Tympanoplasty in Northern Ethiopia: success rates, failure factors, and audiometric improvements

Abstract

Objective

This study aimed to determine the success rate of tympanoplasty, the factors associated with failure rate, and audiometric gains after the surgical procedure in Ethiopia.

Results

There were 82 successful grafts take out of the 92 tympanoplasty procedures with an overall success rate of 89.1%. Females, who had tympanoplasty had 45 (91.8%) successful grafts while 37 (86.0%) successful grafts for male. Of the total tympanoplasties, 50(53.4%) of it were done in the left ear. Ear discharge was found to be a significant predictor of tympanoplasty failure rate [AOR = 9.6, 95% CI (1.1 − 88.8)]. There was a significant gain of air conduction with a p-value of < 0.001 after the tympanoplasty procedure which was about 13 dB mean difference.

Conclusions

In this study, only post-operative ear infection was found to decrease the success rate of tympanoplasty significantly. It also revealed that tympanoplasty improves hearing ability. We recommend a strict application of infection prevention and control strategies, and adequate and comprehensive follow-up of patients including audiometric assessment.

Background

A tympanoplasty is a surgical procedure performed to close perforated tympanic membrane with reconstruction of ossicles whenever necessary. There are 5 types of tympanoplasty. Type 1 is only repair of perforated tympanic membrane the other types (II- V) in addition to repair of perforated tympanic membrane it involves reconstruction of ossicles according to the level of ossicular damage and clear and eliminate of middle ear pathology. This procedure is common after chronic otitis media and trauma [1, 2]. Depending on the site and size of perforation various surgical approaches are utilized for tympanoplasty including endomeatal, endaural, and postauricular approaches. Different surgical approaches affect the surgical outcomes [3,4,5]. The commonest surgical techniques of tympanoplasty include either underlay or overlay techniques of placement grafts over the perforated tympanic membrane. Underlay is widely used, it is quicker, easier to perform, and the graft is placed entirely medial to the remaining drum and malleus [3, 4]. Different types of grafts are used for tympanoplasty, including temporalis fascia, cartilage, perichondrium, fat, and vein [3, 4, 6].

The success rate of tympanoplasty is determined in terms of controlling the disease, restoring anatomy and function as well as preventing recurrent otorrhea [3, 7]. Some studies show a success rate of tympanoplasty ranging from 71 to 80% [8]. Failure of tympanoplasty is mainly represented by the recurrence of tympanic perforation [9]. The surgical outcome depends on several factors including age, gender, and drainage status of the ear at the time of surgery, size and location of the perforation, ossicular status, surgical technique, the experience of a surgeon, type of graft material, and function of the eustachian tube [4, 9,10,11,12]. After tympanoplasty, there is a remarkable improvement in hearing ability [13]. Audiometric gain is found in most of the patients after the first surgery and in 100% of the patients after reoperations [14].

Tympanoplasty is one of the commonest surgical procedures done in the otorhinolaryngology department and knowing the success rate and determinants of success is crucial for service improvement. Information is scarce about the prevalence of middle ear disease in northern region but the overall prevalence of hearing loss among 5 + year old population in Ethiopia showed 19.2% and one cause for this high burden of hearing loss is mentioned due to middle ear infection [15]. Even though there is no specific data that shows prevalence and burden of middle ear disease in norther region this study shows there is high burden nationwide. In addition to that information is scarce about the surgical outcome of tympanoplasty, its associated factors, and audiometric gain in Ethiopia and specifically in this study area. Due to the underdeveloped nature of the country and high poverty level in northern region, there is high rate of chronic ear infection on top of that this region has limited ENT center which provide ear disease treatment. There is only one hospital that provide tympanoplasty or other related ENT services for this large number of populations in this northern region of the country which is around 6 million. This led to occurrence of high number of cases with untreated chronic ear diseased and with a large size perforated tympanic membrane for a long time with limited surgical access. In addition to that there are few ENT surgeons with a limited experience who provide care for this large amount of disease burden. So far, there is no clear data that shows the success rate and effectiveness of this procedure in this region. Therefore, this study aims to investigate the overall success of the tympanoplasty procedure, the factors associated with failure rate, and audiometric gains in this northern region of Ethiopia. This study can be used as a baseline data source for improving the outcome of the surgical procedure in the reconstruction of hearing. Further, knowing the determinant factors helps in preventing complications and improving outcomes of the procedure. This can be achieved by identifying the cause for failure rate of tympanoplasty and improving on weaknesses identified on the study in ways such as improving surgical skill and increasing number of surgeons in the region, expanding ENT centers which provide services and creating awareness on the population to seek treatment early whenever they have ear related problem.

Methods

Study design and area

A facility-based cross-sectional study was conducted in Ayder Comprehensive Specialized Hospital (ACSH), Tigray region, Ethiopia. The hospital is the only tertiary hospital in the Tigray region. The region has around 6 million population and the otorhinolaryngology department is one of the departments that provide both clinical services and post-graduate specialty training. This department is the only area delivering all otorhinolaryngology services in north region which includes the entire Tigray region and neighboring zones of Amhara and Afar regions of Ethiopia.

Study population and sampling

Medical records of patients who visited the ENT clinic and undergone a tympanoplasty procedure during the 2018 to 2019 period were the study subjects. As eligibility criteria, patients with complete medical records and who had follow-up after one month following the tympanoplasty procedure were included in this study. All study participants who fulfilled the eligibility criteria during the two-year study period were included. Based on eligibility criteria cases included were perforated tympanic membrane with or without ossicular loss. Those with middle ear pathology such as cholesteatoma which require further intervention were excluded from the study. All surgeries were performed on dry ear which has no active ear discharge. Other than this, other detail middle ear pathology was not consistently included in the medical records. Due this, severity of middle ear pathology was not considered as inclusion criteria in this study.

Data collection

Data were collected after preparing a checklist from the medical records of patients by two trained medical residents in the ENT department. The content validity of all items in the checklist was assessed by senior staff including their meanings. The data were collected by extracting the required information from the pre-operational, intraoperative and post-operation follow-up notes from chart of each patient who undergone tympanoplasty. There was an ongoing supervision for completeness, consistency, and coding during data collection periods by the investigators. Additionally, the residents were also oriented, well-trained, and supervised to get accurate data. The dependent variables were closure rate of tympanoplasty graft and post-operative air-bone gap (ABG) and air conduction (AC) gain which were obtained by taking pure tone average of audiometric result of the 3 frequency 500, 1000 and 2000 Hz pre and post-operatively. In this study, we have tried to see success rate in relation to the independent variables such as socio-demographics characteristics which included age, sex and whether the procedure was first time or revision surgery after previous failed surgery. Disease factor such as site and size of percentage tympanic perforation identified. Different techniques/procedures were used during the surgery such as including approach used to enter to the middle ear were transcanal, endaural or postauricular. Graft placement was done using either underlay or overlay graft techniques. Underlay is when graft material is placed underneath the remaining tympanic membrane whereas overlay is when it is placed above it [16]. In regard to the types of tympanoplasty performed in this study, in some of the surgeries, only tympanoplasty was done according to the definition it is type I and some of the surgeries also used tympanoplasty with ossiculoplasty in which the types of tympanoplasty were not mentioned in the medical records which supposed to indicate either type II to V. The type graft material used to repair the perforation which could be fascia, cartilage, perichondrium or combination of those was assessed. The level of expertise of surgeon identified whether it was done by ENT resident, who has one up to four years of experience, surgeon, who has more than four years of experience, or otologist, who took special training which focus in ear surgery and has more than six years of experience. Tympanoplasty success was operationally defined as the successful graft take and closure of the perforated tympanic membrane after four weeks of surgery with or without improvement in hearing by objectively measuring audiometry result after 08 weeks of surgery. Thus, the primary purpose of the surgery was to clear the middle ear disease and repair perforated tympanic membrane. Hearing improvement after surgery was considered as a secondary purpose since there are hearing loss due to nerve damage or severe middle ear pathology which cannot be repaired by this surgery. Therefore, in this study success rate was mainly measured by successful graft take. Additionally, we have tried to assess hearing improvement separately. In our setting, surgery done during campaign are identified from the registration book which specifies the cases, date of procedure done and by who it was performed. Campaign in this study was defined as when the procedure is conducted in specific period of time to give hand on training for ENT resident. During this time, many surgeries were conducted in very short period of time with high number of cases as well as most of the surgeries were done by ENT residents as part of their training.

Data processing and analysis

Descriptive and inferential statistics were applied. To identify independent predictors of failure of the tympanoplasty procedure, a binary logistic regression was used. A p-value of 0.25 or less at binary logistic regression was used to screen variables for multivariable regression (operating surgeon, graft material, ear discharge, and campaign). Hosmer-Lemeshow test for the final multivariable model had a p-value of 0.685 saying that the model was a good fit. Also, the maximum Variance Inflation Factor (VIF) found was 1.19 indicating no multicollinearity in the final model. A comparison of ABG and AC between pre-tympanoplasty and post-tympanoplasty was made using a paired-samples t-test. Furthermore, the post-operative AC gain difference between tympanoplasty only and tympanoplasty with ossiculoplasty was evaluated using the independent-samples t-test. Generally, statistical significance was declared at p < 0.05.

Results

In the otorhinolaryngology department, there were a total of 212 medical records with tympanoplasty from 2018 to 2019. However, 92 (43.4%) of the medical records were included in the current analysis, and the rest were excluded due to incomplete records or incomplete follow-up. There were 10 failures out of the 92 tympanoplasty procedures (i.e., the overall success rate of graft take was 89.1% (82/92)).

Socio-demographic and clinical characteristics

The mean age of patients included in this study was 24.4 (SD = 9.1) years with minimum age of 12 years and maximum age of 65 years were included. Of the total of 46 females who had tympanoplasty, 45 (91.8%) had successful graft take and from 43 males who had tympanoplasty 37 (86.0%) had successful graft take. Of the total tympanoplasties, 50(53.4%) of them were done in the left ear and the site of perforations of the tympanic membrane was central (Table 1).

Table 1 Socio-demographic and clinical characteristics of patients and their associations with success of tympanoplasty (n = 92), 2020

Tympanoplasty procedure related characteristics

More than half of the procedures were done by an otology fellow/otologist (64.1%). The failure rate for tympanoplasty done was lowest (6.8%) by otology fellow/ otologist and highest (25.0%) by resident. This high failure rate of tympanoplasty done by residents than those done by otologist shows that skill of the surgeon has a role in success of tympanoplasty. There was a significant association between the type of graft material used and the success rate (p = 0.036). About 88.8% of graft placement was underlay. Ossiculoplasty was done for 26 (28.3%) patients. Post-operative use of antibiotics and ear discharge had a significant association with the success rate of tympanoplasty. The failure rate for those with post-operative ear discharge and those who didn’t take post-operative antibiotics was 28.6% and 21.1% respectively. There were 5 (29.4%) failures during a campaign and 5(6.7%) during routine day surgery. The difference in the proportion of failure was significant (p = 0.007) (Table 2).

Table 2 Tympanoplasty procedure related characteristics and their association with success, (n = 92), 2020

Predictors of tympanoplasty failure

At bivariable logistic regression, the size of perforation, ear discharge, and campaign were found to be statistically significant with p-values of 0.04, 0.043, and 0.012 respectively. Among the variables selected for multivariable analysis, only ear discharge was found to be a significant predictor of tympanoplasty failure. Accordingly, the adjusted odds of tympanoplasty failure were 9.6 times higher in patients who developed ear discharge compared to those who did not develop ear discharge (AOR = 9.6, 95% CI: 1.1–88.9, and p = 0.047) (Table 3).

Table 3 Predictors of tympanoplasty failure among the 92 patients operated in ACSH (a multivariable logistic regression)

Audiometric assessment

Of the 92 patients, only 16 had full pre-op and post-operative air and bone conduction audiometric results. To assess AC and ABG, pure tone average of AC and BC was calculated using the mean of 500 Hz, 1000 Hz and 2000 Hz frequencies. For AC improvement, pre and post operative AC difference was taken. For ABG closure assessment, initially, pre-operative ABG was measured by taking the mean pure tone average of AC and BC difference pre-operatively. Post-operation ABG was measured by taking mean pure tone average of AC and BC difference post-operatively. Difference between pre-operative and post-operative ABG was taken as a measure of ABG closure. Narrowing of the gap (ABG closure) tells, there is improvement in air conduction after operation. Accordingly, the mean pre-operative and post-operative air-bone gap (ABG) was 39.2dB and 27.4 dB respectively with an air-bone gap closure of 5.5 dB. but the closure was insignificant (t = 1.3, df = 15, and p = 0.205). There was a significant (t = 5.2, df = 50, and p < 0.001) improvement of air conduction threshold after tympanoplasty which was about 13 dB (Table 4).

Table 4 Paired sample test to compare audiometric findings before and after tympanoplasty procedure in Northern Ethiopia, 2020

Discussion

The purpose of the study was to assess the success rate of tympanoplasty surgical procedures and associated factors. The overall success of tympanoplasty was 89.1%. This was based on closure rate of the graft. There was also a significant gain of air conduction after tympanoplasty with a mean air conduction gain of about 13 dB. The odds of tympanoplasty failure were significantly higher by 10 times in those who had a sign of infection like ear discharge.

The overall tympanoplasty success in this study is 89.1%. Previous success rates in African settings have been reported ranging from 71 to 92% and other studies also show the same result [8, 17]. On the other hand, higher success rates were seen in other studies with a success rate ranging from 93.3% up to 98.6% in India [3, 4, 6, 18]. This higher result compared to our study could be due to better surgical setup or better experience of the surgeon. In our hospital, most of the tympanoplasty procedures are done by local surgeons which is contrary to the other studies reports where the majority of the surgeries are done by experienced otologists.

Experience is so crucial for fine surgeries like tympanoplasty. For example, according to a study done in Addis Ababa, closure rates of 100 blocks of surgeries done by a single surgeon over successive years show that perforation closure for each successive 100-block increased from 74 to 98%. The closure rates of the second, third, and fourth 100-block were each statistically significantly different from the first 100-block, but not from each other. Likewise, a study done in Brazil found a graft success rate of about 65% in surgeries done by second-year residents. In this study, the low success rate was concluded as being attributed to the surgeon’s experience. These two studies tell that tympanoplasty is skill dependent and those surgeries done by otologist or experienced surgeon graft take is higher due to the delicate nature of the surgery which requires fine hands which is achieved by being exposed to doing more surgeries and by extensive hands on training [12, 14].

In this study, tympanoplasty done during a campaign had a closure rate of 70.6%. According to a study done in another part of Ethiopia on campaign-based tympanoplasty, the closure rate was 83.9% [8]. It is a better result when compared with our findings. The lower success in our setup is due to the majority of procedures during a campaign being done by ENT residents rather than ENT surgeons, otology fellows or otologists.

Our study and some other researchers found that age, sex, size, and site of perforation are not significantly associated with a success rate [4, 9, 14]. However, in a study done in Italy, the size of perforation was significantly associated with a success rate at p < 0.05. According to a study done in Brazil, the type of graft material and surgical technique used had no statistically significant relationship with the success of tympanoplasty [9, 14]. However, in our study, there was a statistically significant association between the type of graft material and tympanoplasty success (p = 0.036).

In our study, only post-operative ear discharge was found to be an independent and negative predictor of overall tympanoplasty success. The odds of failure were 10 times higher in those with ear discharge. This shows that post-operative ear discharge is a poor prognostic factor for the success of tympanoplasty. We recommend a strict application of infection prevention and control strategies. This finding is in line with Italy’s study in which ten patients had a post-operative infection (i.e., ear discharge) and all of them came with failed graft take [9]. This could be because the infection will impair wound healing and graft take.

In this study, the average mean ABG hearing closure across 3 frequencies is 5.5 dB but this closure is not statistically significant. This could be due to the small number of patients who had full pre-operative and post-operative audiometry assessments and it was in line with a study done in India [17]. However, studies done in Nigeria and Brazil reported significant mean ABG hearing improvement ranging from 12.2 dB up to 18.7 dB [3, 13, 14]. On the other hand, the finding of this study is a significant post-operative air conduction gain of about 12.8 dB. This is due to tympanoplasty repairs the perforated eardrum and damaged ossicles thereby restoring air conduction pathway integrity.

Limitation

More than half (56.6%) of the tympanoplasty procedures during the study period were excluded from this study due to incomplete records and follow-up. This may have decreased the statistical power of our analysis. Furthermore, only a few had full audiometric assessments. Moreover, the secondary nature of the data collection may affect the number of factors to be considered. Due to missing important variables such as middle ear pathology, types of tympanoplasty. Moreover, due to lost to follow-up of patients in this study tympanoplasty success of graft take was measured at four weeks and hearing improvement at eight weeks of post-operation which may not be adequate time to assess overall graft take. Therefore, further follow-up study is better recommended to see a variety of factors (middle ear pathology, type of surgeries) related to the success and failure rate of the tympanoplasty procedures. We also recommend for future study to assess success rate after adequate time of post-operation follow-up.

Conclusions

Even though many factors can affect the outcome of tympanoplasty, according to this study, only post-operative ear infection was found to decrease the success rate of tympanoplasty significantly. This study also revealed that tympanoplasty improves hearing ability. We recommend a strict application of infection prevention and control strategies, and adequate and comprehensive follow-up of patients including audiometric assessment. In addition, we recommend tympanoplasty to be done by experienced otologist and by ENT residents under close and strict supervision by a surgeon.