Congenital cholesteatoma: clinical features and growth patterns☆,☆☆

Hyun Woo Lim, MDa, Tae Hyun Yoon, MDa,⁎, Woo Seok Kang, MDb

aDepartment of Otolaryngology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South KoreabDepartment of Otolarnygology, Gangneung Asan Hospital, University of Ulsan College of Medicine, Gangneung, South Korea

Received 11 October 2011

Abstract Objective: The clinical features and patterns of growth, relative to age, were examined in patientswith congenital cholesteatoma to investigate disease progression and site of origin.Patients and methods: We retrospectively reviewed 72 children younger than 15 years withintraoperatively confirmed congenital cholesteatoma diagnosed using the inclusion criteria ofLevenson et al. Patient demographics, history, and otoscopic and operative findings were evaluated.Correlation between mass volumetric and operation age was analyzed in patients with closed-typemasses. Parameters of disease extent of middle ear quadrant, ossicular erosion, attic involvement,invasion of the mastoid cavity, and type of mass were assessed relative to age at operation.Results:Mean patient age was 63.1 months (range, 20–179 months), with 51 patients (70.8%) beingasymptomatic and diagnosed incidentally. Nineteen patients (26.4%) had closed-type cystic masses,and 53 (73.6%) had open-type lesions. The volume of closed-type masses was linearly correlatedwith age at operation. Logistic regression showed that increased age at operation was associated witha lower proportion of anterosuperior quadrant lesions, resulting in uncertainty about the site oforigin. The possibility of open-type masses also increased according to age at operation.Conclusion: Congenital cholesteatoma shows growth and extension over time. Early detection andintervention are necessary to avoid advanced disease.© 2012 Elsevier Inc. All rights reserved.

1. Introduction

Congenital cholesteatoma, defined as a whitish masslesion in the middle ear cleft behind an intact tympanicmembrane early in life, is a relatively rare disease. The termcongenital has been used rather conventionally becausethe pathogenesis of congenital cholesteatoma remainsunclear and suggested hypotheses range from formationduring fetal period to an acquired condition in infant [1].These lesions might have been associated with a historyof acute otitis media, otitis media with effusion, briefotorrhea, or myringotomy [2]. Early diagnosis and inter-

vention have been emphasized because late detection maybe associated with extensive disease [3-5]. More advanceddisease has been observed in older children; and recurrencehas been associated with advanced disease, suggesting thatlesion growth and disease progression are associated withage [3,4]. Passage of time influences not only patientoutcome but understanding of the nature of the disease.Wider lesion extent and infiltration in patients withadvanced disease can result in uncertainly about the siteof origin [6,7].

Although age at operation has been suggested as animportant factor influencing treatment outcome and the ideathat the congenital cholesteatoma grows over time looksquite clear, the effect of aging on congenital cholesteatomagrowth has not been sufficiently demonstrated so far. Thus,the aim of this study was to present how congenitalcholesteatoma changes in size, location, and type accordingto the time passage. We therefore assessed the patterns ofmass growth relative to patient age to support theunderstanding of disease progression and site of origin.

Available online at www.sciencedirect.com

American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 538–542www.elsevier.com/locate/amjoto

☆ Financial disclosure: none.☆☆ Conflict of Interest: none.⁎ Corresponding author. Department of Otolaryngology, Asan Medical

Center, University of Ulsan College of Medicine, 86 Asanbyeongwon-gil,Songpa-gu, Seoul 138-736, South Korea. Tel.: +82 2 3010 3713; fax: +82 2489 2773.

E-mail address: thyoon@amc.seoul.kr (T.H. Yoon).

0196-0709/$ – see front matter © 2012 Elsevier Inc. All rights reserved.doi:10.1016/j.amjoto.2012.01.001

The clinical features of patients with congenital cholestea-toma were also investigated in this study.

2. Subjects and methods

We retrospectively reviewed the medical records of 72consecutive children younger than 15 years with surgicallyconfirmed congenital cholesteatoma, who underwent oper-ations by a single surgeon from January 1999 to September2010. Our study was approved by the Institutional ReviewBoards of Asan Medical Center. Patients were diagnosedusing the inclusion criteria of Levenson et al [8], whichincluded the presence of a whitish middle ear mass behindan intact tympanic membrane, absence of tympanicmembrane perforation, and no previous otologic pro-cedures, such as ventilation tube insertion. Patients with ahistory of acute otitis media, otitis media with effusion,brief otorrhea, or myringotomy were not excluded.Demographic characteristics examined included age atoperation, sex, and chief complaint at the time of initialdiagnosis. Tympanic membranes were examined withototelescopes before surgery; and intraoperative findingsincluded the location of each mass in the middle ear, theextent of disease, and the erosion of ossicles. Masses wereclassified as either open (flat keratinizing epitheliumwithout formation of an epithelial cyst or pearl) or closed(epithelial cyst without exposure of keratin) [9]. Volumetricestimation of each closed-type cholesteatoma was per-formed on 0.6 mm-thickness high-resolution computedtomography images rather than by direct measurement inthe operation fields because the mass was often ruptured orits shape could change during the procedures. Under 6-foldmagnification, mass volume was calculated on the axialimage by volumetric conversion after measuring the areaon each section. The relationship between mass volume andage at operation was analyzed in closed-type cholesteato-mas by nonparametric correlation analysis. Logisticregression analysis was used to assess mass type, location,and extent of disease relative to age at operation for bothtypes of cholesteatoma.

3. Results

3.1. Demographic features and otoscopic findings

The characteristics of the 72 patients are summarized inTable 1. There were 54 boys (75.0%) and 18 girls (25.0%),of mean age 63.1 months (range, 20–179 months), with thelargest number being 4 to 5 years of age (Fig. 1). Fifty-one(70.8%) patients were asymptomatic at the time of diagnosis.The diagnosis was made incidentally in primary care clinicsduring routine tympanic membrane checkup for otherdiseases, mostly common cold. Ten patients (13.9%) visitedfor hearing disturbance, and 9 (12.5%) were referred forprolonged middle ear effusion. In terms of otoscopic

findings, a whitish mass beyond an intact tympanicmembrane was found in 68 (94.6%) patients. Thirteenpatients (18.1%) also had middle ear effusion. Four (5.6%)patients with conductive hearing loss had normal tympanicmembranes on otoscopic examination. They were includedbecause we found definite epithelial pearls during explor-ative surgery for conductive hearing loss.

3.2. Operative findings

When we assessed the location of each mass, asdetermined by the quadrant of the middle ear, all patientsshowed the presence of mass in at least one of anterosuperior(AS) and posterosuperior (PS) quadrants except 1 patientwho had no middle ear lesion (Fig. 2). This 11-year-old boywith conductive hearing loss showed no cholesteatoma inmiddle ear cavity, and operation revealed open-typecholesteatoma in his mastoid antrum with erosion of incusand stapes. We found that 21 patients (29.2%) had lesions inthe AS quadrant, 17 (23.6%) had lesions in the PS quadrant,and 33 (45.8%) showed involvement of both AS and PSquadrants (Table 2). None of patients had lesions limited toanteroinferior (AI) or posteroinferior (PI) quadrant. Weobserved ossicular erosion in 44 patients (61.1%), attic

Table 1Demographic features and otoscopic findings of 72 patients with congenitalcholesteatoma

Demographics No. of patient (%)

SexMale 54 (75.0%)Female 18 (25.0%)Chief complaintIncidental 51 (70.8%)Hearing disturbance 10 (13.9%)Prolonged middle ear effusion 9 (12.5%)Otalgia 1 (1.4%)Facial paralysis 1 (1.4%)Otoscopic findingsWhitish mass 55 (76.4%)Whitish mass with effusion 13 (18.1%)Normal 4 (5.6%)

Fig. 1. Age distribution of 72 patients with congenital cholesteatoma.

539H.W. Lim et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 538–542

involvement in 41 (56.9%), and mastoid invasion in 22(30.6%). All 44 cases showing ossicular erosion had incudalerosion, and 36 (50.0%) patients also had stapedial erosion.Based on morphologic examination at surgery, 19 patients(26.4%) had closed-type cystic masses and 53 (73.6%) hadopen-type lesions. Of the 19 patients with closed-typemasses, 15 had masses confined to the AS and AI quadrants,3 had masses confined to the PS and PI quadrants, and 1 hada mass along both the AS and PS quadrants. Masses wereremoved from 27 patients (37.5%) by tympanotomy withoutmastoidectomy. Forty-five patients (62.5%) underwentmastoidectomy, including one who underwent a canal walldown procedure for extensive cholesteatoma causingunilateral facial paralysis. Mastoidectomy was performedin 36 of 36 patients with stapedial erosion, 40 of 41 with atticinvolvement, and 22 of 22 with mastoid invasion. Limitedattic lesion in 1 patient was completely managed with

atticotomy procedure and attic reconstruction with conchalcartilage. Operative findings are summarized in Table 2.

3.3. Mass dimensions and disease extent

Fig. 3 shows scatter plots of mass volumes relative topatient age at operation. The volumetric size of closed-typemasses was linearly correlated with age at operation(Spearman ρ = 0.653, P value = .01). To evaluate thechange in disease extent over time, we performed logisticregression analysis using variables related to disease extent,including presence in each middle ear quadrant, ossicularerosion, attic involvement, invasion of the mastoid cavity,and presence of an open-type mass (Table 3). We found thatthe presence of an open-type mass (odds ratio [OR] = 1.529,P value = .026) was significantly associated with age atoperation, whereas the presence of a lesion in the ASquadrant lesion (OR = 0.784, P value = .011) significantlydecreased in proportion to age. Ossicular erosion (P value =.061) and the presence of a lesion in the PS quadrant (Pvalue = .162) tended to increase with age, but thecorrelations were not statistically significant.

4. Discussion

We found that the size of closed-type congenitalcholesteatoma tended to increase with age at the time ofsurgery. The linear correlation between mass volume andage suggests that closed-type cystic congenital cholestea-tomas grow steadily and proportionately over time. In

Fig. 2. Ototelescopic findings of an AS quadrant lesion (A) and a PSquadrant lesion (B).

Table 2Disease location and extension on operation findings

Operation findings No. of patients (%)

LocationAS quadrant 21 (29.2%)PS quadrant 17 (23.6%)AS and PS quadrants 33 (45.8%)No middle ear lesion 1 (1.4%)

Ossicular erosionIncus 44 (61.1%)Stapes 36 (50.0%)Malleus 6 (8.3%)None 28 (38.9%)

Mastoid involvementYes 22 (30.6%)No 50 (69.4%)

Type of massClosed 19 (26.4%)Open 53 (73.6%)

Operation methodTympanotomy only 27 (37.5%)Transmeatal approach 21Postauricular approach 6Tympanotomy with mastoidectomy 45 (62.5%)Canal wall up mastoidectomy 44Canal wall down mastoidectomy 1

Fig. 3. The relationship between mass volume and age at operation in closed-type congenital cholesteatoma (Spearman ρ, r = 0.653, P value = .01).

540 H.W. Lim et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 538–542

addition, this study statistically revealed that aging isassociated with increased proportion of the open-type massand relatively decreased proportion of AS quadrant lesionsin mass location.

Several hypotheses have been presented on the patho-genesis of congenital cholesteatoma. Pathogenesis is difficultto determine because the incidence of congenital cholestea-toma is low and direct examination of the whole middle earsof very young, asymptomatic infants is not possible. Therehave also been questions about whether this disease is trulycongenital or acquired. It was suggested that ectodermaltissue from the external acoustic meatus may migrate into themiddle ear cavity due to failure of the inhibitory function ofthe tympanic ring [10]. The epidermoid formation hypoth-esis is based on a finding of cell rest of squamous epitheliumin the fetal middle ear and has been accepted as the mostamenable pathogenesis hypothesis [1,11]. However, inva-sion of the squamous epithelium has been observed afteracquired inflammatory injury, and cholesteatoma duringearly childhood has been described as an acquired inclusionof keratinized squamous epithelium [12,13]. We observedthe linear volume increase in closed-type masses in thisstudy. Although it does not directly support any one of thepathogenesis theories, it may suggest that congenitalcholesteatomas could be present around the very earlyinfancy at least in closed-type lesions.

Most congenital cholesteatomas were reported to arise inthe AS quadrant and grow into the PS quadrant, erodingossicles and finally invading the mastoid [14]. This processmay represent the natural history of congenital cholestea-toma, as shown by serial illustrations of disease extent in astudy with 34 patients [15]. That study, however, did notreport patient ages at operation and did not completelyexclude the possibility of extension in the opposite direction(from PS to AS) because 29 of 35 patients had extensivelesions in several middle ear locations. Other studies havealso reported that the AS quadrant is the most likely origin ofcongenital cholesteatoma [4,5], whereas some have foundthat these lesions originate from posterior quadrants [6,16].Many cases of the latter lesions, however, were quiteextensive; and many of those patients were much older than

those in reports favoring origin from anterior quadrants. Theorigin of disease may be uncertain in many patients withadvanced disease, suggesting that the location of originshould be determined in patients with early disease.

Of our 19 patients with closed-type masses, 15 patientshad anteriorly confined lesions (AS and AI quadrants), 3 hadposteriorly confined lesions (PS and PI quadrants), and only1 showed a cystic mass along both the AS and PS quadrants.Because the extent of disease clearly differed betweenpatients with anteriorly and posteriorly confined lesions, thesite of origin may be both the AS and PS quadrants. Mostearly closed-type lesions seemed to arise from the ASquadrant, although some could also originate from the PSquadrant. However, if the type of mass was not limited toearly closed-type cases, the frequencies of AS and PSquadrant lesions were similar.

This discrepancy in frequency of disease origin may beexplained by the results of logistic regression analysisbetween age at operation and extent of disease. We foundthat the presence of an AS quadrant lesion decreasedsignificantly with patient age, whereas the presence of a PSquadrant lesion tended to increase without significance.Thus, the predominance of AS quadrant lesions in somereports may be due to undetected PS quadrant lesions inyounger patients. In contrast, reports suggesting thepredominance of PS quadrant lesions may involve toomany patients with advanced lesions in the PS quadrant.The timing of detection can therefore change thedistribution of disease location and make disease originuncertain. The origin of congenital cholesteatoma may bedetermined only by a wide screening of middle ears inlarge numbers of infants.

Open-type congenital cholesteatoma has been found tocorrelate with more severe and advanced disease [6,7,9,17],and age has been reported to correlate with degree of diseaseinfiltration [3,5]. In this study, logistic regression analysisshowed that open-type lesions increased significantly withage. We found that age affected the status of congenitalcholesteatoma, not only the size and extent of disease,although it is not clear whether the initial closed-type massruptures changing into the open form or new open-typelesions develop more frequently with aging.

5. Conclusion

To evaluate the pattern of lesion growth, we analyzedpatients by age at operation. We found that lesion sizeincreased linearly with age, as did the likelihood of anopen-type lesion. These findings emphasize the importanceof early detection and management of congenital choles-teatoma to avoid advanced disease. In addition, we foundthat the proportion of patients with lesions in the ASquadrant decreased according to age, suggesting that age atpresentation should be considered when evaluating theorigin of disease.

Table 3The logistic regression analysis between disease extent variables and age atoperation

Parameters No. (%) ofpatients

OR 95% Confidenceinterval

P value

AS quadrant lesion 54 (75.0%) 0.784 0.65–0.95 .011⁎

AI quadrant lesion 37 (51.4%) 0.989 0.84–1.16 .891PS quadrant lesion 50 (69.4%) 1.172 0.94–1.47 .162PI quadrant lesion 41 (56.9%) 0.981 0.83–1.14 .814Ossicular erosion 44 (61.1%) 1.236 0.99–1.54 .061Attic involvement 41 (56.9%) 1.056 0.89–1.25 .526Mastoid invasion 22 (30.6%) 1.125 0.95–1.33 .173Open-type mass 53 (73.6%) 1.529 1.05–2.22 .026⁎

Asterisks mean statistical significance with P value b .05.

541H.W. Lim et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 538–542

References

[1] Persaud R, Hajioff D, Trinidade A, et al. Evidence-based review ofaetiopathogenic theories of congenital and acquired cholesteatoma.J Laryngol Otol 2007;121:1013-9.

[2] Bennett M, Warren F, Jackson GC, et al. Congenital cholesteatoma:theories, facts, and 53 patients. Otolaryngol Clin North Am 2006;39:1081-94.

[3] Lazard DS, Roger G, Denoyelle F, et al. Congenital cholesteatoma:risk factors for residual disease and retraction pockets—a report on 117cases. Laryngoscope 2007;117:634-7.

[4] El-Bitar MA, Choi SS, Emamian SA, et al. Congenital middle earcholesteatoma: need for early recognition—role of computed tomog-raphy scan. Int J Pediatr Otorhinolaryngol 2003;67:231-5.

[5] Potsic WP, Korman SB, Samadi DS, et al. Congenital cholesteatoma:20 years' experience at The Children's Hospital of Philadelphia.Otolaryngol Head Neck Surg 2002;126:409-14.

[6] Kojima H, Tanaka Y, Shiwa M, et al. Congenital cholesteatomaclinical features and surgical results. Am J Otolaryngol 2006;27:299-305.

[7] Richter GT, Lee KH. Contemporary assessment and management ofcongenital cholesteatoma. Curr Opin Otolaryngol Head Neck Surg2009;17:339-45.

[8] Levenson MJ, Michaels L, Parisier SC. Congenital cholesteatomas ofthe middle ear in children: origin and management. Otolaryngol ClinNorth Am 1989;22:941-54.

[9] McGill TJ, Merchant S, Healy GB, et al. Congenital cholesteatoma ofthe middle ear in children: a clinical and histopathological report.Laryngoscope 1991;101:606-13.

[10] Aimi K. Role of the tympanic ring in the pathogenesis of congenitalcholesteatoma. Laryngoscope 1983;93:1140-6.

[11] Michaels L. An epidermoid formation in the developing middle ear:possible source of cholesteatoma. J Otolaryngol 1986;15:169-74.

[12] Ruedi L. Cholesteatoma formation in the middle ear in animalexperiments. Acta Otolaryngol 1959;50:233-40 [discussion 40-2].

[13] Tos M. A new pathogenesis of mesotympanic (congenital) cholestea-toma. Laryngoscope 2000;110:1890-7.

[14] Nelson M, Roger G, Koltai PJ, et al. Congenital cholesteatoma:classification, management, and outcome. Arch Otolaryngol HeadNeck Surg 2002;128:810-4.

[15] Koltai PJ, Nelson M, Castellon RJ, et al. The natural history ofcongenital cholesteatoma. Arch Otolaryngol Head Neck Surg 2002;128:804-9.

[16] Inokuchi G. Congenital cholesteatoma: posterior lesions and thestaging system. Annal Otol Rhinol Laryngol 2010;119:5.

[17] Choi HG, Park KH, Park SN, et al. Clinical experience of 71 cases ofcongenital middle ear cholesteatoma. Acta Otolaryngol 2010;130:62-7.

542 H.W. Lim et al. / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 33 (2012) 538–542

©2012 Elsevier