Otology & Neurotology
30:515Y521 Ó 2009, Otology & Neurotology, Inc.
Asymmetric Hearing Loss: Rule 3,000 for Screening
Vestibular Schwannoma
*Issam Saliba, *Geneviève Martineau, and †Miguel Chagnon
Departments of *Otorhinolaryngology and ÞStatistics, Montreal University, Centre Hospitalier de
l’Université de Montréal, Montreal, Quebec, Canada
Objective: To assess the diagnostic yield of audiograms asso-
ciated to electronystagmography (ENG) for screening vestibular
schwannomas (VSs), to determine what definition of asymmetric
sensorineural hearing loss (ASNHL) fits best for the diagnosis of
VS, and to determine if cochleovestibular symptoms and athero-
sclerotic potential risk factors play a role in the VS screening.
Study Design: Retrospective chart review in a tertiary care
center.
Methods: One hundred twenty-two patients were included in
the study and divided into 2 groups: 1) patients presenting a VS
(n = 74) and 2) patients without VS (n = 48). They had received
an audiometry assessment, an ENG, and a posterior fossa mag-
netic resonance imaging (MRI). In addition, a variety of risk
factors and clinical data were collected. Mean hearing threshold
by frequency, mean asymmetries by frequency, speech discri-
mination score (SDS), ENG results, and presence or absence of
vertigo are studied. Cochleovestibular symptoms and athero-
sclerotic potential risk factors were collected. Characteristics
were studied with analysis of variance, W2 test, or a paired t
test. A receiver operating characteristic curve was obtained. A
logistic regression with a step-wise selection based on the like-
lihood ratio was used to identify the best subgroup of predictors
of the VS.
Acoustic neuroma is the most common tumor (92%)
of the cerebellopontine angle (CPA), accounting for 5 to
10% of all intracranial tumors in adults (1). Because it
is attached to the vestibular branch of the eight cranial
nerve (CN VIII), it is more accurately called vestibular
schwannoma (VS). The mean age at onset is 50 years
except in cases of neurofibromatosis type 2, an autoso-
mal dominant disorder in which the tumors usually occur
before 21 years of age (2).
Address correspondence and reprint requests to Issam Saliba, M.D.,
Centre Hospitalier de l’Université de MontréalYHôpital Notre-Dame, Ser-
vice d’Oto-rhino-laryngologie, 1560 Sherbrooke street east, Montreal
Qc - H2L 4M1, Canada; E-mail: issam.saliba@umontreal.ca
515
Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
Results: The most revealing data were the mean ASNHL at
3,000 Hz ( p G 0.001), the interaural SDS asymmetry ( p G
0.001), the vestibular deficit ( p G 0.049), and the absence of
vertigo ( p G 0.001). The ASNHL at 3,000 Hz was the most rep-
resentative value of all the frequencies and for the SDS asym-
metry. Interaural difference of 15 dB or more at 3,000 Hz is
sufficient to consider hearing loss as asymmetric. When the cut-
off for a positive test was placed at 50% probability, the receiver
operating characteristic curve shows a sensitivity of 73%. The
grade of the tumor was also related with the degree of ASNHL at
3,000 Hz. Caloric test does not predict the localization or the
grade of the VS. Tinnitus and atherosclerotic potential risk fac-
tors were not considered significantly linked with VS.
Conclusion: To reduce the number of negative MRI per-
formed in the investigation of an ASNHL, we propose the
Brule 3,000,[ ASNHL of 15 dB or more at the 3,000-Hz
frequency. In this case, an investigation with MRI is crucial.
If this ASNHL is less than 15 dB, we recommend a biannual
audiometric follow-up. Key Words: AcousticVAsymmetricV
Hearing lossVNeurinomaVNeuromaVRule 3,000V
Schwannoma.
Otol Neurotol 30:515Y521, 2009.
The VS is a slow-growing and well-circumscribed
benign tumor arising from the transitional zone be-
tween Schwann cells and oligodendrocytes at the point
where CN VIII enters the internal auditory meatus, the
Obersteiner-Redlich zone (2). It initially grows within
the bony auditory canal but then expands to the CPA.
As the tumor grows, the next cranial nerve to be affected
is generally the trigeminal nerve, causing facial pain and
sensory loss. If left untreated, the tumor will compress
the fourth ventricle, causing cerebrospinal fluid outflow
obstruction, hydrocephalus, and herniation (3). Clini-
cally, the tumor most commonly presents as an asym-
metric or unilateral sensorineural hearing loss (SNHL)
with or without unilateral tinnitus. However, only 2%
of patients presenting with these classic symptoms will
516 I. SALIBA ET AL.
have VS (4). Disequilibrium is a rather frequent finding,
whereas vertigo is not a classic symptom.
Investigation of asymmetric SNHL is mandatory
to eliminate a retrocochlear lesion. A number of proto-
cols have been published to guide patient selection, but
the disparity reflects a failure to identify the most sui-
table criterias (5). Since the late 1980s, gadolinium-
diethylenetriamine pentaacetic acidYenhanced magnetic
resonance imaging (MRI) was considered the gold stan-
dard for imaging cranial contents (4,5). Unfortunately,
the investigation with MRI is quite expensive and not
always easily available. Furthermore, it does not suit
claustrophobic patients. Because the number of MRI
done is very large for the few VSs found, screening
methods have been studied to determine the best way
of selecting patients with asymmetric SNHL to perform
an MRI while trying to save money and resources.
To decrease the number of negative MRIs in patients
presenting an asymmetric SNHL, the purpose of this
study is first to determine the diagnostic yield and effi-
cacy of audiograms associated to electronystagmo-
graphy (ENG) for screening VS; second, to determine
what definition of asymmetric hearing loss fits best
for the diagnosis of VS; and finally, to determine if
cochleovestibular symptoms (tinnitus, dizziness, and
vertigo) and atherosclerotic potential risk factors play
a role in the VS screening.
MATERIALS AND METHODS
The charts of 176 patients evaluated by an otorhinolaryngol-
ogist at our tertiary care center between December 2003 and
December 2007 were retrospectively reviewed. Cases were
included in the study if 1) patients were evaluated by posterior
fossa MRI; 2) patients received an audiometric assessment
before the first diagnostic MRI; and 3) patients received an
ENG with caloric testing assessment. Most VS cases were
referred cases from other primary care centers. To cover all
the definitions reported in the literature (5Y10), asymmetric
SNHL was defined as 15 dB or greater at 1 or more frequencies
or at least 15% asymmetry in speech discrimination scores. Our
population was then divided in 2 groups: 1) patients presenting
a VS and 2) patients without VS as the control group.
Of all the charts reviewed, 122 patients met the previously
described criteria and were included in the study. Magnetic reso-
nance images were reviewed by radiologists for the presence or
absence of retrocochlear pathology and other abnormalities.
Audiometric assessments were performed by audiologists, and
ENG results were interpreted by an otology/neuro-otology team.
An ENG was considered abnormal if a positional nystagmus or
an anomaly in the saccades, pursuit, optokinetic, or headshaking
test was detected or if a vestibular deficit between both ears of
more than 25% was found. Grades of VS are based on Tos and
Thomsen classification (11).
In addition to the audiograms, the ENG, and the MRI data,
the history taking of each patient was reviewed for demo-
graphic information and for associated symptoms (tinnitus,
dizziness, and vertigo), the presence of history of Ménière’s
disease, and coexisting potential risk factors for atherosclero-
sis (dyslipidemia, diabetes mellitus, blood hypertension, and
smoking) (12).
Otology & Neurotology, Vol. 30, No. 4, 2009
Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
STATISTICAL METHODS
Frequencies (250, 500, 1,000, 2,000, 3,000, and
4,000 Hz) and ear (best or affected) are repeated factors,
whereas the factor group (presence or absence of VS) is
not repeated.
An ANOVA with repeated measures was used for
the mean hearing threshold by frequency, for the
mean asymmetries by frequency, and for the speech
discrimination.
In case of a significant interaction, a local analysis was
done. The others characteristics were studied with a W2 test
or a paired t test (positional nystagmus, pursuit and head-
shaking test result, vestibular deficit, dizziness, vertigo,
tinnitus, dyslipidemia, diabetes, and Ménière’s disease).
For each continuous measure, a receiver operating
characteristic (ROC) curve was obtained. The ROC
curve is the representation of the tradeoffs between sen-
sitivity and specificity. The area under the curve was
used to measure the capacity to predict the VS. A logistic
regression with a stepwise selection based on the like-
lihood ratio was used to identify the best subgroup of
predictors of the VS. The ROC curve was produced with
the predicted probabilities obtained by the regression. To
facilitate the clinical interpretation, the same model was
computed using a categorization of the continuous vari-
able. The analyses were made with SPSS (version 15.0),
and a 5% significance level was considered.
RESULTS
The reports of 122 posterior fossa MRIs of patients
presenting an asymmetric SNHL were reviewed. Forty-
eight patients without VS (34 women, 14 men; mean
age, 41 yr) were identified primarily in our center, and
74 patients with VS were included (36 women, 38 men;
mean age, 52 yr). Of these 74 patients, 1 case of VS was
discovered in our center, and the 73 remaining cases were
referred from other primary care centers. In the VS group,
35 were on the right side and 39 were on the left side.
The mean asymmetries of SNHL, which is the differ-
ence of hearing threshold by frequency between both
ears, are represented in Table 1. The results of audiomet-
ric assessment were most frequently unavailable for the
higher frequencies and consequently not included in the
analysis. Figure 1 shows the estimated marginal means of
the SNHL asymmetries by frequency.
The scores of the speech discrimination test are repre-
sented in Table 2. The difference between VS and con-
trol group is statistically significant ( p G 0.001).
Of the 74 patients with VS, 33 had positional nystag-
mus in the ENG. Three patients had abnormal pursuit test,
and 8 had abnormal headshaking test. The mean vestibular
deficit with the caloric tests was 45%. Of 74, 20 had a
Grade I VS, 28 had a Grade II, 16 had a Grade III, and 10
had a Grade IV VS. Of the 74 patients of the VS group, 40
presented with dizziness, 9 with vertigo, and 59 with tin-
nitus. Dyslipidemia was reported in 13 patients. Diabetes
 ASYMMETRIC HEARING LOSS AND VESTIBULAR SCHWANNOMA 517

TABLE 1. Mean hearing threshold and MASNHL by frequency for the affected and best ear
Control group VS group Control group VS group
MHT of affected MHT of best MHT of affected MHT of best
Frequencies, Hz ear, dB ear, dB ear, dB ear, dB MASNHL, dBa MASNHL, dBa p
250 11.67 T 11.31 12.08 T 13.64 18.33 T 12.75 9.93 T 7.10 9.33 T 11.90 9.55 T 10.15 0.469
500 17.60 T 16.14 20.00 T 17.47 28.04 T 17.32 11.49 T 9.53 12.22 T 16.29 14.20 T 12.93 0.048
1,000 18.54 T 17.68 20.00 T 19.41 33.51 T 21.16 11.50 T 9.53 11.44 T 16.43 19.91 T 15.83 G0.001
2,000 22.60 T 19.52 23.33 T 20.09 44.93 T 22.48 14.80 T 11.39 13.56 T 16.50 28.84 T 19.68 G0.001
3,000 23.80 T 22.48 21.96 T 18.99 45.92 T 19.65 17.11 T 15.58 12.00 T 13.96 30.27 T 18.72 G0.001
4,000 24.48 T 20.19 23.75 T 20.82 47.67 T 21.39 20.54 T 17.45 12.78 T 14.87 26.70 T 19.94 0.043
All statistics are represented by mean T standard deviation.
MHT indicates mean hearing threshold; MASNHL, mean asymmetries of sensorineural hearing loss; VS, vestibular schwannoma.
a
Difference between mean hearing thresholds of affected ear and best ear in absolute value.

was reported in 7 patients. None of the 74 patients had a
superimposed Ménière’s disease (Table 3).
In the control group, 23 of the 48 patients had posi-
tional nystagmus in the ENG. Five patients had abnormal
pursuit test, and 3 had abnormal headshaking test. The
mean vestibular deficit with the caloric tests was 27%. Of
the 48 patients, 34 presented with dizziness, 25 with
vertigo, and 32 with tinnitus. Dyslipidemia was reported
in 12 patients. Diabetes was reported in 6 patients. Of the
48 patients, 5 were diagnosed with Ménière’s disease
(Table 3). The smoking status and the blood hyperten-
sion factor were rarely reported in the charts to be con-
sidered in the analysis. Saccades and optokinetic were
normal in all patients of both groups.
Considering the results of the t test performed and the
p values obtained, the variables linked to the presence of
VS were the following: mean asymmetries of SNHL
between 500 and 4,000 Hz, mean difference of speech
discrimination, vestibular deficit, and absence of vertigo.
The ROC curves and the areas under the curves obtained
from these variables and the logistic regressions were
performed. A subgroup of parameters with the best repre-
sentative capacities was determined: 1, mean asymmetry
of SNHL at 3,000 Hz; 2, absence of vertigo; and 3, ves-
tibular deficit. The larger area under the curve links to the
mean asymmetries of SNHL at the 3,000-Hz frequency
and is equal to 0.799. This is the reason why this hearing
frequency was consequently used in the logistic regres-
sion model.
Categories were then established for the continue vari-
ables (asymmetries of SNHL and vestibular deficit) to
determine the best and most clinically useful model to
predict the probability of VS. A mean asymmetry of
SNHL at 3,000 Hz of 15 dB or more was found statisti-
cally significant ( p G 0.001), as well as a vestibular deficit
of 25% or more, and the absence of vertigo. No statistical
difference was found for an asymmetric SNHL of less
than 15 dB ( p = 0.077).
Following this categorization of the variables, a new
logistic regression was performed. The new model ob-
tained did not include the vestibular deficit anymore. Add-
ing the vestibular deficit information obtained by an ENG
would not improve our predicting capacity once the asym-
metry of SNHL at 3,000 Hz and the absence/presence of
vertigo were considered (Table 4). Because the 3,000-Hz
frequency data were missing in 19 patients, at this stage,

TABLE 2. Speech discrimination score in percentage: the

MV of the affected and the best ear and the mean difference
between both ears for the control and VS groups
Affected ear, Best ear, Mean
MV MV difference
Control group, SDS % 88.94 88.42 11.65
VS group, SDS % 68.04 97.97 29.93
p G0.001
FIG. 1. Mean asymmetries of sensorineural hearing loss by fre-
quency for vestibular schwannoma and control groups; (Mean + /j MV indicates mean value; SDS, Speech discrimination score; VS,
standard error of mean). vestibular schwannoma.

Otology & Neurotology, Vol. 30, No. 4, 2009

Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
518 I. SALIBA ET AL.

TABLE 3. Electronystagmography, symptoms, and potential TABLE 5. Predicted probabilities of VS for each
risk factors for atherosclerosis results combination of categories
Control group, VS group, Predicted
n = 48 n = 74 p Effective, probability
n = 103 average
Positional nystagmus 23 33 0.719
Abnormal pursuit test 5 3 0.165 Vertigo Absence Asymmetry of G15 dB 27 0.48135
Abnormal headshaking test 3 8 0.378 SNHL at Q15 dB 43 0.86054
Vestibular deficit, % 27 45 0.049 3,000 Hz
Dizziness 34 40 0.077 Presence Asymmetry of G15 dB 23 0.13058
Vertigo 25 9 G0.001 SNHL at Q15 dB 10 0.49966
Tinnitus 32 59 0.078 3,000 Hz
Dyslipidemia 12 13 0.448
Diabetes 6 7 0.670 SNHL indicates sensorineural hearing loss; VS, vestibular schwannoma.
Ménière’s disease 5 0 0.005
VS indicates vestibular schwannoma.

tumor. For example, the asymmetries of the Grade I

tumor in comparison with the Grade II were quite differ-
the data from 103 remaining patients were used. A table ent, reaching a 15-dB difference between the averages of
of predicted probabilities of having VS according to asymmetric SNHL ( p = 0.026) of Grades I and II, as well
the testing results for each combination of categories as the differences obtained by comparison of Grades I
was obtained (Table 5). For example, in the absence of and III, reaching 22 dB between the averages of asym-
Ménière’s disease, perilymphatic fistula, acoustic trauma, metric SNHL ( p = 0.004). The difference was not sig-
cochlear hydrops, and cholesteatoma, the predicted prob- nificant between Grades I and IV ( p = 0.209).
ability of a patient who has at 3,000 Hz an asymmetric
SNHL of 15 dB or more without vertigo is 86%.
This predicted probability correlated to our series iden-
tified that 16 cases of VS could be missed. Of the 16
cases, 11 were VS of Grade I, 4 of Grade II, and 1 of
Grade IV.
A new ROC curve with sensitivity and specificity
results for each of its cut points was obtained for this
final model (Fig. 2). The area under the curve is 0.826.
In a descriptive perspective, we cannot state from our
results that vestibular deficit of more than 25% on caloric
test was significantly predictive to the localization of the
VS. We compared the schwannoma reaching the fundus
( p = 0.740), the porus ( p = 0.505), and the CPA ( p =
0.464), but the results were not significant. Moreover, the
grade was not significantly linked with the presence of a
vestibular deficit of more than 25%. The comparison of
Grades I and II ( p = 0.961), Grades I and III ( p = 0.485),
and Grades I and IV ( p = 0.263) did not show signifi-
cant results. On the other hand, the grade of the tumor
was related with the degree of asymmetric SNHL at
3,000 Hz. This asymmetry rises with the grade of the

TABLE 4. Results after categorization of the

continue variables
95% CI for OR
Categorical variables coding p OR Lower Higher
Mean asymmetry of 15 dB or G0.001 6.649 2.555 17.302
SNHL at 3,000 Hz more
(reference, G15 dB)
Vertigo (reference, Absence 0.001 6.179 2.189 17.445
presence)
FIG. 2. Receiver Operating Characteristic (ROC) curve. Sensi-
CI indicates confidence interval; OR, odd ratio; SNHL, sensorineural tivity and specificity values for the final model obtained. The Area
hearing loss. under the curve is 0.826.

Otology & Neurotology, Vol. 30, No. 4, 2009

Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
 ASYMMETRIC HEARING LOSS AND VESTIBULAR SCHWANNOMA
DISCUSSION
This study seeks to determine the most valuable cri-
teria to predict the presence of VS to reduce the number
of negative MRIs performed. The liberal use of MRI
yields a low return: most examinations are negative
due to the low prevalence of VS even in an otologically
selected population. By the definition of asymmetric
SNHL proposed in the literature, we identify 1 case of
VS of 49 MRIs performed in our center, which means
that the criteria for MRI evaluation in patients with asym-
metric SNHL yield a VS in roughly 2% of scanned pa-
tient, similar to the reported value in the literature. The
percentage of patients who underwent MRI and who
were found to have VS is high in this study because
the 73 patients with VS were referred cases from other
primary care centers, where screening and scanning were
done. The results obtained showed us that a clinically
acceptable screening method with symptomatic review
and audiogram assessment could be done first. The
ENG does not raise significantly the predictive probabil-
ities of our model. Indeed, the vestibular deficit was
associated with the tumor but did not give more informa-
tion than the vertigo and the asymmetric SNHL alone for
predicting the presence of the tumor.
Rule 3,000
We need a rule that is simple, practical, and sensitive to
use in our daily consultation for screening VS before ask-
ing for an MRI. We propose the Brule 3,000,[ asymmetric
SNHL of 15 dB or more at the frequency 3,000 Hz. In this
case, an investigation with MRI is crucial. According to
our results, the asymmetric SNHL at 3,000 Hz was the
most representative value of all the frequencies between
250 and 4,000 Hz and for the speech discrimination score
asymmetry. This rule 3,000 can therefore be used as a
reference when the clinician wants to estimate the unilat-
eral SNHL of his patients, which simplifies the clinical
screening process in comparison to definitions proposed in
earlier studies (Table 6). In VS, higher frequencies are
more affected than lower frequencies; anatomically, mid-
and high-frequency nerve fibers lie on the outer surface,
whereas low-frequency nerve fibers are found at the cen-
tral core of the cochlear nerve (13), which supports the
proposition that compression neuropathy or nerve conduc-
tion block on the cochlear nerve leads to mid- or high-
frequency hearing loss. The optimal test for excluding VS
is a gadolinium-enhanced T1 MRI. It has virtually 100%
sensitivity and specificity (5). On MRI, VSs are frequently
uniformly enhanced, dense, and expand the internal audi-
tory meatus (14,15).
Because the consequences of an undiagnosed VS can
be severe (deafness, facial paralysis, other cranial nerves
injuries, and brainstem injuries), we are first aiming for a
high sensitivity, defined as the proportion of patients
with VS well identified by the test performed. The spe-
cificity, defined as the proportion of patients without
disease identified as free of VS by the test, also needs
to be acceptable to reduce the number of negative MRIs
Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
519
TABLE 6. Definitions of asymmetric SNHL reported
in the literature
Asymmetries Source
Q20 dB at any single frequency between 0.5 Department of
and 4 kHz Health (5)
Q20 dB at 2 neighboring frequencies Sunderland (6)
Q15 dB between the average of 0.5, 1, 2, and 3 kHz AAO-HNS (5)
Q15 dB between the average of 0.5 and 8 kHz Oxford (7)
Q15 dB between the average of 1 and 8 kHz Seattle (8)
Q15 dB at any single frequency between 0.5 Nashville (9)
and 4 kHz
Q10 dB at 2 or more frequencies; OR Q15 dB AMCLASS (10)
at any single frequency
Q15 dB at 2 or more frequencies; OR Q15% Cueva (5)
difference between speech discrimination
AAO-HNS indicates American Academy of Otolaryngology Head
and Neck Surgery; AMCLASS, audiogram classification system; OR,
odds ratio; SNHL, sensorineural hearing loss.
performed. When the cutoff for a positive test was placed
at 50% probability, the ROC curve shows a sensitivity of
73% (95% confidence interval, 54Y79%) and a specifi-
city of 76% (95% confidence interval, 78Y97%). This
corresponds to 16 of 74 false negatives and 14 of 48
false positives.
In our study, 16 patients with VS have less than 15 dB
at 3,000 Hz; they do not fit to the rule 3,000. Of these
16 VSs, 13 were not identified by the rule 3,000, neither
by the other definitions of asymmetric SNHL reported in
the literature. Eleven VSs of Grade I, 4 VSs of Grade II,
and 1 VS of Grade IV of our series could be underdiag-
nosed by our screening method. However, 10 of 11
patients of Grade I and 2 of 4 patients of Grade II had
no hearing asymmetry greater than 10 dB at the remaining
frequencies. This conclusion was satisfying: on the one
hand, the other definitions in the literature (Table 6) can
fail to notice these cases too, and on the other hand, we
considered that the VSs of Grade I or II were less harmful,
and an asymmetry could probably be developed later and
identified by a biannual hearing test. Therefore, the con-
sequences of a delayed diagnosis with false reassurance
would be less severe than if it is a Grade III or IV. Teppo
et al. (16) evaluated the delays in VS diagnosis. They
reported that prolonged diagnostic delay did not seem to
have significant consequences in tumor size, symptoms at
the time of diagnosis, or posttreatment morbidity (16). The
unique case of Grade IV who did not respond to the rule
3,000 did not correspond either to the other definitions of
asymmetric SNHL (Table 6). This patient had no hearing
difference between both ears. Her hearing was normal
bilaterally. The MRI was done because of other unspecific
symptoms such as headache and neck pain. In the control
group (n = 48), the number of patient with a SNHL asym-
metry less than 15 dB at the 3,000-Hz frequency between
the 2 ears is 34. That means we can reduce by 71% the
number of negative MRIs performed. The application of
this rule would result in significant savings of cost,
resources, and patient morbidity.
Otology & Neurotology, Vol. 30, No. 4, 2009
520 I. SALIBA ET AL.
Grade of VS
The grade of the tumor was also related with the
degree of asymmetric SNHL at 3,000 Hz. This asymme-
try rises with the grade of the tumor and is statistically
significant for Grades I, II, and III. This difference was
not significant between Grades I and IV ( p = 0.209)
because of the small number of cases of Grade IV
(n = 10) in the study and because 1 of the 10 patients
of Grade IV had normal hearing in both ears. The con-
troversy between normal hearing and large tumor can be
related to tumor invading into the CPA without com-
pressing CN VIII. By correlating these clinical features,
one can explain why the cisternal VSs are more likely to
be diagnosed late and present when large in size (17).
Vestibular Deficit, Dizziness, and Vertigo
Although the vestibular deficit more than 25% on the
caloric test does not predict the localization or the grade
of the VS, Tringali et al. (18) found that caloric deficits
were higher in Grade IV VS compared with other grades
probably because of the higher number of cases (n = 734)
included in their study and the statistical power conse-
quently generated. The vestibular symptoms that were
included in the analysis were dizziness and vertigo. Diz-
ziness has an incidence approaching 50% in the presence
of VS (19), but it was not found to be predictive of this
disease probably because dizziness has a high prevalence
in the general population. On the opposite, the absence of
vertigo had a strong association with VS in the presence
of asymmetric SNHL of 15 dB or more. The incidence of
vertigo is 9% in this population (20). The presence of
vertigo can generally point toward other diagnosis such
as Ménière’s disease, cochlear hydrops, perilymphatic
fistula, or cholesteatoma. In patients with asymmetric
SNHL of 15 dB or more, the predicted probability of
VS declines from 86% in the absence of vertigo to
49% if vertigo is an associated symptom.
Tinnitus and VS
Tinnitus was not considered a symptom significantly
linked with VS. This finding was also reported in a study
by Lustig et al. (21) reviewing 542 patients with VS and
symmetric hearing. He found only 4 patients with asym-
metric tinnitus as the unique symptom; these results are
supported by Obholzer and Harcourt (4) comparing 36
patients with VS and 92 without. These studies consid-
ered that this symptom might have been more revealing
if its variability was better clarified (unilateral or bilat-
eral, consistently present or not, present in quiet or noisy
environment, relationship with exogenous stress and dis-
turbed sleep).
Application of the Rule 3,000 to the Sudden
Hearing Loss
A sudden hearing loss has also been associated with
VS. Sudden hearing loss is defined as an asymmetry of
more than 30 dB on 3 consecutive frequencies appearing
in less than 3 days (22). Unfortunately, the number of
Otology & Neurotology, Vol. 30, No. 4, 2009
Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
sudden hearing loss cases in our study was too small to
evaluate statistically this parameter. In a study from
Takebayashi et al. (23), 506 patients with sudden hearing
loss were investigated by audiometry. Thirteen patients
(2.8%) consisted with VS, and 493 patients consisted
with idiopathic sudden hearing loss. Ten of 13 patients
with VS (77%) and 40 of 493 patients with idiopathic
sudden hearing loss (8%) presented with the main hear-
ing loss on the high frequencies. The group of patients
with VS had a less important hearing loss ( p G 0.05) on
the 125-, 250-, and 500-Hz frequencies than the group
with idiopathic acute hearing loss. In another study by
Sauvaget et al. (24), 28 patients with a history of sudden
hearing loss were found beyond 138 cases of VS (20%).
These studies point toward a same conclusion: a sudden
hearing loss occurring mostly on the higher frequencies
leads us toward VS. On the other hand, 25 of the 28
patients with VS presented as sudden SNHL are classi-
fied as Grade I or II.
The average duration of hearing loss or other unilat-
eral symptoms in patients with VS is less than 3 years,
with a range of 6 months to 7 years (5). Even if there is
stability of hearing loss for more than 3 years, in the
presence of asymmetry at the 3,000-Hz frequency, we
recommend doing a posterior fossa MRI. Vestibular
schwannoma can grow just in the CPA without hearing
deterioration.
Limitation of the Study
This study is hampered by the limited number of our
population and the important number of VS-referred
cases. As we have shown, by the rule 3,000 screening
method, there is a small risk to reject an MRI and to
underdiagnosed VS; this will not have immediate grave
implications but is nevertheless undesirable. A prospec-
tive study with a high number of patients chosen from a
population who seeks first-line care by the general prac-
titioner instead of referred patients to a specialized care
facility in otolaryngology would improve the validity of
our study.
CONCLUSION
To reduce the number of negative MRIs performed in
the investigation of an asymmetric SNHL and the costs
generated, we propose the rule 3,000. When the cutoff
for a positive test was placed at 50% probability, the
ROC curve shows a rule 3,000 sensitivity and specificity
of 73% and 76%, respectively. The selection of patients
who need MRI to exclude retrocochlear pathology
should be based on an asymmetric SNHL of 15 dB or
more at the 3,000-Hz frequency. This should raise suspi-
cion of a retrocochlear disease. If the asymmetric SNHL
at 3,000 Hz is less than 15 dB, we recommend a biannual
audiometry testing follow-up. The ENG can be done for
documenting the case and for postoperative prognosis
purposes, but the vestibular deficit obtained does not
help significantly in predicting the presence of a VS.
 ASYMMETRIC HEARING LOSS AND VESTIBULAR SCHWANNOMA
REFERENCES
1. Hochberg F, Pruitt A. Neoplastic diseases of the central nervous
system. In: Isselbacher ICJ, Braunwald E, Wilson JD, Martin JB,
Franci AS, Kasper DL, eds. Harrison’s Principles Of Internal
Medicine. 13th ed., Chapter 369. New York: McGraw-Hill, 1994.
2. Kumar V, Abbas AK, Fausto N. Robbins and Cotran Pathologic
Basis of Disease. 7th ed., Chapter 28. Philadelphia: Elsevier, 2005:
1411Y2.
3. Blumenfeld H. Neuroanatomy Through Clinical Cases. Chapter
12. Sunderland, MA: Sinauer Associates, 2002:491Y2.
4. Obholzer RJ, Harcourt JP. Magnetic resonance imaging screening of
vestibular schwannoma: analysis of published protocols. J Laryngol
Otol 2004;118:329Y32.
5. Cueva RA. Auditory brainstem response versus magnetic reso-
nance imaging for the evaluation of asymmetric sensorineural hear-
ing loss. Laryngoscope 2004;114:1686Y92.
6. Dawes PJD, Jeannon JP. Audit of regional screening guidelines for
vestibular schwannoma. J Laryngol Otol 1998;112:860Y4.
7. Sheppard IJ, Milford CA, Anslow P. MRI in the detection of acoustic
neuromasVa suggested protocol for screening. Clin Otolaryngol
Allied Sci 1996;21:301Y4.
8. Mangham CA. Hearing threshold difference between ears and risk
of acoustic tumor. Otolaryngol Head Neck Surg 1991;105:814Y7.
9. Welling DB, Glasscock ME 3rd, Woods CI, Jackson CG. Acoustic
neuroma: a cost-effective approach. Otolaryngol Head Neck Surg
1990;103:364Y70.
10. Margolis RH, Saly GL. Asymmetric hearing loss: definition, vali-
dation, and prevalence. Otol Neurotol 2008;29:422Y31.
11. Tos M, Thomsen J. Synopsis on: disagreements in measuring tu-
mour size at the Copenhagen Acoustic Neuroma Conference. In:
Tos M, Thomsen J, eds. Acoustic Neuroma, Proceedings of the
First International Conference on Acoustic Neuroma. Amsterdam,
The Netherlands: Kugler Publications, 1992:975Y8.
Copyright @ 2009 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
521
12. Sabini P, Sclafani AP. Efficacy of serologic testing in asymmetric
sensorineural hearing loss. Otolaryngol Head Neck Surg 2000;122:
469Y76.
13. Sando I. The anatomical interrelationship of the cochlear nerve
fibers. Acta Otolaryngol (Stockh) 1965;59:417Y36.
14. Donnelly MJ, Daly CA, Briggs RJ. MR imaging features of an
intracochlear acoustic schwannoma. J Laryngol Otol 1994;108:
1111Y4.
15. Sheppard IJ, Carrasco V, Whaley R, Pillsbury H. GadoliniumVthe
new gold standard for diagnosing cerebellopontine angle tumors.
Arch Otolaryngol Head Neck Surg 1989;115:1244Y7.
16. Teppo H, Heikkinen J, Laitakari K, Alho OP. Diagnostic delays in
vestibular schwannoma. J Laryngol Otol 2009;123:289Y93.
17. Day AS, Wang CT, Chen CN, Young YH. Correlating the cochleo-
vestibular deficits with tumor size of acoustic neuroma. Acta Oto-
laryngol 2008;128:756Y60.
18. Tringali S, Dubreuil C, Zaouche S, Ferber-Viart C. Are stage IV
vestibular schwannomas preoperatively different from other stages?
Otol Neurotol. 2008;29:46Y9.
19. Morrison G, Sterkers J. Unusual presentations of acoustic tumors.
Clin Otolaryngol 1996;21:80Y3.
20. Kentala E, Pyykkö I. Clinical picture of vestibular schwannoma.
Auris Nasus Larynx 2001;28:15Y22.
21. Lustig LR, Rifkin S, Jackler RK, Pitts LH. Acoustic neuromas
presenting with normal or symmetrical hearing: factors associated
with diagnosis and outcome. Am J Otol 1998;19:212Y8.
22. O’Malley MR, Haynes DS. Sudden hearing loss. Otolaryngol Clin
North Am 2008;41:633Y49.
23. Takebayashi S, Asai Y, Nozue M, et al. Study of trough type
audiogram of sudden hearing loss. Nippon Jibiinkoka Gakkai
Kaiho 1997;100:295Y8.
24. Sauvaget E, Kici S, Kania R, Herman P, Tran Ba Huy P. Sudden
sensorineural hearing loss as a revealing symptom of vestibular
schwannoma. Acta Otolaryngol 2005;125:592Y5.
Otology & Neurotology, Vol. 30, No. 4, 2009