Congenital Aural Stenosis: Clinical Features and Long-term Outcomes

The aim of the present study was to comprehensively evaluate the clinical features and long-term outcomes of congenital aural stenosis (CAS). This study presents a retrospective review of patients who underwent meatoplasty for CAS at a tertiary referral hospital from 2008 to 2015. A total of 246 meatoplasty procedures were performed on 232 patients in the present study. We performed multivariate regression analysis. Except in the age < 6 years group, no significant difference was observed among different age groups for cholesteatoma formation, p > 0.05. Except for the stenosis of the external auditory canal (EAC) (>4 mm) group, the other stenosis of EAC groups were not associated with cholesteatoma formation, p > 0.05. Postoperative air-bone gaps (ABG) less than 30 dB occurred in 77.3% (99/128) of the patients, and the Jahrsdoerfer score was associated with postoperative ABG, p < 0.001. The complication rate of CAS was 13.8% (20/144), and males showed a higher risk for postoperative complications (OR, 6.563; 95% CI, 1.268–33.966, p = 0.025). These results indicate that meatoplasty was an effective surgical intervention for CAS, showing a stable hearing outcome with prolonged follow-up. There was no significant difference between the cholesteatoma and no cholesteatoma groups for hearing outcomes, p > 0.05.

cholesteatomas and 85 females with 39 cholesteatomas; 140 ears were affected on the right side with 90 cholesteatomas, and 106 ears were affected on the left side with 46 cholesteatomas; 57.3% (78/136) of the cholesteatoma patients also presented with infection, and 4.5% (5/110) of the patients without cholesteatoma presented with infection. To analyse the relationship between age and cholesteatoma formation, we divided the patients into four groups: age < 6, 6 ≤ age ＜ 12, 12 ≤ age ＜ 18 and age ≥ 18 years. Nearly half of the patients underwent surgery at 6 to 12 years old. A total of 32 meatoplasty procedures were performed in patients less than 6 years old. Except for 3 ears without cholesteatoma, the remaining 29 ears underwent surgery for cholesteatoma (Table 1). We conducted a multivariate logistic regression analysis for cholesteatoma formation. The results showed that males (OR, 2.401; 95% CI, 1.054-5.469, p = 0.037), the right side group (OR, 3.531; 95% CI, 1.618-7.706, p = 0.002); individuals aged < 6 years (OR, 6.494; 95% CI, 1.263-33.396, p = 0.025), and the presence of infection (OR, 24.946; 95% CI, 8.432-73.803, p < 0.001) all showed a higher risk for cholesteatoma formation. Other age groups and degrees of microtia were not associated with cholesteatoma formation, p > 0.05.
HRCT data. A total of 207 HRCT DICOM datasets were imported into MIMICS, including 94 ears without cholesteatoma and 113 ears with cholesteatoma. The mean stenosis of EAC at surgery was 3.1 ± 1.2 mm (range: 1-9 mm) (Fig. 1). A significant difference was observed between the cholesteatoma and no cholesteatoma groups for the stenosis of EAC using an unpaired t test, p < 0.01. A chi-square test also detected a significant difference among subgroups of stenotic EAC for cholesteatoma formation, χ 2 = 12.4, p < 0.05, whereas the cholesteatoma group might exhibit a larger stenotic EAC reflecting the erosion of cholesteatoma. Among the 21 patients with cholesteatoma, individuals with stenosis of EAC (> 4 mm) and a defective bony canal were observed. When we removed this group, there was no significant difference between the cholesteatoma and no cholesteatoma groups for the stenosis of EAC using an unpaired t test, p > 0.05, and the chi-square test detected no significant difference among the subgroups of stenotic EAC for cholesteatoma formation, χ 2 = 2.5, p > 0.05. We conducted a multivariate logistic regression analysis for cholesteatoma formation. The results also show that the stenosis of EAC (> 4 mm) group was associated with cholesteatoma formation (OR, 5.337; 95% CI, 1.444-19.735, p = 0.012), but the other stenosis of EAC groups were not associated with cholesteatoma formation, p > 0.05. These differences indicated that the stenosis of EAC group (> 4 mm) with cholesteatoma might represent a special EAC state.
The mean Jahrsdoerfer score at surgery was 8.1 ± 1.6 (range: 4-10); the Jahrsdoerfer score for the no cholesteatoma group was 8.6 ± 1.2, and for the cholesteatoma group was 7.7 ± 1.7 (Table 2). A significant difference was detected between the cholesteatoma and no cholesteatoma groups in the Jahrsdoerfer scores using an unpaired t test, p < 0.01. For patients with more severe stages of EAC cholesteatoma, the Jahrsdoerfer score decreased, reflecting the erosion of cholesteatoma.
Hearing outcomes. A total of 207 patients showed preoperative PTA, and 128 patients showed postoperative PTA. A total of 103 patients had complete data, including HRCT, preoperative and the latest postoperative PTA. Fifty-seven ears without cholesteatoma and 46 ears with cholesteatoma were analysed for hearing improvement (Table 3). There was no significant difference between the cholesteatoma and no cholesteatoma groups using an unpaired t test, including Δ ABG, preoperative and postoperative ABG, p > 0.05.
We conducted a multivariate linear regression analysis for postoperative ABG. The results showed that the Jahrsdoerfer score was associated with postoperative ABG, p < 0.001; other factors, such as gender, age, laterality, presence of infection, cholesteatoma and stenosis of EAC, were not associated with postoperative ABG, p > 0.05. We also conducted a multivariate linear regression analysis for Δ ABG. The results showed that the female group had better results for Δ ABG, p = 0.032; other factors, such as age, laterality, presence of infection, cholesteatoma, stenosis of EAC and Jahrsdoerfer score, were not associated with Δ ABG, p > 0.05. A latest postoperative ABG less than 30 dB was observed in 77.3% (99/128) of the CAS cases, and an ABG less than 10 dB was observed in 16.4% of the CAS cases (21/128) ( Table 4). There was no significant difference between the cholesteatoma and no cholesteatoma groups in ABG < 30 dB (χ 2 = 3.1, p > 0.05) or ABG < 10 dB (χ 2 = 0.2, p > 0.05). To evaluate the stability of the hearing outcomes, a total of 106 patients underwent short-term (< 1 yr) postoperative PTA, and 55 patients underwent long-term (> 1 yr) postoperative PTA (Fig. 2). There was  Jahrsdoerfer score, mean ± SD 8.2 ± 1.6 7.9 ± 1.5 6.1 ± 1.5 7.7 ± 1.7 8.6 ± 1.2 8.1 ± 1.6 (4-10)

Discussion
The problems associated with CAS include conductive hearing loss, cerumen impactions and EAC cholesteatoma 7 . The high risk of cholesteatoma formation suggests that more attention should be paid to patients with CAS. Cholesteatoma is a crucial factor in surgical indication, thus, patients were divided into cholesteatoma and no cholesteatoma groups in the present study. Compared with other studies, this study included the largest sample of CAS patients with comprehensive analysis of the clinical features and long-term outcomes. There were significant differences between males and   CAS patients without cholesteatoma typically underwent meatoplasty at ages greater than 6 years, when these patients were old enough to understand the rationale behind the meatoplasty procedure and assist in the postoperative care. Age was not an exclusion criterion in CAS patients with cholesteatoma, except patients aged < 6 years, and there was no significant difference among different age groups for cholesteatoma formation. A previous study suggested that pediatric EAC cholesteatoma is less aggressive than adult, but age was not a crucial factor in surgical indication in the present study 15 . This result challenges the previous viewpoint that the appropriate time for surgery is late childhood or early adolescence, prior to the occurrence of irreversible damage 1 . There was some correlation between microtia and temporal bone malformation, evaluated using Marx grading systems 16 . Sixty-six percent of the individuals with deletions of the distal long arm of chromosome 18 had aural stenosis or atresia, and none of these subjects had microtia 17 . However, we also evaluated the relationship between the degree of microtia and EAC cholesteatoma, and there was no significant difference using the Marx grading system.
HRCT is indicated in patients with suspected congenital malformation of the external and middle ear, and preoperative planning is absolutely necessary 3,18 . The analysis was more accurate using MIMICS software 19 . The stenotic EAC was larger in the cholesteatoma group than in the no cholesteatoma group, reflecting the erosion of cholesteatoma. Among all 21 patients with cholesteatoma, the stenosis of EAC (> 4 mm) group showed a defective bony canal. When we removed this group, there was no significant difference between the cholesteatoma and no cholesteatoma groups for the stenosis of EAC, and no significant difference was observed among the subgroups of stenotic EAC for cholesteatoma formation. We conducted a multivariate logistic regression analysis for cholesteatoma formation. The results also showed that the stenosis of EAC (> 4 mm) group was associated with cholesteatoma formation (OR, 5.337; 95% CI, 1.444-19.735, p = 0.012), but the other stenosis of EAC groups were not associated with cholesteatoma formation, p > 0.05. These results challenged the previous viewpoint that patients with stenosis of 2 mm or less are at high risk of developing cholesteatoma and should undergo surgery 1 . Moreover, the differences in the results indicated that the stenosis of EAC group with cholesteatoma (> 4 mm) might represent a special state of EAC, namely, the blockage of EAC. This state is a new concept based on HRCT, in which the EAC bony segment is expanded through cholesteatoma, and the stenotic EAC is larger than 4 mm. This result also challenged the previous viewpoint that CAS has been defined as an EAC with a diameter of 4 mm or less, as none of the patients with canal openings larger than 4 mm developed cholesteatoma 1 .
Except for some patients with stage IV EAC cholesteatoma, the middle ear was not eroded by cholesteatoma. Thus, there was no significant difference in the same surgical technique between the cholesteatoma and no cholesteatoma groups, including Δ ABG, preoperative and postoperative ABG 20 . A previous study suggested that the Jahrsdoerfer score (modified) was less useful in terms of predicting long-term hearing prognosis after canal tympanoplasty for CAS in 25 patients 6 . Nevertheless, the mean Jahrsdoerfer score in the no cholesteatoma group was higher than that in the cholesteatoma group, and the Jahrsdoerfer score is a factor that affects postoperative ABG.
Modified meatoplasty with endoaural-conchal incision is an effective surgical intervention for CAS. In the present study, a stable hearing outcome was observed with prolonged follow-up. However, in CAA cases, the degradation of hearing outcomes was observed with prolonged follow-up. This difference might reflect the status of postoperative EAC, TM and ossicles 21 . Another interesting finding was increased BC hearing thresholds of 5 to 6 dB at 0.5, 1 and 2 kHz after surgery. Animal experiments demonstrated that the inertia of the ossicular chain contributed to partial BC hearing, and this inertia produced greater effects on high frequencies compared with low frequencies. Patients with CAS presented BC hearing loss, which could primarily reflect the absence of the inertia of the ossicular chain 22 .
None of the patients examined in the present study showed facial nerve palsy or total deafness. Postoperative complications, included soft tissue stenosis (4.1%), perforation of TM (3.4%), bony regrowth of EAC (2.7%), lateralization of TM (2.0%), granulation tissue (0.6%) and infection of the new EAC (0.6%). It has been suggested that normal epithelial migration from the tympanic membrane and EAC is an essential phenomenon to maintain the cleanliness of the outer ear. Predisposing factors, such as bony external ear irregularities and local inflammation, prevent or slow the normal migration of the squamous cells in the external ear canal, leading to a build-up of keratin debris, ultimately forming a cholesteatoma 18  cholesteatoma formation but also a higher rate of postoperative complications. The role of chronic inflammation during disease progression and recurrence might trigger cholesteatoma onset and are important in guiding clinical intervention 25 .
In conclusion, the results of the present study indicate that meatoplasty is an effective surgical intervention for CAS, and a stable hearing outcome was observed with prolonged follow-up. The Jahrsdoerfer score affected postoperative ABG, and age was not a crucial factor in surgical indication. Except for the stenosis of EAC (> 4 mm) group, there was no significant difference among subgroups of stenotic EAC for cholesteatoma formation, and no significant difference between the cholesteatoma and no cholesteatoma groups for hearing outcomes was observed. The stenosis of EAC with cholesteatoma (> 4 mm) might represent a special state of EAC, namely, the blockage of EAC. This blockage is a new concept based on HRCT, in which the EAC bony segment is expanded through cholesteatoma, and the stenotic EAC is larger than 4 mm.

Methods
Patient Selection. This study was a retrospective review of patients who underwent meatoplasty for CAS at a tertiary referral hospital, from April 2008 to August 2015. The inclusion criteria were patients with CAS who underwent meatoplasty. The exclusion criteria were acquired aural stenosis, otological surgery history, temporal bone fibrous dysplasia, or benign or malignant tumours in EAC. Reflecting the particularity of meatoplasty, most of the patients followed up regularly. Ultimately, a total of 246 meatoplasty procedures were performed on 232 patients who became the subjects in the present study. Fourteen patients were bilateral, and for the purpose of this study, each of the 246 procedures was considered an individual case. The study was approved through the institutional review board.

Data Collection.
A structured form was used to obtain patient data, including anamneses, pure-tone audiometry (PTA), HRCT of the temporal bones, operation notes and videos, pathology reports, postoperative follow-up records and detailed contact information. There is a complete medical record system at the hospital, and used a custom database software to process the data of CAS and CAA patients.
The demographic data for each patient, including age, gender, laterality, infection, stage of EAC cholesteatoma and degree of microtia, were collected. The CAS infections primarily manifested as otorrhea or postauricular subperiosteal abscess. The stage of EAC cholesteatoma was determined using the Naim classification, in which stages I and II could not easily be distinguished; therefore, we combined these stages together for analysis 26 . The degree of microtia was determined using Marx grading systems 27 .
HRCT images were reconstructed with 0.75 mm thick sections at 0.5 mm increments. Digital imaging and communication in medicine (DICOM) datasets were imported into MIMICS 15.0 software (Materialize, Belgium) for image processing 19 . This software enables the simultaneous viewing of HRCT datasets using a set of 2D images and a 3D rendered image for each dataset. The Jahrsdoerfer score, for preoperative evaluation and outcome expectations, was evaluated using MIMICS 28 . With respect to measuring the EAC diameter, an analytical method was used. Sagittal scans were used to measure the diameter of EAC on MIMICS; sagittal reconstruction is the most useful image for defining the involvement of anterior, posterior, and inferior walls by the EAC cholesteatoma 18 . Only bony segments were measured, reflecting actual CAS conditions, with or without cholesteatoma. The minor axis of EAC was measured in each slice, and the minimum value was considered as the stenosis size (Fig. 1).
PTA was performed in a soundproof booth. Frequencies of 0.5, 1, 2 and 4 kHz were analysed in the present study. We collected the preoperative and postoperative PTA and calculated the air-bone gaps (ABG), and the mean of the four frequencies was calculated as the ABG value 9 . The short-term (< 1 yr) and long-term (> 1 yr) hearing results were reviewed to evaluate the stability of the hearing outcomes 29 .
Postoperative complications were also collected during follow-up. We recorded the number and the type of different complications arising from meatoplasty, including stenosis, bony regrowth of EAC, infection, lateralization of tympanic membrane (TM), perforation of TM, granulation tissue of TM or EAC, total deafness and facial nerve palsy. Some patients showed more than one of these complications, and each of the complications was separately analysed in the present study.
Surgical Technique and Follow-up. The same surgeon performed all procedures, and CAS patients without cholesteatoma typically underwent meatoplasty at ages greater than 6 years 1 . For CAS patients with cholesteatoma, age was not an exclusion criterion. We used a modified meatoplasty procedure with an endoaural-conchal incision, in which two local rotation flaps and a transposition split-thickness scalp flap were used to widen the stenotic EAC and reconstruct the TM. When necessary, i.e., Jahrsdoerfer score 6 or greater, tympanoplasty or ossicle mobilization was performed 20,30 . The long-term care of the ear is important because reconstructed EACs do not typically clear squamous debris 29 . Most of the patients required regular follow-up every 3 to 6 months. The postoperative follow-up records, including PTA, HRCT and complications, were archived using custom database software.
Statistical Analysis. Descriptive and inferential statistical analyses were performed using parametric and nonparametric tests, as appropriate. For continuous variables, independent groups were compared using t tests. The relationships between categorical variables were assessed using χ 2 analyses. Multivariate logistic/linear regression analysis were preformed to see whether CAS is independently associated with clinical features and long term outcomes. All analyses were performed using SPSS software (version 20.0; IBM, New York). For all comparisons and analyses, a p value of < 0.05 was used as the cut-off point for statistical significance.