European Archives of Oto-Rhino-Laryngology

https://doi.org/10.1007/s00405-023-07996-w

OTOLOGY

EAONO/JOS classification for acquired cholesteatoma: evaluating

the impact of the number of affected sites on treatment and outcomes
B. Renner1 · A. Vasconcelos Craveiro1 · M. Balk1 · M. Allner1 · M. Sievert1 · S. K. Mueller1 · K. Mantsopoulos1 · H. Iro1 ·
R. Rupp1 · J. Hornung1 · A. O. Gostian1

Received: 18 December 2022 / Accepted: 24 April 2023

© The Author(s) 2023

Abstract
Purpose The European and Japanese system for cholesteatoma classification proposed an anatomical differentiation in five
sites. In stage I disease, one site would be affected and in stage II, two to five. We tested the significance of this differentia-
tion by analyzing the influence of the number of affected sites on residual disease, hearing ability and surgical complexity.
Methods Cases of acquired cholesteatoma treated at a single tertiary referral center between 2010-01-01 and 2019-07-31
were retrospectively analyzed. Residual disease was determined according to the system. The air–bone gap mean of 0.5, 1,
2, 3 kHz (ABG) and its change with surgery served as hearing outcome. The surgical complexity was estimated regarding
the Wullstein’s tympanoplasty classification and the procedure approach (transcanal, canal up/down).
Results 513 ears (431 patients) were followed-up during 21.6 ± 21.5 months. 107 (20.9%) ears had one site affected, 130
(25.3%) two, 157 (30.6%) three, 72 (14.0%) four and 47 (9.2%) five. An increasing number of affected sites resulted in higher
residual rates (9.4–21.3%, p = 0.008) and surgical complexity, as well poorer ABG (preoperative 14.1 to 25.3 dB, postopera-
tive 11.3–16.8 dB, p < 0.001). These differences existed between the means of cases of stage I and II, but also when only
considering ears with stage II classification.
Conclusion The data showed statistically significant differences when comparing the averages of ears with two to five affected
sites, questioning the pertinence of the differentiation between stages I and II.

Keywords Cholesteatoma · EAONO/JOS classification · Tympanoplasty

Introduction In stage I, the cholesteatoma affects only one site, with two
possibilities: “A” referring to a pars flaccida cholesteatoma
The Japanese Otological Society (JOS) published in 2008 a (PFC) or “T” for a pars tensa cholesteatoma (PTC). In stage
classification system for middle ear cholesteatoma [1], which II, two to five sites are affected. The presence of extracra-
was reviewed in 2010 [2] and 2015 [3]. The European Acad- nial or intracranial complications results respectively in a
emy of Otology and Neurotology (EAONO) presented its stage III or IV classification. So stages I and II only refer to
own system in 2015, which also featured the disease recidi- disease extension and to some extent localization, exclud-
vism [4]. Seeking an international uniformization, both parts ing any complications. Recidivism was the only described
joined efforts and released in 2016 the EAONO/JOS con- treatment outcome and a differentiation between residual
sensus system [5]. Five anatomical sites were categorized: disease (incomplete surgical removal with no contact to the
supratubal recess (S1), sinus tympani (S2), tympanic cavity tympanum) and recurrence (after complete surgical removal,
(T), attic (A) and mastoid (M). Four stages were defined. reformation of the retraction pocket and new cholesteatoma)
was made. An international validation study of nine centers
and 1482 cases demonstrated statistically significant dif-
B. Renner and A. Vasconcelos Craveiro are co-first authors.
ferences in 5 years residual and recurrence rates between
* A. Vasconcelos Craveiro stages I and II, respectively 3–13% and 4–10% [6]. Another
[email protected] retrospective single-center study with 34 patients with PFC
found a relation between better postoperative audiological
1
Department of Otolaryngology, Head and Neck Surgery, outcome and three factors: lower stage (stages I, II and III),
University Hospital of Erlangen, Erlangen, Germany

13
Vol.:(0123456789)
 European Archives of Oto-Rhino-Laryngology

better condition of the stapes and a better development of the and II; complete medical records including surgery reports,
mastoid cells [7]. A single-center cohort of 125 patients with documented pre- and post-surgery, bone and air conduction
retraction pocket cholesteatomas (RPC) showed no direct audiograms; minimum follow-up of 3 months. The consid-
correlation between stage and recidivism. However, involve- ered exclusion criteria were: congenital and unclassifiable
ment of S1 led to a more frequent recidivism. Higher rates of cholesteatoma; presence of complications (disease stages
residual cholesteatoma were associated with a defect greater III and IV); prior TPL for cholesteatoma treatment in the
than half of the posterior auditory canal and patients younger concerning ear. In patients with bilateral cholesteatoma,
than 15 years [8]. In addition, a single-center analysis of each ear represented an independent case. Recidivism was
290 patients with a mean follow-up of 4 years showed that differentiated in residual and recurrent disease according
a higher disease stage resulted in worse hearing outcome to the EAONO/JOS classification. Hearing ability is pre-
and more frequent canal wall down procedures, while a sented by the mean pure-tone air-bone-gap (ABG) of the
correlation between stage and recidivism was not registered frequencies 0.5, 1, 2 and 3 kHz as described by the Ameri-
[9]. The available research considering the EAONO/JOS can Academy of Otolaryngology—Head and Neck Surgery
consensus shows conflicting findings about the relationship Foundation (AAO-HNS) [10]. Pre- and postoperative ABG
between stage and different pathophysiological/epidemio- were analyzed, as well the difference between both. The type
logical characteristics, as well as treatment outcomes. More- of Wullstein’s TPL and the surgical approach were consid-
over, these studies analyzed several characteristics inherent ered as indicators of the surgical complexity. The surgical
to the disease that were not featured in the EAONO/JOS approach was classified in increasing difficulty as: transcanal
proposal, as hearing ability, age, ossicle state, among others. (TC), canal wall up (CWU) and canal wall down (CWD).
To address these questions, further developing a consensus There was a differentiation of the CWD procedures: with-
about cholesteatoma, but also to uniformize the report of out reconstruction, simply CWD, or with reconstruction of
otologic data, the International Otology Outcome Group the posterior canal wall using auricular or tragal cartilage
(IOOG) was established in 2017 (www.​IOOG.​net). Our (CWR, in the same surgery). The influence of the number of
study intended to contribute for these purposes. The primary affected sites (independent variable) on recidivism, hearing
objective was to assess the pertinence of the differentiation ability and surgical complexity was analyzed.
between stages I and II. For this purpose, the impact of the
number of affected sites on residual rate, hearing ability and Statistical analyses
surgical complexity was analyzed. We hypothesized that an
increasing number of affected sites would result in more fre- Continuous variables were tested for normal distribution via
quent residual rate, poorer hearing outcomes and greater sur- the Kolmogorov Smirnov Test and by QQ-Plots and his-
gical complexity. We assumed that these differences would tograms. These variables are presented with the mean ± 1
exist between cases with one affected site (stage I) and two standard deviation and range (min–max). Nominal variables
to five affected sites (stage II), but also within stage II. As are presented as absolute and relative frequencies (N/%).
only recidivism was discussed in the EAONO/JOS system, The means of the parameters recidivism, hearing ability and
the secondary objective was to demonstrate the relevance of surgical complexity according to the number of affected sites
the hearing outcomes and surgical procedure in the evalua- are always the subject of the different comparisons, even
tion of cholesteatoma and its treatment. when this is not explicitly written (to simplify the text).
Changes of the pre- to postsurgical ABG were compared
with a dependent t-test. Differences in ABG between groups
Methods were analyzed with a between-subjects ANOVA with the
main effect “number of affected sides”. To test the hypoth-
The medical records of patients that underwent a tympa- esis that the patients with more affected sides showed higher
noplasty (TPL) due to cholesteatoma at the Department of ABG, linear contrasts were performed. The partial ­Eta2 was
Otorhinolaryngology and Head and Neck Surgery, Friedrich- reported as the effect size of the main effect “number of
Alexander-University Erlangen-Nürnberg (FAU), Germany, affected sides” of the ANOVA. Here, an ­Eta2p of 0.01 equals
a tertiary referral medical center, between 2010-01-01 and a small effect, 0.06 a medium effect and 0.14 a strong effect.
2019-07-31 were retrospectively analyzed and classified Bonferroni correction was performed for all explorative pair-
according to the EAONO/JOS system (5). The present wise post-hoc comparisons. Differences between groups
study was performed in fulfillment of the requirements for regarding nominal variables were tested with cross tables
obtaining the degree “Dr. med.”, approved by the Local Eth- and the ­Chi2-test. If cross tables with more dimension than
ics Committee (Nr. 371_20 Bc) and carried out according 2 × 2 were significant, effects were further investigated with
to the Declaration of Helsinki. The following inclusion cri- 2 × 2 cross tables for pairwise comparisons. Effect sizes are
teria were applied: acquired cholesteatoma; disease stages I reported for t-tests in terms of r and for the C­ hi2-test with

13
European Archives of Oto-Rhino-Laryngology

phi (2 × 2 cross tables) or Cramer’s V (> 2 × 2 cross tables). IBM SPSS Statistics Version 28.0. All statistic computations
An r, phi or Cramer’s V of 0.1 displays a small effect, are shown in the publication supplement.
0.3 represents a moderate effect and 0.5 a strong effect.
A p ≤ 0.05 was considered as statistically significant, but
P-values ≤ 0.05 marked with an asterisk (*) should be inter- Results
preted as only a trend towards significance after correction
for multiple comparisons. Data analysis was performed with Between 2010-01-01 and 2019-07-31, a total of 2005
TPLs to treat cholesteatoma were performed. Out of
these, 513 TPLs (right ears 254/49.5%, female 185/35.9%)
met the inclusion/exclusion criteria. 41 patients (female
Table 1  Demographic und clinical data of the patients 10/24.4%) were treated on both sides, so that 472 subjects
(female 175/37.1%) were analyzed. The mean age was
Number of cases 513
39.0 ± 20.7 years (4–82 years) and the mean follow-up time
Number of patients 472
21.6 ± 21.5 months (3–112 months). A total of 107/20.9%
Patients treated both sides (N/%) 41/8.7
ears showed one affected site, 130/25.3% two sites,
Age (years, mean ± SD, min–max) 39.0 ± 20.7, 4–82
157/30.6% three, 72/14.0% four and 47/9.2% five (Table 1).
Follow-up (months, mean ± SD, min–max) 21.6 ± 21.5, 3–112
All surgical procedures were performed by ten experienced
Side (N (%))
otologic surgeons.
Right 254 (49.5)
A total of 403/78.6% ears developed no recidivism,
Left 259 (50.5)
73/14.2% developed a residual disease and 37/7.2% a recur-
Gender (N (%))
rence (Fig. 1). The rate of residual disease increased with the
Male 529 (64.1)
number of involved sites (p = 0.008, Cramer’s V = 0.163).
Female 184 (35.9)
Post-hoc comparisons showed that on average ears with one
Number of affected sites (N(%))
affected site developed less often a residual disease com-
One 107 (20.9)
pared to the cases with three (9.3% vs. 21.0%, p = 0.012*,
Two 130 (25.3)
phi = 0.155) and five (21.3%, p = 0.043*, phi = 0.163). Simi-
Three 157 (30.6)
larly, cases with two affected sites accounted for less resid-
Four 72 (14.0)
ual disease compared to three (8.5% vs 21.0%) (p = 0.003,
Five 47 (9.2)
phi = 0.065) and five (21.3%, p = 0.020*, phi = 0.175).

Recidivism by number of affected sites

86.9%
85.0%

100%
79.2%

90%
71.4%

80%
63.8%

70%
Recidivism (%)

60%
50%
40%
21.0%

21.3%

14.9%

30%
12.5%
9.4%

8.5%

8.3%
7.6%

20%
5.6%

4.6%

10%
0%
1 2 3 4 5
Number of affected sites
No recidivism Residuum Recurence

Fig. 1  Cholesteatoma recidivism (%) according to the number of ual disease (residuum) (p = 0.008), but not on recurrence (p = 0.196).
affected sites considering the EAONO/JOS classification. The impact However, the results concerning the recurrence rate are conditioned
of the number of affected sites was statistically significant on resid- by the brief follow-up (average of 21.6 months and minimum of 3)

13
 European Archives of Oto-Rhino-Laryngology

During the short follow-up time, the number of affected between ears with one and two affected sites (p = 0.007*).
sites did not significantly influence the recurrence rates Regarding the postoperative ABG, there was a significant
(p = 0.196, Cramer’s V = 0.109). difference between cases with one and three/four (p < 0.001)
The overall preoperative ABG was 19.1 ± 11.7 or five (p = 0.003) affected sites. A trend to significance was
(0.3–55.0) dB, which improved postoperatively to 14.6 ± 9.5 registered between ears with one involved site and to two
(0.6–101.0) dB by 4.4 ± 11.29 (-46.0–45.3) dB (p < 0.001, (p = 0.024*), as well when comparing the means of two and
r = 0.363, Fig. 2). The main effect “number of affected sites” four affected sites (p = 0.019*).
led to statistically significant increase of the preoperative Looking at the surgical approach, TC was performed
ABG (p < 0.001, ­Eta2p = 0.080; linear contrast: p < 0.001) in 259/50.5% of the cases, CWU in 103/20.1%, CWR
and postoperative ABG (p < 0.001, ­Eta2p = 0.045; linear in 145/28.3% and CWD in 6/1.2% (Fig. 3). In general,
contrast: p < 0.001). The change of the ABG was not statis- a higher number of involved sites required a more com-
tically significant (p = 0.057, E­ ta2p = 0.018; linear contrast: plex surgical approach (p < 0.001, Cramer’s V 0.337). An
p = 0.003). Regarding the preoperative ABG, a significant increasing number of affected sites resulted in less fre-
difference was found between the means of cases with one quent TC (p < 0.001, Cramer’s V = 0.565). Patients with
affected site and three/four/five (p < 0.001), as well when one and two affected sides received TC respectively in
comparing ears with two affected sites to four (p = 0.004) 90.7% and 70.0% of the cases, while TC was less often
or five (p < 0.001) and between cases with three affected performed in patients with three (32.5%), four (19.4%)
sites and five (p = 0.004). A trend to significance was present and five sites (12.8%) involved. A significant difference

Fig. 2  Means of the preopera- ABG by number of affected sites

tive, postoperative and change
30
of the ABG due to surgery 25.26
according to the number of 25 22.79
affected sites by cholesteatoma 19.74
ABG mean (dB)

considering the EAONO/JOS 20 17.98

classification. The preoperative 17.39 16.77
and postoperative differences 14.12
15 15.46
were statistically significant, but
not the change. ABG—mean 14.2
10 11.27 8.49
pure-tone air-bone-gap of the
5.39
frequencies 0.5, 1, 2 and 3 kHz 5 3.78 4.28
2.86
0
1 2 3 4 5
Number of affected sites
Preoperative Postoperative Change

Fig. 3  Performed surgical Surgical approach by number of affected sites

approach considering the
number of affected sites with
90.7%

cholesteatoma according to the

EAONO/JOS classification. 100%
70.0%
Surgical approach (%)

TC transcanal, CWU​canal wall 90%

up, CWR​canal wall down with 80%
54.2%

51.1%

reconstruction, CWD canal wall 70%

down without reconstruction 60%
36.2%
36.3%
32.5%
30.6%

50%
25.0%
19.4%

40%
12.8%
15.4%
12.3%

30%
2.3%
4.7%
3.7%

1.4%

20%
0.9%

0.0%
0.6%

10%
0%
1 2 3 4 5
Number of affected sites
TC CWU CWR CWD

13
European Archives of Oto-Rhino-Laryngology

was computed between ears with one affected site and ears and five (p < 0.001, phi = 0.365). No trend or significance
with more than one affected site (p < 0.001, phi = 0.254), was found considering only CWD (without reconstruc-
as well between cases with two and three to five sites tion), probably because of the low number of such cases.
(p < 0.001, phi = 0.254). Between ears with three sites and TPL type I according to Wullstein was performed
four (p ≤ 0.042*, phi = 0.134) or five (0.008*, phi = 0.185) in 182/35.5% of all cases, type II in 4/0.8%, type IIIa in
only a trend was shown. The opposite was observed 203/39.6%, type IIIb in 74/14.4% and type IV in 50/9.7%.
for CWU (p < 0.001, Cramer’s V = 0.287) and CWR No type V surgery was conducted. (Fig. 4) An increasing
(p < 0.001, Cramer’s V = 0.397), i.e. these approaches number of affected sites demanded a more complex TPL
were more frequently performed the more sites were (p < 0.001, Cramer’s V = 0.557). TPL I was more com-
involved. Ears with one and two affected sites required mon when one site was affected (79.4%) and decreased to
a CWU in 3.7% and 12.3% of the cases (CRW 4.7% and 4.3% in ears with five affected sites (p < 0.001, Cramer’s
15.4%). On the other hand, when three, four or five sites V = 0.526). A significant difference was identified between
were affected, CWU was performed in 30.6%, 25.0%, ears with one affected site and two (p < 0.001, phi = 0.390),
35.2% and CWR in 36.3%, 54.2% and 51.1% of the cases, three (p < 0.001, phi = 0.590), four (p < 0.001, phi = 0.631)
respectively. Considering CWU, a trend to significance and five (p < 0.001, phi = 0.698), as well between cases with
was shown when comparing ears with one affected site two affected sites and three (p < 0.001, phi = 0.230), four
and two (p = 0.018*, phi ≥ 0.153) and two affected sites (p < 0.001, phi = 0.262) and five (p < 0.001, phi = 0.348).
and four (p = 0.021*, phi = 0.162). A statistically signifi- Comparing three sites to five, a trend towards signifi-
cant difference was present when comparing cases with cance was shown (p = 0.011*, phi = 0.177). TPL IIIa also
two affected sites and three (p < 0.001, phi = 0.331), four differed significantly according to the number of affected
(p < 0.001, phi = 0.318) or five (p < 0.001, phi = 0.435). sites (p < 0.001, Cramer’s V = 0.306). It was less frequently
For CWR, a trend to significance was found between cases performed in patients with one affected site (14.0%), ris-
with one affected site and two (p = 0.008*, phi = 0.174), ing to 35.4% by two affected sites and further increasing
as well between three and four (p = 0.011*, phi = 0.168). when three to five sites were involved (46.8–54.8%). A
The data showed a significant difference between ears with significant difference was present when comparing cases
one affected site and three (p < 0.001, phi = 0.366), four with one affected site to two (p < 0.001, phi = 0.243), three
(p < 0.001, phi = 0.564) and five (p < 0.001, phi = 0.546), (p < 0.001, phi = 0.412), four (p < 0.001, phi = 0.365) and
as well when comparing cases with two involved sites and five (p < 0.001, phi = 0.353). Additionally, the difference
three (p < 0.001, phi = 0.235), four (p < 0.001, phi = 0.408) between ears with two a three affected sites was statistically

Tympanoplasty by number of affected sites

79.4%

80%
Type of Wullstein's tympanoplasty (%)

70%
54.8%

60%
47.2%

46.8%
40.8%

50%
35.4%

34.0%

40%
23.6%
19.7%

30%
15.3%

14.9%
14.0%

14.0%

13.9%
11.5%
10.8%

10.8%

20%
4.7%

4.3%
1.5%
0.9%

0.9%

0.6%

0.0%

0.0%

10%

0%
1 2 3 4 5
Number of affected sites
1 2 3a 3b 4

Fig. 4  Type of performed Wullstein’s tympanoplasty considering the number of affected sites with cholesteatoma according to the EAONO/JOS
classification

13
 European Archives of Oto-Rhino-Laryngology

significant (p = 0.001, phi = 0.194). The frequency of TPL Ears with one affected site and two presented a residual
IIIb (34.0%) was higher in cases with five affected sites and rate under 10%, while the rest of the cases had a superior
decreased with less affected sites (4.7–23.6%). Overall, one. Regarding the preoperative ABG, statistically signifi-
these differences were statistically significant (p < 0.001, cant differences were seen between cases with one affected
V = 0.239). Concretely, significant differences were found site and three to five, but also between ears with two affected
between cases with one and four (p < 001, phi = 0.283) or sites and four or five and even between three and five sites.
five (p < 001, phi = 0.394) affected sites, as well between Looking at the postoperative ABG, statistically significant
ears with two and five affected sites (p < 0.001, phi = 0.274) differences were found between cases with one affected
and when comparing cases with three and five involved sites site and cases with three to five. Similarly, the differences
(p = 0.002, phi = 0.217). A trend to significance was found regarding the surgical approach were also evident between
between cases with one and three affected sites (p = 0.014*, cases with one affected site and all others, but also between
phi = 0.151), as well between two and four (p = 0.015*, two affected sites and all others. Furthermore, the distri-
phi = 0.171). The number of affected sites also influenced bution of the tympanoplasties according to the Wullstein’s
the frequency how TPL IV was performed (p = 0.012, classification also did not support the distinction into stages I
Cramer’s V = 0.159). Here, cases with one affected site and II. For type I, IIIa and IV there were statistically signifi-
received on average the procedure systematically less often cant differences between cases with one affected site and all
(0.9%) than others (11.5–14.9%). The analysis showed the others, but for instance in type I these differences were also
following significances: comparing cases with one affected present between ears with two affected sites and all others.
site to two (p = 0.001, phi = 0.210), one to three (p = 0.002, As our independent variable was the number of affected
phi = 0.193), one to four (p < 0.001, phi = 0.264) and one to sites, the comparison of our results with the ones of other
five (p < 0.001, phi = 0.290). publications is challenging, but possible to some extent.
James et al. demonstrated a higher residual rate between
stages I and II [6], contrary to Angeli et al. [8] and Ardiç
Discussion et al. [9]. This last publication and the one from Fukuda
et al. [7] correlated a higher disease stage with poorer audio-
513 cases of acquired cholesteatoma, treated between 2010- logical results. However, none of the two studies specified a
01-01 and 2019-07-31, in a single tertiary center, were clas- difference between stage I and II. Moreover, both publica-
sified according to the EAONO/JOS consensus and retro- tions considered an ABG of ≤ 10 or ≤ 20 dB as successful
spectively analyzed. On average an increasing number of outcomes, therefore not allowing a meaningful comparison
affected sites led to higher residual rates, poorer preoperative with our results.
and postoperative ABG and higher complexity regarding the The EAONO/JOS classification discusses recidivism
surgical approach and TPL according to the Wullstein’s clas- as a treatment outcome, which we also acknowledge, as
sification. The effect of surgery on the ABG was independ- extremely important. Nevertheless, the hearing ability can
ent of the number of affected sites, encouraging outcome also be considered as equally relevant. A future revision of
that reflects the excellency of the modern treatment, no mat- the EAONO/JOS system could also feature the hearing abil-
ter how extensive the disease might be. In fact, even if not ity and set indications how to evaluate it, for instance follow-
statistically significant, a worse starting ABG resulted on ing the AAO-HNS recommendations [10].
average in a greater improvement, as the change of the ABG Focusing on our data, grouping cases with one affected
increased with the number of affected sites. site and two in stage I and cases with three to five in stage
This is the first study to focus on the number of affected II would be an option to consider. A group of interna-
sites defined in the EAONO/JOS system (stages I and II) tional surgeons responded to the EAONO/JOS classifica-
as an independent variable, excluding any complications tion with an alternative system called ChOLE [11], which
(stages III and IV). Statistically significant differences were could eventually better correlate with our findings. In
demonstrated between the means of cases with one and two the ChOLE, four disease stages based only on the chole-
to five affected sites, reflecting a difference between stages I steatoma extension were described. In the first two, only
and II. However, these differences were also present between the middle ear is affected, in this aspect similar to the
ears both in stage II, with greater relevance when comparing EAONO/JOS classification. In stage 1a, the epitympanic
cases with two involved sites and cases with three to five, space is affected and in 1b, the epitympanic space and the
questioning the significance of the differentiation between sinus tympani. In stage 2a, the cholesteatoma would extend
stages I and II, the main objective of our study. The confir- further to the attic and antrum and in stage 2b obligatory to
mation of the influence of the number of affected sites on the the supratubal recess, optionally to the protympanum and/
hearing ability and surgical complexity met our secondary or sinus tympani. However, returning to our results, there
objective. were also significant differences between cases with one

13
European Archives of Oto-Rhino-Laryngology

and two affected sites and between ears with three, four approach and postoperative care. For this, further multi-
or five, raising the question if the classification should not center prospective trials are desirable to obtain reliable
always specify which sites are affected. This was suggested results of higher evidence.
in the STAMCO system from a multicenter nationwide
Dutch study group, which intended to further develop the
EAONO/JOS classification [12]. The EAONO/JOS group Conclusion
responded by welcoming this Dutch study, as the progres-
sive evolution of the classification based in international Considering the EAONO/JOS system for cholesteatoma,
consensus is itself the main goal [12]. the data showed that the number of affected sites with dis-
As expected, and considering our data, the number of ease influences the residual rate, hearing outcomes, surgi-
affected sites also correlated with the surgical approach. cal approach and type of Wullstein’s TPL. The results did
Another publication showed that a CWD was performed not support the purposed differentiation in stages I and II.
more often with increasing stage (p < 0.001) [9]. In our pop- We advocate for the further development of the definition
ulation, a statistically significant influence of the number of and classification of the pathology, as well of the uniformi-
affected sites on the performed TPL according to the Wull- zation of the description of the treatment and its outcomes.
stein’s classification was registered, which reflects indirectly This should be based on prospective randomized trials,
the state of the ossicles. Fukuda et al. [7] reported a better always seeking international consensus.
state of the stapes for lower disease stages, without directly
Supplementary Information The online version contains supplemen-
comparing the different stages. The former JOS classifica- tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00405-0​ 23-0​ 7996-w.
tion of 2015 [3], the already mentioned ChOLE [11] and
STAMCO [12] also featured the ossicles state. The IOOG Funding Open Access funding enabled and organized by Projekt
proposed a classification of tympanomastoid surgery, featur- DEAL.
ing for instance the type of intervention and state of different Data availability The data supporting the study are available within
anatomical structures, although not the hearing outcomes the article and supplement.
[13]. So, an international consensus statement could con-
sider all factors that can apply to tympanomastoid surgery Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
in general, like the type of procedure, state and reconstruc- tion, distribution and reproduction in any medium or format, as long
tion of the ossicles, hearing outcomes, among others. The as you give appropriate credit to the original author(s) and the source,
EAONO/JOS system would continue to approach aspects provide a link to the Creative Commons licence, and indicate if changes
exclusively related to cholesteatoma, namely its definition, were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
classification and recidivism. otherwise in a credit line to the material. If material is not included in
The presented study is limited due to the inevitable the article's Creative Commons licence and your intended use is not
bias inherent to a retrospective study design of one single- permitted by statutory regulation or exceeds the permitted use, you will
center. However, when interpreting the results, it should be need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
taken into consideration that only ten experienced surgeons
were involved, as well as the considerably high number of
included patients and the homogeneity of the performed
procedures, according to the institution surgical guidelines.
Another critical point is the follow-up time, with an average
References
of 21.6 months and a minimum of 3. In such a brief period, 1. Tono T, Okamaoto M, Sakagami M, Okuno T, Hinohira Y,
conclusions regarding recurrence rates, if any, should be Mishiro Y (2008) Staging of middle ear cholesteatoma 2008.
drawn with caution. The fact that the patients ages ranged Otol Jpn 18:611–615
from 4 to 82 years should also be pointed out, as cholestea- 2. Tono T, Aoyagi M, Ito T, Okuno T, Kojima H, Hinohira Y
et al (2010) Staging of middle ear cholesteatoma 2010. Otol Jpn
toma differs between children and adults. In fact, a publica- 20:743–745
tion regarding the EAONO/JOS consensus showed that cho- 3. Tono T, Sakagami M, Kojima H, Yamamoto Y, Matsuda K,
lesteatoma is more aggressive and relapses more frequently Komori M et al (2016) Staging and classification criteria for
in pediatric age (defined as < 16 years old) [14]. middle ear cholesteatoma proposed by the Japan Otological
Society. Auris Nasus Larynx. https://​doi.​org/​10.​1016/j.​anl.​2016.​
Our results indicate that the existing classification is cer- 06.​012
tainly useful but may be improved in clinical application 4. Olszewska E, Rutkowska J, Ozgirgin N (2015) Consensus-based
and in reflecting the cases to be treated. This is especially recommendations on the definition and classification of chole-
true, when considering the study objectives chosen here steatoma. J Int Adv Otol 11:81–87. https://​doi.​org/​10.​5152/​iao.​
2015.​1206
and that the classification of acquired cholesteatoma should
be widely used for preoperative preparation, intraoperative

13
 European Archives of Oto-Rhino-Laryngology

5. Yung M, Tono T, Olszewska E et al (2017) EAONO/JOS joint 11 Linder TE, Shah S, Martha AS, Röösli C, Emmett SD (2019)
consensus statements on the definitions, classification and stag- Introducing the ‘“ChOLE”’ classification and its comparison
ing of middle ear cholesteatoma. J Int Adv Otol 13:1–8. https://​ to the EAONO/JOS consensus classification for cholesteatoma
doi.​org/​10.​5152/​iao.​2017.​3363 staging. Otol Neurotol 40(1):63–72. https://​doi.​org/​10.​1097/​
6. James AL, Tono T, Cohen MS, Iyer A, Cooke L, Morita Y, MAO.​00000​00000​002039
Matsuda K, Yamamoto Y, Sakagami M, Yung M (2019) Interna- 12. Merkus P, Ten Tije FA, Stam M, Tan FML, Pauw RJ (2017)
tional collaborative assessment of the validity of the EAONO- Implementation of the “EAONO/JOS definitions and classifica-
JOS cholesteatoma staging system. Otol Neurotol 40(5):630– tion of middle ear cholesteatoma” from STAM to STAMCO. J Int
637. https://​doi.​org/​10.​1097/​MAO.​00000​00000​002168 Adv Otol 13:272–275. https://​doi.​org/​10.​5152/​iao.​2017.​4049
7. Fukuda A, Morita S, Nakamaru Y, Hoshino K, Fujiwara K, 13. Yung M, James A, Merkus P, Philips J, Black B, Tono T et al
Homma A (2019) Short-term hearing prognosis of ossiculo- (2018) International otology outcome group and the international
plasty in pars flaccida cholesteatoma using the EAONO/JOS consensus on the categorization of tympanomastoid surgery. J Int
staging system. J Int Adv Otol 15(1):2–7. https://​doi.​org/​10.​ Adv Otol 14(2):216–226. https://​doi.​org/​10.​5152/​iao.​2018.​5553
5152/​iao.​2019.​5983 14. Lima AF, Moreira FC, Menezes AS, Costa IE, Azevedo C, Breda
8. Angeli S, Shahal D, Brown CS, Herman B (2020) Predicting MS, Dias L (2020) Is pediatric cholesteatoma more aggressive in
recidivism for acquired cholesteatoma: evaluation of a current children than in adults? A comparative study using the EAONO/
staging system. Otol Neurotol 41(10):1391–1396. https://​doi.​ JOS classification. Int J Pediatr Otorhinolaryngol 138:110170.
org/​10.​1097/​MAO.​00000​00000​002823 https://​doi.​org/​10.​1016/j.​ijporl.​2020.​110170
9. Ardiç FN, Mengi E, Tümkaya F, Kara CO, Bir F (2020) Correla-
tion between surgical outcome and stage of acquired middle ear Publisher's Note Springer Nature remains neutral with regard to
cholesteatoma: revalidation of the EAONO/JOS staging system. J jurisdictional claims in published maps and institutional affiliations.
Int Adv Otol 16(1):34–39. https://​doi.​org/​10.​5152/​iao.​2020.​7598
10. Gurgel RK, Jackler RK, Dobie RA, Popelka GR (2012) New
standardized format for reporting hearing outcome in clinical
trials. Otolaryngol Head Neck Surg 147(5):803–807. https://​doi.​
org/​10.​1177/​01945​99812​458401

13