Introduction
Endoscopic ear surgery (EES) has an established role in otologic surgery as a standalone operative technique or an adjunct to traditional binocular microscopy. Many surgeons now employ the technique for widened visibility and postural ergonomics.Reference Arrighi-Allisan, Fan, Wong, Wong, Garvey and Schwam1, Reference Kozin, Kiringoda and Lee2 EES has gained traction for a wide variety of otologic applications including chronic ear, otosclerosis and, more recently, skull base tumors.Reference Ho, Patel, Ballard, Rosenfeld and Chandrasekhar3–Reference Koukkoullis, Tóth, Gede, Szakács, Hegyi and Varga7 Since the surgeon must use one hand to drive the endoscope, bleeding management requires special attention due to the inherent limitations of single-handed instrumentation. Furthermore, the endoscopic camera is situated close to the operative tissue which can result in an obscured view from excessive bleeding. As such, if bleeding is not well controlled, surgical misadventure could occur or surgical progress slowed.
Despite this possible pitfall, EES has been shown to have similar operative time, intra-operative blood loss and lead to similar outcomes when compared to the traditional microscopic approach.Reference Kiringoda, Kozin and Lee8, Reference Kozin, Gulati, Kaplan, Lehmann, Remenschneider and Landegger9 Strategies to keep the field dry include canal injection of vasoconstrictor agents like diluted epinephrine, topical application of cottonoids soaked in vasoconstrictors, direct electrocautery, use of instruments equipped with a self-suction attachment and/or use of an endoscope holder to facilitate two-handed instrument manipulation.Reference Anschuetz, Bonali, Guarino, Fabbri, Alicandri-Ciufelli and Villari10
While numerous strategies exist to manage intra-operative bleeding in standard EES, concern can arise when faced with a patient on anticoagulation (AC) or antiplatelet (AP) medications. The safety of performing microscopic otologic surgery without pause of AC/AP has been reported, but the impact in EES is not well understood.Reference Hunter, Carlson, Sweeney, Tombers, Wanna and Driscoll11 EES is often performed electively; therefore, an increase in morbidity that could result from AC/AP pause should be avoided when possible.Reference Moster and Bolliger12 While EES can reduce the need for a postauricular incision which may reduce overall bleeding, due to the unique procedural strategies employed for bleeding management in EES, we sought to better understand the impact AC/AP may have on procedural progress and success.
Methods
This single-center, single surgeon, retrospective case–control study was approved by the New York University Institutional Review Board (IRB; #22-01408). Paediatric and adult patients who underwent EES from 1 April 1 2018, to 1 January 2023 were included. At the study center, it is standard practice to keep patients on pre-existing AC/AP medications during otologic surgery except in select few cases, so those who remained on AC/AP were identified by close review of pre-operative medical clearance documentation and post-operative medication reconciliation records; these patients were selected as cases. Each case was then randomly matched to two control patients who underwent EES over the same time period. Matching of case and control was done based on procedure and indication: tympanoplasty for traumatic perforation, tympanoplasty for chronic otitis media (COM)/cholesteatoma, stapedectomy, tympanomastoidectomy with intact canal wall for COM/cholesteatoma and canal wall down (CWD) tympanomastoidectomy for chronic otitis media/cholesteatoma.
Surgical videos were obtained from the operating surgeon for all cases and controls and edited down to three clips of key portions of the procedure (3-minutes each) as follows. (1) For COM/cholesteatoma tympanoplasty, they were: (i) tympanomeatal flap elevation, (ii) middle ear scar/cholesteatoma dissection and (iii) TM/ossicular chain reconstruction placement. (2) For traumatic tympanoplasty, these steps were: (i) perforation preparation, (ii) tympanomeatal flap elevation and (iii) tympanic membrane (TM) reconstruction placement. (3) For stapedectomy they were: (i) tympanomeatal flap elevation, (ii) footplate fenestration and (iii) stapes prosthesis placement. (4) Finally, for canal wall up (CWU) and CWD tympanomastoidectomy, they were: (i) tympanomeatal flap elevation, (ii) middle ear scar/cholesteatoma dissection and (iii) TM/ossicular chain reconstruction placement.
Case patients were excluded if a control patient could not be identified for matching or full surgical videos were not available. When a procedure utilised combined visualisation techniques, only the endoscopic portions were selected.
Two surgeons blinded to patient identifiers and case details reviewed the clips for the impact of bleeding on procedural progress using the Modena Bleeding Scale (MBS), a validated instrument designed to evaluate the impact of bleeding in EES which serves as the primary endpoint for this study.Reference Alicandri-Ciufelli, Pingani, Mariano, Anschuetz, Molinari and Marchioni13 The surgeons that scored videos were separate from the individuals that selected case/control patients and edited the videos for review.
The MBS scoring rubric ranges from a score of 1 for no bleeding, 2 for bleeding easily controlled by ordinary measures (suctioning, washing, packing) without any significant modification/slowing of the procedure, 3 for bleeding slowing the surgical procedure, 4 if the majority of manoeuvers performed are dedicated solely to bleeding control or, finally, 5 for bleeding that prevents every surgical procedure except those dedicated to bleeding control.
All case and control patients included had their charts reviewed to extract basic demographic and clinical information, case length, estimated intra-operative blood loss based on surgeon judgement, intra- or post-operative complications, audiologic outcomes and post-operative success, all analysed as secondary outcomes. Post-operative success was defined based on procedure and indication; further details are available in Table 1. Patient data were extracted up to 1.5 years from surgery, utilising the most recent follow-up information if 1.5-year data were not available.
Five pathologies observed in patients who underwent endoscopic ear surgery while on anticoagulation or antiplatelet medication along with corresponding criteria deemed adequate for surgical success
Table 1 Long description
The table lists five ear surgery case types and the criteria used to judge whether surgery was successful. For tympanoplasty done for chronic otitis media with cholesteatoma, success requires closure of the eardrum perforation, improvement in the air to bone hearing gap, and no remaining disease on follow-up exam or MRI. For tympanoplasty after trauma, success is defined by eardrum perforation closure and improvement in the air to bone hearing gap. For stapedectomy, the only listed success criterion is improvement in the air to bone hearing gap. For tympanomastoidectomy using either canal wall up or canal wall down approaches, success is defined as no residual disease on follow-up exam or MRI. Overall, hearing improvement criteria apply to tympanoplasty and stapedectomy, while disease clearance on follow-up is the key criterion for mastoidectomy approaches.
ABG = air–bone-gap; COM = chronic otitis media; CWD = canal wall down; CWU = canal wall up; MRI = magnetic resonance imaging; TM = tympanic membrane.
MBS scores and secondary outcomes were compared between case and control patients utilising descriptive statistics and paired t-tests assuming equal variance using Microsoft Excel version 16.81 and GraphPad Prism version 10.1.0. Additionally, equivalence analysis was conducted to compare MBS scores in cases vs. controls. The two one-sided tests (TOST) procedure was employed with a significance level of α = 0.05 for each test.Reference Schuirmann14 The equivalence margin was pre-specified as minus 0.4 to plus 0.4, with 0.4 MBS points representing the largest difference we considered clinically insignificant for the purpose of the present study.
Results
Seventeen cases of patients who remained on their pre-existing AC/AP medications during EES were identified over the reviewed time period. Two cases were excluded due to an inability to match with paired controls: one with a middle ear glomus tumor and one with an external ear canal osteoma. A third case was excluded as the patient had not had their procedure recorded to facilitate video review. The remaining 14 cases were successfully matched based on surgical procedure and indication. All included cases in the study were performed by a single neurotologist. The primary strategies employed to manage intra-operative bleeding at the center include thorough local pre-operative injection of 1 per cent lidocaine with 1:100,000 epinephrine into the external ear canal (EAC) at the bony-cartilaginous junction, in addition to the use of ¼ × ¼ patties soaked in 1:1,000 epinephrine which are replaced throughout the procedure as needed. No suction elevators were used in the studied cases.
Demographics
A list of patient demographic information and case details can be found in Table 2. Cases were significantly older than controls, with cases having a median age of 54.9 years (range 10.6 to 77.3) compared to 37.8 (range 4.3 to 69.7; p < 0.01). Cases also had significantly more males (case: 85.7 per cent male; control: 50.0 per cent male; p < 0.01). Eleven cases (78.6 per cent) were on aspirin 81 mg daily, two (14.2 per cent) on coumadin (4 mg and 6 mg daily) and one (7.1 per cent) on aspirin 325 mg daily. When available, pre-operative medical clearance was reviewed with attention to coagulation factor assessment. The two patients on anticoagulation (patients 5 and 14 on Table 2) had an average INR of 2.35 (SD 0.07), falling within the therapeutic target of 2–3.
Demographics and surgical details of included cases and matched controls
Table 2 Long description
Rows list 14 surgical cases, each followed by two matched controls, with age, sex, ear pathology, procedure type, antiplatelet or anticoagulant use, operative time, estimated blood loss, and brief notes. Most entries involve chronic otitis media with cholesteatoma treated with tympanoplasty and middle ear exploration or canal-wall-up or canal-wall-down tympanomastoidectomy; three cases involve traumatic tympanic membrane perforation and two involve otosclerosis treated with stapedotomy and ossicular chain reconstruction. Cases are typically on aspirin once daily, with two cases on warfarin; controls generally have no listed antiplatelet or anticoagulant therapy. Operative time ranges from 42 to 269 minutes, with the longest times appearing in tympanoplasty and middle ear exploration and canal-wall-up tympanomastoidectomy groups. Estimated blood loss is low overall, most often 5 ml, occasionally 10 to 15 ml, and reaching 20 ml in one case and one control. Notes highlight a few special circumstances, including procedures performed with Eustachian tube balloon dilation, prior outside-institution surgery, planned second-look operations, and one control with post-operative day 2 surgical site bleeding. The table provides descriptive case-control details and does not establish causal links between medication use and outcomes.
Note: Each case listed with two matched controls below.
ASA = aspirin; mg = milligrams; COM = chronic otitis media; CWD = canal-wall-down; CWU = canal-wall-up; ETBD = Eustachian tube balloon dilation; MEE = middle ear exploration; M&T = myringotomy and tympanostomy tube; POD = post-operative day; OSH = outside institution; QD = once daily.
Cases and controls
Six (42.8 per cent) cases were categorised as tympanoplasty for COM/cholesteatoma, three (21.4 per cent) CWU tympanomastoidectomy, two (14.3 per cent) traumatic tympanoplasty, two (14.3 per cent) stapedectomy and one (7.1 per cent) CWD tympanomastoidectomy. One control for the CWD tympanomastoidectomy did not have an endoscopic video clip available for step 3 as that step of the case was completed under the microscope. The control patient was still included to evaluate steps 1 and 2, as no additional control patient could be identified due to the rarity of CWD tympanomastoidectomy occurring over the time period studied.
One case and one control in the tympanoplasty for COM/cholesteatoma category also had Eustachian tube balloon dilation (ETBD) performed during the same procedure. Another case in this category (case 2) had a prior otologic surgery at another institution two years before the examined intervention, and a control for CWU tympanomastoidectomy (case 11) had a prior cholesteatoma resection at our institution 17 months before.
MBS
Overall, cases had an average MBS score of 2.31 (SD 0.50) across all three steps compared to 2.23 (standard deviation [SD] 0.48) for controls (p = 0.63). When each case was compared with its two matched controls, there was no significant difference in MBS score both overall and per surgical step. MBS analysis per case and two controls is summarised in Figure 1. This was also true when cases/controls were grouped with others within their surgery category (Figure 2). Scores for the two patients on anticoagulant medications (cases 5 and 14) and the patient on a higher dose of aspirin (case 2) did not differ significantly from the remaining cases who were on the lower dose of aspirin.
Mean Modena bleeding score (MBS) for cases and controls with all steps included as points. Error bars represent one standard deviation. X-axis denotes each case with its controls plotted together to the right.
Figure 1 Long description
A point plot with vertical error bars compares Mean Modena bleeding scores for cases and controls across 14 cases. The x-axis is labeled ‘Case’ and the y-axis is labeled ‘MBS Score’, ranging from 0 to 5. Cases are marked in one style and controls in another, with error bars representing one standard deviation. The legend indicates ‘Cases (n=14): 2.3 (0.8)’ and ‘Both Controls (n=38): 2.2 (0.9)’, with a p-value of 0.63, suggesting no significant difference overall. The cases are grouped under different surgical categories: Tympanoplasty COM/Cholesteatoma (Cases 1-4), Tympanoplasty Traumatic (Cases 5-6), Stapes (Cases 7-8), CWU Tmastoid (Cases 9-12) and CWD Tmastoid (Cases 13-14). The plot shows that scores for cases and controls are similar across most cases, with no notable outliers or clusters. The layout is a two-series point plot with error bars at each case position, providing a clear comparison of scores between the two groups.
MBS separated by five pathologies observed. Each pathology is grouped with its respective cases and controls. Error bars represent one standard deviation.
Overall MBS score differences (between each case and the mean of the two corresponding controls) averaged 0.08, higher for cases than for controls. The 90 per cent confidence interval was [-0.16 to 0.31], completely within the chosen equivalence range of plus-or-minus 0.4 MBS points, supporting the conclusion of equivalence in outcomes (p < 0.03 for both one-sided tests).
When comparing MBS scores for each intra-operative step, controls had a significantly higher score for step 1 (mean 2.6; SD 0.9) than for steps 2 (2.2; 0.9) or 3 (1.9, 0.8; p < 0.01) but cases did not (step 1 mean 2.6, SD 0.7; step 2 mean 2.3, SD 0.7; step 3 mean 2.0, SD 0.9; p = 0.06). Score for step 3 was also significantly lower for controls when compared to the other two steps (p = 0.04) but the same was not true for cases (p = 0.08). Step 2 score was not significantly different for either cohort.
When comparing case and control cases based on surgical step and surgery type/indication and surgical step, the score for step 1 within the control group for stapedectomy (mean 3.0; SD 0.8) was significantly higher than the other two steps (step 2: 1.2, 0.5; step 3: 1.5, 0.6; p < 0.01), and the score for step 3 within the control group for tympanoplasty for trauma was significantly lower than steps 1 and 2 (step 1; 2.7, 0.5; step 2; 2.5, 0.6; step 3; 1.5, 0.6; p < 0.01). Analysis of other pathology groups did not show any significant differences either overall or by surgical step. See details in Figure 3.
MBS for all cases and controls per surgical step. Error bars represent one standard deviation. X-axis denotes each step as described in the body of the text. Asterisk denotes significant difference.
Secondary outcomes
Estimated intra-operative blood loss (EBL) for cases was comparable to that for controls, with an average of 7.8 ml (SD 5.1) for cases and 8.7 ml (SD 7.6) for controls (p = 0.70). Procedure duration was also similar (p = 0.95). See Table 2 for full details of all cases.
For the two traumatic tympanoplasty cases and their four controls, all perforations were closed at three-month post-operative follow-up and audiogram demonstrated minimal residual air–bone gap (ABG), at an average of 2.5dB (SD 3.5) for cases and 5.7 (SD 4.3) for controls (p=0.64). Similar outcomes were seen for tympanoplasty for COM/cholesteatoma with no persistent disease at 18-months post-operatively or the most recent follow-up available, as well as no difference in residual ABG (p = 0.50). Stapedectomy cases had an average residual ABG of 12 (SD 8.5) which was comparable to 6 (SD 7.6) for controls (p = 0.27). Two CWU patients (case 12 and one of its controls) have planned second-look procedures. One is planned due to the original extent of disease invading the tegmen and an increase in ABG from 15dB at three months post-operatively to 20dB one year post-operatively and the other due to residual TM retraction. The remaining CWU and CWD tympanomastoidectomy patients all had no residual disease on follow-up exam or magnetic resonance imaging (MRI) when available. One control experienced low-volume bloody otorrhea which resolved with local vasoconstrictor medication (oxymetazoline) and additional packing for pressure. No other peri- or post-operative complications were observed.
Discussion
EES provides unique visualisation properties which can be hindered in the presence of significant surgical bleeding. While multiple strategies exist to manage bleeding, previous literature has not directly assessed the impact of AC/AP medications that may increase the risk and severity of intra- and post-operative bleeding, and there are currently no standardised guidelines for peri-operative AC/AP for patients undergoing otologic surgery. One study by Hunter et al. addressed the safety of performing traditional microscopic surgery, specifically cochlear implants, on 46 adults while on AC/AP, which found no increased risk of peri- or post-operative complications and, thus, recommended continuing any pre-operative medications through surgery.Reference Hunter, Carlson, Sweeney, Tombers, Wanna and Driscoll11 In cholesteatoma surgery, Kiringoda et al. found that more bleeding was correlated with the presence of residual disease on follow-up, although extent of original pathology was the most significant factor.Reference Kiringoda, Kozin and Lee8 Lastly, a recent study by Molinari et al. used the MBS to evaluate the effects of inhalational anesthesia compared to total intravenous anesthesia on bleeding in EES.Reference Molinari, Reale, Alicandri-Ciufelli, Villari, Presutti and Feminò15 Notably, this study excluded patients on any AC/AP.
Our study is the first to evaluate multiple pathologies and find no significant difference in intra-operative bleeding score or outcomes for those on AC/AP undergoing EES and also establish equivalence for the main outcome measure of average MBS score. It is also useful to note that the two patients on warfarin, an anticoagulant medication, and the one on a higher dose of aspirin, an antiplatelet medication, did not show any difference in score or outcome compared to the other cases or controls. Based on mechanism, AC theoretically poses a larger risk by inhibiting the entire coagulation cascade rather than accumulation of the platelet plug. Prior studies in microlaryngeal and endoscopic sinus surgery have demonstrated no significant difference in risk between patients who are on AC/AP, however other fields such as Ophthalmology have observed increased peri-operative bleeding for those on AC.Reference Oh, Smiddy and Kim16–Reference Le Nobel, Cushing, Papsin and James19 Further study is needed to elucidate if any differences exist in this patient population.
In addition to comparing the cases and control scores overall, we also broke down the scores per surgical step, as outlined in the Methods section. The first step for each procedure/indication was universal: the raising of a tympanomeatal flap. This step consistently scored higher on MBS than the two subsequent steps for all cases and controls. As raising a tympanomeatal flap involves the most direct soft tissue dissection, this may inherently explain these findings. While concern may arise later in the case that bleeding may worsen after vasoconstrictor injection wears off, particularly in the setting of patients with reduced clotting ability, this was not observed. This is especially favourable for critical portions of the case such as placing a prosthesis, in which the average MBS was only 2.3 in our six cases.
While our study demonstrated equivalence between treatments using the plus-or-minus 0.4 MBS range, our sample size (n = 14 cases) was very small, which may have limited our ability to detect subtle differences between outcomes. We also acknowledge that our chosen equivalence margin may have been too liberal. To address this concern, we conducted a post-hoc analysis using a stricter equivalence range of plus-or-minus 0.3 MBS points. With this more stringent criterion, we were unable to establish equivalence between the treatments. The 90 per cent confidence interval for the MBS difference (-0.16 to 0.31) extended beyond the 0.3 threshold, indicating that we cannot rule out potentially meaningful differences at this more conservative level. These findings highlight the importance of careful consideration when selecting equivalence margins and underscore the need for larger, more highly powered studies to detect and rule out smaller but potentially important treatment differences. Future research should aim to recruit larger sample sizes and consider using more conservative equivalence margins to provide more definitive evidence of therapeutic equivalence in this context.
It is also important to contextualise our findings by remembering that our study grouped together multiple procedures/indications to allow for more statistical power and generalisability; however, this does limit the ability to compare across multiple pathologies since the steps examined were different. Despite this, the trend of decreasing bleeding as the case progresses over time is still a useful finding.
Lastly, another limitation of our study is the limits of the rating scale used. Though the MBS is useful to objectively describe bleeding, it is inherently limited by the judgement of the rating surgeon and short duration of the clip examined. It also does not capture the variability based on operating surgeon, surgical technique and patient pathology which undeniably contributes to the amount of bleeding in each clip analysed. It is reassuring that the lack of significant difference appreciated between cases and controls based on MBS is echoed in their similar outcomes; however, to fully describe best practices, more studies are needed. Other limitations include the retrospective nature of the study and the possible difference in characteristics between cases and their controls despite the efforts towards matching. It is also important to note that the two patients in our cohort who were on anticoagulation had INR values in the expected therapeutic range, so our results may not be applicable to patients with higher INRs. Additionally, since only two cases were on anticoagulation, the ability to draw conclusions from this subgroup is limited.
Future studies should also include studies with a wider variety of patients and pathologies, especially as the applications of EES continue to expand, and consideration of the relationship between bleeding ratings and relevant co-morbidities, such as hypertension or renal disease. There continues to be a lack of high-quality studies comparing hemostatic techniques which would be practically insightful, although practically difficult to develop studies for.
• Endoscopic ear surgery is particularly vulnerable to procedural slowing and thus possibly surgical complications in the setting of uncontrolled bleeding. The patient on anticoagulant (AC) or antiplatelet (AP) medication may pose such a situation. There are currently no guidelines for peri-procedural management of these medications
• The current retrospective case–control cohort study uses a previously validated scoring system to compare intra-operative procedural progress and post-operative outcomes between those on AC/AP and normal controls
• No additional risk was found for intra-operative visibility, procedural progress, complications or outcomes for those on AC/AP during their surgery across a variety of pathologies
Conclusion
Overall, the results of our study show the relative safety of performing EES in the context of continued AC/AP therapy throughout the peri-operative period, as there were no statistically significant differences found between the case and control groups in terms of MBS score, surgical time or surgical outcomes based on indication, and equivalence in MBS score was established for a plus-or-minus 0.4 point equivalence range. This is favourable as it reduces the chance of surgical complications which can arise from holding AC/AP medications in anticipation of surgery.
Funding
This study was not funded.
Competing interests
None.
IRB approval
New York University IRB #22-01408.
