Toxic inner ear lesion following otitis media with effusion: a comparative CT-study regarding the...

OTOLOGY

Toxic inner ear lesion following otitis media with effusion:a comparative CT-study regarding the morphologyof the inner ear

Thomas Wilhelm • Tim Stelzer • Susanne Wiegand •

Christian Guldner • Afshin Teymoortash •

Thomas Gunzel • Rudolf Hagen

Received: 8 July 2014 / Accepted: 28 November 2014

� Springer-Verlag Berlin Heidelberg 2014

Abstract Viral infections of the upper respiratory air-

ways can lead to a delayed viral otitis media (VOM) caused

by a diffusion of viruses/virus particles through the round

window membrane and resulting in sensorineural hearing

loss. The treatment of choice is immediate paracentesis,

evacuation of all fluids from the middle ear cavity, and

haemorrheological infusions. However, in some cases,

persistent symptoms may be an indication for a surgical

approach using mastoidectomy. In high-resolution com-

puted tomography, an extended small-sized pneumatisation

of the mastoid cells with complete shading was found in

these non-responsive cases. Therefore, a direct means of

inner ear affliction through weak parts of the labyrinthine

bone may be hypothesised. Patients suffering from a toxic

inner ear lesion (TIEL) following a common cold, treated

over a 10-year period in a Tertiary Care Centre (N = 52,

57 ears), were identified and the morphological character-

istics of the temporal bones of affected patients were

examined by means of high-resolution computed tomog-

raphy (hrCT). The findings were compared with a matched

control group of 64 normal ears (CONT). Measurements

included the grade of pneumatisation, distances within the

temporal bones and Hounsfield units (HU) at defined

anatomical structures. In the TIEL group, we found a

small-sized pneumatisation in 79.4 % and a medium-sized

pneumatisation in 10.9 %, thus differing from the CONT

group and the literature data. Thickness of the bone wall of

the lateral semicircular canal (LSC) and distances within

the aditus ad antrum were significantly reduced in the TIEL

group. HU’s were markedly lower in the TIEL group at the

precochlea, the LSC, and dorsolateral to the promentia of

the LSC. There was a correlation between the HU’s at the

prominentia of the LSC and the hearing loss (p = 0.002).

Persisting interosseous globuli, as described in 1897 by

Paul Manasse, form an osseochondral network within the

otic capsule and may be responsible for a direct means of

toxic inner ear infection. The CT-morphometric results

support this thesis. In the group of these patients (TIEL) a

CT-scan and in non-responders to conservative treatment a

surgical approach by mastoidectomy is recommended.

Keywords Toxic lesion � Inner ear � Subacute

mastoiditis � Otitis media with effusion � Computed

tomography � Sensorineural hearing loss

Introduction

Viral otitis media in most cases is an acute disease of the

middle ear mucosa caused by ascending infection from the

Thomas Wilhelm and Tim Stelzer are both first authors since idea,

conception, and execution of this project derive from both of them

alike.

T. Wilhelm (&) � T. Stelzer

Department of Otolaryngology, Head, Neck and Facial Plastic

Surgery, Kliniken Leipziger Land, Klinikum Borna,

Rudolf-Virchow-Str. 2, 04552 Borna, Germany

e-mail: thomas.wilhelm@kliniken-leipziger-land.de

URL: http://www.kliniken-leipziger-land.de

S. Wiegand � C. Guldner � A. Teymoortash

Department of Otolaryngology, Head and Neck Surgery,

Philipps-University, Marburg, Germany

T. Gunzel

Department of Otolaryngology, Head and Neck Surgery,

Borromaus Hospital, Leer, Germany

R. Hagen

Department of Otolaryngology and Facial Plastic Surgery,

Julius-Maximilians-University, Wurzburg, Germany

123

Eur Arch Otorhinolaryngol

DOI 10.1007/s00405-014-3425-4

nasopharynx. Clinically, patients suffer from ear ache and

hearing loss and otoscopic findings show an inflamed

tympanic membrane with effusions in the middle ear, in

some cases a typical bullous myringitis leading to direct

diagnosis of a viral origin of the disease [1, 2]. A high virus

concentration within the middle ear may lead to a con-

comitant sensorineural hearing loss which, in up to 60 % of

the cases, is caused by penetration of viruses through the

round window membrane [3, 4]. The incidence of viral

otitis is estimated at between 0.5 and 2 % in all common

colds in adults [5].

However, in rare cases middle ear symptoms occur with

some delay after acute rhinosinusitis. Patients report a

‘‘sudden’’ hearing loss up to 4 weeks after a respiratory

tract infection. Otoscopic findings demonstrate middle ear

effusions without acute bacterial infection and audiological

tests reveal a combined sensorineural and conductive

hearing loss. The treatment of choice is a paracentesis with

suction of the middle ear fluid and, in the course of this

particular study period between 2001 and 2011, a so-called

‘‘haemorrheological’’ regime similar to that used in cases

with sudden hearing loss of unknown origin [6] to improve

inner ear function. In cases with persistent symptoms and

missing improvement of inner ear function, computed

tomography demonstrates the involvement of the antrum

and the perilabyrinthine spaces (Fig. 1a).

The commonly accepted path of infection of the inner

ear in viral otitis media is a diffusion of viruses or toxic

products through the round window membrane or the

annular ligament in the oval niche [4]. In our case series,

the removal of middle ear effusions by paracentesis did not

result in a successful recovery of the inner ear function in

all cases, subsequently leading to the decision to improve

the healing process by removing diseased mucosa via

mastoidectomy (Fig. 1b, c).

The purpose of the present study is the evaluation of

the radiological findings in computed tomography of the

temporal bones in a case series of toxic inner ear lesion

following otitis media with effusion due to viral infec-

tions of the upper airways to identify a typical config-

uration of periantral and perilabyrinthine structures,

which may be responsible for the persistent toxic inner

ear lesion.

Materials and methods

Since 2001, patients suffering from a toxic inner ear lesion

(TIEL) with subacute mastoiditis following a viral infec-

tion of the upper airways 3–4 weeks prior to the hearing

impairment have been treated according to the diagnostic

and therapeutic workflow outlined in Fig. 2. All of these

cases of ‘‘toxic inner ear lesion’’ were included and

evaluated with respect to the hearing outcome, following

conservative or surgical therapy in a Tertiary Care Centre

in a retrospective chart review. Identification of cases

started with a search within the clinical information system

according to the ICD (International Classification of Dis-

eases) classes H83.0 labyrinthitis, H91.0 ototoxic hearing

loss, H65.0/1 acute otitis media with effusion, H66.9 acute

otitis media, H73.0 acute myringitis, J11.8 viral otitis, and

H70.0 acute mastoiditis. Identified cases were checked for

the following inclusion criteria:

Fig. 1 a High-resolution computed tomography of the temporal

bones shows shading of the mastoid cells around the labyrinth.

b Intraoperative findings. c Surgical situs after complete exenteration

of all cells of the temporal bone

Eur Arch Otorhinolaryngol

123

1. Otitis media with effusion following a viral upper

respiratory tract infection with pathological findings in

tympanometry.

2. Combined conductive and sensorineural hearing loss in

the initial audiometric testing.

3. At least one additional audiometric examination during

the treatment period.

4. Either conservative or surgical treatment.

5. No prior history of hearing impairment.

6. No acute bacterial infection (earache, enflamed and

protruding ear drum, purulent secretion by spontaneous

perforation) of the middle ear or the eardrum.

All patients were followed up and re-examined via pure

tone audiometry. Vestibular testing following caloric

stimulation was performed by means of electronystag-

mography in patients reporting dizziness. The directional

and side preponderance were calculated according to the

formula of Jongkees et al. [7]; directional preponderance

exceeding 25 % and side preponderance [15 % were

judged as ‘‘pathological finding’’.

A high-resolution computed tomography (CT) of the

temporal bones was performed on all patients within the

first 5 days. All patients demonstrated signs of an

involvement of the mastoidal mucosa in the CT (shading

and reduced opacity without bone destruction).

A matching group of patients (CONT) from the Picture

Archiving and Communication System of the clinic (PACS

Centricity Enterprise Web, General Electric Medical Sys-

tems, Fairfield/Connecticut, US) was selected for the pur-

pose of comparing morphological features of the temporal

bone configuration. Under the search term ‘‘high-resolution

temporal bone computed tomography’’ a comparable group

with the following exclusion criteria was established: acute

or chronic mastoiditis, previous middle ear or mastoid

surgery, chronic otitis media with meso- or epitympanal

tympanic membrane defect, fracture of the temporal bone.

Indications for the CT-scans in the CONT group were:

exclusion of temporal bone fractures (N = 15), exclusion

of neurological disorders in cases of paresis or tumour

(N = 7), diagnostic workup in cochlear disorders or diz-

ziness of unknown origin (N = 5) or navigational CT-

scans prior to sinus surgery or microvascular decompres-

sion according to JANNETTA (N = 5).

All CT-scans were available as digital scans within the

PACS system from 2002. CT-scans were performed with a

HiSpeed NX/i dual slice CT-scanner and later on with a

BrightSpeed scanner from General Electric Company. The

resolution was 512 9 521 pixels with a pixel size of 0.332

and 0.488 mm, respectively. All CT-scans were done

without contrast medium.

The extent of pneumatisation was rated according to the

recommendations of Allam and Schuknecht [8, 9]: mas-

toidal cell diameters below 1.2 mm were rated as small

cells, diameters between 1.2 and 2.5 mm as medium sized,

and all cells with a diameter of more than 2.5 mm as large

cells. The pneumatisation grade was determined for dif-

ferent regions (Fig. 3a) and in each area the smallest cell

diameters were used to determine the classification as a

small-, medium- or large-sized pneumatisation grade.

A second measurement included different distances

within the temporal bone (Fig. 4). The following distances

were evaluated in both the groups (TIEL and CONT):

1. Distance from the suprameatal spine (Henles spine) to

the sigmoid sinus (SpiSin),

2. Thickness of the prominentia of the lateral semicircu-

lar canal (LSC) at the level of the short process of the

incus (PromLat1),

3. Thickness of the labyrinthine bone at a 90� angle

dorsolateral to PromLat1 (PromLat2),

4. Distance from the tympanic sulcus (fibrocartilaginous

sulcus) to the posterior cranial fossa (SulFos),

5. Distances within the aditus ad antrum: medial isthmus

from the prominentia of the LSC to the short process of

the incus (IsthMed),

Fig. 2 Diagnostic and

therapeutic workflow

Eur Arch Otorhinolaryngol

123

6. Distances within the aditus ad antrum: lateral isthmus

from the short process of the incus to the lateral wall of

the epitympanic recess (IsthLat),

7. Shortest distance from the LSC to the external auditory

canal (PromCan).

To get an appraisal of the density of the temporal bone

in different anatomical structures (Fig. 3b) and to compare

these values to previously published data [10, 11], the

HOUNSFIELD units (HU) were recorded additionally at

the following points: (1) fissula ante fenestram (FAF), (2)

cochlear apex (AC), (3) Precochlea (PC), (4) anterior edge

of the inner auditory canal (ACAI), (5) posterior wall of the

inner auditory canal (DCAI), (6) medial to the vestibular

vestibule (SAC), (7) dorsal of the confluence of the crus

commune into the vestibular vestibule (SCC-LP), (8) pos-

terior border of the LSC (PromLat3), (9) at the prominentia

of the LSC at the level of the short process of the incus

(PromLat1), and (10) at the confluence of the lateral and

the posterior semicircular canal (PromLat2).

Fig. 3 a Areas for determination of the cellular pneumatisation of the

temporal bone. b Measurement points of the Hounsfield units (for

detailed descriptions please refer to the ‘‘Materials and methods’’

section)

Fig. 4 Distances measured within the hrCT-scans of the study and

control group (for details please refer to the ‘‘Materials and methods’’

section)

Eur Arch Otorhinolaryngol

123

In the first step, all HU were measured at the described

points: the highest value was chosen for calculation of the

average. In cases of a marked difference between the

averages (PromLat1, PromLat2, and PromLat3), an area

measurement (circular area with at least 5 pixel but without

measuring the outlying areas like mastoidal cells or peri-

and endolymphatic spaces) was performed for calculation

purposes.

Approval for this retrospective study was obtained from

the Ethical Committee of the Saxonian Chamber of Phy-

sicians (22.06.2011, Sign: Prof.Dr.H/Di). The study has

been performed in accordance with the ethical standards

prescribed in the 1964 Declaration of Helsinki in line with

its latest revision. All patients gave their informed consent

to scrutiny of their medical files prior to their inclusion in

the study. Data handling was carried out in accordance

with local data protection laws.

Statistical methods and data analysis

The data were documented in Microsoft Excel 2003�(Redmont, Washington, USA) and evaluated statistically

with SPSS� 14 (Armonk, New York, USA). Standard

deviation and mean variation were calculated for

descriptive statistical means. Data distribution was

checked for normal distribution with the Kolmogorov–

Smirnov test: a p value [0.05 constituted normal distri-

bution. In the case of rejection of the hypothesis of normal

distribution, paired samples were tested using the Wilco-

xon signed-rank test and unpaired samples were tested by

means of a Mann–Whitney test (MWT). Only the com-

parisons of the overall change in hearing level at the

follow-up between the two treatment groups were per-

formed in a confirmatory sense (MWT, significance level

was set to 0.05). A classical Bonferroni transformation

was applied for the purpose of multiple testing: p value

was set at 0.005 for the comparison of the HU (punctual

measurement) and at 0.025 (area measurement); p value

was set at 0.00714 for the distance measurements. The

correlation according to Kendall and Pearson was calcu-

lated to correlate morphological abnormalities with clini-

cal symptoms.

Results

52 patients, including 5 patients affected bilaterally (57

ears), fulfilling the criteria of toxic inner ear lesion fol-

lowing a common cold of the upper air way (TIEL group)

were identified over a 10-year period. 20 of these patients

(21 ears) with a mean age of 54 ± 18.3 years (9 women,

11 men) were treated conservatively (TIEL-CONS) and 32

patients (36 ears) with a mean age of 49 ± 20.1 years (10

women, 22 men) were treated surgically (TIEL-SURG,

Fig. 2). All patients had had a paracentesis on initial pre-

sentation and conservative treatment consisted of haemor-

rheological infusions (500 ml hydroxyethyl starch 6 %

over 4–5 h intravenously ? 300 mg pentoxifylline)

according to the national treatment guidelines which were

valid during the study period (Fig. 2, [6]). All patients were

examined by high-resolution computed tomography during

the first 5 days. The treatment schedule was changed to a

surgical treatment (mastoidectomy) as outlined in the

diagnostic and therapeutic work flow (Fig. 2) in the case of

non-response of the inner ear function.

The duration from the onset of symptoms was

5.9 ± 3.7 days in the TIEL-CONS group and

14.7 ± 12.8 days in the TIEL-SURG group. The audio-

metrical testing results of the inner ear function (bone

conduction thresholds) are displayed in Fig. 5 and the

changes within the hearing classes according to the AAO/

HNS reporting criteria [12] in both groups initially, at the

end of the primary therapy, and at the follow-up are

listed in Table 1. On calculation of the average hearing

loss (250–8,000 Hz), significant improvement in the

TIEL-SURG group was displayed in comparison to the

TIEL-CONS group (Fig. 3, p = 0.025, Mann–Whitney

Test).

14 patients in the TIEL-CONS- and 18 patients in the

TIEL-SURG group complained of dizziness during the

initial phase. In computed nystagmography analysis, 7

patients in the TIEL CONS and 9 patients in the TIEL-

SURG group revealed pathological results: the directional

preponderance was 26.6 ± 15.4 (mean ± standard devia-

tion) in the CONS and 42.9 ± 30.6 in the SURG group;

side preponderance was 22.9 ± 15.4 and 34.2 ± 21.0,

respectively.

As a control group (CONT) patients from the PACS

system of the clinic were matched as described in the

methods ‘‘Materials and methods’’: 32 patients (16 women:

mean age 56.1 ± 24.6 years and 16 men: 52.8 ± 15.4)

were chosen. Both sides were evaluated, giving a total of

64 ears.

A small-sized pneumatisation was found in 79.4 % of

the TIEL group and in 10.9 % of the CONT group while

medium-sized pneumatisation was visible in 20.6 and

62.5 %, respectively. Whereas a large-sized pneumatisa-

tion was absent in the TIEL group, it was present in 26.6 %

of the control group. When compared to literature data

(36.7 % small/41.7 % medium/21.6 % large-sized cells),

the control group matched the data of Turgut and Tos and

the TIEL group showed an increased small-sized pneu-

matisation [13].

The distances measured within the temporal bones are

listed in Table 2. Notable differences between the TIEL

group and the control group were found with regards to

Eur Arch Otorhinolaryngol

123

• The thickness of the bone wall of the LSC at the level

of the short process of the incus (PromLat1,

p = 0.027),

• The thickness of the bone wall of the labyrinthine bone

at a 90� angle dorsolateral to PromLat1 (PromLat2,

p \ 0.005),

• The medial isthmus between the prominentia of the

LSC and the short process of the incus (IsthMed,

p \ 0.005),

• And the shortest distance from the LSC to the external

auditory canal (PromCan, p = 0.008).

The results of the Hounsfield units (HU) measured in

different areas (Fig. 3b) are displayed in Table 3: differ-

ences (two-sided Student’s t test for independent samples

with Bonferroni transformation) with reduced HU’s in the

TIEL group (bold values) were found in the precochlea

(p = 0.016), in the LSC contralateral to the prominentia

(p = 0.014), and dorsolateral of the LSC (p \ 0.005).

Pathologic electronystagmography recordings were

registered in 12 ears in patients reporting dizziness during

their initial visit (N = 32, 2 patients affected bilaterally

resulting in 34 ears). Comparing the HU in the evaluated

CT-scans showed a reduced thickness (bold values) of the

prominentia of the LSC in the group with pathological

results (p = 0.046, Table 4). Correlation analysis

between analysed Hounsfield units, bone thickness at the

LSC and hearing loss, showed a correlation (bold values)

between the HU’s in the prominentia of the LSC and

hearing loss (p = 0.002, Table 5).

Discussion

Toxic inner ear lesions following acute inflammation of

the middle ear are mainly attributed to a permeation of

Fig. 5 Pure tone audiometry pre- and postoperatively following

conservative and surgical treatment (bone conduction thresholds are

displayed): outcome differs markedly (p = 0.025 Mann–Whitney

test,filled circle pre-therapeutic sensorineural hearing threshold, open

square sensorineural hearing threshold on follow-up)

Table 1 Hearing results according to the AAO-HNS reporting cri-

teria (bone conduction thresholds are used for analysis)

Initial EOT Follow-up

CONS

Aver. HL 32.4 ± 15.6 24.3 ± 14.7 20.1 ± 15.0

Class A 10 14 12

Class B 8 5 1

Class C 3 1 1

SURG

Aver. HL 35.4 ± 12.0 28.3 ± 15.1 14.3 ± 9.0

Class A 11 19 21

Class B 20 13 2

Class C 4 3 0

CONS conservative therapy, SURG surgical therapy, EOT end of

primary therapy, aver. HL averaged hearing loss, N patients

Eur Arch Otorhinolaryngol

123

ototoxic substances through the round window membrane

[4, 14, 15]. Yellon and colleagues separated the cytokines

tumour necrosis factor-alpha (TNF-a), gamma-interferon

(IFN-c), interleukin-1-beta (IL-1b), and interleukin-2 (IL-

2) within the inflammatory effusions of chronic otitis

media [16]. Consequently bacteria themselves as well as

metabolic products may transmigrate the round window

membrane and cause direct damage of sensorineural cells

within the cochlea [14].

Moreover, it seems that bacterial infections of the

middle ear are not the leading cause in the development of

Table 2 Distances in the temporal bone (mm) and comparison of affected and control ears

Distancea Study Group (N = 34) Control group (N = 64) Mean difference SD 95 % CI p value

PromLatl 0.651 ± 0.114 0.703 ± 0.101 -0.053 0.023 -0.099 to -0.006 0.027

PromLat2 1.080 ± 0.329 1.405 ± 0.382 -0.326 0.074 -0.473 to -0.178 0.000

IsthMed 1.218 ± 0.331 1.459 ± 0.386 -0.241 0.075 -0.389 to -0.093 0.002

IsthLat 1.139 ± 0.405 1.199 ± 0.438 -0.060 0.088 -0.236 to 0.116 0.499

SpiSin 15.342 ± 3.005 15.646 ± 2.724 -0.304 0.618 -1.539 to 0.931 0.625

SulFos 12.909 ± 2.085 12.390 ± 1.958 0.518 0.433 -0.347 to 1.384 0.236

PromCan 6.481 ± 0.908 7.033 ± 1.057 -0.552 0.204 -0.958 to -0.145 0.008

p value tested with the two-sided Student’s t test for independent samples

SD standard deviation, CI confidence intervala For description of the different distances and applied statistical methods please refer to the ‘‘Materials and methods’’ section)

Table 3 Hounsfield units at different measurement points within the temporal bone

Study group

(N = 34)

Min Max Control group

(N = 64)

Min Max Mean

difference

SD 95 % CI p value

Fissula ante fenestram 2,018.4 ± 129.2 1,763 2,306 2,044.5 ± 163.2 1,767 2,511 -26.1 30.1 -86.0 to 33.9 0.390

Apex cochleae 1,861.0 ± 151.1 1,595 2,188 1,898.3 ± 210.5 1,521 2,703 -37.3 36.9 -110.7 to 36.1 0.315

Precochlea 1,948.9 ± 115.6 1,728 2,316 2,024.7 ± 189.9 1,035 2,393 -75.8 30.9 -137.2 to -14.4 0.016

Anterior edge internal

auditory canal

2,034.0 ± 136.4 1,741 2,352 2,012.9 ± 212.8 1,064 2,583 21.1 35.4 -49.2 to 91.4 0.553

Posterior edge internal

auditory canal

1,697.0 ± 85.0 1,527 1,870 1,717.0 ± 108.9 1,452 2,044 -20.0 19.9 -59.6 to 19.7 0.320

Sacculus 1,880.0 ± 120.7 1,500 2,045 1,911.8 ± 167.5 1,675 2,400 -31.8 29.4 -90.3 to 26.7 0.284

Posterior SCC 1,895.2 ± 143.8 1,689 2,482 1,944.6 ± 172.4 1,687 2,367 -49.4 32.8 -114.6 to 15.8 0.136

Lateral SCC

contralateral

to the prominentia

1,661.1 ± 111.0 1,501 1,960 1,718.8 ± 98.8 1,544 2,058 -57.7 22.7 -103.1 to -12.3 0.014

Prominentia of the

lateral SCC

1,410.8 ± 179.8 957 1,738 1,461.4 ± 242.6 948 2,189 -50.6 43.3 -136.6 to 35.4 0.245

Dorsolateral of the

lateral SCC

1,615.1 ± 89.0 1,464 1,935 1,779.0 ± 115.0 1,481 2,051 -163.9 21.0 -205.6 to

-122.2

0.000

p value tested with the two-sided Student’s t test for independent samples; please refer for details of the Bonferroni transformation to the

‘‘Materials and methods’’ section)

SD standard deviation, CI confidence interval, SCC semicircular canal

Table 4 ENG findings related to thickness and Hounsfield units

(HU) of the lateral semicircular canal (LSC) (Prom) and distance to

the outer ear canal (Can; two-sided t test, mean ± standard deviation;

SCC: semicircular canal)

Normal

(N = 22)

Pathologic

(N = 12)

p value

Thickness prominentia

lateral SCC

1.108 ± 0.301 0.883 ± 0.295 0.046

Distance PromCan 6.344 ± 0.960 6.732 ± 0.781 0.214

HU PromLat2 1,613 ± 66.6 1,618 ± 123.7 0.897

Eur Arch Otorhinolaryngol

123

acute otitis media [1]. Infections of the upper airways from

various viruses are accompanied by a virus replication

within the mucosa of the Eustachian tube (ET) with an

expression of cytokines which leads to cytotoxic effects.

This results in a dysfunction of the ET including secretion,

mucosal swelling, impaired mucociliary clearance, and

direct mucosal damage followed by a decrease of middle

ear pressure and a secondary bacterial infection [5].

15 years ago, a subset of patients, who had developed an

otitis media with effusion without acute inflammatory

reaction following a viral infection of the upper airways

and leading to a concomitant sensorineural hearing loss,

was identified. Typically, about 4 weeks following these

common colds, the patients had developed a combined

conductive and sensorineural hearing loss. On admission,

an immediate paracentesis with evacuation of the typically

amber fluids from the middle ear was performed under

local anaesthesia followed by haemorrheological infusions

according to the national treatment guidelines valid during

the study period [6]. In high-resolution CT-scanning of the

temporal bones within the first 5 days, an extended small-

sized pneumatisation of the temporal bone with a shading

of the cells, especially around the semicircular canals, was

found in all cases. In contrast to solely conservative

treatment, surgical therapy in the case of non-response to

conservative treatment resulted in a significant improve-

ment of the inner ear function with complete recovery.

Therefore, the question arose whether a direct perme-

ation through the labyrinthine bone may be responsible for

the toxic inner ear lesion in addition to the classical means

of penetration of toxic agents through the round window or

the ligament of the oval window. This led to a comparison

of the morphological characteristics of the temporal bone

configuration of the affected patients to normal temporal

bone CT-scans displayed by a matched control group.

This analysis revealed characteristic differences

regarding the thickness of the labyrinthine bone at the LSC

and also reduced Hounsfield units within this area. A

notable difference regarding the extent of pneumatisation

and the cell size was also found: the affected patients

showed significant differences in a predominance of small

and medium-sized mastoid cells when compared to the

control group and also to literature data [13].

During the embryologic development of the otic capsule

(consisting of the endosteal, the enchondral, and periosteal

layer) the middle layer is rebuilt by enchondral ossification;

this process of bony reorganisation is incomplete resulting

in interosseal globuli (bone with chondral islands) in var-

ious extensions. Paul Manasse described these interglobu-

lar spaces as remnants of the embryological bone within

the otic capsule in 1897 [17]. In 1933, Lotte Steinberg

published her results on the distribution of the interosseal

globuli [18]: she described a similar arrangement of these

‘‘weak’’ spots around the labyrinth complex and the

cochlea to what was found in the actual study using mea-

surement of Hounsfield Units in CT. These interosseal

globuli are also described in recent literature: Rauchfuss

showed these chondroosseous canals to be persistent after

completion of the ossification and building of the complex

canal system within the otic capsule [19–21]. The

hypothesis, that these channels are a ‘‘weak’’ point for a

possible infectious permeation is at least a reasonable

explanation for the group of patients, who do not improve

following conservative treatment despite immediate para-

centesis with evacuation of all fluids from the middle ear

cavity.

Limitations of the study

A major limitation of this study is the retrospective design

and the missing randomisation of the treatment groups

(conservative vs. surgical treatment); an allocation of

patients to an ongoing conservative treatment in case of

missing recovery seemed to be unethical—therefore, this

retrospective and comparative approach was chosen. The

result of the hearing outcome in favour of the surgically

treated patients supports the point of view of the author’s

opinion.

Another major criticism may be the use of Hounsfield

units to estimate the ‘‘density’’ of the bony structures of the

labyrinthine capsule. Hounsfield units describe an attenu-

ation of X-radiation in the tissue. A density measurement

in a physical definition (kg/m3) is, therefore, not possible.

However, since it is an attenuation of radiation and this

attenuation is compared between the two groups, the value

of the Hounsfield units should correlate with the density of

the measured tissue and probably fulfil the prerequisites for

a realistic comparison. Definitive results regarding the real

density of the bony structures can only be made by quan-

titative computed tomography [22]. Therefore, the

Table 5 Correlation between Hounsfield units (HU), bone thickness

at the lateral semicircular canal (LSC) and hearing loss (HL: bone

conduction thresholds) according to Kendall and Pearson

HU

PromLat2

HL

(dB)

Thickness

LSC

HU

PromLat2

1 Correlation

p value

34 N

HL (dB) -0.561 1 Correlation

0.002 p value

25 25 N

Thickness

LSC

-0.005 0.219 1 Correlation

0.49 0.146 p value

34 25 34 N

Eur Arch Otorhinolaryngol

123

statements regarding the conventional CT-scans used in the

study are only approximate.

Conclusion

On the basis of the embryological development of the otic

capsule with persistent ‘‘interosseous globuli’’ to various

extents and the results of the actual CT-morphological

study, the hypothesis of a direct means of viral infections

through weak parts of the bony labyrinth causing inner ear

afflictions may be realistic. This hypothesis may justify the

surgical approach of mastoidectomy in non-responders to

conservative treatment, particularly in cases of virus-rela-

ted otitis media and persistent sensorineural symptoms

despite of removal of infectious middle ear effusions by

paracentesis. However, these findings should be confirmed

by additional imaging studies of the temporal bone on the

basis of high-resolution CT-scans.

Conflict of interest The authors declare that there is no conflict of

interest.

References

1. Chonmaitree T (2006) Acute otitis media is not a pure bacterial

disease. Clin Infect Dis 43:1423–1425

2. Murray LN (2004) Myringitis Bullosa. In: Alper CM, Bluestone

CD, Casselbarnt ML, Dohar J, Mandel EM (eds) Advanced

therapy of otitis media. BC Decker Inc, Hamilton, pp 49–51

3. Hariri MA (1990) Sensorineural hearing loss in bullous myrin-

gitis. A prospective study of eighteen patients. Clin Otolaryngol

Allied Sci 15:351–353

4. Cureoglu S, Schachern PA, Rinaldo A, Tsuprun V, Ferlito A,

Paparella MM (2005) Round window membrane and labyrinthine

pathological changes: an overview. Acta Otolaryngol 125:9–15

5. Hayden FG (2001) Influenza virus and rhinovirus-related otitis

media: potential for antiviral intervention. Vaccine 19(Suppl

1):S66–S70

6. Michel O (2011) The revised version of the German guidelines

‘‘Sudden idiopathic sensorineural hearing loss’’. Laryngo Rhino

Otol 90:290–293

7. Jongkees LB, Maas JP, Philipszoon AJ (1962) Clinical nystag-

mography. A detailed study of electro-nystagmography in 341

patients with vertigo. Pract Otorhinolaryngol (Basel) 24:65–93

8. Allam AF (1969) Pneumatisation of the temporal bone. Ann Otol

Rhinol Laryngol 78:49–64

9. Schuknecht BF (1974) Pathology of the ear. Harvard University

Press, Cambridge, pp 79–92

10. Huizing EH, de Groot JA (1987) Densitometry of the cochlear

capsule and correlation between bone density loss and bone

conduction hearing loss in otosclerosis. Acta Otolaryngol

103:464–468

11. Grayeli AB, Yrieix CS, Imauchi Y, Cyna-Gorse F, Ferrary E,

Sterkers O (2004) Temporal bone density measurements using

CT in otosclerosis. Acta Otolaryngol 124:1136–1140

12. Monsell EM, Balkany TA, Gates GA, Goldenberg RA, Meyer-

hoff WL, House JW (1995) Committee on Hearing and Equi-

librium guidelines for the evaluation of hearing preservation in

acoustic neuroma (vestibular schwannoma). Otolaryngol Head

Neck Surg 113:179–180

13. Turgut S, Tos M (1992) Correlation between temporal bone

pneumatisation, location of lateral sinus and length of the mastoid

process. J Laryngol Otol 106:485–489

14. Kubo T, Anniko M, Stenqvist M, Hsu W (1998) Interleukin-2

affects cochlear function gradually but reversibly. ORL J

Otorhinolaryngol Relat Spec 60:272–277

15. Kim CS, Cho TK, Jinn TH (1990) Permeability of the round

window membrane to horseradish peroxidase in experimental

otitis media. Otolaryngol Head Neck Surg 103:918–925

16. Yellon RF, Leonard G, Marucha PT, Craven R, Carpenter RJ,

Lehmann WB, Burleson JA, Kreutzer DL (1991) Characteriza-

tion of cytokines present in middle ear effusions. Laryngoscope

101:165–169

17. Manasse P (1897) Ueber knorpelhaltige Interglobularraume in

der menschlichen Labyrinthkapsel. Z Ohrenheilk 31:1–10

18. Steinberg L (1933) Uber die Verteilung und Lage der knorpel-

haltigen Interglobularraume in der menschlichen Labyrinthkap-

sel. Arch Ohren Nasen Kehlk Heilk 135:187–196

19. Hildmann H, Haack M (1979) Autoradiographic investigations on

the developing otic capsule of hamsters. Arch Otorhinolaryngol

224:177–186

20. Rauchfuss A (1980) Some morphological details of the endo-

chondral layer of labyrinthine bone. A comparative anatomical

light and transmission electron microscopy study. Arch Otorhi-

nolaryngol 226:239–250

21. Rauchfuss A (1981) On the enchondral ossification of the otic

capsule: formation of the globuli ossei and the interglobular

spaces. Arch Otorhinolaryngol 233:237–250

22. Todisco M, Trisi P (2005) Bone mineral density and bone his-

tomorphometry are statistically related. Int J Oral Maxillofac

Implants 20:898–904

Eur Arch Otorhinolaryngol

123