Otology & Neurotology

33:379Y386 Ó 2012, Otology & Neurotology, Inc.

Functional Correlations of Tympanic Membrane

Perforation Size

*Bob Lerut, †Alain Pfammatter, Johnny Moons, and †Thomas Linder

*Department of Otorhinolaryngology, Head and Neck Surgery, AZ Sint Jan BruggeYOostende AV,
campus Brugge, Belgium; and ÞDepartment of Otorhinolaryngology, Head and Neck Surgery,
Kantonspital Luzern, Luzern, Switzerland

Objective: The correlation between tympanic membrane per- Winverted V-shapeW pattern, additional pathology behind the
forations and hearing loss was studied. eardrum must be assumed and addressed.
Study Design: Prospective data from 220 patients, who un- Conclusion: We propose using standardized photographs or
derwent primary surgery for simple chronic otitis media with a drawings to document preoperative perforation sizes. A linear
perforated eardrum, were analyzed. relationship between the size of a perforation and the conduc-
Setting: Tertiary referral center. tive hearing loss does exist. Umbo involvement at the perfora-
Patients: One hundred fifty-one patients with 155 eardrum tion margin may worsen the hearing by 5 to 6 dB, whereas the
perforations, which were checked for correct diagnosis, nor- position of the perforation itself does not play a role. The least
mal middle-ear status, and integrity of the ossicular chain, were impact of a perforation is seen at the resonance frequency
included. of 2 kHz. An Binverted V-shape[ pattern, above and below
Interventions: All patients underwent primary myringoplasty. 2 kHz, of the air-bone gap is a consistent finding. If the air-
Main Outcome Measures: Preoperative conductive hearing loss bone gap exceeds this pattern, additional pathology behind
due to eardrum perforations. the eardrum perforation must be assumed and addressed.
Results: Hearing loss shows a linear relationship with increas- Key Words: Conductive hearing lossVFrequency of hear-
ing eardrum perforation size. Umbo involvement shows a wor- ing lossVLocationVResonance frequencyVRelation between
sening of the hearing by 5 to 6 dB (p G 0.0001). The least impact air-bone gap and perforation sizeVTympanic membrane
of a perforation is seen at the resonance frequency of 2 kHz. perforationVUmbo involvement.
Above and below 2 kHz, an Winverted V shapeW of the air-bone Otol Neurotol 33:379Y386, 2012.
gap is a consistent finding. If the air-bone gap exceeds the

Chronic otitis media (COM) has been defined as a a TMP is said to result in a low-tone conductive hearing
longstanding inflammatory condition of the middle ear loss, which increases with perforation size and is influ-
and mastoid associated with or without a perforation of enced by the location of the perforation (worse hearing
the tympanic membrane (1). COM with a perforation if the TMP is within the posterior half of the eardrum).
often is accompanied by a past and present history of Recently, 2 other factors have been added: the effect of
intermittent otorrhea and conductive hearing loss. His- malleus involvement (4Y6) and the influence of the volume
torically, a lot of presumptions about perforations of the of the middle ear cavity on hearing (7,8).
tympanic membrane have been made. However, the ef- As ENT physicians, we evaluate tympanic membrane
fects of a tympanic membrane perforation (TMP) on pathologies on a daily basis; however, accurately pre-
hearing were not studied until the 20th century. Three dicting hearing loss in tympanic membrane perforations
notions have been described and passed on in textbooks: is not always straightforward. Moderators of roundtable
sessions tend to surprise the audience with preoperative
Address correspondence and reprint requests to Bob Lerut, M.D., AZ otoscopic pictures not matching the subsequent audio-
Sint Jan Hospital Brugge-Oostende AV, campus Brugge, Bruges, West- gram, and in the discussion with colleagues, there often is
Vlaanderen, Belgium; E-mail: [email protected] quite a discrepancy in the estimation of a perforation size.
Dr. Lerut was an otology fellow supported by the Fisch International Relying on our prospective otologic database, we studied
Microsurgery Foundation (http://www.fimf.ch) Neuro-otology 2010 at
Kantonsspital Lucerne, Switzerland.
the relationship between hearing loss and the size and
The other authors disclose no conflicts of interest. location of a TMP in simple chronic otitis media patients,
Supplemental digital content is available in the text. analyzed other possible interfering factors, and reviewed

379

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
380 B. LERUT ET AL.

the literature. Tools for measuring tympanic membrane egory contains patients with relative TMP sizes less than 2.5%,
perforations were studied, and a simple method will be which corresponds to a microperforation. The second category
proposed. contains patients with relative TMP sizes between 2.5% and
12.5%, which corresponds to a one quadrant TMP. The third
category contains patients with relative TMP sizes between
MATERIALS AND METHODS 12.5% and 32.5%, consisting of kidney-shaped TMP, and fi-
nally, the last category contains all patients with relative TMP
Recruitment of Patients sizes larger than 32.5%, including all subtotal TMP. These cut-
Using a prospective database (Innoforce ENTstatistics, offs were made after careful examination of the shapes of TMP
www.innoforce.com), a search was performed for all patients, and our measurements, irrespective of patient numbers within
who underwent primary myringoplasty for chronic otitis media each category.
simplex at the Kantonspital Luzern, Switzerland. A preopera-
tive audiogram and a perioperative drawing or otoscopic pic- Audiometric Data Analysis and Statistical Analysis
ture and the operating report had to be available. Overall, 220 Using the Innoforce ENTstatistics software, the preopera-
primary operations were checked for correct diagnosis, normal tive air-bone gap (ABG) was calculated for each patient, for
middle-ear status, and integrity of the ossicular chain. Sixty-five the whole patient group and for each category. The ABG was
of 220 TMP (29.5%) were rejected from the study because of calculated for each frequency from 500 to 4,000 Hz, as well
middle-ear pathology (e.g., granulation tissue, malleus adhesive as for the pure tone average (PTA) from 500 to 4,000 Hz. For
to the promontory, synechiae at the level of the ossicular chain) one of the statistical analysis, the patient group was subdivided
or bad quality of photographs or drawings, leaving 151 patients into patients with normal bone conduction (BC) and abnormal
with 155 TMP on which all statistical calculations were per- (920 dB) BC thresholds on preoperative audiogram. Statistical
formed. The study was approved by the institutional review calculations were performed using JMP Statistical Discovery
board of the Kantonsspital Luzern. Software 8.0 (SAS Institute Inc., Cary, NC, USA). The relation
between the perforation size, mean ABG, average ABG for each
frequency, and each size group were studied using regression
Analysis of the Data methods. A linear model, estimated by least squares, was fitted
through the points. The least-squares method minimizes the sum
Method of Measuring Tympanic Membrane of squared differences from each point to the line. p G 0.05 was
Perforation Sizes considered statistically significant.
Fifteen percent of the 155 included patient files had good peri- The relative position measurements of the TMP and umbo
operative photographs, and all had surgical drawings. After val- involvement were used to evaluate the influence of position of
idating the correct correlation between surgical drawings and TMP on hearing. Using the Innoforce ENTstatistics database, a
preoperative photographs (see below), we chose to use consis- search was performed within our dataset on patients with TMP
tently the perioperative drawings, made by 3 senior otologists. restricted for 90% to the anteroinferior quadrant, as well as the
They were all measured with the Cyclops Auris software (gener- posteroinferior quadrant. The same was done for TMP restricted
ously provided by Dr. Sady da Costa, Porto Alegre, Brazil). After for 90% to the anterior and posterior half of the eardrum. The
reviewing the literature, this computational semiautomated and number of patients with TMP in a certain quadrant was counted,
user-friendly software was selected as it measures the relative and the PTA for these patient subgroups was calculated and
sizes of the TMP to the whole eardrum as well as their relative compared for statistical significant differences.
location in percentages (9,10), (Fig. 1).

Validation of TMP Drawings Versus RESULTS

Photographic Records
Because not all patients had accurate preoperative photo- General Data
graphs, a separate study was performed within our department Of 151 patients (155 TMP), there were 81 men (54%)
to check the accuracy and validity of perioperative drawings and 70 women (46%) with an average age of 38 years
in comparison to preoperative endoscopic pictures of TMP. (7Y89 yr). Size measurements showed perforation sizes
Otoscopic photographs of tympanic membrane perforations of from 0.34 to a maximum of 67%, which were divided into
different sizes, sites, and shapes were showed to 14 clinicians of 4 size categories. BC was normal in 92 patients (95 ear-
all grades in our ENT department. They were asked to give es- drums) and abnormal with mainly presbyacusis in 59 pa-
timations of the perforations sizes as a percentage of the total tients (60 eardrums). In total, 35 of 155 TMP showed umbo
drumhead and to draw the perforations from their memory. We involvement, divided over the perforation size categories.
then measured and compared the perforations on the photographs
and the drawings using the Cyclops Auris software. Any invol-
vement of the umbo of the malleus into the TMP was identified Validation of TMP Drawings Versus
and added as a further parameter for data evaluation. Student’s Photographic Records
t-test was used to compare the results. The results of the internal study are visualized in
Figure 2. There is a strong correlation between the rela-
Correlation Between Clinical Assessment of tive sizes of the TMP of the otoscopic images and their
TMP and Computer Data corresponding drawings. When numeric estimation was
The correlation between the clinical assessment of TMP and asked, significant overestimation is apparent with increas-
our computer measurements was determined. The dataset was ing TMP size. Accuracy seemed not to depend on the ex-
divided into 4 categories of relative perforation sizes. They were perience level of the observer as experienced otologists
chosen according to our daily clinical judgments. The first cat- and junior ENT clinicians rated similarly.

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
 TYMPANIC MEMBRANE PERFORATION SIZE 381

FIG. 1. Example of tympanic membrane perforation measurement (with umbo involvement). The measurement shows a 32.88 % per-
foration and the repartition between every single quadrant of the tympanic membrane.

Relation Between TMP Size and Hearing Loss bigger. Figure 5 gives an overview of the linear relation-
Figure 3 shows the outcome of all the measurements. ship for each frequency between the perforation size and
With increase in TMP size, the mean ABG increases. We the ABG from 500 to 4,000 Hz and the mean ABG of
found a clear distribution of umbo involvement over all all frequencies. It is clear that 2,000 Hz is the most sta-
the size categories. The relation between TMP size and ble frequency with only minimal change in ABG, whereas
mean ABG is linear (Fig. 3), with the following formula: 4,000 Hz is the most subject to changes in perforation size.
y = (0.35  p) + 10.84, with p being the TMP size. R2 is
0.43, which means that 43% of the variability is explained Relation Between Hearing Loss, Frequency,
by the model. Figure 4 shows the relationship between and Umbo Involvement
tympanic membrane perforation size and hearing loss for The ABG for each frequency was plotted for the whole
each frequency. At each frequency, a clear increase of the patient group and for the 4 size categories, with a 95%
hearing loss is seen, as the size of the perforation gets confidence interval (CI). In every size group, the ABG

FIG. 2. A, A good correlation between drawings made by ENT physicians (irrespective of their level of training) and photographic
TMP images is seen (p 9 0.05). B, A clear overestimation of the actual size of photographic TMP images (p G 0.05): as the tympanic
membrane perforation gets bigger, the margin of error increases.

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
382 B. LERUT ET AL.

the audiogram with the largest ABG found at 0.5 kHz for
every size category and the smallest at 2 kHz. Frequencies
above 2 kHz again show a bigger ABG (Fig. 6, A and B).
The significance of umbo involvement on hearing loss
was studied for individual frequencies and on the PTA.
There is a significant difference (varying from 5 to 12 dB)
for the ABG at all frequencies and, in a multivariate
model (combined with perforation size), remained sig-
nificant at all frequencies. Overall, this hearing loss due to
umbo involvement was 5.5 dB for the whole group.
However, the effect of the umbo involvement was much
smaller than the effect of the perforation itself.

Relation Between Position of TMP

FIG. 3. Mean air-bone gap in decibels from 500 to 4,000 Hz and the Hearing Loss
versus tympanic membrane perforation size. The solid line re- When comparing the ABG for each quadrant, no sig-
presents the linear regression curve with the formula; the dotted
lines correspond to the 95% CI. nificant difference became obvious (Fig. 7). In addition,
all perforations restricted to one quadrant or one-half of
the eardrum were grouped. When comparing one quadrant
decreases with an increase in frequency until 2 kHz. perforations in the anteroinferior quadrant versus the pos-
Above 2 kHz, there is an increase in ABG with increas- teroinferior quadrant, there was no significant difference
ing frequency. There is a consistent Binverted V shape[ of in mean ABG ( p = 0.93). The same is true for one-half

FIG. 4. Relationship between tympanic membrane perforation size and hearing loss for each frequency: A, The ABG in decibels for 500 Hz
for every size category, with a 95% CI. B, ABG for 1,000 Hz for every size category, with a 95% CI interval. C, ABG for 2,000 Hz for every
size category, with a 95% CI interval. D, ABG for 4,000 Hz for every size category, with a 95% CI interval.

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
 TYMPANIC MEMBRANE PERFORATION SIZE 383

FIG. 5. Summary of linear relationship for each frequency between the perforation size and the mean ABG from 500 to 4,000 Hz;
2,000 Hz shows the least increase in ABG with increasing perforation size, and 4,000 Hz shows the most increase in ABG with increasing
perforation size.

perforations ( p = 0.69). The difference between high and can be expected because of presbyacusis; however, there
low frequencies also was checked and showed no differ- was no relation between age and size of the ABG in TMP.
ence in ABG related to the location of the TMP.

Relation Between Bone Conduction and DISCUSSION

Hearing Loss Because of TMP
The BC levels were checked for influence on the ABG Evaluating patients with COM includes a careful oto-
at each frequency. There is no significant difference be- scopic assessment of the perforation size, location, and
tween the group with normal BC and the one with abnormal an estimate of its effect on sound transmission. Before
BC, at each frequency, showing that a TMP causes a con- checking the patient’s audiogram, the otologist should be
ductive hearing loss without influencing the BC. There able to predict the expected hearing impairment from his
was a significant increase in BC thresholds with age, as clinical judgment. If the audiogram does not correspond

FIG. 6. A, The mean ABG per frequency for the whole patient group showing an overall inverted V shape. B, The mean ABG per fre-
quency for every size category showing a clear inverted V shape within each group. The ABG for each frequency or the Binverted V[
gets bigger with increasing size of the tympanic membrane perforation.

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
384 B. LERUT ET AL.

observers showed similar accuracy of TMP drawings.

The objective computerized analysis of a perforation size
from a surgeon’s drawing seems to be a correct and prag-
matic methodology.

Size of TMP and Hearing Loss

Previous studies comparing hearing loss and perfora-
tion size used air conduction (AC) levels or ABG.
Because elderly patients often present with presbyacusis,
AC levels give a wrong estimate of the high-frequency
impairment because of the perforation itself. Most studies
concur that conducive hearing worsens with increasing
perforation size; however, a linear relationship remained
debatable. As shown in this study, a linear relationship
between the mean ABG 0.5 to 4 kHz and the size of the
tympanic membrane perforation could be verified. This
linear regression line and the 95% CI are depicted in
Figure 3. A similar linear regression line was observed for
each frequency (Fig. 4). Furthermore, if the 4 most fre-
quent types of perforations were sorted out as groups, a
similar pattern was found.
FIG. 7. Statistical analysis revealing no significant difference
between the conductive hearing loss at different locations of the Frequency-Dependent Hearing Loss in TMP
perforation. The diagram shows the incidence of a certain quad- The focus of previous studies has been on frequencies
rant being involved, whenever the tympanic membrane in general below 2 kHz. Animal experiments in cats revealed greater
is perforated. This can be restricted to one quadrant or involve hearing loss in the lower frequencies, also increasing
several quadrants.
in magnitude with increasing size of the TMP (2,3,6).
Anthony et al. examined 103 patients with TMP and de-
to the expected finding, further pathology within the mid- scribed a pronounced hearing loss in the lower frequen-
dle ear must be expected. Our goal was to offer some basic cies, gradually becoming less in the high frequencies (4).
rules and tools to make this evaluation easier. Ahmad and Ramani (5) performed a clinical study on
70 patients and confirmed a greater hearing loss in the
Determination of the Size and Position of TMP lower frequencies increasing with the size of the per-
How large is a tympanic membrane perforation? This foration. Conversely, Austin (13) speculated that hear-
simple question is not easily answered. In textbooks and ing losses due to perforations of the tympanic membrane
older literature, affected quadrants are used to describe were unrelated to frequencies. Our results from 155 TMP
TM pathologies. Absolutes sizes (mm) are rarely used revealed a consistent frequency pattern, similar to an Bin-
and require photographs or microscopic images with mil- verted V shape[ of the audiogram with a turning point
limeter reference lines (5,8,11). More frequently, per- around 2 kHz. Below 2 kHz, the ABG is larger for the lower
centages are used. However, the human eye seems to frequencies, and above 2 kHz, the ABG gets bigger again
overestimate the relative size, adhering to the fact that a in the higher frequencies. There are minimal changes in
one-quadrant perforation equals 25%, and thus, the esti- ABG at 2 kHz, independent of the size of the TMP. At 0.5
mated values tend to vary greatly (4,12Y14). Hampal et al. and 4 kHz, the changes are maximal, with greater variabil-
(15) and Hsu et al. (16) showed in their studies substantial ity with increasing size of TMP. Recent studies by Voss
differences between objective photographs of eardrum et al. (19Y21) disclosed the same Bturning point[ around
perforations and the visual estimates of clinicians. To over- 2 kHz on cadaver temporal bones. The impairment of
come this subjective bias, computer-based methods are sound transmission was greatest at the lowest frequencies
preferred for accurate measurements. Different types of and decreased toward zero at 1 to 2 kHz with further sound
software have been proposed to measure eardrum perfo- transmission loss above 2 kHz typically less than 10 dB.
rations (9,15Y18). The semi-automated software Cyclops They did not comment on this high-frequency observation.
Auris has been validated and proved to be user-friendly Throughout the whole study, our data confirmed a con-
and accurate to assess eardrum perforations (9,10). Ide- sistent turning point at 2 kHz, even when subdividing the
ally, standardized digital endoscopic photographs should entire group into size categories, bone conduction issues,
be used (18). As they are not always available, we tested umbo involvement, or other variables. Considering anat-
the accuracy of physician’s drawings of TMPs compared omy and physiology of the outer ear canal and the middle
with objective otoscopic photographs within the ENT ear, 2 kHz is known as their dominant resonance frequency.
department and found a strong correlation between the Lokberg et al. (22) have calculated the resonance frequency
drawings and otoscopic photographs, in contrast to the of the eardrum to be around 2 kHz, and Stasche et al.
subjective rough estimates. Surprisingly, less experienced (23) found the highest umbo displacement of the tympanic

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
 TYMPANIC MEMBRANE PERFORATION SIZE 385

FIG. 8. Impact of increasing perforation size on hearing at 4,000 Hz.

membrane in a cadaver temporal bone study at 2 kHz on the hearing loss. Almost every textbook still mentions
and the resonance frequency between 2 and 3 kHz. Gyo worse hearing if the posterior half is involved. Reviewing
et al. (24) discovered the highest lever ratio of the ossicu- human studies, Anthony and Harrison (4) and Ahmad and
lar chain at 2.2 kHz using laser Doppler vibrometry. We Ramani (5) supported this idea, as have many others since
therefore conclude that the human middle ear has the least then. However, in most studies, position was assigned
loss of sound transmission (or best hearing) around 2 kHz, or estimated but never exactly measured, as can be done
no matter which pathology is at hand. In contrast with the with the Cyclops Auris software (Fig. 7). There are as
stability and consistency found for 2,000 Hz, there is a many studies contradicting this Blocation[ theory. Voss et al.
maximum of variation found at 500 and 4,000 Hz in our (7,20) made it clear with several studies that the theory of
patient group. These frequencies were the most sensitive sound pressures acting directly onto the round and oval
to changes within the eardrum (Fig. 8). windows in posterior TMP and, thus, inhibiting pressure
difference between the windows has only a very small
The Influence of Umbo Involvement in a TMP on influence on the hearing loss caused by TMP. Our data
the Hearing Loss show no significant difference of location of TMP on the
Whereas most studies do not mention or even exclude hearing loss, when comparing anteroinferior and postero-
malleus or umbo involvement in chronic otitis media, inferior quadrant perforations or when comparing ante-
Ahmad and Ramani (5) demonstrated a clear influence of rior half and posterior half perforations and thus affirm the
malleus involvement on the hearing. They made obvious latest data. We feel that the phase cancellation theory must
that hearing got worse once the malleus was involved in be abandoned once and for all.
the TMP. Voss et al. (19) mentioned a 5-dB hearing loss
in case of reduced coupling of the tympanic membrane The Influence of Middle Ear and Mastoid Volume on
to the ossicular chain. It would be a false presumption the Hearing Loss Caused by TMP
to believe that umbo involvement only occurs in large Recently Voss et al. (7,21) introduced the idea of
(subtotal) perforations. Our data show a clear distribu- middle ear and mastoid volume as an influencing factor
tion of umbo involvement in 3 size categories, but it is on hearing loss in chronic otitis media. Several studies
obvious that the bigger the TMP, the more likely the were performed using cadaver temporal bones and theo-
umbo is involved. Because of the conical shape and the retical models to calculate the influence of a varying
radial fibers in the eardrum, the sound is directed toward mastoid volume on hearing loss assuming that the middle
the umbo of the malleus. Gundersen (25) has shown that ear volume is almost constant. They state that, for equal
there is a difference between the acoustical driving force sizes of TMP, the bigger the mastoid volume, the smaller
at the umbo and at the neck of the malleus. All these the conductive hearing loss and vice versa. This idea is
data align with our findings that umbo involvement in based on a model, which shows that the increased sen-
TMP worsens the hearing significantly with 5.5 dB on sitivity (or better hearing) found around the 2-kHz region
mean ABG, meaning that whenever the umbo is involved, shifts toward a lower frequency region (e.g., 1,500 Hz)
5.5 dB can be attributed to the loss of coupling, and the with an increase in mastoid volume. This means that hear-
rest of the ABG is caused by the fact that there is a per- ing loss should change with varying mastoid volumes for
foration in the eardrum. an equal TMP size. How to accurately measure middle ear
and mastoid volume in patients is still a matter of debate.
The Influence of Position of a TMP on the Serial sections of CT scans may be used; however, the exact
Hearing Loss volumetric measurements are tedious to obtain (personal
Since the earliest experiments on TMP, there has been study, not published), and opacifications (e.g., effusion
debate on whether the position of TMP has an influence or granulation tissue) within the pneumatized air cells

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.
386 B. LERUT ET AL.

may vary over time. Tympanometry has been suggested and their effects upon middle-ear transmission. Arch Klin Exp
to predict the total volume. Mehta et al. (8) conducted a Ohren Nasen Kehlkopfheilkd 1968;192:145Y62.
4. Anthony WP, Harrison CW. Tympanic membrane perforation. Ef-
clinical study, using tympanometry to determine mid- fect on audiogram. Arch Otolaryngol 1972;95:506Y10.
dle ear and mastoid volume in TMP, which confirmed the 5. Ahmad SW, Ramani GV. Hearing loss in perforations of the tym-
results found by Voss et al. (21). For the same size of a panic membrane. J Laryngol Otol 1979;93:1091Y8.
perforation, they found that Bthe bigger the mastoid vol- 6. Bigelow DC, Swanson PB, Saunders JC. The effect of tympanic
membrane perforation size on umbo velocity in the rat. Laryngo-
ume, the smaller the hearing loss.[ In our current study, scope 1996;106(1 Pt 1):71Y6.
this parameter was not included as CT scans or volume 7. Voss SE, Rosowski JJ, Merchant SN, et al. How do tympanic-
recordings of tympanometries were not adequately avail- membrane perforations affect human middle-ear sound transmis-
able. It also appears (Ahn et al. [26]) that tympanometry sion? Acta Otolaryngol 2001;121:169Y73.
8. Mehta RP, Rosowski JJ, Voss SE, et al. Determinants of hearing
is overestimating the true volume and that some tympa-
loss in perforations of the tympanic membrane. Otol Neurotol 2006;
nometry devices do not measure more than 3 to 5 ml of 27:136Y43.
volume (Mehta et al. [8]). 9. Comunello E, Wangenheim A, Junior VH, et al. A computational
The impact of Breal[ volume and Bacoustic relevant[ method for the semi-automated quantitative analysis of tympanic
volume on hearing still remains debatable and needs to membrane perforations and tympanosclerosis. Comput Biol Med
2009;39:889Y95.
be further examined. 10. Heck JV, Wangenheim Av, Abdalla DD, et al. Computational
techniques for accompaniment and measuring of otology patholo-
gies. IEEE Computer Society 20th Symposium 2007;53Y8.
CONCLUSION 11. Nepal A, Bhandary S, Mishra SC, et al. The morphology of cen-
tral tympanic membrane perforations. Nepal Med Coll J 2007;9:
Previous studies do not adequately measure but rather 239Y44.
guess the perforation size or use inadequate quadrant rules 12. Gleeson MJ, Jones NS, eds. Scott-Brown‘s Otorhinolaryngology
Head and Neck Surgery. 7th ed, Vol. 1. Great Britain: Hodder
to describe a tympanic membrane perforation. We propose Arnold, 2008.
to use standardized endoscopic photographs or at least 13. Austin DF. Sound conduction of the diseased ear. J Laryngol Otol
well performed drawings to document preoperative per- 1978;92:367Y93.
foration size. Surgical drawings seem to correlate well 14. Griffin WL Jr. A retrospective study of traumatic tympanic mem-
with photographs. brane perforations in a clinical practice. Laryngoscope 1979;89
(2 Pt 1):261Y82.
The impact of a Bsimple[ perforation (with an intact 15. Hampal S, Padgham N, Bunt S, et al. Errors in the assessment
ossicular chain) on hearing should be predictable by the of tympanic membrane perforations. Clin Otolaryngol Allied Sci
otologist. A linear relationship between the size of a per- 1993;18:58Y62.
foration and the conductive hearing loss does exist as a 16. Hsu CY, Chen YS, Hwang JH, et al. A computer program to cal-
culate the size of tympanic membrane perforations. Clin Otolar-
general rule. The involvement of the umbo at the perfo- yngol Allied Sci 2004;29:340Y2.
ration margin may worsen the hearing by 5Y6 dB, whereas 17. Ibekwe TS, Adeosun AA, Nwaorgu OG. Quantitative analysis
the position of the perforation itself does not play a role. of tympanic membrane perforation: a simple and reliable method.
The impact of middle ear and mastoid volume on con- J Laryngol Otol 2009;123:e2.
ductive hearing loss needs further evaluation. 18. Jaisinghani VJ, Hunter LL, Li Y, Margolis RH. Quantitative ana-
lysis of tympanic membrane disease using video-otoscopy. Lar-
The least impact of a perforation is seen at the reso- yngoscope 2000;110(10 Pt 1):1726Y30.
nance frequency of 2 kHz. Above and below 2 kHz, an 19. Voss SE, Rosowski JJ, Merchant SN, et al. Middle-ear function with
Binverted V shape[ of the air-bone-gap is a consistent tympanic-membrane perforations. I. Measurements and mechanisms.
finding. If the air-bone-gap does not follow this pattern J Acoust Soc Am 2001;110(3 Pt 1):1432Y44.
20. Voss SE, Rosowski JJ, Merchant SN, et al. Non-ossicular signal
and exceeds 16 dB for one-quadrant perforations, 25 dB transmission in human middle ears: experimental assessment of the
for kidney-shaped TMP and 37 dB for subtotal TMP at Bacoustic route[ with perforated tympanic membranes. J Acoust
4 kHz, additional pathology behind the tympanic mem- Soc Am 2007;122:2135Y53.
brane perforation must be assumed and addressed. An 21. Voss SE, Rosowski JJ, Merchant SN, et al. Middle-ear function
evaluation tool for TMP is provided in pdf-format (see with tympanic-membrane perforations. II. A simple model. J Acoust
Soc Am 2001;110(3 Pt 1):1445Y52.
pdf diagram, Supplemental Digital Content 1, available at 22. Lokberg OJ, Hogmoen K, Gundersen T. Vibration measurement
http://links.lww.com/MAO/A102, which can be used in of the human tympanic membraneVin vivo. Acta Otolaryngol 1980;
daily use to evaluate the expected hearing loss of a TMP). 89:37Y42.
23. Stasche N, Foth HJ, Hörmann K, et al. Middle ear transmission
disordersVtympanic membrane vibration analysis by laser-Doppler-
REFERENCES vibrometry. Acta Otolaryngol 1994;114:59Y63.
24. Gyo K, Aritomo H, Goode RL. Measurement of the ossicular
1. Fisch U, May JS, Linder T, et al. Tympanoplasty, Mastoidectomy vibration ratio in human temporal bones by use of a video mea-
and Stapes Surgery. 2nd ed. Stuttgart, Germany: Thieme Verlag, suring system. Acta Otolaryngol 1987;103:87Y95.
2007. 25. Gundersen T. Prostheses in the ossicular chain. I. Mechanics of
2. Payne MC Jr, Githler FJ. Effects of perforations of the tympanic movement. Arch Otolaryngol 1972;96:416Y22.
membrane on cochlear potentials. AMA Arch Otolaryngol 1951; 26. Ahn JY, Park HJ, Park GH, et al. Tympanometry and ct measure-
54:666Y74. ment of middle ear volumes in patients with unilateral chronic
3. McArdle FE, Tonndorf J. Perforations of the tympanic membrane otitis media. Clin Exp Otorhinolaryngol 2008;1:139Y42.

Otology & Neurotology, Vol. 33, No. 3, 2012

Copyright © 2012 Otology & Neurotology, Inc. Unauthorized reproduction of this article is prohibited.