Abstract
Objective:
Determine the effect of complete malleus removal during canal wall up tympanomastoidectomy for cholesteatoma on ossiculoplasty success and rate of residual cholesteatoma.
Methods:
We reviewed the operative, audiogram, and clinical reports of patients who underwent canal wall up tympanomastoidectomy for cholesteatoma between 2009 and 2016 at a tertiary academic medical center with at least 8 months of follow-up after surgery. To control for extent of disease, we independently catalogued the subsites of the middle ear and mastoid that cholesteatoma involved from each operation. We performed multivariate logistic regression to determine the independent effect of complete removal of the malleus on the rate of residual disease and success of ossiculoplasty.
Results:
One hundred eighty surgeries were included in the analysis. For ossiculoplasty success, the adjusted odds ratio of complete malleus removal was 1.7 (95% CI, 0.43-7.0, P = .43), which was not statistically significant. For residual disease, the adjusted odds ratio of complete malleus removal versus not was 0.29 (95% CI, 0.074-1.1, P = .076), which approached but did not meet statistical significance.
Conclusion:
Though complete malleus removal does not independently decrease the rate of residual cholesteatoma, it may be a safe technique as it did not compromise ossiculoplasty success.
Introduction
Cholesteatoma is an ectopic focus of squamous keratinizing epithelium in the middle ear that can cause hearing loss, otorrhea, and intracranial complications. 1 The goal of surgical treatment is to eradicate disease and create a dry, “safe” ear without recurrent/residual disease and secondarily to restore/improve hearing. Surgery for cholesteatoma can be broadly categorized into canal wall down (CWD) and canal wall up (CWU) techniques, depending on whether the bony wall of the external auditory canal is left intact during surgery. Advantages of the CWD technique include improved visualization of the middle ear, particularly the anterior epitympanum, and lower rates of recurrent or residual disease. 2 The CWD procedure creates an open cavity requiring regular cleaning. In general, hearing outcomes after ossiculoplasty are believed to be worse compared to CWU techniques, 3 although some studies have found similar hearing outcomes. 4 The CWU technique avoids an open cavity, but visualization during surgery and postoperatively is more difficult, and recidivism is higher. 2 Due to the high rate of residual disease with the CWU technique, particularly in extensive cases, otologists often perform a planned “second-look” procedure to reexamine the middle ear several months after the original surgery, with a staged ossiculoplasty performed at that time.5,6 Most otologists feel comfortable removing the head of the malleus during CWU surgery if the cholesteatoma involves this portion of the ossicle, but many otologists are reluctant to completely resect the malleus (CMR) for fear of causing significant distortion of the tympanic membrane, resulting in a significant hearing loss. This notion was supported by a meta-analysis by Blom et al, 7 which examined ossiculoplasty outcomes with varying degrees of ossicular chain damage. Though pooled data showed absence of the malleus was a predictor of poor postoperative outcomes, some of the individual studies included found no difference in ossiculoplasty success rates, 8 and many of the studies were specific to primary surgery and/or did not examine ears involved by cholesteatoma. 7 In regard to recidivism with CMR, many studies have analyzed cholesteatoma recurrence, though only 1 has evaluated the effect of CMR on residual disease, and it found that CMR with cartilage reinforcement of the tympanic membrane had similar residual disease rates compared to preservation of the manubrium of the malleus and partial cartilage reinforcement. In that study, there was no controlling for extent of disease between the 2 groups. 9 Due to selection bias (CMR is generally utilized in cases of more extensive cholesteatoma), a direct comparison of CMR to other techniques for residual disease and ossiculoplasty success rates is not meaningful. At our institution, we remove the malleus completely as part of a CWU procedure in certain cases of extensive disease involving the anterior middle ear to obtain improved visualization and removal of the cholesteatoma. To our knowledge, no study has controlled for extent of disease and type of ossiculoplasty to evaluate the efficacy of a CMR as part of a CWU with facial recess technique, which was the objective of this article.
In this study, we catalogued the subsites of the middle ear and mastoid that we used as a proxy for extent of disease and performed multivariate logistic regression to control for disease burden and other potential risk factors to determine the independent effect of CMR on residual disease and ossiculoplasty success. This is the first study to evaluate the efficacy of CMR in a manner controlling for extent of disease and the second to examine the effect of CMR on residual disease at all.
Methods
Patient Selection
The study was deemed exempt from full protocol review by the West Virginia University Office of Research Integrity and Compliance’s Institutional Review Board in February 2016 (protocol No. 1601980796). We identified patients who had undergone tympanomastoidectomy by a current procedural terminology (CPT) code search of the electronic records of the billing department of the Department of Otolaryngology–Head and Neck Surgery at West Virginia University (CPT codes 69501, 69502, 69505, 69511, 60530, 69601-5, 69635-7, and 69641-6). We reviewed the operative and clinical reports of these patients and only included those in the study that had a canal wall up surgery with facial recess approach for cholesteatoma and had at least 8 months of follow-up after surgery. Eight months was selected as the minimum follow-up period because this is the minimum time period between primary surgery and a planned second-look surgery at our institution. If ossiculoplasty was performed, only those performed with a titanium prosthesis were included.
Assessing Extent of Cholesteatoma
During tympanomastoidectomy for cholesteatoma, our surgeons routinely describe in detail in the operative report the areas the cholesteatoma involved, whether the cholesteatoma was recurrent or residual, and the status of the ossicles. To define the extent of disease, we divided the middle ear into 10 specific subsites: the anterior and posterior epitympanum, anterior and posterior mesotympanum, anterior and posterior hypotympanum, the facial recess, sinus tympani, mastoid (including the antrum), and the tympanic membrane. Two of the authors independently catalogued the subsites of the middle ear and mastoid involved by the cholesteatoma. If there was any discrepancy, the authors reviewed the operative note together until they reached an agreement. If the 2 reviewers agreed that the operative note was unclear in regard to extent of cholesteatoma, status of ossicles, residual versus recurrent disease, or the ossicles that were removed, that surgery was excluded from the study. We used the number of subsites the cholesteatoma involved as a proxy for extent of disease.
Hearing and Recidivism Outcomes
All patients routinely had a hearing test prior to surgery, 3 months after surgery, and then as needed based on their symptoms and exam findings. Patients were seen 2 weeks, 1 month, 3 months, and then every 6 months after surgery for surveillance. If there was no evidence of disease with a good hearing result after 1 year of follow-up, surveillance intervals were lengthened to once per year. If at any time during follow-up there was suspicion for recidivism, computed tomographic imaging of the temporal bone was obtained. Patients underwent a second surgery if it was a planned second-look surgery, or if symptoms, exam findings, imaging findings, and/or hearing tests suggested recurrent or residual disease. We defined residual or recurrent disease based on a surgically confirmed presence of cholesteatoma after the initial surgery. Recurrence versus residual disease was defined at the time of surgery in the operative report based on timing and location of disease compared to the previous surgery and presence of new retraction pockets. The follow-up period began at the time of surgery and was cut off if recurrence or residual disease was found, even if that patient was followed for longer after revision surgery.
Statistical Analysis
Statistical analysis was performed in SAS 9.3 (SAS Institute Inc, Cary, North Carolina, USA). The threshold for statistical significance was set at α = 0.05 for all analyses. Continuous variables were summarized with means and standard deviations, and categorical variables were summarized with proportions or percentages. Key demographic and clinical characteristics were compared based between the CMR and non-CMR group (Table 1). Continuous variables were compared with a student t test, and categorical variables were compared with Fisher’s exact test.
Summary Demographic and Clinical Characteristics of Study Population.
Abbreviation: CMR, complete malleus removal.
Reached statistical significance.
For patients who had no recurrence or residual disease during entire length of follow-up.
Multivariate logistic regression analysis was performed with the primary outcomes of ossiculoplasty success and residual disease (Table 2). Ossiculoplasty analysis was limited to those patients without recurrent or residual disease during follow-up and who had undergone ossiculoplasty. The potential risk factors studied for residual disease were complete malleus removal, number of subsites involved by cholesteatoma, and right versus left ear. The variables included in the model were based on a priori suspicion for association with either ossiculoplasty success or residual disease. Univariate regression was not performed prior to multivariate because of suspicion for significant confounding effects, and with multivariate regression, the individual effect (while controlling for all other variables included in the model) of each variable could be determined. The potential risk factors studied for ossiculoplasty success were complete malleus removal, number of subsites involved by cholesteatoma, revision surgery, and partial versus total ossicular replacement prosthesis.
Multivariate Logistic Regression for Ossiculoplasty Success (Air-Bone Gap <20dB).
Abbreviations: CI, confidence interval; PORP, partial ossicular replacement prosthesis; TORP, total ossicular replacement prosthesis.
Results
Two otologists performed 201 operations on 113 patients, which were part of the initial analysis. Twenty-one surgeries were excluded due to an unclear operative note. The remaining 180 surgeries were included in the study (Figure 1). Thirty-one (17%) surgeries included CMR. One hundred thirty one (73%) surgeries included ossiculoplasty, though those that did not often underwent ossiculoplasty as part of a second-stage surgery several months later.
Flow diagram of patient selection and type of surgery.
Summary demographic and clinical characteristics of the study population, broken down by CMR and non-CMR, are displayed in Table 1. Comparison between CMR and non-CMR groups revealed no statistically significant differences in age, sex, right versus left ear, preoperative air-bone gap, revision versus primary surgery, residual disease, or closure of air-bone gap to less than 20 dBHL. However, there was a statistically significant difference in the number of subsites (CMR: 4.1 vs 3.2 for non-CMR, P = .002). In aggregate, patients ranged in age from 4 to 78 years with a mean of 30.4 years (SD = 21). Ninety-four surgeries (52%) were primary surgeries, 39 (21%) were planned second-look procedures, and 47 (26%) were revision surgeries for recurrent or residual disease. Sixty patients (33%) had undergone 1 previous surgery prior to their operation, 18 (10%) underwent 2 previous surgeries, and 3 (1.6%) underwent 3 previous surgeries. No patient had undergone more than 3 previous surgeries. The length of follow-up was 8 to 60 months with a mean of 19.2 months (SD = 7.8). The overall residual disease rate was 17.3%, and the recurrence rate was 23.1%. For patients that underwent ossiculoplasty and had no recurrence or residual disease, the mean postoperative air-bone gap was 22.3 dB HL (SD = 14). The success rate (<20 dB air-bone gap) of ossiculoplasty for those without disease after 9 months was 47%. The percentage of those with <30 dB air-bone gap after ossiculoplasty without residual or recurrent disease that underwent ossiculoplasty was 79%.
Multivariate logistic regression results are summarized in Table 3. For residual disease, the adjusted odds ratio (AOR) of complete malleus removal was 0.29 (95% CI, 0.074-1.10, P = .076), which at a threshold α = 0.05 approached but did not meet statistical significance. An increase in the number of subsites involved by cholesteatoma predicted an increased odds of residual disease, which was statistically significant with an AOR of 1.53 (95% CI, 1.3-1.9, P < .00010). Whether the right versus left ear was involved had no statistically significant effect on residual disease with an AOR of 0.75 (95% CI, 0.32-1.8, P = .52).
Multivariate Logistic Regression for Residual Disease.
Abbreviations: CI, confidence interval; PORP, partial ossicular replacement prosthesis; TORP, total ossicular replacement prosthesis.
Reached statistical significance.
CMR, the number of subsites involved, revision surgery, or use of a partial versus total ossicular reconstruction prosthesis did not have a statistically significant effect on ossiculoplasty success when defined at less than or equal to 20 dB air-bone gap (P = .43, .77, .15, .18, respectively).
Discussion
Due to the high rate of residual cholesteatoma after canal wall up surgery,5,6 there is an interest in identifying techniques that reduce the rate of residual disease without compromising ossiculoplasty success. CMR can improve exposure of the middle ear and assist in removing cholesteatoma involving the malleus, which can be tedious and difficult to dissect. In the authors’ experience, during second-look procedures, we have found cholesteatoma “hiding” behind an intact malleus handle. Thus, its removal may be associated with lower recurrence rates, though there is a general fear of CMR due to disruption of the tympanic membrane, and meta-analysis data suggest absence of the malleus is associated with worse postoperative hearing outcomes. 7 Regarding studies of CMR, most studies are not specific to cholesteatoma surgery, are limited to primary ossiculoplasty, and the malleus tended to be absent rather than deliberately removed. Also, these studies have not controlled for extent of disease, which is a significant confounder in the relationship between malleus removal and ossiculoplasty success and postoperative hearing outcomes. 7 Similarly, only 1 study has examined the effect of CMR on residual disease 9 ; it also did not control for extent of disease. Due to these study limitations, in the authors’ opinion, the current literature has not adequately examined the independent effect of CMR (controlling for extent of disease and other confounding factors such as revision surgery), which was the impetus for this study.
Comparison of the CMR and non-CMR groups in our study showed that CMR was used in cases where cholesteatoma involved more subsites of the middle ear and mastoid. This is consistent with our institution’s use of CMR generally for more extensive cases. This further supports the need for a study that controls for extent of disease.
The most notable finding from our analysis was that CMR was not independently predictive of ossiculoplasty success. This differs from previous uncontrolled analyses, which suggested that it is associated with worse hearing outcomes. There may be several reasons for this. It may be that CMR does not compromise ossiculoplasty success and that findings from previous studies are confounded by selection bias (the malleus is most likely absent or completely removed in situations of extensive disease and ossicular damage, which result in worse postoperative hearing outcomes). It also may be that our study was not powered to detect a small difference in ossiculoplasty success between CMR, although if there was a small difference that went undetected, it was likely not large enough to be clinically significant based on the AOR of 1.75 (95% CI, 0.43-7.0, P = .43) for ossiculoplasty success. It is possible that the techniques of the 2 otologists who perform the surgeries may be slightly different from other studies, but the fact that the overall recurrence and residual disease rates were similar to other studies suggests techniques used were probably similar.10,11
Our hypothesis that CMR would independently decrease the rate of residual disease was not statistically supported with an AOR of 0.285 for residual disease (95% CI, 0.074-1.1, P = .076), but it did approach significance, and larger studies may be able to detect a difference between the techniques. This suggests that complete removal of the malleus is a safe technique, which could be useful if there is extensive disease on the malleus or to improve exposure of the anterior middle ear space, despite not independently reducing the rate of residual disease in our study.
The main limitation of our study was the retrospective collection of data and limitation to 2 otologists at 1 center. We relied on the surgeons’ operative reports to define most of the variable characteristics of surgery, although we standardized this as much as possible by using 2 independent reviewers and a rigorous, predefined system for classification and excluding reports that were ambiguous. We also had 2 surgeons who may have had variations in their description of the surgery or surgical techniques. There was also potential for misclassification of indolent recurrent or residual cholesteatomas for patients who did not have a second-look surgery that may not have manifested inside the 8-month minimum follow-up period, though overall this would only affect a small subset as overall follow-up was much longer. There is also possibility that some residual cholesteatomas could have been classified as recurrent, particularly for patients with longer follow-up and those that developed new retraction pockets after surgery.
We did not find that partial versus total ossicular reconstruction prosthesis had a statistically significant effect on ossiculoplasty success at either threshold (<20 or 30 db HL), which has been shown in some previous studies.12,13 The overall success rate for ossiculoplasty at threshold for success <20 db HL was lower than most other studies have reported.14,15 We speculate that this may be because we based success on a postoperative audiogram done 3 months postoperatively, which was earlier than most other studies have reported, as audiograms 1 year after surgery were often not available to us, and this finding may reflect persistent middle ear effusion postoperatively. There was also a large number of patients in our study who had undergone previous surgeries, which may predispose to ossiculoplasty failure due to damage to stapes and oval window structures from surgery or disease as well chronic changes of the middle ear spaces from long-term chronic otitis media and cholesteatoma.
Conclusion
Overall, our study, which represents the first analysis of CMR in a manner that controls for confounding risk factors for outcomes of residual disease and ossiculoplasty success, suggests that CMR does not independently decrease the rate of residual disease but is a safe technique as it does not affect ossiculoplasty success. Surgeons may consider CMR for treatment of cholesteatomas involving the manubrium of the malleus and/or the anterior middle ear without compromising hearing outcomes. Larger, prospective, and/or randomized studies can better clarify the validity of CMR in cholesteatoma surgery.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.