Cholesteatoma is a common health problem in Europe with a yearly incidence of about 3/100,000 in children and 9.2/100,000 in adults. Most cases are acquired, while only 4% of cholesteatomas in children are congenital.1,2 Acquired cholesteatoma results from a retraction pocket, from invasion of squamous cell epithelium from the outer ear canal3 or as a consequence of prior surgery or infection.4 Cholesteatoma generally requires complete surgical removal which is done by Canal-Wall-Up (CWU) or Canal-Wall-Down (CWD) surgery. In CWD procedures the posterior wall of the auditory canal is partially removed, usually allowing a good visualization of the middle ear. Furthermore, the detection of recurrences in the follow-up period should be facilitated. However, CWD surgery often results in reduced earwax and dander cleaning and subsequent recurrent cavity infections with otorrhea.5 In CWU procedures the posterior auditory canal wall remains intact, on the one hand accelerating postoperative healing and reducing the need of frequent cleaning of the outer ear canal. On the other hand, the intraoperative visualization for complete cholesteatoma removal might be limited and it might be more difficult to identify residual disease.5 CWD surgery can be combined with a reconstruction of the previously removed parts of the posterior auditory canal wall. With this approach a good intraoperative overview should be guaranteed and the postoperative problems from a classic CWD surgery should be avoided.

Cholesteatomas bear a relevant risk of residual disease. In the literature rates of residual cholesteatoma considerably vary and range from 12.3% to 67.0%, mainly due to inconsistent study designs.6–14 Despite the variety of trials, various risk factors for residual cholesteatoma seem to have been identified, including ossicular erosion,6,10–12,15 growth toward the sinus tympani10,11,13 or the posterosuperior region,6,8 extended size of cholesteatoma,11,15,16 inexperience of the surgeon11 and cholesteatoma in children.13,17 Even though CWD procedures seem to have a lower recurrence rate Mishiro et al. proposed an individual decision depending on patient and cholesteatoma extension whether to go with CWD, CWU or CWD with reconstruction of the posterior auditory canal wall.18 Patients with a high risk of residual disease usually undergo a second-look surgery. However, the recommendation for this operation some months after cholesteatoma removal remains a matter of controversy. On the one hand, it offers the possibility of an early detection of residual disease, while on the other hand, it bears the risk that patients are exposed to the danger of a second surgery and general anesthesia for no benefit. In this context, great efforts have been made in recent years to improve sensitivity and specificity of radiological follow-up examinations, e.g., diffusion-weighted MRI scans.19 However, these radiological scans are relatively cost-intensive. The recommendation of a postoperative radiological control or a second-look operation is still based on the surgeon's impression during cholesteatoma removal as to date there exist no guidelines. To this end, we wanted to identify anamnestic, clinical, and surgery-related risk factors for residual cholesteatoma by a follow-up investigation of all patients having undergone a second-look operation or revision tympanoplasty after initial cholesteatoma removal in our department. Including both children and adults, our study design also allows the comparison of residual cholesteatoma rates in patients of different age.

The only effective therapy for cholesteatoma is complete surgical removal. However, involuntarily remaining cholesteatoma tissue leads to an early recurrence of the disease.18 Different risk factors have been described in the literature. Unfortunately, studies are hard to compare as they were conducted in different countries with different surgical standards and surgical concepts and partially have observation periods reaching back to the 1970ies. Furthermore, most previous studies only investigated infantile cholesteatomas. Therefore, our current retrospective study included children as well as adults in order to identify risk factors for residual cholesteatoma irrespectively of the patients’ age. Risk factors were subdivided into preoperative, intraoperative, and postoperative parameters referring to the first surgery.

Spilsbury et al. has recently documented on an impressively high number of 45,980 children that a history of ventilation tubes was associated with a higher risk of initial cholesteatoma formation.4 This is in line with the relatively high number of patients (about one third) with previous ventilation tube insertion in our current study. However, we could not identify previous ventilation tubes as a significant risk factor for residual disease. The decision in favor or against the insertion of ventilation tubes and the indication of a semi-prophylactic tympanoplasty due to an adhesive otitis media vary considerably among different countries and even among different surgeons, making a comparison of different studies difficult.21 Additionally, our study involved patients independent of their age. Consequently, we also analyzed many cholesteatomas in adults where usually the incidence of a previous ventilation tube insertion is lower than in an infantile cohort. In contrast, previous adenoidectomy came out as a statistically significant risk factor for residual cholesteatoma in our current study. A possible explanation is the risk of scar formation after adenoid removal which could lead to a chronic ventilation disorder and consequently to an atrophy of the tympanic membrane, an adhesive otitis media, and more difficult surgical conditions in initial surgery.22 Furthermore, adenoidectomy is usually performed in cases of chronic ventilation problems with recurrent otitis and formation of retraction pockets thereby presumably reflecting the increased risk of cholesteatoma formation in already compromised middle ears. The intraoperative extension of the cholesteatoma to the sinus tympani, the antrum, and the mastoid was associated with a statistically higher risk of residual disease in accordance to the literature.6,8,10,11,13,14 This seems to be self-explanatory as the larger a cholesteatoma is, the greater is the risk of residual cholesteatoma tissue. Furthermore, the visualization of the middle ear compartments affected by the cholesteatoma can be insufficient in advanced disease. The hypothesis that the erosion of ossicles is a risk factor for recurrence,6,10–12,15 can neither be approved nor disproved by our data as there were few patients with a completely intact ossicular chain. This phenomenon is due to the specific patient population of a tertiary-referral ENT hospital where predominantly patients with an extended cholesteatoma were treated. The different techniques of cholesteatoma removal have also been discussed as being associated with different rates of residual disease. We found similar rates of residual cholesteatoma in cases of CWD surgery (33%) and CWD+R 2 ways surgery (41%), while all other surgical techniques were associated with higher rates of residual disease. Recently, many authors have stated that the CWU surgery is not a relevant risk factor for residual cholesteatoma.6,10–13,23 In the present study, however, the CWD+R 2 ways-approach was clearly superior to the CWU technique in avoiding residual cholesteatoma, except in cases of small epitympanic cholesteatomas. These small cholesteatomas could be sufficiently resected without partial removal of the external ear canal wall. However, the number of patients with limited cholesteatoma growth in our patient population was too small to allow a valid statistical analysis. All in all, cholesteatoma removal should be adapted to its extension and a rather aggressive removal seems favorable regarding residual disease. On the one hand, a consecutive reconstruction of the previously removed parts of the posterior wall of the auditory canal in CWD+R 2 ways approaches does not seem to be disadvantageous regarding residual cholesteatoma, although there is in principle an imminent risk of burying residual epithelia behind the inserted cartilage. On the other hand, it restores the natural anatomy of the external auditory canal, thus avoiding the problems coming along with extensive postoperative cavities. When compared to the rates of residual disease in literature, which are usually indicated as lowest with 4–25% after CWD surgery, with 9–70% after CWU surgery, and with 10–27% after reconstruction of the posterior auditory canal wall, if named separately at all,7–10,12,13,15,16,24,25 the rates of residual cholesteatoma in this study were generally higher. This can be explained by the high number of extended cholesteatomas in our study, reflected by the high rate of patients undergoing CWD+R surgery (about 60.0%) and the low number of patients receiving a CWU operation (about 30%). Furthermore, this might also be an effect of the selected study design. In most studies, patients who are not operated for a second time are considered as free of cholesteatoma while in this study we evaluated only patients who were effectively operated for a second time, due to either a recommendation for a second-look surgery or a suspicion of residual disease during follow-up, thus already representing a group of high risk patients. Another interesting approach for consideration is endoscopic cholesteatoma removal. As for now, there seem to be promising results regarding residual or recurrent pathology for this minimally invasive treatment.26

The type of tympanoplasty after cholesteatoma removal did not have any influence on the rate of residual cholesteatoma. For ossicular chain reconstruction the rates of residual disease were slightly higher for autologous implants when compared to alloplastic material. With respect to the comparable hearing results27 the re-implantation of auto-ossicles after cholesteatoma removal should therefore be performed restrictively.

Both the postoperative retraction and perforation of the upper quadrants of the reconstructed tympanic membrane were identified as risk factors for residual disease, which is in accordance to the literature.13,28 Otorrhea was often present in relapses,13,28 but usually together with a re-perforation of the eardrum and not as an isolated symptom. Although progressive postoperative hearing loss was identified as an independent risk factor in the present study, it is self-evident that hearing impairment should result in an accurate inspection of the eardrum. Importantly, stable postoperative hearing and especially a Type A tympanogram indicated a reduced probability of residual cholesteatoma, presumably reflecting normal middle ear ventilation.

In the present study children and adults had comparable rates of residual cholesteatoma (60.0% vs. 55.4%). The current literature has not been able to determine to date if age is really an independent risk factor for residual cholesteatoma.13,17,29 Histopathologically, there is no difference between cholesteatomas in adults and children,30 but cholesteatomas are considered to have a larger extension17 and to be more often super-infected in children, thus mimicking aggressive growth.31 Despite the often delayed diagnosis, we did not identify congenital cholesteatomas as a statistically significant risk factor for residual disease in this study. This is in accordance with the recent literature8,25,32 and might be due to the fact that children with congenital cholesteatomas usually do not have chronic ventilation problems and thus are not affected by an atrophy of the tympanic membrane and the associated difficult surgical conditions at initial surgery. In recent years, with advances in modern imaging, the second-look surgery is more and more a topic of controversial discussion.33 In several studies imaging was compared with intraoperative findings. As a result, both initially diagnosed cholesteatomas and residual disease can be detected with the aid of non-echo planar diffusion weighted magnetic resonance imaging (and computed tomography) with high sensitivity and specificity.19,34,35 In order to avoid financial burdens caused by the expense of modern sectional image analysis, it is crucial to identify parameters which suggest high rates of residual disease in order to filter out those patients who really need a follow-up with imaging and/or second-look surgery. In this context we developed a score by which residual disease can be predicted with relatively high sensitivity and specificity. The score, including the intraoperative extension of the cholesteatoma to the sinus tympani and/or the antrum, the development of a postoperative retraction or perforation of the reconstructed tympanic membrane, and the type of postoperative tympanogram, provides the opportunity to assign each patient to a group with low or high risk of residual cholesteatoma. On the one hand, the score is quite simple, allowing unpretentious application in clinical routine. On the other hand, it highlights the great clinical importance of post-operative retraction pocket formation as a risk factor for residual cholesteatoma, as patients with post-operative retractions alone are statistically prone to residual disease. We suggest that “high-risk” patients, as identified by the score, should undergo sectional imaging and, in uncertain cases, consecutive second-look surgery.

The only treatment of cholesteatomas remains surgical removal, mostly combined with a consecutive reconstruction of the hearing apparatus. However, cholesteatomas tend to relapse. This study aimed at identifying anamnestic, clinical, and surgery-related risk factors for residual disease. Cholesteatoma extension to the sinus tympani and/or the antrum and a retraction or perforation of the reconstructed tympanic membrane were strongly associated with residual cholesteatoma, while a regular type A tympanogram makes residual disease unlikely. For extended cholesteatomas the CWD+R 2 ways approach seems to be safe without the disadvantages of an extensive postoperative cavity, thus being more convenient for the patients concerned. We developed a score including the intraoperative extension of the cholesteatoma, the type of the postoperative tympanogram and the postoperative retraction and perforation of the tympanic membrane to assess the risk of residual cholesteatoma. This score might be more objective than the personal recommendation of a surgeon after initial cholesteatoma removal and should be used to filter out high-risk patients needing further diagnosis as well as regular clinical follow-up.

The authors declare no conflicts of interest.