Hearing Dysfunction After Treatment with Teprotumumab for Thyroid Eye Disease

Abstract

Purpose:

To characterize the frequency, severity and resolution of hearing dysfunction in patients treated with teprotumumab for thyroid eye disease (TED).

Design:

Prospective observational case series

Methods:

Ophthalmic examination and adverse event assessment, including otologic symptoms, were performed at baseline, after infusions 2, 4, and 8, and at 6-month follow-up in consecutive patients who received at least 4 teprotumumab infusions. Labs were collected at baseline and during treatment. Audiometry, patulous Eustachian tube (PET) testing and otolaryngology evaluation were obtained for patients with new or worsening otologic symptoms, with a subset obtaining baseline and post-treatment testing.

Results:

Twenty-seven patients were analyzed (24 females, 3 males, average 56.3-years-old). Twenty-two patients (81.5%) developed new subjective otologic symptoms, after a mean of 3.8 infusions (SD 1.8). At 39.2 week average follow-up after the last infusion, most patients with tinnitus (100%), ear plugging/fullness (90.9%), and autophony (83.3%) experienced symptom resolution, while only 45.5% (5 of 11) of patients with subjective hearing loss/decreased word comprehension experienced resolution. Six patients underwent baseline and post-treatment audiometry, 5 of whom developed teprotumumab-related sensorineural hearing loss (SNHL) and one patient also developed PET. Three of the 5 patients with teprotumumab-related SNHL had persistent subjective hearing loss at last follow-up. A prior history of hearing loss was discovered as a risk factor for teprotumumab-related SNHL (p=0.008).

Conclusions:

Hearing loss is a concerning adverse event of teprotumumab and its mechanism and reversibility should be further studied. Until risk factors for hearing loss are better understood, we recommend baseline audiometry with PET testing and repeat testing if new otologic symptoms develop. Screening, monitoring and prevention guidelines are needed.

Table of Contents Statement

This study describes the onset, severity, and reversibility of various otologic symptoms in patients with thyroid eye disease treated with teprotumumab. Changes in audiometry and patulous Eustachian tube testing are characterized in a small subset. The majority of patients developed new otologic symptoms with treatment. Our findings raise awareness of this adverse effect, support audiometric testing around treatment, and invite development of guidelines for screening, monitoring, and management of hearing dysfunction in teprotumumab-treated patients.

Introduction

Thyroid eye disease (TED) is an unpredictable autoimmune disease that can significantly impact quality of life (QoL). Conventional treatments often have limited response and high relapse rates with unfavorable side effect profiles. Teprotumumab, an insulin-like growth factor-I (IGF-I) receptor antibody, improves proptosis, diplopia, soft tissue inflammation and QoL,^1,2 and is the first medication approved by the U.S. Food and Drug Administration for the treatment of TED.

Due to systemic IGF-I signaling inhibition with teprotumumab, side effects are expected. In the ear, IGF-I plays a critical physiologic role³ in both animals^4–9 and humans.^10–12 In the teprotumumab randomized controlled trials (RCTs), 10% of patients experienced mild-to-moderate hearing dysfunction,^1,2 which were reported to be reversible. Details regarding the timing, extent, time to resolution, and formal audiology evaluation were not provided. The mechanism of hearing impairment is unclear. It is known, however, that IGF-I deficiency is characterized by sensorineural hearing loss (SNHL).^13–17 In animal models, IGF-I trophic support is vital in the cochleovestibular ganglion’s early development,⁷ while lack of IGF-I leads to auditory neuron apoptosis, and IGF-I null mice have been proposed as a model of syndromic deafness.⁶ Patients with Laron syndrome, a rare congenital disorder characterized by growth hormone (GH) resistance and low IGF-I levels, develop SNHL when young, while treatment with recombinant IGF-I preserves hearing.¹³ Low IGF-I is also associated with SNHL in Turner and Noonan syndromes.^10,11 In longitudinal aging studies, lower IGF-I levels are associated with an increased risk of developing age-related hearing loss.¹⁶ Interestingly, recombinant IGF-I successfully protects against noise-induced hearing loss in animal models^18,19 and can treat sudden onset SNHL in humans.^20–22 In one RCT of patients with sudden SNHL, IGF-I delivered through gelatin hydrogels in the middle ear outperformed intratympanic dexamethasone injections in restoring hearing.²²

The effect of pharmacologic IGF-I inhibition on hearing has not been previously studied,²³ however case reports are beginning to emerge.^24–27 As teprotumumab use expands, an understanding of otologic side effects is crucial, since significant hearing loss can profoundly affect QoL,^28–30 while autophony interferes with daily performance.³¹ There are currently no guidelines to screen for, monitor, or treat auditory dysfunction, though surveillance has been recommended.²⁴ Here we describe our single center experience with hearing dysfunction in patients treated for TED with teprotumumab.

Methods

This is a prospective observational case series of otologic symptoms in consecutive TED patients treated with teprotumumab between February 2020 and May 2021. Institutional Review Board (IRB)/Ethics Committee approval was obtained. This study adheres to the tenets of the Declaration of Helsinki as amended in 2013 and the Health Insurance Portability and Accountability Act. Informed consent was obtained from all patients.

Patients received intravenous teprotumumab (10 mg/kg for the first infusion then 20 mg/kg) every three weeks, with a plan for 8 infusions, as per protocol.^1,2 Patients were evaluated at baseline, after infusion number 2, 4, and 8, and at 6-month follow-up, with complete eye exam and adverse event (AE) assessment at each visit. Due to government-mandated COVID-19 vaccine production orders, teprotumumab production was halted in December 2020. Infusions were, therefore, interrupted and resumed a few weeks later for 10 patients who were monitored at baseline, throughout treatment, and after therapy. Only patients who received ≥4 teprotumumab infusions were included.

Baseline audiometry and patulous Eustachian tube (PET) testing were recommended for all patients but not enforced due to COVID-19 restrictions. PET testing is a tympanometry pressure test checking for tympanic membrane movement with regular and with forced nasal breathing through one nostril. Testing was considered baseline if performed within one month prior or within 3 weeks after the first infusion. Audiometry and PET testing were also recommended to patients with new or worsening otologic symptoms during treatment. Due to COVID-19 restrictions, some patients refused testing. If numerous audiograms were available during therapy, only the latest audiogram was considered. At each follow-up, patients were asked about otologic symptoms, and symptom onset/resolution (if applicable). Audiometry included both pure tone testing (thresholds at 500, 1000, 2000, 4000, and 8000 Hz) as well as speech testing with word recognition scores (percentage correct of single-syllable words presented to patients at an audible level — 30 dB louder than their speech reception thresholds). Patients who met American Speech-Language-Hearing Association (ASHA) criteria for ototoxicity (>20 dB pure-tone threshold shift at one frequency, >10 dB shift at each of two consecutive frequencies, or threshold response shifting to “no response” at three consecutive test frequencies)³² or who developed abnormal PET testing were evaluated by an otolaryngologist. If testing revealed new significant hearing loss, infusions were halted, and the patient was closely monitored.

Patient gender, ethnicity, thyroid diagnosis and treatments, presence of diabetes, hypertension, smoking status, ototoxic medication use and history of hearing dysfunction were documented. Baseline and mid-treatment (after 4 doses) laboratory data included thyroid-stimulating hormone (TSH), free thyroxine (FT4), total triiodothyronine (T3), hemoglobin A1c (HbA1c), GH, and IGF-I. Patients were classified according to FT4 and T3 levels in three categories: normal (normal FT4 and T3), high (high FT4 or T3), or low thyroid hormone levels (low FT4), based on the reference ranges provided by each respective laboratory. TSH levels were not used in this classification, as they lag behind hyperthyroidism correction.

Patient characteristics were described using means and standard deviations for continuous and proportions for categorical variables. Differences between patients with and without subjective otologic symptoms, and differences between patients with audiometry data suggestive of teprotumumab-related SNHL and asymptomatic patients were characterized with descriptive analyses. Categorical variables were compared using Pearson’s chi-square or Fisher’s exact tests and continuous variables were compared using Student t-test. P-value <0.05 was considered statistically significant. All statistical analyses were conducted using R 3.6.1 (R foundation for Statistical Computing, Vienna, Austria).

Results

Patient characteristics and risk factor evaluation

Twenty-seven patients were analyzed, 24 females (88.9%) and 3 males (11.1%) with an average age of 56.3 years (standard deviation (SD) 15.1). Eighteen patients underwent audiometric and PET testing throughout the study: 6 baseline and post-treatment, 8 post-treatment only and 4 baseline only. Patients had an average of 7.1 infusions (SD 1.5) and were followed for an average of 55.9 weeks (SD 15.5) after teprotumumab initiation and 39.2 weeks (SD 14.2) after completion of infusions. Treatment for Graves’ disease varied. Eleven (40.7%) patients received anti-thyroid drugs (ATD) alone and 2 (7.4%) patients did not require treatment as they were euthyroid. Eight (29.6%) patients were treated with a combination of modalities (ATDs, radioactive iodine ablation (RAI), and/or thyroidectomy). Six (22.2%) patients underwent a thyroidectomy and 9 (33.3%) received radioactive iodine ablation (RAI) alone or in combination with other modalities. Pertinent medical history included: diabetes, pre-diabetes, or steroid-induced diabetes (n=9, 33.3%), hypertension (n=7, 25.9%), and subjective baseline hearing loss (n=10, 37%). Most patients had a prior history of ototoxic medication use (n=25, 92.6%) (ibuprofen, aspirin, naproxen, meloxicam, acetazolamide, furosemide, hydrochlorothiazide, paroxetine, duloxetine, vancomycin) (Table 1).

Table 1:

Demographic and Disease Characteristics of Patients With and Without New or Worsening Subjective Otologic Symptoms

	Total (n=27)		With New or Worsening Otologic Symptoms (n=22)		Without New or Worsening Otologic Symptoms (n=5)		p-value a
	Mean / N	SD / %	Mean / N	SD / %	Mean / N	SD / %
Age	56.3	15.1	56.5	13.3	55.4	23.4	0.924
Sex							0.474
Male	3	11.1%	2	9.1%	1	20.0%
Female	24	88.9%	20	90.9%	4	80.0%
Number of Infusions	7.1	1.5	7.3	1.4	6.4	1.7	0.326
Follow-up Time (since first infusion, weeks) b	55.9	15.5	56.3	14.1	54.4	21.8	0.862
Follow-up Time (since last infusion, weeks) b	39.2	14.2	39.6	13.4	37.9	18.7	0.856
Prior Graves’ Disease Treatment
Anti-thyroid medication alone	11	40.7%	8	36.4%	3	60.0%
Thyroidectomy alone	2	7.4%	2	9.1%	0	0.0%
Radioactive iodine ablation alone	4	14.8%	4	18.2%	0	0.0%
Multiple therapies	8	29.6%	7	31.8%	1	20.0%
None, Euthyroid	2	7.4%	1	4.5%	1	20.0%
Thyroidectomy c							1.000
Yes	6	22.2%	5	22.7%	1	20.0%
No	21	77.8%	17	77.3%	4	80.0%
Radioactive iodine ablation c							0.136
Yes	9	33.3%	9	40.9%	0	0.0%
No	18	66.7%	13	59.1%	5	100.0%
Diabetes							0.174
Yes	2	7.4%	2	9.1%	0	0.0%
Pre-diabetic	7	25.9%	5	22.7%	2	40.0%
No	18	66.7%	15	68.2%	3	60.0%
Hypertension							1.000
Yes	7	25.9%	6	27.3%	1	20.0%
No	20	74.1%	16	72.7%	4	80.0%
Ototoxic Medication Use							1.000
Yes	25	92.6%	20	90.9%	5	100.0%
No	2	7.4%	2	9.1%	0	0.0%
Subjective Baseline Hearing Loss							0.124
Yes	10	37.0%	10	45.5%	0	0.0%
No	17	63.0%	12	54.5%	5	100.0%
Onset of Otologic Symptoms (after # of infusions)			3.8	1.8
Resolution of Otologic Symptoms d
Yes			13	68.4%
No			6	31.6%

^aP-values calculated using Pearson chi-square or Fisher’s exact tests for categorical variables and Student t-test for continous variables. P<0.05 was considered significant.

^bThree patients were lost to follow-up

^cAlone or in combination with other therapies

^dThree (13.6%) patients were lost to follow-up, therefore the denominator for resolution of otologic symptoms is 19.

Subjective otologic symptoms (Figure 1)

Figure 1. Otologic Symptom Trajectory.

Circles denote timing of teprotumumab infusions. Bars represent the onset and duration of the most common otologic symptoms: Ear fullness/pressure/plugging (blue), tinnitus/popping (orange), hearing loss/word recognition/muffled hearing (yellow), and autophony (green). “x” indicates the time of last follow-up. Patients 1, 4, and 11 were lost to follow-up shortly after completing therapy. Patients 10, 13, 17, 18, 20 and 22 had pre- and post-treatment audiologic evaluation.

Twenty-two of 27 (81.5%) patients, 20 females (90.9%) and 2 males (9.1%), developed new or worsening otologic symptoms after starting teprotumumab; while 5 (18.5%) patients, 4 females (80%) and 1 male (20%), remained asymptomatic throughout therapy. The mean age of symptomatic patients was 56.5 years (range 30 to 84 years) compared to 55.4 years (range 30 to 92 years) in the asymptomatic group. On average, new or worsening otologic symptoms started after 3.8 infusions (SD 1.8). Onset ranged from after the first infusion to after completion of 8 infusions. The most prevalent symptoms included ear plugging/fullness/pressure (n=13, 48.1%), muffled hearing/hearing loss/diminished word recognition (n=11, 40.7%), tinnitus/ear popping (n=10, 37.0%) and autophony (n=7, 25.9%) (Table 2).

Table 2:

Subjective Otologic Symptom Presentation and Resolution

	Number of patients who had symptom	Average number of infusions at onset of symptom	Resolution after last treatment	Time to Resolution (weeks)
Ear fullness/Plugging/Pressure	13 of 27	3.8 (SD 2.1)	90.9% (n=10)a	22.6 (SD 13.2)
Hearing loss, muffled hearing or worse word comprehension	11 of 27	3.3 (SD 1.3)	45.5% (n=5)	29.0 (SD19.1)
Tinnitus/ear popping	10 of 27	4.0 (SD 1.6)	100% (n=10)	23.1 (SD 12.3)
Autophony	7 of 27	2.7 (SD 1.6)	83.3% (n=5)b	18.1 (SD 11.9)

^aTwo patients lost to follow up

^bOne patient lost to follow up

There were no significant differences in demographic or clinical characteristics between symptomatic and asymptomatic patients (Table 1). In the symptomatic group, 45.5% (10 of 22 patients) had a history of baseline hearing loss compared to 0% (0 of 5 patients) in the asymptomatic group (p=0.124). Baseline (p=0.566) and mid-treatment FT4 and T3 hormone levels (p=1.000), mid-treatment HbA1c (p=0.683), GH (p=0.715), and IGF-I (p=0.714) values also did not differ (Supplemental Table 1). There was no significant difference in ATD use between symptomatic and asymptomatic patients. In the symptomatic group, 63.6% (14/22) had a history of treatment with methimazole (MMI) and 9.1% (2/22) with both MMI and propylthiouracil (PTU), compared to 60% (3/5) and 20% (1/5) in the asymptomatic group, respectively. Of the patients who had a history of ATD treatment, 9 of 22 symptomatic patients (40.9%) and 2 of 5 asymptomatic patients (40%) were concomitantly treated with ATD (all MMI) during teprotumumab therapy (p=0.970).

Three symptomatic patients were lost to follow-up after week 24, leaving 19 patients who were evaluated for resolution analysis on average 39.2 weeks after the last infusion. Thirteen of 19 (68.4%) patients had complete resolution of otologic symptoms, whereas 6 of 19 (31.6%) patients had persistence of at least one otologic symptom (Figure 2). The symptoms most likely to resolve, or return to baseline, were tinnitus/ear popping (10/10, 100%), ear fullness/plugging/pressure (10/11, 90.9%), and autophony (5/6, 83.3%). The symptom least likely to resolve was subjective hearing loss/muffled hearing/word comprehension (5/11, 45.5%). Of the 6 patients with persistent hearing loss symptoms, all had post-treatment audiometric testing which showed decreased hearing thresholds but only 3 of these patients had baseline audiometric testing to compare to. The latter 3 were diagnosed with teprotumumab-related SNHL and will be discussed as part of the teprotumumab-related SNHL cohort. Four of the six patients with persistent hearing loss symptoms required hearing aids. In patients whose symptoms resolved, the average time to resolution for each otologic symptom after completion or discontinuation of teprotumumab was 23.1 weeks for tinnitus/ear popping, 22.1 weeks for ear fullness/plugging/pressure, 18.1 weeks for autophony, and 29.0 weeks for hearing loss/muffled hearing/word comprehension (Table 2).

Figure 2. Subjective Symptom Rate and Resolution.

Time course of otologic symptom development and resolution among 27 patients following initiation of teprotumumab. Lines illustrate the frequency and resolution of ear fullness/pressure/plugging (blue), tinnitus/popping (orange), hearing loss/word recognition/muffled hearing (yellow), and autophony (green) among the 27 patients following initiation of teprotumumab. “x” indicates the time of loss to follow-up for 3 patients.

Audiogram and PET testing

Patients with baseline and post-treatment audiometric and PET testing

Six of 27 patients (T-013, T-018, T-024, T-025, T-028, T-031) had both baseline and either mid-treatment or post-treatment audiometry testing (Table 3) performed after an average of 5 infusions (SD: 1.7 infusions). Of these, 5 of 6 patients (7 of 12 ears) developed teprotumumab-related SNHL on audiometry and met ASHA criteria for ototoxicity (6 ears demonstrating >10 dB decline at 2 consecutive frequencies, 4 ears demonstrating >20 dB decline at 1 frequency) after teprotumumab compared to baseline. Two of these 5 patients required hearing aids. The average age of patients with teprotumumab-related SNHL was 70.8 years (range 59–84 years, SD 11.4) with 4 females and 1 male. All 5 patients in the teprotumumab-related SNHL cohort had a history of baseline hearing loss compared to none in the asymptomatic group (p=0.008). There were no significant differences in the remainder of demographic and clinical characteristics between asymptomatic patients and patients who developed teprotumumab-related SNHL (Supplemental Table 2).

Table 3:

Characteristics of Patients with Pre- and Post-Treatment Audiogram and Patulous Eustachian Tube Testing

Subject	Age	Sex	Baseline Hearing Loss	Subjective Symptoms	Objective Findings	Resolution of New Symptoms
T-013	50	F	N	Plugging	No worsening	Y
T-018	59	F	Y	Popping	SNHL	Y
T-024	82	F	Y	Worse Word Comprehension	SNHL	N
T-025	64	F	Y	Plugging	SNHL + PET	Y
				Popping
				Hearing Loss
				Fullness
				Autophony
T-028	65	M	Y	Hearing Loss	SNHL	N
				Muffled Hearing
				Tinnitus
T-031	84	F	Y	Hearing Loss	SNHL	N
				Autophony
				Ear Pressure

Demographic and disease characteristics of 6 patients who received both pre-treatment and post-treatment audiology and PET testing.

SNHL, sensorineural hearing loss; PET, patulous Eustachian tube

The average four-frequency pure tone average (PTA) decreased by a mean of 3.2 dB (95%CI −6.77 to 0.4, p=0.08) and there were small, but statistically significant differences in 3 frequencies with average threshold declines ranging from −1 to 8.5 dB (Figure 3A). The average word recognition score declined by 14% (95%CI −9.1 to 37.1, p=0.22) after treatment with teprotumumab (Figure 3B). Two patients (4 of 12 ears) demonstrated significant decline in word recognition beyond test-retest variation. Patient T-024 had a decrease from 88% and 92% at baseline, to 32% and 56% in the right and left ears, respectively, and patient T-028 had a decline in word recognition from 84% and 100% at baseline, to 48% and 56% in the right and left ears, respectively, at final testing.

Figure 3.

Pure Tone Audiometry (A) and Word Recognition Scores (B) for 12 ears with both pre- and post- treatment hearing testing. Overall changes were modest, with no statistically significant differences in overall world recognition or pure tone average (average at 500, 1000, 2000 and 4000 Hz). Seven of twelve ears (5 patients) demonstrated a statistically significant decline (p < 0.05) at the tested frequency (*). Four of twelve ears (2 patients) demonstrated individual word recognition score decline beyond test-retest differences (**). Bars indicate standard error.

Patient T-025 developed both teprotumumab-related SNHL and PET, based on abnormal tympanometry testing (Supplemental Figure 1). The patient reported bilateral ear popping, ear plugging, autophony, ear fullness and hearing loss after 2 infusions. Endoscopy demonstrated normal left Eustachian tube (ET) prior to initiating teprotumumab (Figure 4A). Upon repeat post-treatment endoscopy, there was abnormal patency of the left ET (Figure 4B). No significant associated risk factors were identified. Notably, this patient had resolution of ear popping and hearing loss symptoms with normal ear canal pressure measurements on follow-up PET testing (Supplemental Figure 1) and improved PTA in each ear (back to baseline) on an audiogram performed 34.5 weeks after discontinuing therapy.

Figure 4. Eustachian Tube (ET) on Rigid Nasal Endoscopy of a Symptomatic Patient.

(A) Normal left ET. (B) The same patient demonstrated abnormally patent left ET after treatment with teprotumumab. Arrow denotes the reduced cartilaginous portion of the ET and star denotes the reduced fat and glandular tissue.

Patients T-024 and T-025 discontinued teprotumumab due to teprotumumab-related SNHL after infusions 6 and 4, respectively. Patient T-031 also had teprotumumab-related SNHL and discontinued therapy after infusion 4 due to other unrelated AEs.

Patient T-018 had teprotumumab-related SNHL, with post-treatment audiogram demonstrating a 10 dB decline, compared to baseline, across two consecutive frequencies in one ear, thereby meeting ASHA criteria for ototoxicity. However, she did not report subjective hearing loss during therapy and instead reported ear popping after the fifth infusion that resolved after treatment.

Patient T-013 did not demonstrate SNHL or PET on audiometric testing. She also did not report hearing loss symptoms during therapy but experienced mild ear plugging that started after the second infusion and resolved after discontinuation of teprotumumab.

Patients without both baseline and post-treatment audiometric testing

Twelve patients had only baseline or post-treatment testing and were therefore not included in the SNHL or PET objective analysis, as it could not be objectively determined if any hearing changes were pre-existing or related to teprotumumab. Of these, 4 had baseline testing only and 8 had mid or post-treatment testing only. The 4 with only baseline testing declined post-treatment testing, as 2 developed hearing loss symptoms that resolved, one remained asymptomatic, and one was lost to follow-up. In the group with post-treatment testing only, 4 of 8 patients developed new subjective hearing loss with mid-treatment audiometric evidence of SNHL (3 had persistent subjective hearing loss at last follow-up: 2 required hearing aids, 1 also developed PET), 1 patient reported autophony and ear plugging and developed post-treatment PET, 2 symptomatic patients had normal mid-treatment audiometric testing, and one patient remained asymptomatic. Of note, 2 of the 4 patients with both new subjective hearing loss and abnormal mid-treatment audiometric testing demonstrated partial audiometric improvement at 1-year follow-up (20–25 dB improvement at two consecutive frequencies in one patient with a history of baseline hearing loss and 10–15 dB improvement across three consecutive frequencies in each ear in the other patient without a history of baseline hearing loss), though overall functional and objective hearing loss remained significant. This partial post-treatment improvement suggests a likely true temporary hearing decrement related to teprotumumab. Interestingly, both patients who developed PET had resolution of subjective ear plugging and autophony at 27 and 32 week follow-up, but the one patient who developed both SNHL and PET had persistent subjective hearing loss at last follow-up.

Finally, 9 study patients did not undergo audiometric or PET testing. Six developed otologic symptoms, 2 of whom were lost to follow-up and 4 had complete resolution of all otologic symptoms at a mean follow-up of 15.9 weeks; 3 remained asymptomatic throughout therapy.

Discussion

Teprotumumab has recently emerged as a promising therapy for TED, providing significant improvement in proptosis, diplopia and QoL. Hearing dysfunction has been described as a side effect in 10% of patients in the RCTs. However, the true prevalence of new otologic symptoms in teprotumumab-treated patients as well as their severity and resolution have not yet been described. Our study is the first to characterize the onset, severity and reversibility of various hearing abnormalities in teprotumumab-treated patients. This is also the first study to compare baseline and post-treatment audiometric data, in a subset of patients, showing abnormalities in both audiometry and PET testing. Our multidisciplinary team included an oculoplastic surgeon, endocrinologist, audiologists, an otolaryngologist who specializes in PET and an otologist/neurotologist.

Our results demonstrate a much higher percentage of teprotumumab-treated TED patients with new otologic symptoms than previously reported (81.5% versus 10%). Symptoms ranged from mild (the sensation of ear popping once per week) to significant declines in word recognition resulting in a change from serviceable to non-serviceable hearing. Subjectively, most patients studied were not functionally impacted by their otologic symptoms, though 4 patients began wearing hearing aids during the study. Three patients discontinued therapy due to otologic symptoms, one of whom, as previously mentioned, required hearing aids. The vast majority of patients with symptoms of tinnitus, ear fullness, and autophony had resolution of symptoms at an average follow-up of 39.2 weeks after the last teprotumumab infusion, whereas less than half of patients with hearing loss symptoms experienced resolution. Importantly, our study found that baseline hearing loss was a significant risk factor for teprotumumab-related SNHL. Therefore, patients with baseline hearing loss should be counseled on this potential risk and strongly advised to undergo baseline audiometric testing with frequent monitoring.

Of the six patients who received pre- and post-teprotumumab audiometric testing, 5 patients met ASHA criteria for ototoxicity. While as a group the overall change in sensorineural thresholds did not meet statistical significance, the two patients who had subjective worsening of word comprehension also had significant decline in word recognition scores on audiometry. When comparing patients with pre- and post-treatment audiometry who had teprotumumab-related SNHL to asymptomatic patients, patients with a history of baseline hearing loss were at a significantly higher risk of developing SNHL after treatment. Additionally, the group with teprotumumab-related SNHL was older than the asymptomatic group with a mean age of 70.8 compared to 55.4 years, respectively, however this was not statistically significant in our small cohort. Future studies should evaluate the risk of teprotumumab-related SNHL with advanced age, which is a known risk factor for hearing loss in the literature.^33,34

The value of baseline audiometric and PET testing is underscored by three additional observations. First, lack of such testing precludes interpretation of potential teprotumumab-related audiometric changes. In our study, 3 of the 6 patients with persistent subjective hearing loss at last follow-up all had mid or post-treatment audiograms that showed the presence of mixed or sensorineural hearing loss, but were not included in the teprotumumab-related SNHL cohort, due to lack of baseline audiometric testing. Therefore, we could not ascertain whether their subjective hearing loss and objective hearing threshold decline were related to teprotumumab exposure or due to other etiologies. Second, there can be discordance between subjective symptoms and objective findings, as has been previously reported.³⁵ Our study included one patient who did not report subjective hearing loss throughout therapy, however, she met ASHA criteria for ototoxicity based on differences between baseline and post-treatment testing. Third, without baseline audiometric testing, it is challenging to determine if teprotumumab therapy should be discontinued or if treatment for hearing loss should be implemented should significant hearing loss symptoms develop.

Though the high incidence of hearing dysfunction in our cohort is likely multifactorial, the inhibition of IGF-I receptor signaling may be pivotal. Patients treated with teprotumumab experience a prolonged state of IGF-I deficiency given a 6-month therapy duration and a 20-day medication half-life that results in some inhibition of IGF-I signaling for another 100 days. IGF-I has both neurotrophic effects on cochlear hair cells^4,5,36 and synapses,³⁷ which may explain SNHL in its absence. The marked decrease in word recognition in two patients with relative preservation of pure tones in particular raises the question of an induced synaptopathy, or loss of neuronal connections between the cochlea and the brain, leading to hearing loss. Another possibility is that high GH levels, which result from loss of IGF-I feedback inhibition, may potentially contribute to otologic dysfunction, through direct GH actions. Such may be the case in acromegaly (GH and IGF-I excess), in which SNHL has been noted.³⁸ Our results confirmed that high GH levels developed with teprotumumab treatment, as expected based on its mechanism of action, yet there was no significant difference in GH between symptomatic and asymptomatic patients.

Other risk factors for hearing dysfunction include medications, noise, tobacco, alcohol,³⁹ male sex, older age,^33,34 hypertension,⁴⁰ thyroid dysfunction,^41–43 and diabetes/prediabetes.^44–46 Ototoxicity is a common medication side effect, with the most common culprits being aminoglycoside and macrolide antibiotics, loop diuretics, certain chemotherapeutic agents, and nonsteroidal anti-inflammatories.³² Antithyroid drugs have been rarely associated with hearing loss, mainly involving PTU in the setting of antineutrophil cytoplasmic antibody-associated vasculitis.^47–52 Thyroxine depletion and low T3 levels after RAI or total thyroidectomy have also been associated with increased risk of acute hearing loss.^41,42 Notably, patients with TED have several risk factors for hearing loss including age, thyroid dysfunction, history of ATD use, and a potential history of acute hypothyroidism. It is possible these factors play a role in hearing loss independent of, synergistic with, or exacerbated by the effect of IGF-IR inhibition with teprotumumab. Our results did not find a significant difference in thyroid status, history of RAI and thyroidectomy, or ATD use between symptomatic and asymptomatic patients.

Mechanisms for PET development are likely different. PET is a middle ear condition with symptoms of plugged ears, fullness and autophony due to ET patency.^53–56 Potential etiologies include patency within the ET cartilaginous portion⁵⁷ or loss of fat or glandular tissue around the ET opening.^58,59 Our patient with new onset PET demonstrated reduced fat and glandular tissue as well as less cartilaginous thickness around the ET compared to normal after treatment with teprotumumab (Figure 4). As IGF-I has trophic effects on cartilage, its inhibition may impact closure of the cartilaginous ET at rest^60–63 or IGF-I inhibition could reduce Ostmann’s fat (similar to orbital fat modulation), glandular tissue, or both. In our study, 26% and 48% of patients complained of new onset autophony and ear plugging/fullness/pressure, respectively, which is higher than previously published. Notably, the majority of these patients had resolution of symptoms at final follow-up. PET was attributed to teprotumumab in only 1 of 5 patients with baseline and post-treatment PET testing. Importantly, this patient’s PET symptoms normalized 29.4 weeks after therapy, and repeat PET testing at 34.5 weeks after completion of infusions demonstrated a return to baseline. Two additional patients were diagnosed with PET during teprotumumab therapy; however, they were not included in our PET analysis due to lack of baseline testing, and both had resolution of PET symptoms at follow-up. Our study suggests that the majority of patients that develop middle ear dysfunction symptoms with teprotumumab therapy will improve after treatment. More extensive study of middle ear dysfunction with teprotumumab is needed.

Until further studies are available, providers should educate patients on the risk of hearing loss and assess for a history of hearing loss symptoms prior to initiating therapy. We recommend offering baseline audiometric and PET testing to all patients, particularly those with baseline hearing loss, with repeat testing should symptoms develop, and post-treatment testing if symptoms persist. Speech testing and speech in noise testing may be of particular importance going forward due to possible synaptopathy induced by teprotumumab. The utility of ultra-high frequency threshold (>8 kHz) testing in the monitoring of teprotumumab therapy needs to be determined, as it has been found to be an early ototoxicity marker in other settings.⁶⁴ Teprotumumab discontinuation should be considered in patients meeting ASHA ototoxicity criteria or for any hearing loss that is worrisome to the patient or provider. While no targeted therapy is currently commercially available, future studies might examine the use of trans-tympanic injections of IGF-I in either the prevention or treatment of suspected teprotumumab-related SNHL. Should sudden SNHL develop, oral prednisone 60 mg daily for 1 week with a short taper or trans-tympanic corticosteroid injections should be considered per current practice guidelines.⁶⁵ However, the applicability of steroids in cases of subacute teprotumumab-related hearing loss is not clear.^25,66 Treatment options for PET include therapies to temporarily increase swelling around the ET orifice (hypertonic saline, estrogen nasal drops and acidic drops) and avoidance of medications that further exacerbate ET patency such as nasal decongestants.

Study limitations include the absence of pre-treatment audiometry and PET testing for most patients. Due to COVID-19 restrictions, only 6 of 27 patients in the study received pre- and post-treatment audiology and PET testing, limiting our sample size and the interpretation of teprotumumab-related audiometric changes. This also introduced selection bias, as symptomatic patients were more likely to undergo testing. Additionally, patients who obtained baseline testing during the COVID-19 pandemic may have been more likely to have pre-existing hearing issues and more concerned about this side effect. Patients were questioned on all common AEs at each visit, which may have led to recall bias and inflation of the incidence of hearing dysfunction. Mean follow-up was 39.2 weeks after completing teprotumumab, which may be insufficient to evaluate for symptom resolution in all patients. Finally, this study did not include a control group of patients which limits the interpretation of the results. This study serves primarily for hypothesis generation. Hypothesis testing is currently underway in a prospective study with baseline, mid-treatment and post-treatment audiometry and PET testing, in addition to including a control group of patients.

In conclusion, our study demonstrates that the majority of TED patients treated with teprotumumab experienced otologic symptoms. Importantly, less than half of patients with subjective hearing loss experienced symptom resolution at mean follow-up of 39.2 weeks, while the majority of other otologic symptoms resolved. Patients with a history of baseline hearing loss are at increased risk of developing SNHL. These findings invite research to further characterize this adverse event, identify risk factors, and create guidelines for screening, monitoring and management of auditory side effects in teprotumumab-treated patients. Until then, the authors recommend baseline audiometric testing with repeat testing should symptoms develop or persist. Partnering with audiology and/or otolaryngology may allow early identification and management of otologic dysfunction.

Highlights.

• The majority of teprotumumab-treated patients experienced otologic symptoms.

• Most otologic symptoms resolved after therapy, with the exception of hearing loss.

• Baseline hearing loss portends higher risk of worsening sensorineural hearing loss.

Biography

Connie Martin Sears, MD, is an Ophthalmology Resident at Stanford University School of Medicine. Dr. Sears earned her medical degree from Harvard Medical School and holds a Bachelor of Science degree in Biology from the University of California, Riverside where she graduated magna cum laude. She will be completing a fellowship in Ophthalmic Plastic and Reconstructive Surgery at UCLA. Her research interests include the use of biologics in thyroid eye disease and orbital oncology.

Andrea Lora Kossler, MD, is Director of the Oculofacial Plastic, Reconstructive Surgery and Orbital Oncology Service at the Byers Eye Institute at Stanford University. Andrea specializes in aesthetic oculofacial surgery, thyroid eye disease and orbital oncology. Dr. Kossler is Co-Director of the Stanford Multidisciplinary Thyroid Eye Disease Center. Her research interests include the proteomic evaluation of orbital disease, modulation of neural reflexes via neurotization and neurostimulation and evaluating targeted therapies for cutaneous and orbital malignancy.