Surgical outcome in class 4 congenital anomalies of the ossicular chain: a systematic review of the literature

Objective To review hearing and surgical outcomes after reconstructive middle ear surgery in class 4 congenital middle ear anomalies (CMEA), e.g., patients with oval- or round window atresia of dysplasia. Data sources Pubmed/Medline, Embase and Cochrane library. Review methods Articles containing data on hearing outcomes and complications after reconstructive ear surgery in class 4 anomalies were analyzed and critically appraised. The following data were included and reviewed: patient demographics, audiometric testing, surgical techniques, complications, revision surgeries and their outcomes. Risk of bias was determined, and GRADE certainty of evidence was assessed. Primary outcomes were postoperative air conduction thresholds (AC), change in AC, and success rates (closure of the ABG to within 20 dB), the occurrence of complications (most importantly sensorineural hearing loss) and the long-term stability of hearing results (> 6-month follow-up) and occurrence of recurrence of preoperative hearing loss. Results Success rates varied from 12.5 to 75% at long-term follow-up with larger cohorts reporting success rates around 50%, mean postoperative gain in AC varied from 4.7 to 30 dB and − 8.6 to 23.6 dB at, respectively, short- and long-term follow-up. No postoperative change in hearing occurred in 0–33.3% of ears, and recurrence of hearing loss occurred in 0–66.7% of ears. SNHL occurred in a total of seven ears across all studies of which three experienced complete hearing loss. Conclusion Reconstructive surgery can be an effective treatment option which should be considered in patients with very favorable baseline parameters, while also considering the substantial risk of recurrence of hearing loss, the possibility of unchanged hearing despite surgery and the rare occurrence of SNHL. Level of evidence 2c. Supplementary Information The online version contains supplementary material available at 10.1007/s00405-023-08091-w.


Introduction
Congenital middle ear anomalies (CMEA) are a very rare cause of conductive hearing loss (CHL) in children (1-2% of diagnosed middle ear hearing losses) [1]. Usually children with CMEAs present with a concomitant otitis media with effusion (OME), which is the most common cause of hearing loss in 1-6-year-old children. Therefore, initial age as the prevalence of recurrent otitis media is high and may lead to disappointing outcomes or complications.
Non-favorable hearing outcomes after surgery occur in the minority of CMEA cases and are more likely in ears with concomitant syndromal diagnoses (around 25% of CMEA patients) and in class 4 anomalies, which are ovaland/or round window anomalies with a possible abnormal course of the facial nerve [10]. In class 4 anomalies, reconstructive surgery consists of creating an opening from middle to inner ear which has been shown to result in more (and more severe) complications compared to class 1-3 anomalies, i.e., considerable recurrence of conductive hearing loss by re-obliteration of the constructed connection to the inner ear and most importantly complete-or partial sensorineural hearing loss (SNHL) [11][12][13]. To overcome these risks, otologists have suggested opting for audiological hearing amplification instead of surgery. Because this uncertainty around surgery in class 4 CMEA patients remains, it is of value that clinicians are acquainted with the audiometric results and surgical complications reported in the literature. This systematic review provides an overview of the scientific literature in this challenging group of patients and offers evidence for optimal decision-making in treatment strategies.

Methods
This systematic review was reported according to the PRISMA guidelines. Our protocol was registered in the National Institute for Health Research's PROSPERO before commencing (Prospero ID: CRD42021229898) [14]. All methodological aspects of this review were carried out independently by two reviewers (SH and JR). Disagreement between reviewers was solved by consensus.
The reviewers searched PubMed/Medline, Embase and Cochrane library, most recently in July 2022. The search terms are available in the supplementary materials.

In-and exclusion of studies
Studies were selected using the in-and exclusion criteria as defined in the study protocol. Studies were included if: reporting hearing outcome on reconstructive surgery in patients with congenital absence or dysplasia of the oval or round windows, published between 1985 and July 2022 in English, French, Spanish and Dutch. Studies were excluded if surgical methods were not specified or inadequately described, if conducted on animals or cadavers, if less than 5 patients were included or if surgical results were not reported separately from results of non-class 4 CMEA patients.
The reviewers independently selected articles from the search by screening titles and abstracts and analyzed the full texts of all remaining relevant articles.

Critical appraisal and level of evidence
Assessment of risk of bias and overall study quality was performed using the National Heart, Lung and Blood institute's Quality Assessment Tool for Before-After (Pre-Post) Studies With No Control Group [15][16][17][18]. Criteria 8 'Where the people assessing the outcomes blinded to the participants' exposures/interventions?' and 12 'If the intervention was conducted at a group level (e.g., a whole hospital, a community, etc.) did the statistical analysis take into account the use of individual-level data to determine effects at the group level?' were not used in the critical appraisal as these questions were not applicable to the general literature regarding CMEAs. The studies were classified as good, fair or poor according to pre-specified criteria created to apply to general literature on CMEAs. Studies classified as 'good' were considered to be of high methodological quality and have a low risk of bias, 'fair' as average quality with moderate risk of bias and poor as low quality and high risk of bias. The certainty of evidence was assessed using Grading of Recommendations, Assessment, Development and Evaluations (GRADE) [19].

Data extraction
Data were extracted independently by the reviewers using a preset data extraction sheet, disagreement was solved by consensus.
The following data were extracted: study and patient demographics, surgical techniques, results of preoperative and postoperative audiometric testing, i.e., air conduction (AC) pure-tone average (PTA), bone conduction (BC) PTA and air-bone gap (ABG), peri-and postoperative complications, need for revision surgery and findings during revision surgery. For each study, audiometric results were made complete using two out of AC, BC and ABG reported by authors. Audiometric data were extracted in accordance to the committee of hearing and equilibrium guidelines averaging audiometric measures at 0.5, 1, 2, and 3 or 4 kHz. As stated in these guidelines, 3 and 4 kHz were equally accepted and applied accordingly to calculate the PTA [20].

Outcomes and measures of effect
The primary outcome measures were: postoperative AC and postoperative gain in AC. Audiometric results measured at ≤ 3-month and ≥ 6-month follow-up were used as shortand long-term follow-up. Surgical success was defined by closure of the ABG to within 20 dB. Other definitions of surgical success used in the literature: postoperative AC ≤ 30dB and gain in AC ≥ 15 dB were also extracted if available.
As secondary outcome measures, the incidence of the following complications was defined: postoperative sensorineural hearing loss (SNHL) due to damage to the inner ear during surgery, perilymph gusher, persistent or temporary vertigo/dizziness, postoperative facial nerve weakness, extrusion of prosthesis material and wound infection such as otitis media with or without concomitant labyrinthitis. SNHL was defined as a postoperative decline in BC ≥ 10 dB.
Furthermore, data on stability of hearing results were collected. If reported, results of audiometric testing shortly after surgery (within three months) and at ≥ 6 months were compared. Studies that did not report individual but generic hearing outcome data were also included. Hearing was defined as stable if AC PTA remained within 5 dB of fluctuation from AC PTA measured at short-term follow-up. Other outcome parameters were: (1) need for revision surgery and (2) the incidence of fenestral re-obliteration (diagnosed during revision surgery or on radiologic imaging such as HR-CT).

Statistical analysis
P values < 0.05 were considered as statistically significant. A meta-analysis was not performed as we expected included studies to have a low number of patients due to the rarity of the condition and a high heterogeneity in reported data. Collected data were analyzed in a descriptive manner.

Critical appraisal of risk of bias
Outcome of critical appraisal is displayed in Table 2. One study was classified as good by scoring yes on all criteria within the quality assessment tool. Five studies were

Identification of studies via databases and registers
Identification Screening Included 1 3 classified as fair either by scoring low on how representative the study was, i.e., small patient group and more strict than usual in-and exclusion criteria or scoring low on several criteria. Five studies classified as poor, these all scored moderate to low on how outcome was measured and on other previously stated criteria.

Complications and revision surgery
Complications are displayed in Table 4. One study (Han et al.)    Re-obliteration of the oval window was discovered in 6.1% (2/33)-15.4% (2/13) of patients in three different studies. Revision surgery was performed on 16 ears reported by six different studies; findings and outcomes are available in Table 5.

Discussion
In this systematic review, we reported hearing outcomes of reconstructive surgery in class 4 CMEA patients. A total of 188 ears were included across 11 different studies. Overall, the reviewed data showed promising results. Six of seven studies that reported mean gain in AC ≥ 6 months reported AC gain > 15 dB, four of six studies that reported success rates (closure of the ABG to ≤ 20 dB or AC < 30 dB) ≥ 6 months reported rates > 50%, thereby proving that fenestration surgery can greatly improve hearing in some patients.
However, there also seems to be a varying percentage of ears in which hearing remains unchanged after surgery (in total 16/188 ears, 8.5%) and in which hearing loss recurs after improvement by surgery (in total 14/188 ears, 7.4%). Major permanent complications, i.e., decreased AC and SNHL reportedly occurred in eleven of 188 ears (5.9%) and SNHL specifically in seven of 188 ears (3.7%) ranging from 0 to 50% [11] in individual studies.
Our results negate the belief that surgery in class 4 ears should be avoided entirely, but confirm the belief that fenestration surgery in class 4 ears is less favorable than reconstructive surgery in class 1-3 ears, where success rates vary from 56 to 75% [4,6,32].

Complications and disappointing results
A recurrence of hearing loss and unchanged hearing was reported by several studies in varying percentages. Recurrent hearing loss can be caused by several factors like prosthesis displacement, adhesions in the middle ear, or re-obliteration of the artificially fenestrated area. Surgery is performed to restore the normal ossicular chain conducted acoustic hearing. However, in ears with severe anatomic variations/ deformities, as is often the case in syndromal ears, the success rates might be lower [10]. Similar observations are seen in the surgical opening of a congenital bony ear canal atresia (major ear anomalies). In cases with low scores on the Jahrsdoerfer grading scale (more pre-existing anatomical deformities, on a scale from 0 to 10), surgical treatment is not advocated due to its low success rates and high odds of recurrence [33]. Therefore, given the abnormal anatomy in class four CMEA ears, i.e., aberrant facial nerve course, obliterated oval fenestra, it is not surprising that successful outcome is less frequent than in class 1-3 ears and that the middle ear dynamics might be prone to restore the preoperative condition resulting in recurrent fenestral obliteration. Unfortunately, the individual causes of complications and disappointing results could not be included in all the cases because of lack of reporting in the included studies.
The most feared complication, complete SNHL or 'deaf ears' after surgery occurred in three ears reported by two studies. Partial SNHL occurred in an additional four ears, reported by two studies. Although the incidence of complete SNHL might indeed be low, it is essential to obtain patient informed consent including considering the alternatives, i.e., hearing amplification devices (air and bone conduction hearing aids). Revision surgeries were performed revealing a myriad of different operative findings and varying postoperative results from complete SNHL to great improvement of AC thresholds. Due to disappointing long-term results (see Table 6), surgery could be considered in the minority of cases.
Despite not being able to investigate the effect of a syndromal diagnosis on outcome in class 4 ears due to the lack of data provided in the literature, it is probable that outcome in these ears will be worse than in non-syndromal ears due to non-surgical factors like possible associated inner-ear anomalies as has been shown in syndromal class 1-3 ears [10]. Another factor to consider is recurrent otitis media as middle ear infection is known to be detrimental in operated ears and might result in total SNHL as was the case in one ear reported on by Thomeer et al. [11].

Alternative treatment modalities
Patients with aforementioned unfavorable factors that might lead to an increased risk of unsuccessful surgery should be counseled for other options such as bone conduction devices (BCD), the vibrant soundbridge and regular hearing aids. Especially in children, alternative treatment modalities (hearing aids) should be considered to start treatment as soon as possible given that surgery is not recommended during the first 6-8 years of life due to the higher prevalence of recurrent otitis media. Other reasons for choosing alternative strategies are ears with an extensive history of middle ear problems (or previous surgeries) as these are at risk for disappointing results due to adhesions and overall rigidity in the ossicular chain and are renowned for their challenging surgical management.

Decision-making and surgical preparation
It is of importance to note that despite the mentioned risks, surgery in class 4 ears can be of great benefit to the patient by either improving hearing thresholds to functional levels or to make hearing rehabilitation using hearing aids more  accessible. Our treatment strategy would be to attempt reconstructive surgery in patients with very favorable baseline parameters, i.e., non-syndromal, no middle ear history and with an accessible operating area (access not completely blocked by the facial nerve and no ear canal atresia). It is, therefore, of utmost importance to obtain a high-resolution computed tomography scanning (HR-CT) of the temporal bone before discussing surgery as an option, to be acquainted with the middle-and inner-ear anatomy for optimal preparation and surgical feasibility. During the decision-making process, patients (and their families) should be informed about the successful long-term hearing outcome rate of around 50%, based on the findings of this review, that some patients might benefit from reconstructive surgery while in other cases surgery might have to be aborted due to unfavorable anatomy and that in some cases hearing is not improved after surgery. Furthermore, patients and family should be informed about the occurrence of re-obliteration, prosthesis misplacement, need for revision surgery and that facial nerve damage or vestibular complaints might happen more frequently than in other CMEAs or in the general population needing middle ear surgery. Lastly, patients should be informed that, although rare, complete hearing loss and postoperative dizziness (vertigo), are a real risk.

Limitations
The following limitations should be noted: (1) the studies included were heterogenous in patient size, class of surgery conducted and duration of follow-up. Also, audiometric assessment, outcomes and reporting thereof were variable and often incomplete; (2) most of the included studies reported on small patient cohorts or case series; (3) certain important primary and secondary outcome measures reviewed here might be underreported due to vague, or lack of, reporting of individual hearing results in some of the included studies; (4) some studies only mentioned stability of results without providing the most recent hearing outcomes and (5) it was not possible to carry out a metaanalysis due to aforementioned limitations. Lastly, it should be noted that many studies aborted surgery in ears in which the facial nerve completely blocked the original oval window site; therefore, data on surgery in these kinds of ears are very limited.

Quality of the evidence
We used the GRADE method to assess the certainty of the evidence and strength of recommendation in this review, see Table 7. The evidence regarding all outcomes was of very low certainty which means uncertainty remains of the effects of the intervention, reconstructive surgery, on the outcomes defined in this review. The quality of evidence was influenced by aforementioned limitations, the high to moderate risk of bias identified in all but one study, inconsistency of results in the different studies and imprecision caused by the small sample sizes, incomplete reporting and lack of statistical analysis in most studies.
a Several studies showed moderate to high risk of bias due to inconsistent and incomplete reporting of outcome measures b This outcome measure was variable and, thus, inconsistent across the different studies c Few studies included sufficient patients to provide precise results d Several studies mentioned this outcome without providing individual patient data e Several studies provided mean audiometric outcomes without mentioning this outcome measure f One study provided this outcome measure without providing individual patient data