Typewriter Tinnitus: Value of ABR as a Diagnostic and Prognostic Indicator

Abstract

Introduction: 

Typewriter tinnitus refers to a special kind of staccato tinnitus, which is mostly described by patients as Morse code, popcorn, or machine-gun. It has been accepted that the mechanism of typewriter tinnitus is caused by the neurovascular compression of the cochleovestibular nerve. Patients who suffered from typewriter tinnitus have exhibited a good response to carbamazepine or oxcarbazepine, but there is a risk of recurrence after treatment cessation. The present study aims to determine the value of auditory brainstem response (ABR) in diagnosing typewriter tinnitus and predicting relapse after drug withdrawal.

Methods: 

Patients who presented with typewriter tinnitus from March 2019 to March 2022 were included for the present retrospective study. The auditory and vestibular test results and drug treatment effects were collected and analyzed. Patients with idiopathic unilateral subjective tinnitus, who were matched by age to patients with typewriter tinnitus at a ratio of 2:1, were consecutively recruited for the control group.

Results: 

Eighteen patients with typewriter tinnitus and 38 controls were included. Ears with typewriter tinnitus had longer interpeak latency (IPL) I-III, and wave III and V latencies, and a higher ratio of IPL I-III ≥2.3 ms based on ABR, when compared to the unaffected side and controls (p<0.05). Seventeen patients with typewriter tinnitus responded positively to medication. Among these patients, seven patients had a relapse after drug cessation, while 10 patients did not have a relapse. The relapse group had significantly longer IPL I-III and wave V latency, older age, and poorer hearing, when compared to the nonrelapse group (p < 0.05). Furthermore, IPL I-III had the largest area under the receiver operating characteristic curve, and the optimal cutoff was 2.4 ms (sensitivity, 100.0%; specificity, 66.7%). There were no significant differences in other demography or other clinical test results between the relapse and nonrelapse groups (p > 0.05). Ramsay Hunt syndrome and neuromyelitis optica spectrum disorders were identified in two cases.

Conclusion: 

Prolonged IPL I-III based on ABR can help in the diagnosis of typewriter tinnitus and its prognosis after treatment cessation. Patients with IPL I-III greater than 2.4 ms, older age and poorer hearing are more likely to relapse. In addition to the neurovascular conflict of the cochleovestibular nerve, the etiologies of neuroinflammation and demyelinating diseases are also possible for typewriter tinnitus.

INTRODUCTION

Typewriter tinnitus refers to a special kind of staccato tinnitus, which is mostly described by patients as Morse code, popcorn, or machine-gun. This was first reported by Mardini (¹⁹⁸⁷). Typewriter tinnitus shares several clinical characteristics with neurovascular compression of the cranial nerves or NVCC (trigeminal neuralgia, vestibular paroxysmia [VP], and hemifacial spasm), including unilateral, staccato and intermittent attacks, and effective response to medications, such as carbamazepine or oxcarbazepine (^{Levine 2006; Sunwoo et al. 2017; Huh et al. 2020; Koo et al. 2021}). Typewriter tinnitus has been classified as a type of cochleovestibular compression syndrome (CVCS), which is caused by the compression of a loop of intracranial vessels, such as the anterior-inferior cerebellar artery, posterior inferior cerebellar artery, or dilated vertebrobasilar artery (^{De Ridder et al. 2007; Huh et al. 2020}). The cisternal segment of the eightth cranial nerve has been reported to range within 14.2-19.2 mm (^{Guclu et al. 2012}), which is particularly longer than other cranial nerves, making it more vulnerable to compression by the neighboring blood vessels. This neurovascular conflict may lead to demyelination (^{Devor et al. 2002}) and endoneurial fibrosis (^{Schwaber & Whetsell 1992}), creating foci of ectopic excitation. Typewriter tinnitus should be differentially diagnosed from objective tinnitus (crackling or buzzing objective tinnitus due to muscle contraction), including palatal myoclonus and middle ear myoclonus. The observation of tympanic membrane movement or palatal tremor, and good response to medication (i.e., muscle relaxants, sedatives and anticonvulsants) or surgery (i.e., tensor tympani, tensor veli palatini, or stapedius muscle tenotomy) provides strong evidence for the diagnosis of objective tinnitus. Carbamazepine and oxcarbazepine have been used in NVCC to release symptoms by inhibiting the voltage-gated sodium channel and suppressing the aberrant ephaptic axonal transmission of the cranial nerve. These drugs have also been used to treat VP, which has been accepted to be a type of CVCS, as described by the Bárány Society (^{Strupp et al. 2016}). Although not licensed for use for typewriter tinnitus at present, most patients with typewriter tinnitus have a good response to carbamazepine or oxcarbazepine, as reported in the literature, suggesting that these two drugs have good potential for the treatment of typewriter tinnitus. However, typewriter tinnitus may recur after drug withdrawal (^{Sunwoo et al. 2017}). At present, there are no instructions or guidelines for treatment using carbamazepine and oxcarbazepine in typewriter tinnitus, and no indicator has been reported as a possible predictor of relapse after drug cessation.

Auditory brainstem response (ABR) has been described as a useful tool to diagnose CVCS, and screen patients for microvascular decompression surgery (^{Møller 1990; De Ridder et al. 2007; Sun et al. 2022}). To our knowledge, few studies have reported the characteristics of ABR in typewriter tinnitus. Han et al. reported that for patients with typewriter tinnitus, a decrease in wave II amplitude was more likely, when compared to patients with tinnitus, due to middle ear myoclonus. Furthermore, although no clear difference was observed, patients with typewriter tinnitus presented with a longer interpeak latency (IPL) of wave I-III and a higher rate of IPL I-III ≥2.3 ms, when compared to patients with middle ear myoclonus (^{Han et al. 2020}). ^{Brantberg (2010}) reported that two of four patients had ABR abnormalities in a study conducted for paroxysmal staccato tinnitus, while the other two patients had normal ABR results. Interestingly, the ABR results were normal in cases reported by ^{Mathiesen and Brantberg (2015}) and ^{Reynard et al. (2020}). To date, there is no consensus on the clinical values for ABR in the diagnosis or prediction of relapse of patients with typewriter tinnitus. Thus, the present study aimed to investigate the application of ABR in the diagnosis of typewriter tinnitus, and its prognosis after medical treatment is stopped.

MATERIALS AND METHODS

The present study was exempted from the Institutional Review Board review by the Medical Ethics Committee of Peking Union Medical College Hospital due to the retrospective design of the study (No. I-22PJ211).

The medical information of patients who presented to the Department of Otolaryngology-Head and Neck Surgery of Peking Union Medical College Hospital from March 2019 to March 2022, and were diagnosed with unilateral typewriter tinnitus, was retrospectively analyzed. Typewriter tinnitus was defined as unilateral staccato sounds, such as Morse code, popcorn, or machine-gun. To exclude objective tinnitus (palatal myoclonus and middle ear myoclonus), physical examination was performed to exclude the rhythmic movement of the tympanic membrane and the tremor of the soft palate, and long-time-based tympanometry was conducted to exclude those with rhythmic changes in middle ear compliance. Pulsatile tinnitus consistent with the heartbeat was also excluded. The other exclusion criteria were otitis media, internal acoustic canal (IAC) or cerebellopontine angle (CPA) tumor, and head or temporal bone trauma. Age-matched control subjects with idiopathic unilateral subjective tinnitus were consecutively recruited from the same hospital during the same enrollment period as patients with typewriter tinnitus at a ratio of 2:1. The same exclusion criteria were applied to the control subjects.

The demographic characteristics of all the patients were recorded. Vertigo that met the criteria for VP (≥10 attacks of spontaneous spinning or non-spinning vertigo, duration <1 minute, stereotyped phenomenology, response to treatment with carbamazepine/oxcarbazepine, and exclusion of other diagnoses), as suggested by the Bárány Society (^{Strupp et al. 2016}), was documented as VP. Otherwise, it was documented as non-VP. Pure-tone audiometry (PTA) was used to evaluate the hearing level, and the average hearing threshold was calculated at 0.5, 1.0, 2.0, and 4.0 kHz. For patients with typewriter tinnitus, auditory and vestibular electrophysiological tests, including ABR, ocular vestibular evoked myogenic potential (oVEMP), and cervical vestibular evoked myogenic potential (cVEMP), were performed. The relative location of the vessel and eighth cranial nerve at the CPA and IAC on magnetic resonance imaging (MRI) was recorded. Neurovascular contact was defined as the disappearance of the cerebrospinal fluid signal gap between the vessel and nerve (^{Best et al. 2013}). This was classified into three types, according to the Chavda system (^{McDermott et al. 2003}): (1) type I, the vascular loop lies at the CPA, but does not enter the IAC; (2) type II, the vascular loop lies within the medial half of the IAC; (3) type III, the vascular loop extends to the lateral half of the IAC. The PTA and ABR results were recorded for the control group.

Carbamazepine (400–800 mg/day) or oxcarbazepine (300–600 mg/day) was administered for patients with typewriter tinnitus for a period of 3 months, along with the active monitoring of side effects, such as tiredness, dizziness, ataxia, and nausea. The therapeutic effectiveness was recorded and classified into three categories: (1) nonresponse (NR), no reduction of typewriter tinnitus; (2) partial remission (PR), partial relief of typewriter tinnitus, including lower frequency and shorter duration of attacks; (3) complete remission (CR), complete elimination of typewriter tinnitus (^{Sunwoo et al. 2017}). If a relapse of typewriter tinnitus occurred after drug cessation, the same medication was prescribed again at the effective dose. Follow-up was achieved via subsequent on-site visit or telephone interview, up to July 2022. Patients who responded positively to the medication (PR and CR) were further divided into two groups, based on the recurrence or exacerbation of typewriter tinnitus after discontinuation of medication: relapse group and nonrelapse group. Then, the clinical characteristics were compared between these two groups, and analyzed for risk of recurrence.

Statistical Analysis

The data were analyzed using SPSS v.23.0 (Chicago, IL, USA). Normally distributed continuous variables were reported as mean ± standard deviation, and analyzed using Student’s t-test, while non-normally distributed continuous variables were reported in median (interquartile range), and calculated using the Mann–Whitney U-test. Categorical data were reported in absolute and relative frequencies, and calculated using chi-squared test or Fisher’s exact test. A two-tailed p value of <0.05 was considered statistically significant. Receiver operating characteristic (ROC) curves were drawn for the positive results in the univariate analysis between the relapse and nonrelapse groups. The area under the curve (AUC) and best cutoff values were determined by maximizing the Youden index for sensitivity and specificity optimization.

RESULTS

Demographic and Testing Characteristics of the Typewriter Tinnitus Group

Eighteen patients with typewriter tinnitus (7 male and 11 female patients) were enrolled for the present study. For patients with typewriter tinnitus, the average age was 49.4 ± 13.6 years, and the average duration of typewriter tinnitus was 1.0 ± 1.1 years. No patient was lost to follow up, and the mean follow-up time was 13.6 ± 7.9 months. Furthermore, the left ear was affected in eight patients, and the right ear was affected in 10 patients. Moreover, six patients merely had typewriter tinnitus, while 12 patients had accompanying dizziness/vertigo. Among these 12 patients, 11 patients met the criteria for VP (^{Strupp et al. 2016}). In addition, 16 patients had CR, one patient had PR, and one patient had NR after medication. Oxcarbazepine, which is generally administered by our hospital as a first-line therapy for this condition, was prescribed for 14 patients who had no obvious side effects from the medication, while carbamazepine was prescribed for three patients who presented with nausea and vomiting after taking oxcarbazepine. In addition, 10 patients had no relapse after treatment cessation, while seven patients had a relapse (Table 1 and File in Supplemental Digital Content 1, https://links.lww.com/EANDH/B147).

TABLE 1. - Demographic characteristics of all patients with typewriter tinnitus.

No.	Gender	Age (y)	Lateral	Possible Trigger	Symptom	Neurovascular Contact Type	Duration (y)	Pharmacological Treatment Scheme	Follow-up Time (mo)	Therapeutic Response	Relapse
1	F	56	L	/	TT+non-VP	I	0.2	Oxcarbazepine, 600 mg/d	16	CR	Relapse
2	F	43	L	/	TT	None	0.6	Oxcarbazepine, 600 mg/d	13	CR	No relapse
3	M	52	L	/	TT+VP	I	1.0	Oxcarbazepine, 600 mg/d	12	PR	No relapse
4	F	56	L	Neuromyelitis optica spectrum disorders	TT+VP	II	0.1	Oxcarbazepine, 600 mg/d	12	CR	Relapse
5	F	31	R	/	TT+VP	II	2.0	Oxcarbazepine, 600 mg/d	4	CR	No relapse
6	M	35	R	/	TT+VP	II	1.0	Oxcarbazepine, 600 mg/d	24	CR	No relapse
7	M	61	L	/	TT+VP	III	1.5	Carbamazepine, 800 mg/d	34	CR	Relapse
8	F	73	R	/	TT+VP	II	1.0	Oxcarbazepine, 600 mg/d	14	CR	Relapse
9	F	74	R	/	TT+VP	II	1.0	Carbamazepine, 400 mg/d	25	CR	Relapse
10	M	33	L	/	TT+VP	I	0.1	Oxcarbazepine, 600 mg/d	17	CR	No relapse
11	M	51	L	/	TT	I	0.6	Oxcarbazepine, 600 mg/d	7	CR	No relapse
12	M	44	R	/	TT	I	4.0	Oxcarbazepine, 600 mg/d	16	CR	Relapse
13	F	56	R	Ramsay Hunt Syndrome	TT+VP	I	0.4	Oxcarbazepine, 600 mg/d	6	CR	No relapse
14	F	65	R	/	TT	I	0.5	Carbamazepine, 400 mg/d	13	CR	No relapse
15	F	38	L	/	TT	I	0.8	Oxcarbazepine, 600 mg/d	6	CR	No relapse
16	F	37	R	/	TT	None	0.1	Oxcarbazepine, 600 mg/d, and carbamazepine, 400 mg/d	6	NR	/
17	F	33	R	/	TT+ VP	II	3.0	Oxcarbazepine, 600 mg/d	14	CR	No relapse
18	M	52	R	/	TT+VP	I	0.3	Oxcarbazepine, 600 mg/d	6	CR	Relapse

CR = complete remission; F = female; ; L = left; M = male; PR = partial remission; R = right; TT = typewriter tinnitus; VP = vestibular paroxysmia.

In the ABR analysis, the wave III and V latencies on the affected side were more prolonged, when compared to the unaffected side (wave III, p = 0.002; wave V, p = 0.030). Furthermore, IPL I-III was notably longer in the affected side, when compared to the unaffected side (p < 0.001, Table 2). However, no difference was observed in wave I latency or IPL III-V between the affected and unaffected sides (wave I latency, p = 0.321; IPL III-V, p = 0.572). In addition, IPL I-III ≥2.3 ms was noted for all affected ears, and this was consistent with Møller’s criteria for NVCC, in which IPL I-III ≥2.3 ms is an important criterion for diagnosing the microvascular conflict of the eighth cranial nerve (^{Møller 1990}).

TABLE 2. - Comparison of auditory and neuro-electrophysiological results between the affected side (with typewriter tinnitus) and unaffected side (without typewriter tinnitus) of all patients (n = 18)

Valuables	Unaffected Side (n = 18)	Affected Side (n = 18)	p
PTA (median [IQR], dB HL)	15.6 (10.0–27.8)	16.3 (10.9–27.8)	0.815
ABR
I wave latency (median [IQR], ms)	1.6 (1.6–1.7)	1.7 (1.6–1.7)	0.321
III wave latency (mean ± SD, ms)	3.9 (3.8–4.1)	4.1 (4.0–4.3)	0.002 *
V wave latency (median [IQR], ms)	5.7 (5.6–5.8)	5.9 (5.8–6.1)	0.030 *
IPL I-III (median [IQR], ms)	2.3 (2.2–2.3)	2.4 (2.3–2.5)	<0.001 *
IPL III-V (mean ± SD, ms)	1.8 ± 0.1	1.8 ± 0.1	0.572
IPL I-III ≥2.3 ms (n, %)
No	14 (77.8)	0 (0.0)	<0.001 *
Yes	3 (16.7)	18 (100.0)
Absent response	1 (5.5)	0 (0.0)

^*P < 0.05.

ABR, auditory brainstem response; IPL, interpeak latency; IQR, interquartile range; PTA, pure-tone audiometry; SD, standard deviation.

There was no significant difference in hearing level between the affected side and unaffected side (affected side, 16.3 (10.9–27.8) dB HL; unaffected side, 15.6 (10.0–27.8) dB HL, p = 0.815; Table 2). The exceptions were two patients who had no evidence of neurovascular contact on the MRI, while nine patients had type I, six patients had type II, and one patient had type III, respectively.

Comparison of Patients With Typewriter Tinnitus and Controls With Idiopathic Unilateral Subjective Tinnitus

A total of 38 controls (19 male and 19 female patients) were enrolled. There were no differences in age, gender, duration of tinnitus, laterality, and hearing level of the affected side between the typewriter tinnitus group and control group (p > 0.05). Based on the ABR, the wave III and V latencies, and IPL I-III of the affected side were significantly prolonged in the typewriter tinnitus group, when compared to the control group (wave III latency, p < 0.001; wave V latency, p = 0.003; IPL I-III, p < 0.001). Furthermore, the ratio for IPL I-III ≥2.3 ms of the affected side was significantly higher in patients who suffered from typewriter tinnitus, when compared to the controls (p < 0.001, Table 3).

TABLE 3. - Comparison of auditory and neuro-electrophysiological results between patients with typewriter tinnitus and controls with idiopathic unilateral tinnitus

Valuables	Sensorineural Tinnitus (n = 38)	Typewriter Tinnitus (n = 18)	p
Age (mean ± SD, y)	45.2 ± 13.8	49.4 ± 13.6	0.282
Gender (n, %)
Male	19 (50.0)	7 (38.9)	0.436
Female	19 (50.0)	11 (61.1)
Duration of tinnitus	1.0 (0.4–5.0)	0.7 (0.3–1.1)	0.122
Lateral (n, %)
Left	23 (60.5)	8 (44.4)	0.258
Right	15 (39.5)	10 (55.6)
PTA (median[IQR], dB HL)	17.5 (11.3–28.8)	16.3 (10.9–27.8)	0.752
ABR
I wave latency (median[IQR], ms)	1.7 (1.6–1.7)	1.7 (1.6–1.7)	0.209
III wave latency (median[IQR], ms)	3.8 (3.7–3.9)	4.1 (4.0–4.3)	<0.001 *
V wave latency (median[IQR], ms)	5.7 (5.6–5.8)	5.9 (5.8–6.1)	0.003 *
IPL I-III (median[IQR], ms)	2.2 (2.1–2.3)	2.4 (2.3–2.5)	<0.001 *
IPL III-V (mean ± SD, ms)	1.9 ± 0.1	1.8 ± 0.1	0.055
IPL I-III ≥2.3 ms (n, %)
No	35 (92.1)	0 (0.0)	<0.001 *
Yes	2 (5.3)	18 (100.0)
Absent response	1 (2.6)	0 (0.0)

^*P<0.05.

ABR, auditory brainstem response; IPL, interpeak latency; IQR, interquartile range; PTA, pure-tone audiometry; SD, standard deviation.

Comparison of Patients With Typewriter Tinnitus Between The Relapse and Nonrelapse Groups

After treatment cessation, seven patients experienced a recurrence at a mean follow-up time of 17.6 ± 9.2 months, while 10 patients did not have a recurrence in 11.6 ± 6.1 months follow-up. The IPL I-III and wave V latency based on the ABR were significantly prolonged in the relapse group, when compared to the nonrelapse group (IPL I-III, p = 0.004; wave V latency, p = 0.018; Table 4). Furthermore, patients with relapse had poorer hearing level and older age, when compared to patients who did not relapse (hearing level, p = 0.009; age, p = 0.013; Table 4). Moreover, IPL I-III had the largest AUC under the ROC curve (0.87, 95% CI: 0.69–1.00), and the optimal cutoff level for IPL I-III was 2.4 ms (sensitivity, 100.0%; specificity, 66.7%; Fig. 1). No differences were detected between the two groups, in terms of gender, duration of typewriter tinnitus, presence of vertigo, neurovascular contact type, or oVEMP and cVEMP (p > 0.05, Table 4).

TABLE 4. - Comparison of relapsed and nonrelapsed patients after medication

Valuables	Nonrelapse (n = 10)	Relapse (n = 7)	p
Gender (n, %)			1.000
Male	4 (40.0)	3 (42.9)
Female	6 (60.0)	4 (57.1)
Age (mean ± SD, y)	43.7 ± 11.7	59.4 ± 10.9	0.013 *
Duration of typewriter tinnitus (median [IQR], y)	0.7 (0.5–1.3)	1.0 (0.2–1.5)	0.922
Follow-up period (mean ± SD, mo)	11.6 ± 6.1	17.6 ± 9.2	0.127
Vertigo (n, %)			0.304
No	5 (50.0)	1 (14.3)
Yes	5 (50.0)	6 (85.7)
Neurovascular contact type (n, %)
Type I	6 (60.0)	3 (42.9)	0.579
Type II	3 (30.0)	3 (42.9)
Type III	0 (0.0)	1 (14.2)
None	1 (10.0)	0 (0.0)
PTA (mean ± SD, dB HL)	15.5 ± 6.0	32.5 ± 16.7	0.009 *
OVEMP (n, %)			1.000
Abnormal	5 (50.0)	3 (42.8)
Normal	3 (30.0)	2 (28.6)
Absent response	2 (20.0)	2 (28.6)
CVEMP (n, %)			1.000
Abnormal	5 (50.0)	4 (57.1)
Normal	3 (30.0)	1 (14.3)
Absent response	2 (20.0)	2 (28.6)
ABR of affected side
I wave latency (median [IQR], ms)	1.7 (1.6–1.8)	1.7 (1.6–1.7)	0.784
III wave latency (mean ± SD, ms)	4.1 ± 0.2	4.3 ± 0.3	0.140
V wave latency (mean ± SD, ms)	5.8 ± 0.2	6.2 ± 0.2	0.018 *
IPL I-III (median [IQR], ms)	2.3 (2.3–2.4)	2.5 (2.4–2.8)	0.004 *
IPL III-V (mean ± SD, ms)	1.8 ± 0.2	1.9 ± 0.1	0.144
IPL I-III ≥2.3 ms (n, %)			/
No	0 (0.0)	0 (0.0)
Yes	10 (100.0)	7 (100.0)

^*P<0.05.

ABR, auditory brainstem response; cVEMP, cervical vestibular evoked myogenic potential; IPL, interpeak latency; IQR, interquartile range; oVEMP, ocular vestibular evoked myogenic potential; PTA, pure-tone audiometry; SD, standard deviation.

Fig. 1.: Receiver operating characteristic curves for IPL I-III, age, PTA, and wave V latency. The optimal cutoff point was calculated according to the max Youden Index (Youden Index = Sensitivity + Specificity − 1).

DISCUSSION

The present retrospective study compared the clinical characteristics of 18 patients with typewriter tinnitus and 38 controls, and determined the value of ABR in diagnosing typewriter tinnitus and predicting relapse after drug cessation. Among the patients with typewriter tinnitus, 16 patients had CR, one patient had PR, and one patient had no response to medical therapy. As a preliminary study, these results suggest that ABR can be used as a diagnostic indicator for typewriter tinnitus, and might serve as an objective prognostic indicator for relapse after treatment cessation, with relatively high sensitivity and specificity.

To date, the diagnosis of typewriter tinnitus is mainly based on symptoms and the degree of response to carbamazepine (^{Sunwoo et al. 2017}), and there is a lack of diagnostic indicators. The present study compared the ABR results between the side affected with typewriter tinnitus and the unaffected side, with reference to the characteristics of the ABR findings in NVCC or CVCS. The results revealed that the affected side had significantly longer IPL I-III, and wave III and V latencies, and a higher ratio of IPL I-III ≥2.3 ms, but had similar IPL III-V and wave I latency. Furthermore, the ABR results for patients with typewriter tinnitus were compared with those for the controls who had idiopathic unilateral subjective tinnitus. More prolonged IPL I-III, and wave III and V latencies, and a higher ratio of IPL I-III ≥2.3 ms were observed, but there were no differences in wave I latency and IPL III-V between the two groups. Since IPL III-V exhibited no differences, the elongation of wave V latency possibly resulted from the prolonged IPL I-III. These outcomes suggest that extended IPL I-III (especially ≥2.3 ms) and wave III latency are the characteristic findings for typewriter tinnitus, as supported by the ABR. These are the possible diagnostic indicators that concur with the clinical values of ABR in NVCC and CVCS.

The present study also explored the applicable value of ABR in predicting the relapse of typewriter tinnitus after drug cessation. The results indicated that the IPL I-III and wave V latency were significantly longer in the relapse group, when compared to the nonrelapse group. In the ROC analysis, patients with IPL I-III greater than 2.4 ms had a higher risk of recurrence, when compared to patients with IPL I-III of less than 2.4 ms, and this was highly sensitive. To our knowledge, no previous study has analyzed the efficacy of ABR in screening patients who have a high risk of relapse of typewriter tinnitus after discontinuation of medication. The present study was the first to discuss this issue. As explored in the literature, IPL I-III prolongation has a significantly positive correlation with the severity of auditory nerve impairment and tinnitus caused by CVCS (^{De Ridder et al. 2007}). Substantially in line with this early conclusion, the present study suggests that the more prolonged the IPL I-III, the more serious the damage to the auditory nerve, and the more chances of relapse after drug withdrawal.

In the present study, 58.8% of patients had a limited course of medical therapy that appeared to have a lasting effect. This phenomenon was also observed in patients with VP. ^{Chen et al. (2022}) reported that 20.7% of patients with VP became attack-free after a follow-up period that ranged within 24 to 43 months, and ^{Steinmetz et al. (2022}) reported that 74% of patients became attack-free after a mean follow-up period of 4.8 years. The underlying mechanism for this may be that some abnormal discharges caused by the neurovascular compression of the cranial nerve disappeared after medication, or the neurovascular conflict was resolved with changes in the hemodynamic status. Another possible aspect is that the neuroinflammation or demyelination observed in typewriter tinnitus may be cured when the cause is removed over time. The present research did not arrive at a definite conclusion for this issue. Thus, dynamic and long-term observations are suggested in the future to answer this intriguing question.

Patients in the relapse group were older, and had poorer hearing, when compared to patients in the nonrelapse group. A possible explanation is that older age increases the risk for hypertension and atherosclerosis of the affected vessel, leading to more severe damage to the eighth cranial nerve. Furthermore, the poorer hearing levels further support the damage to the auditory nerve fibers caused by the neurovascular conflict. The review conducted by ^{Nash et al. (2017}) revealed that among tinnitus-only cases, surgical decompression for neurovascular conflict is only effective when performed early in the course (<5 years), indicating that the longer the duration of the disease, the more severe the injury to the nerve. The present limited data did not reveal any relationship between the course of symptoms and its recurrence. This difference may be attributed to the short course of the included cases, among which merely one case had symptoms with a duration of more than 5 years.

It has been considered that typewriter tinnitus arises from the neurovascular compression of the eighth cranial nerve. However, in the present study, two patients were found to have no neurovascular contact in either the CPA or the IAC on the MRI. It has been reported that some patients with typewriter tinnitus does not present with radiological evidence of neurovascular conflict (^{Bae et al. 2017}), which is similar to the findings in NVCC (i.e., VP, hemispasmus facialis, and trigeminal neuralgia) (^{Girard et al. 1997; Lee et al. 2014; Steinmetz et al. 2022}). IAC osteoma and basilar invagination have been reported to be associated with the incidence of typewriter tinnitus (^{Nam et al. 2010; Reynard et al. 2020}). Apart from CVCS, some possible predisposing factors for typewriter tinnitus were found in two patients in the present study. One female patient developed typewriter tinnitus and VP at three months after Ramsay Hunt syndrome. The other female patient had neuromyelitis optica spectrum disorder, and she gradually developed optic neuritis, extremity numbness, and Sjogren’s syndrome by the age of 37. As previously reported in the literature, neuroinflammatory and demyelinating disorders may induce ectopic neural electrical activity (^{Smith & McDonald 1982; Schwaber & Whetsell 1992; Ori et al. 2018; Laakso et al. 2020}). The present study results suggest that the neurovascular conflict was responsible for the majority of, but not all, cases of typewriter tinnitus. Furthermore, the ectopic neural electrical activity of the cochleovestibular nerve, which was caused by inflammation, autoimmune disease, neurovascular conflict, etc., was the main mode of pathogenesis of typewriter tinnitus.

There were some limitations in the present study. The present study was retrospective in nature, and carried some of the disadvantages typically observed in a retrospective study, such as selection and observer bias. In addition, due to the rare incidence of typewriter tinnitus, the sample size of the present study was small, and the individual differences could not be ignored. Furthermore, the cutoff value for IPL I-III based on ABR needs to be verified in future randomized controlled trials with a larger sample size.

CONCLUSION

ABR is a useful tool for the diagnosis of typewriter tinnitus. The IPL I-III of most ears with typewriter tinnitus was longer than 2.3 ms. Furthermore, almost half of the patients with typewriter tinnitus had a risk of relapse after the cessation of medical treatment, and patients with IPL I-III greater than 2.4 ms were more likely to have recurrence. In addition to the neurovascular conflict of the cochleovestibular nerve, neuroinflammation and demyelinating diseases that involve the eighth cranial nerve were the other possible etiologies for the occurrence of typewriter tinnitus.

ACKNOWLEDGMENTS

We would like to thank all the patient for allowing us to review their medical findings and records.

This work was supported by National High Level Hospital Clinical Research Funding (2022-PUMCH-C-041) and Beijing Natural Science Foundation (7222313).

This study was exempted from the Institutional Review Board (IRB) review by the Medical Ethics Committee of Peking Union Medical College Hospital (No. I-22PJ211).

H.W. designed the study, reviewed, and revised the manuscript. H.S. drafted the initial manuscript and completed the statistical analysis. R.Y., X.T., and Y.Z. collected the data. Z.G. critically reviewed the manuscript. All authors read and approved the final manuscript.