Viale Umbria 6, 30019 Chioggia, Italy

e-mail: [email protected]

E. Muzzi

Otorhinolaryngology Unit,

University Hospital S. Maria della Misericordia, Udine, Italy

F. Trabalzini

Department of Sense Organs, Otology and Skull Base

Surgery Unit, Siena University Hospital, Siena, Italy

123

Eur Arch Otorhinolaryngol (2012) 269:12411249

DOI 10.1007/s00405-011-1761-1

upper aerodigestive tract flora and includes both aerobic

and anaerobic microorganisms. As a consequence, the

microbiology of DNIs is similar and no correlation usually

exists between the anatomical region and microbiology of

the infection [3].

The management of DNIs requires a multidisciplinary

approach including head and neck surgeon, thoracic sur-

geon, infectious disease specialist, and radiologist.

The aims of this investigation were to review the clinical

behavior of DNIs treated in our institution in order to

propose valuable recommendations for management and

identify the predisposing factors of life-threatening

complications.

Patients and methods

This is an observational descriptive retrospective study of

all cases of DNIs treated at the Department of Surgery,

Treviso Regional Hospital over a period of 15 years

(between May 1995 and November 2010).

Clinical charts, imaging and bacteriologic studies were

reviewed. Patients with head and neck cancer, peritonsillar

cellulitis or abscess, and post-traumatic infections were not

included in the study. The following variables were

reviewed: demographic and clinical data, associated sys-

temic diseases, bacteriology, imaging studies, source, site,

and character of the infections, medical and surgical

treatment, complications, and outcome.

The infection was categorized according to the character

of infections (cellulitis vs. abscess) and to the involved

spaces (submandibular space, lateral pharyngeal space,

retropharyngeal space, prevertebral space, parotid space,

masticatory space, temporal space, visceral vascular space,

anterior visceral space) according to Levitt [4]. Patients

with involvement of two or more spaces were classified as

having multiple spaces infection.

The reference ranges for standard values at our labora-

tory were 4 9 10311 9 103/mm3 for white blood cell

count (WBC), 1.8 9 1038 9 103/mm3 for neutrophil

count, 1 9 1034.5 9 103/mm3 for lymphocytic count,

010 mm/h for erythrosedimentation rate (ESR), and less

than 0.5 mg/dL for C-reactive protein (CRP).

Descriptive data are reported as median, range, and

percentages, as appropriate. Data were recorded from all

patients unless otherwise specified. Following parameters

were analyzed in order to identify potential risk factors for

life-threatening complications: gender, age, body temper-

ature, WBC, diabetes mellitus, character of infection,

multiple space involvement. A multivariate logistic

regression analysis was undertaken using a forward step-

wise technique, in which including significant risk factors

in univariate analysis, in order to identify independent risk

factors for complications. Statistical analysis was per-

formed using the SPSS/PC software package (SPSS Inc.,

Chicago, IL, USA).

Results

Demographic and clinical data

A total of 365 adult patients with DNI were identified for

this evaluation. The 365 patients consisted of 205 males

(56.2%) and 160 females (43.8%) ranging in age from 18

to 96 years (median 52).

Patients were symptomatic for a median of 5.5 days

prior to admission to our institution ranged from 1 to

22 days.

On admission neck swelling (n = 340; 93.2%), throat

pain (n = 205; 56.2%), and dysphagia (n = 201; 55.1%)

were the most common symptoms. Other symptoms and

signs included fever (n = 257; 70.4%), swelling of the

upper aero-digestive tract (n = 218; 59.7%), dyspnea

(n = 54; 14.8%), neck stiffness (n = 54; 14.8%), trismus

(n = 51; 14.0%), dysphonia (n = 50; 13.7%), and otalgia

(n = 19; 5.2%).

The total WBC count (median 11.8 9 103/mm3; range

2.533.6 9 103/mm3) was increased above the upper limit

of normal in 171 cases (46.8%), normal in 192 cases

(52.6%), and under the lower limit of normal in two cases.

Neutrophil count (median 8.7 9 103/mm3, range

1.126.9 9 103/mm3, rate of unknown data 16.4%) was

increased above the upper limit of normal in 172 cases

(56.4%), normal in 129 cases (42.3%), and under the lower

limit of normal in 4 cases. Lymphocytic count (median

1.4 9 103/mm3, range 0.16.2 9 103/mm3, rate of

unknown data 16.4%) was decreased under the lower limit

of normal in 76 cases (24.9%), normal in 224 cases

(73.4%), and above the upper limit of normal in 5 cases.

ESR (median 50 mm/h, range 2140 mm/h, rate of

unknown data 26.8%) and CRP concentration (median

17.5 mg/dL, range 245 mg/dL, rate of unknown data

56.2%) were elevated above the upper limit of normal in

260 cases (97.4%) and in all cases, respectively.

Comorbidity

Eighty-two patients (22.5%) had relevant associated sys-

temic disorders including cardiovascular diseases (n = 53),

diabetes mellitus (n = 52), pulmonary diseases (n =

14), liver diseases (n = 13), hematological diseases

(n = 13), renal diseases (n = 5), connective tissue diseases

(n = 3).

1242 Eur Arch Otorhinolaryngol (2012) 269:12411249

123

Diagnostic investigations

All patients underwent otolaryngological examination with

fiber-optic, b-mode ultrasonography of the neck and/or

contrast-enhanced computed tomography (CECT)/mag-

netic resonance imaging (MRI) of the neck. CECT and

MRI were performed in 321 (87.9%) and 23 (6.3%)

patients, respectively. 3-mm slides from skull base to the

superior mediastinum were obtained before and after con-

trast injection using either the spiral or multi-slice tech-

nique. The CECT scan was interpreted as demonstrating an

abscess in presence of the enhancing rim around non-

enhancing central density consistent with fluid. The initial

CECT scan was extended to include the chest in cases of

suspected descending infection. Acquisition of high-reso-

lution axial scans of the jaw together with curved and or-

thoradial multiplanar reconstructions was performed in

patients with submandibular space infections and/or sus-

pected odontogenic infection. Follow-up CECT was the

diagnostic procedure of choice to evaluate response to

medical and/or surgical treatment and was performed in

286 cases (78.3%). Overall, the median number of imaging

examinations was two per patient (range 19). No signifi-

cant differences were found in number of imaging proce-

dures between patients who were immediately operated

and in patients selected for observation (P = 0.670). On

the other hand, a higher number of imaging procedures,

particularly CECT, was performed in patients developing

complications (median 4, range 39).

Source, site, and character of DNIs

The source of infection was identified in 297 patients

(81.4%): the most common cause was a pharyngitis

(n = 119; 32.6%), followed by dental infection (n = 102;

27.9%), submandibular sialadenitis (n = 39; 10.7%), par-

otitis (n = 23; 6.3%), cervical lymphadenitis (n = 7;

1.9%), otitis (n = 4; 1.1%), epiglottitis (n = 2; 0.5%). One

patient developed deep neck abscess with descending

mediastinitis secondary to cervical intravenous drug abuse.

The pathogenesis of DNI was not determined in 69 patients

(18.9%).

According to the source of infection, the most common

primary site of DNI was submandibular space followed by

parapharyngeal space (Table 1). In 191 cases (52.3%), a

multiple space involvement was observed. An abscess was

present in 213 patients (58.4%), a cellulitis in 152 patients

(41.6%).

Microbiology

Microbiological analysis included aerobic anaerobic

cultures and were performed from blood samples, material

obtained from the primary source of infection, the neck or

the mediastinum, using either a sterile swab or suction trap.

Microbiological diagnosis (Table 2) was successful in

177 patients (48.5%); 15.8% positive cultures were

polymicrobial.

Table 1 Site and character of deep neck infections

Involved spaces No. of patients (%)

(N = 365)No.

cellulitis

No.

abscess

Submandibular 220 (60.3) 108 111

Parapharyngeal 211 (57.8) 88 123

Parotid 48 (13.1) 8 40

Retropharyngeal 36 (9.9) 10 26

Visceral anterior 29 (7.9) 7 22

Visceral vascular 12 (3.3) 0 6

Masticatory 11 (3.0) 3 8

Prevertebral 9 (2.4) 1 8

Temporal 3 (0.8) 3 0

Table 2 Isolated pathogens from 177 patients with deep neckinfections

No.

Aerobic/facultatives

Streptococcus viridans not typed 37

Coagulase-negative staphylococcus 33

Staphylococcus aureus 23

Klebsiella pneumoniae 18

Staphylococcus epidermidis 11

Haemophilus influenzae 11

Streptococcus pneumoniae 6

Streptococcus, b-hemolytic, group A 5

Streptococcus constellatus 5

Proteus mirabilis 4

Streptococcus group F 3

Pseudomonas aeruginosa 3

Acinectobacter baumanii 2

Gemella morbillorum 2

Stenotrophomonas maltophilia 1

Streptococcus oralis 1

Anaerobic

Bacteroides spp 19

Peptostreptococcus spp 15

Fusobacterium spp 5

Prevotella melaninogenica 4

Propionibacterium acnes 2

Veillonella spp 2

Others

Candida spp 4

Aspergillus spp 2

Eur Arch Otorhinolaryngol (2012) 269:12411249 1243

123

Treatment

All patients received empirical broad-spectrum intravenous

antimicrobial therapy on admission in order to eradicate

both aerobic and anaerobic microorganisms. The first-line

therapy was later modified according to microbiological

findings if the isolated microorganisms revealed resis-

tance towards the empiric therapy. The most frequently

provided treatment regimens, alone or in combination,

were amoxicillin/clavulanate potassium (58.9%), second-

and third-generation cephalosporins (37.3%), ampicilline/

sulbactam (12.9%), clindamicyn (11.4%), metronidazole

(3.6%), and vancomycin (2.4%).

Patients who were clinically unstable (airway obstruction,

signs and symptoms of sepsis); patients with descending

infection; patients with anterior visceral space involvement,

with abscess involving more than two deep neck spaces; and

patients with abscess larger than 3.0 cm, underwent imme-

diate surgical drainage. Gas-forming infections were not in

itself an absolute indication for immediate surgery unless

large amount of tissue were involved. In all the other cases,

patients were observed for 48 h. If the patients symptoms

and signs worsened or if no clinical improvement was noted

after 48 h, surgical drainage was performed. On the other

hand, if clinical response was seen, a radiographic study was

repeated to confirm clinical judgment. If the repeat imaging

did not confirm a regression of collection of pus, surgical

intervention was anyway considered. In selected cases,

therapeutic needle aspiration of abscess was considered an

alternative to conventional open surgery.

One-hundred and thirty-nine patients (38.1%) under-

went surgical drainage. Of the abscess group (n = 213),

111 patients (52.1%) underwent surgical drainage. Of the

cellulitis group (n = 152), 28 patients (18.4%) underwent

surgical drainage.

In 112 cases (30.7%), an open surgical drainage was

performed under general anesthesia. An exclusively

transoral approach was used in 21 cases. An external or

combined approach was necessary in 91 patients. In all

cases, a wide exposure of the abscess cavity was performed

including blunt avulsion of any loculations, the devitalized

tissue was debrided, and the wound was irrigated with half-

strength hydrogen peroxide. In patients with extensive

tissue necrosis, the cervical incision was packed with plain

gauze and left open to allow oxygenation of the tissue and

daily irrigations with antiseptic solutions. In other cases,

wounds were closed after placement of large-bore drains

for irrigation. Twenty-seven patients (7.4%) underwent

needle aspiration of abscess, with CT-scan guidance in five

cases. Intraoperative findings confirmed the CECT diag-

nosis of abscess in 87.1%. Duration of symptoms (\5 daysvs. C5 days) was not found to be predictive of necessity of

surgical drainage (P = 0.566).

Overall, 226 patients (61.9%) responded effectively to

intravenous antimicrobial therapy only.

Sixty-five patients (17.8%) underwent tooth extraction.

On discharge, tonsillectomy was proposed to all patients

treated for DNI secondary to pharyngotonsillitis.

Complications

There were 67 patients (18.4%) developing life-threatening

complications (Table 3). Forty-three were men (64.2%)

and 24 were women (35.8%) with a median age of 59 years

(range 1889 years). Diabetes mellitus occurred in 27

patients (40.3%). An abscess was present in 54 patients

(80.6%) and a multiple-space involvement was diagnosed

in 52 cases (77.6%).

Sixteen patients (4.4%) developed descending necrotiz-

ing mediastinitis with a median of 6 days (range 312 days)

after onset of first symptoms of cervical infection. Most

common symptoms and signs included neck and/or upper

aero-digestive tract swelling (n = 16), dysphagia (n = 10),

throat pain (n = 11), neck stiffness (n = 5). Acute onset of

dyspnea and thoracic pain were seen in three and four

patients, respectively. Neck swelling was the only clinical

finding in five patients. In most cases (n = 10), the diag-

nosis of mediastinitis was made on the basis of CECT

findings in absence of clinical signs of mediastinum

involvement. Twelve patients underwent external drainage

of the cervical abscess in conjunction with posterolateral

thoracotomy, four patients with infection limited to the

upper mediastinal spaces above the tracheal carina under-

went transcervical thoracic drainage. Among patients with

descending mediastinitis, a microbiological diagnosis was

obtained from 10 patients (62.5%). A polymicrobial infec-

tion was identified in six patients. The isolated aerobic

bacteria were Streptococcus spp (n = 5), Coagulase-nega-

tive staphylococcus (n = 3), Acinectobacter baumanii

(n = 1), Gemella morbillorum (n = 2), Stenotrophomonas

maltophilia (n = 1), and Klebsiella pneumoniae (n = 1).

Bacteroides spp (n = 3), Fusobacterium spp (n = 2),

Table 3 Life-threatening complications

Complications No. of

patients (%)

(N = 365)

No.

deaths

Airway obstruction 31 (8.5) 0

Sepsis 22 (6.0) 1

Descending mediastinitis 16 (4.4) 0

Pneumonia 12 (3.3) 0

Jugular vein thrombosis septic embolism 11 (3.0) 0

Pleural effusion 4 (1.1) 0

Disseminated intravascular coagulation 1 (0.3) 1


123

Peptostreptococcus spp (n = 2), and Veillonella spp

(n = 1) were detected in anaerobic cultures.

Among patients with critical airway narrowing, 15

patients (48.4%) had a bilateral diffuse gangrenous cellu-

litis of the submandibular and sublingual spaces (Lud-

wigs angina) with or without extension to the visceral

anterior space. In order to resolve airway obstruction, 8

patients underwent emergency tracheostomy, 17 patients

underwent fiberoptic guided awake endotracheal intuba-

tion, and 6 patients achieve relief after successful treatment

by intravenous corticosteroids.

Eleven patients developing visceral vascular space

infection with jugular vein thrombosis secondary to oro-

pharyngeal infection, underwent long-term antibiotic ther-

apy in association with anticoagulant therapy (enoxaparin

100 IU/kg twice daily for 3 months). All patients with

jugular vein thrombosis complained of neck stiffness.

Positive blood cultures for Fusobacterium spp. and Strep-

tococcus constellatus were documented in four and one

patient, respectively. Septic embolization to multiple sites

(lung, liver, spleen, and joints) was observed in five cases

(Lemierres syndrome). Two patients with associated pa-

rapharyngeal abscessone of those with concomitant

descending mediastinitisunderwent surgical drainage of

the pus collection. No patients underwent ligation of the

internal jugular vein. Less severe complications account for

osteomyelitis (n = 2) and vocal cord palsy (n = 2).

Predictors of complications

In univariate logistic-regression analysis, we assessed fac-

tors associated with life-threatening complications. The

strongest predictor of life-threatening complications was

diabetes mellitus [odd ratio 7.37 (95% CI 3.9013.94);

P \ 0.001]. Other variables significantly associated withcomplications are shown in Table 4. Factors that were

independently associated with life-threatening complica-

tions on the basis of a multinomial regression model, are

shown in Table 5: diabetes mellitus [odd ratio 5.43 (95%

CI 2.5611.53); P \ 0.001] and multiple deep neck spaces

involvement [odd ratio 4.92 (95% CI 2.3810.16);

P \ 0.001] were the strongest independent predictors ofcomplications.

Outcome

One patient with diabetes mellitus and liver dysfunction

who have developed sepsis and disseminated intravascular

coagulation died from severe hepatic insufficiency 16 days

after successful drainage of bilateral submandibular

abscess with extension to the anterior visceral space. All

other patients were discharged in stable condition after a

median length of inpatient stay of 11 days (range

673 days).

Discussion

This is the largest series of DNIs reported in Western lit-

erature. Although DNIs can affect all age-group, most of

cases in the present series were concentrated between the

fifth and seventh decade of life. About one-fourth of

patients have relevant associated comorbidities with dia-

betes mellitus being the most frequent. Diabetes mellitus is

commonly reported in patients with DNIs [5]. Several

authors have identified diabetes mellitus as a significant

risk factor for infection-related morbidity and mortality

[5, 6]. The results of our study confirmed diabetes mellitus

as the strongest independent predictors of complications.

Peripheral vascular disease in diabetics may predispose

patients to anaerobic infection [7]. Furthermore, patients

with a hyperglycemic state have functional leukocyte,

macrophage, and fibroblast impairments that increase their

susceptibility to serious infections [8, 9]. Therefore, opti-

mal control of diabetes mellitus play a critical role in DNIs

management: insulin use is the best option due to flexibility

of timing and dose.

Most DNIs are mixed polymicrobial infections includ-

ing aerobes and anaerobes. Bacteroides fragilis, Prevotella,

Porphyromonas, and Fusobacterium spp resist penicillin

through the production of beta-lactamase. Overall, more

Table 4 Univariate associations with life-threatening complications

Factor Odds ratio

(95% CI)

P value

Sex: male vs. Female 1.50 (0.872.60) 0.145

Age: per 10-year increase 1.20 (1.041.39) 0.012

Body temperature: per increase of 1C 1.41 (1.101.80) 0.006White blood cell count: per increase

of 1 9 103/mm31.11 (1.051.17) \0.001

Diabetes mellitus 7.37 (3.9013.94) \0.001Evidence of colliquation 3.63 (1.906.93) \0.001Multiple space involvement 3.96 (2.147.36) \0.001

Table 5 Factors associated with life-threatening complications in thestepwise multivariate model

Factor Odds ratio

(95% CI)

P value

Body temperature: per increase of 1C 1.49 (1.092.04) 0.012White blood cell count: per increase

of 1 9 103/mm31.09 (1.021.17) 0.005

Diabetes mellitus 5.43 (2.5511.53) \0.001Evidence of colliquation 2.51 (1.225.15) 0.012

Multiple space involvement 4.92 (2.3810.16) \0.001


123

than two thirds of DNIs contain beta-lactamase-producing

microorganisms. The low tissue oxygen tension in the

loose areolar tissue of the cervical spaces favor the syn-

ergistic growth of aerobic and anaerobic bacteria. Strep-

tococcus spp and Bacteroides spp were the most prevalent

microorganism in aerobic and anaerobic bacterial cultures,

respectively, reflecting the predominant pharyngeal source

of DNIs in the present series. No bacterial growth was

recorded in 188 patients and anaerobes were isolated in

minority of cases. Use of antibiotics before admission,

high-dosage intravenous empiric antibiotic therapy prior to

surgical drainage, improper collection of specimen, no

routine use of anaerobic cultures, and difficult in culturing

anaerobes may affect and may have affected the result of

microbiological tests in this series. Increase in the inci-

dence of anaerobic bacteremias with multiple-drug-resis-

tant organisms is emerging as a significant health problem

as there is an increasing population with multiple comor-

bidities and compromised immune system [10]. Anaerobes

express significant virulence factors including adherence

and spreading factors as hyaluronidase, collagenase, and

fibrolysin that may promote the dissemination of a local-

ized infection [7]. Anaerobes also have the ability to pro-

duce the enzyme beta-lactamase protecting themselves and

other penicillin-susceptible organisms from the activity of

penicillins [11].

Therefore, all efforts should be directed to maximize

successful isolation of anaerobes. In order to increase the

chances of effective microbiological diagnosis, the speci-

men for anaerobic cultures should be an aspirate obtained

by needle and syringe, transferred into anaerobic culturette,

avoiding exposure to oxygen, and transported to the labo-

ratory within 23 h [7]. Tissue samples and biopsies placed

in a sterile container are also adequate specimens for

anaerobic cultures. The high rate of coagulase-negative

Staphylococcus positive culture may reflect the collection

of contaminated specimens. In this sense, when the mate-

rial for microbiological cultures is transmucosally col-

lected, it is essential to decontaminate the mucous

membrane. Although no methicillin-resistant strains were

identified, as community-associated methicillin-resistant

Staphylococcus aureus (MRSA) isolation is increasingly

common among out- and inpatients with suppurative

infections, MRSA may play an increasing role in DNIs in

the next future [12, 13].

CECT was the modality of choice in the evaluation of

DNIs. Taken into account that trismus may significantly

limit an accurate inspection of the upper aerodigestive tract

and that clinical examination may underestimate the extent

of infection in about two-third of cases, CECT plays a

critical role in confirming the clinical suspect of DNIs, in

the differentiation of deep neck abscesses from cellulitis, in

the delineation of the involved spaces, in the diagnosis of

complications, such as descending necrotizing mediastini-

tis and internal jugular vein trombosis, and in monitoring

the evolution of the infection [14]. Although CECT scan

has a good sensitivity in detecting infection and delineating

the cervical spaces involved, its accuracy is lower in dif-

ferentiating abscess from cellulitis [15, 16]. A single or

multiloculated low density area with a complete circum-

ferential rim of enhancement, surrounded by soft tissue

swelling, is considered the hallmark of abscess. Also, the

presence of an air-fluid level and subcutaneous air are

findings suggesting an abscess formation [17]. Deep neck

cellulitis presents as a mass with low-density core and

surrounding edema without enhancing rim or air-fluid level

[16]. On the other hand, lymphadenitis presents as a soft

tissue swelling obliterating adjacent fat planes. It is lapal-

issian that, as the diagnosis of deep neck abscess is based

on subjective findings, the accuracy of CECT is dependent

upon the experience of the radiologist and may be con-

siderably lower in the transition stages from cellulitis to

abscess. In the present series, intraoperative findings did

not confirm the CT diagnosis of abscess in 13%. It has been

reported that pus may not be intraoperatively found in up to

one-fourth of cases with CECT scans suggestive of deep

neck abscess [18]. A scalloped contour of the ring-

enhancement, was more recently found to have a positive

predictive value of 94% in predicting the presence of pus

[19]. In order to identify periapical infections in patients

with suspected odontogenic DNIs, acquisition of high-

resolution axial scans of the jaw together with curved and

orthoradial multiplanar reconstructions are desirable [20].

On CECT, internal jugular vein thrombosis appears as an

enlarged vein with a low-density lumen surrounded by a

sharply defined wall [21] (Figs. 1, 2). In patients with

descending mediastinitis (Figs. 3, 4), CECT may show

fluid collection with gas formations, soft tissue thickening

and enhancement with loss of the normal fat planes, pleural

or pericardial effusion [22]. As descending mediastinitis

may be clinically silent [22], we suggest to routinely

extend the CT scans to the superior mediastinum in all

cases of DNI.

The mainstay of treatment of DNIs consists of airway

control, effective antibiotic therapy, and, when appropriate,

surgical incision and drainage of the pus collection.

The maintenance of a secure airway, a challenging task

both for surgeon and anesthesiologist, is the first step in

the treatment of patients with DNIs and airway compro-

mise. Upper airway obstruction may result from laryngeal

edema secondary to anterior visceral space involvement

or tongue pushing against the roof of the mouth and the

posterior pharyngeal wall secondary to extensive sub-

mandibular space infection. In the present series about

half of patients with critical airway were affected by

Ludwigs angina, a potentially life-threatening bilateral


123

diffuse gangrenous cellulitis of the submandibular and

sublingual spaces (Fig. 5). In these patients, who are not

rarely diabetic, conventional endotracheal intubation and

tracheotomy under general anesthesia may be made even

more difficult by morbid obesity. In our experience,

fiberoptic guided awake endotracheal intubation is an

appropriate procedure both allowing a safe and atraumatic

intubation in cooperative patients and enabling the sur-

geon to explore an anatomically distorted upper aerodi-

gestive tract [23].

On the basis of the above considerations, empirical

antibiotic therapy with a combination of a penicillin plus a

beta-lactamase inhibitor (amoxicillin/clavulanate, ticarcil-

lin/clavulanate, piperacillin/tazobactam), cefoxitin, carba-

penem, or clindamycin should provide sufficient coverage

for both anaerobic and aerobic bacteria. Metronidazole has

excellent activity only against strict anaerobic bacteria and

therefore is poorly effective as a single-agent in DNIs [7].

Fig. 1 CECT findings of a deep neck spaces abscess with left jugularvein trombosis

Fig. 2 Angio-MRI showing the absence of venous drainage from theleft internal jugular vein

Fig. 3 Necrotizing descending mediastinitis: histological sectionshowing agglomerates of neutrophil cells and bacteria (Streptococcusoralis) in a contest of muscular necrosis

Fig. 4 CECT findings of a deep neck space abscess descending in themediastinum


123

Macrolides or ketolides plus metronidazole should be

considered in patients with a penicillin allergy. Clinda-

mycin resistance among strains of Bacteroides fragilis has

increased over 10 years, and current resistance rates reach

2050% or more worldwide [24]. Take into account that in

the present and other series [22] Bacteroides spp were

among the most frequently isolated anaerobic pathogens

both in uncomplicated and complicated DNIs, clindamycin

may no longer be considered a first-line antibiotic in DNIs.

First intention antibiotic therapy should be reviewed 48 h

later and potentially adjusted according to the microbio-

logical- and drug-resistance patterns. A prolonged antibi-

otic therapy should be advisable as anaerobic infections are

frequently chronic. After resolution of clinical signs of

DNIs, oral therapy can replace parenteral one.

Open surgical incision and drainage are considered the

mainstay of treatment for deep neck abscesses. Almost

two-third of the patients responded satisfactorily to medical

therapy only. We and several authors have demonstrated

previously that a trial of intravenous antibiotic treatment

associated with an aggressive CECT-based wait-and-watch

policy may result in a significant number of selected

patients (patients with cellulitis, abscesses \3 cm notinvolving danger spaces or more than one space, stable

general condition) avoiding an unnecessary surgical

drainage [15, 17, 25].

This policy did not result in significantly higher number

of imaging procedures in patients selected for observations

mainly because imaging investigations were routinely

performed also after surgical drainage of deep neck abscess

in order to confirm the resolution of the infections.

In the present series, about one-fourth of patients

required an extensive external cervical approach. This

approach is mandatory for drainage of large abscesses,

multiple space abscesses, and impending complications. In

patients with Ludwigs angina, an external surgical

approach is justified, even if areas of colliquation are not

usually evident [6].

No correlation was found between duration of symptoms

and the necessity of surgical drainage. Considering that a

short duration of symptoms may correlate with a more

aggressive infection, one should have expected a higher

prevalence of surgical drainage in these patients. On the

other hand, an inappropriate antibiotic therapy and anaerobic

infections may be responsible for a slow course of disease

and longer duration of symptoms in patients who finally

develop complications requiring surgical procedures.

About 7.5% of patients were successfully drained by

needle aspiration. Minimally invasive techniques are

attractive options in patients with well-defined, unilocular

abscess without airway compromise. Draining an abscess by

needle aspiration reduces the morbidity of open surgery by

limiting surgical trauma, reducing healing time, minimizing

the risk of contaminating the surrounding healthy tissue. CT

or ultrasound guidance may improve the efficacy and safety

of percutaneous abscess drainage. In selected retro- and

parapharyngeal abscesses without involvement of visceral

vascular space, endo-oral aspiration and/or incision should

be considered in order to reduce patient morbidity, economic

burden and avoid aesthetic complications.

About 18% of patients developed life-threatening com-

plications. Diabetes mellitus was confirmed to be the

strongest predictor of life-threatening complications [5].

Airway obstruction and spread of infection to the medias-

tinum are the most troublesome complications in patients

with deep neck space infections. In our study population,

most patients with mediastinitis had not shown any

symptoms and signs of mediastinum involvement with

symptoms of neck infection being common. Therefore,

prompt diagnosis of descending mediastinitis may be

missed in the absence of a high index of suspicion and

routine CECT through the mediastinum. On the basis of

our multivariate analysis, patients with diabetes mellitus,

multiple space involvement, evidence of colliquation, high

WBC, or high body temperature should be considered to

potentially have a descending mediastinitis until proven

otherwise. Descending necrotizing mediastinitis requires

an aggressive multidisciplinary management. Delay in

diagnosis as well as inadequate drainage of the mediasti-

num are considered to be the most significant factors

responsible for mortality [22]. Transcervical drainage of

the mediastinum should be reserved for patients with

infection limited to the upper mediastinal spaces above the

tracheal carina. On the other hand, cervicotomy along with

posterolateral thoracotomy incision is the standard of care

in patients with inferior mediastinum involvement.

Lemierres syndrome is an uncommon seen and often

forgotten complication of acute oropharyngitis affecting

healthy adolescents and young adults. Central to the

Fig. 5 A case of Ludwigs angina


123

pathogenesis of this disease is the internal jugular vein

thrombophlebitis. Septic metastases may occur and fre-

quently affect the lungs. Clinically, the onset of septic

symptoms often coincides with the end of oropharyngeal

symptoms. Septic fever, tension of the sternocleidomastoid

muscle and a stiff neck are the most frequent symptoms

plus those connected with the site of the secondary local-

izations (chest pain, dyspnea, hemoptysis, and more

uncommonly joint pains, abdominal pain with possible

acute abdomen) [26]. Broad-spectrum therapy should be

given for more than 3 weeks. On the other hand, the role of

anticoagulation has remained controversial [27]. Ligation

and resection of the internal jugular vein, which was fre-

quent in the pre-antibiotic era, is now recommended by

some authors only in the case of persistent sepsis with

embolism.

Conclusion

The availability of effective antibiotics and improved oral

hygiene have dramatically modified the epidemiology of

DNIs making them less common today than in the past.

However, even in this era of antibiotic therapy and modern

imaging techniques, DNIs remain a constant challenge.

Airway obstruction and descending mediastinitis are the

most troublesome complications of DNIs. In selected

patients, a trial of intravenous targeted or broad-spectrum

empiric antibiotic therapy associated with an intensive

CECT-based wait-and-watch policy may avoid an unnec-

essary surgical procedure. However, about one-fourth of

patients present significant comorbidities, which may

negatively affect the course of the infection. In these cases

and in patients with large or multiple spaces infections, a

more aggressive surgical strategy is mandatory.

Conflict of interest There are no potential conflicts of interest.

References

1. Daramola OO, Flanagan CE, Maisel RH et al (2009) Diagnosis

and treatment of deep neck space abscesses. Otolaryngol Head

Neck Surg 141:123130

2. Reynolds SC, Chow AW (2007) Life-threatening infections of the

peripharyngeal and deep fascial spaces of the head and neck.

Infect Dis Clin North Am 21:557576

3. Brook I (2004) Microbiology and management of peritonsillar,

retropharyngeal, and parapharyngeal abscesses. J Oral Maxillofac

Surg 62:15451550

4. Levitt GW (1970) Cervical fascia and deep neck infections.

Laryngoscope 80:409435

5. Lin HT, Tsai CS, Chen YL et al (2006) Influence of diabetes

mellitus on deep neck infection. J Laryngol Otol 120:650654

6. Boscolo-Rizzo P, Da Mosto MC (2009) Submandibular space

infection: a potentially lethal infection. Int J Infect Dis

13:327333

7. Brook I (2007) The role of anaerobic bacteria in upper respiratory

tract and other head and neck infections. Curr Infect Dis Rep

9:208217

8. Leibovici L, Yehezkelli Y, Porter A (1996) Influence of diabetes

mellitus and glycaemic control on the characteristics and out-

come of common infections. Diabetic Med 13:457463

9. Slovenkai MP (1998) Foot problems in diabetes. Med Clin North

Am 82:949971

10. Lassmann B, Gustafson DR, Wood CM et al (2007) Reemergence

of anaerobic bacteremia. Clin Infect Dis 44:895900

11. Brook I (2010) b-Lactamase-producing bacteria in upper respi-ratory tract infections. Curr Infect Dis Rep 12:110117

12. Moran GJ, Krishnadasan A, Gorwitz RJ et al (2006) Methicillin-

resistant S. aureus infections among patients in the emergencydepartment. N Engl J Med 355:666674

13. Popovich KJ, Weinstein RA, Hota B (2008) Are community-

associated methicillin-resistant Staphylococcus aureus (MRSA)strains replacing traditional nosocomial MRSA strains? Clin

Infect Dis 46:787794

14. Nagy M, Backstrom J (1999) Comparison of the sensitivity of

lateral neck radiographs and computed tomography scanning in

pediatric deep-neck infections. Laryngoscope 109:775779

15. Boscolo-Rizzo P, Marchiori C, Zanetti F et al (2006) Conserva-

tive management of deep neck abscesses in adults: the impor-

tance of CECT findings. Otolaryngol Head Neck Surg

135:894899

16. Elden LM, Grundfast KM, Vezina G (2001) Accuracy and use-

fulness of radiographic assessment of cervical neck infections in

children. J Otolaryngol 30:8289

17. Wang LF, Tai CF, Kuo WR et al (2010) Predisposing factors of

complicated deep neck infections: 12-year experience at a single

institution. J Otolaryngol Head Neck Surg 39:335341

18. Smith JL, Hsu JM, Chang J (2006) Predicting deep neck space

abscess using computed tomography. Am J Otolaryngol

27:244247

19. Kirse DJ, Roberson DW (2001) Surgical management of retro-

pharyngeal space infections in children. Laryngoscope

111:14131422

20. Gahleitner A, Watzek G, Imhof H (2003) Dental CT: imaging

technique, anatomy, and pathologic conditions of the jaws. Eur

Radiol 13:366376

21. Giyanani VL, Mirfakhraee M, Gerlock AJ et al (1985) Computed

tomography of internal jugular thrombosis. J Comput Assist

Tomogr 9:3337

22. Ridder GJ, Maier W, Kinzer S et al (2010) Descending necro-

tizing mediastinitis: contemporary trends in etiology, diagnosis,

management, and outcome. Ann Surg 251:528534

23. Allan AG (2004) Reluctance of anaesthetists to perform awake

intubation. Anaesthesia 59:413

24. Boyanova L, Kolarov R, Mitov I (2007) Antimicrobial resistance

and the management of anaerobic infections. Expert Rev Anti

Infect Ther 5:685701

25. Sichel JY, Dano I, Hocwald E et al (2002) Non surgical man-

agement of parapharyngeal space infections: a prospective study.

Laryngoscope 112:906910

26. Boscolo Rizzo P, Scotton PG, Da Mosto MC et al (2003) Sin-

drome di Lemierre da Streptococcus constellatus. Caso clinico e

rassegna della letteratura. Giorn It Mal Inf 9:167170

27. Syed MI, Baring D, Addidle M et al (2007) Lemierre syndrome:

two cases and a review. Laryngoscope 117:16051610


123

Deep neck infections: a study of 365 cases highlighting recommendations for management and treatmentAbstractIntroductionPatients and methodsResultsDemographic and clinical dataComorbidityDiagnostic investigationsSource, site, and character of DNIsMicrobiologyTreatmentComplicationsPredictors of complicationsOutcome

DiscussionConclusionConflict of interestReferences

Documents

10.1007_s00405-011-1761-1