Article Text

Download PDFPDF
Research paper
Clonazepam Quiets tinnitus: a randomised crossover study with Ginkgo Biloba
  1. Seon-Sook Han1,
  2. Eui-Cheol Nam2,3,
  3. Jun Yeon Won2,
  4. Kang Uk Lee4,
  5. Wanjoo Chun3,5,
  6. Hyun Kyung Choi6,
  7. Robert Aaron Levine7,8
  1. 1Department of Internal Medicine, Kangwon National University, School of Medicine, Chuncheon, Republic of Korea
  2. 2Department of Otolaryngology, Kangwon National University, School of Medicine, Chuncheon, Republic of Korea
  3. 3Institute of Medical Science, Kangwon National University, School of Medicine, Chuncheon, Republic of Korea
  4. 4Department of Psychiatry, Kangwon National University, School of Medicine, Chuncheon, Republic of Korea
  5. 5Department of Pharmacology, Kangwon National University, School of Medicine, Chuncheon, Republic of Korea
  6. 6Department of Otolaryngology, Kangwon National University Hospital, Chuncheon, Republic of Korea
  7. 7Eaton-Peabody Laboratory, Department of Otolaryngology, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts, USA
  8. 8Departments of Otology and Laryngology and Neurology, Harvard Medical School, Boston, Massachusetts, USA
  1. Correspondence to Dr Eui-Cheol Nam, Department of Otolaryngology, School of Medicine, Kangwon National University, 1 Kangwondaehak-gil, Chuncheon, Kangwon-do 200-701, Republic of Korea; birdynec{at}


Objective To assess the effect of Ginkgo biloba and clonazepam, a γ-aminobutyric acid (GABA)-receptor agonist, upon tinnitus.

Methods This was an open-label, randomised, crossover study. 27 men and 11 women (aged 16–80 (mean 58)) with tinnitus for more than 2 months were enrolled. Participants were randomised to either clonazepam or G biloba for the first 3 weeks. For the next 2 weeks of washout no medication was taken. For the final 3 weeks, subjects were given the other drug. The initial dose of clonazepam and G biloba was one tablet daily (clonazepam 0.5 mg; G biloba 40 mg). Subjects were instructed to increase the dose by one tablet every 3 days to a maximum of four tablets daily until they perceived a satisfactory decrease in tinnitus loudness or intolerable side effects. Tinnitus was assessed with pitch and loudness matching, tinnitus handicap inventory, and visual analogue scales of loudness, duration and annoyance.

Results Comparing before and after each drug, clonazepam significantly improved tinnitus loudness (74% of subjects), duration (63%), annoyance (79%), and tinnitus handicap inventory score (61%), whereas the G biloba showed no significant differences on any of these measures.

Conclusion Clonazepam is effective in treating tinnitus; G biloba is ineffective.

  • Clonazepam
  • benzodiazepine
  • Ginkgo biloba
  • tinnitus
  • neurootology
  • stress
  • psychiatry
  • pharmacology
  • pain

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Tinnitus, the perception of sound in the absence of external acoustic stimulation, has become a serious societal issue because of its increasing prevalence, now estimated to be 10.5% (in Korea). It can be debilitating with a negative impact upon quality of life, such as social function and leisure activity. In the more distressed patient there can be cognitive impairment, concentration difficulty, insomnia, anxiety and depression.1–3 A recent review of tinnitus medications concluded that no drugs have been demonstrated to provide replicable long-term reduction of tinnitus in excess of placebo effects.4

A prospective, double-blind placebo-controlled study of the short-acting benzodiazepine, alprazolam reduced tinnitus loudness in 76% of subjects compared with 5% of the control group.5 However, the study was criticised because of its small sample size and lack of replication, despite a Japanese study reported to have a similar result using bromazepam, a short-to-intermediate-acting benzodiazepine.6 A more recent alprazolam report has confirmed the 1993 study. In this second alprazolam study tinnitus loudness (using a visual analogue scale (VAS)) was reduced by 20.83 units compared with 1.5 units for the placebo, chlorpheniramine (p<0.001).7 Other less rigorous studies also lend support to the use of benzodiazepine for tinnitus. An open-label study of Lechtenberg and Shulman8 found that several benzodiazepines were more effective in treating tinnitus than an antihistamine and that clonazepam was the most effective. A large retrospective study of clonazepam found it improved tinnitus in 32% of over 3000 subjects.9 Despite methodological shortcomings, a small prospective, randomised study of subjects with ‘severe disabling tinnitus’ found that clonazepam with or without gabapentin, significantly reduced tinnitus loudness and annoyance compared with an unspecified ‘placebo’.10

Prompted by the hints from prior clonazepam studies, this study was undertaken to definitively assess the effect of the long half life benzodiazepine, clonazepam, upon tinnitus. This study was done prospectively using systematic comprehensive measurements without the flaws of the prior studies. Other reasons for studying this benzodiazepine were related to its long half life; compared with their shorter-acting counterparts, long-acting medications have less withdrawal issues and better compliance due to convenient once-a-day dosing. A crossover design was used with Ginkgo biloba, since G biloba has been widely promoted as an effective tinnitus treatment. However, multiple, large, well controlled, double-blind, placebo-controlled clinical studies have found that G biloba is no more effective than placebo in alleviating tinnitus.4 ,11 ,12

Materials and methods


This randomised, crossover trial was undertaken at the Kangwon National University Hospital between May 2007 and May 2009 with subjects from our tinnitus clinic. This study was approved by the review board and the ethics committee of Kangwon National University Hospital. Inclusion criteria were patients aged over 16 and under 80 with tinnitus that had been present for at least 2 months. Exclusion criteria were use of sedating or antidepressant medications, tinnitus with a curable cause, health issues precluding the use of either medication, pregnancy or planned pregnancy, mental retardation, psychosis and severe cognitive disorders. After each patient's initial evaluation, all appropriate treatment options were discussed, such as oral or intratympanic medications, tinnitus retraining therapy, sound therapy and hearing aids. Patients who chose oral medications were invited to participate in our study comparing the relative effectiveness of two commonly used tinnitus medications, which they would take sequentially, blinded to their identities. The purported mechanisms of action and adverse effects of each medication were explained. All patients signed a statement of informed consent prior to participation.

Study design and procedures

This was an 8-week crossover study (figure 1). Two 3-week treatment periods were separated by a washout period of 2 weeks. Four clinic visits were required (figure 2). The following tests were performed at the first visit: VAS of tinnitus loudness where 0 is no tinnitus and 10 is the loudest tinnitus imaginable; similarly, VAS of tinnitus annoyance; VAS of tinnitus awareness, indicating the proportion of the day patients are aware of their tinnitus; Tinnitus Handicap Inventory (THI) to assess the impact of tinnitus on everyday life, including concentration, sleep, and other psychological parameters;13 pure tone and speech audiometry; tinnitus pitch and loudness matching (‘tinnitogram’). Subjects twice listened to the following pure tones presented at a comfortable level: 250, 500, 1000, 2000, 3000, 4000, 5000, 6000, 7000 and 8000 Hz. The tones were presented once in ascending order and once in descending order. Subjects chose the closest match each time. If the subject chose the same pure tone by both orders, then that frequency was considered the match. Otherwise a third presentation was made using the two previously chosen frequencies. For loudness assessment, subjects adjusted the level of a tone presented at the pitch-match frequency to match their tinnitus loudness. Tinnitus loudness was considered to be the tone level yielding a match (expressed relative to pure tone threshold at the pitch-match frequency). For subjects with non-lateralized tinnitus, tinnitus loudness was taken as the mean of the loudness measurement of each ear. When referring to tinnitus loudness from matching, the term decibel in sensation level ‘dB SL’ loudness is used, while loudness from using the VAS is referred to as ‘VAS loudness’. Somatic modulation testing was performed to identify the somatic properties of each subject's tinnitus. Using the 10-point scale subjects reported changes in the quality and loudness of their tinnitus from maximal isometric contractions of jaw, head and neck muscles.14 After patients had undergone the baseline assessment of tinnitus, they were randomly assigned to begin the first treatment with either clonazepam or G biloba. The dose of clonazepam and G biloba was started at one tablet per day (clonazepam 0.5 mg, orally at bedtime; G biloba 40 mg, orally at bedtime), and the patients were instructed to increase the dose by one tablet every 3 days until their tinnitus was satisfactorily reduced or there were intolerable side effects. If there were significant adverse effects such as drowsiness or dizziness, patients were instructed to reduce the amount of medication back to the previous dose. The maximum dose of either medication was four tablets: clonazepam 2 mg (four tablets at bedtime), G biloba 160 mg (two tablets twice daily). The G biloba used was the standardised G biloba extract, EGB 761, adjusted to 24% ginkgo flavonoids and 6% terpen lactones with <1% ginkgolic acid. A G biloba dose of 160 mg per day was chosen because it is the same dose used in three studies that claimed a benefit for EGB 761 upon tinnitus.15 Likewise the clonazepam dose was following the report of Bahmad et al.10

Figure 2

Schematic flow chart of the study. VAS, visual analogue scale.

On the assumption that clonazepam and G biloba would reduce VAS tinnitus loudness by 50–60% and 10–20% respectively, using statistical power analysis we calculated that 17–29 subjects were required for each group.

Statistical analysis

Efficacy analyses were performed using non-parametric tests because data were not normally distributed. Efficacy variables were described by the median and IQR. Differences between baseline and treatment values of each drug were compared by the Wilcoxon signed-rank test. The effectiveness of clonazepam was compared with that of G biloba by the t-test, Pearson's χ2 test or the Mann–Whitney test. The factors associated with an effect of clonazepam were evaluated using multiple regression analysis by a stepwise method. P values of <0.05 were considered statistically significant. All statistical analyses were performed using SPSS software, V.17.0 (SPSS Inc).


Baseline characteristics

The first drug for each subject was determined by full randomisation using a digital random number generator. We finished initial enrolment when the number of subjects exceeded 29 for each group (35 for clonazepam first, 30 for ginkgo biloba first). Twenty-seven subjects dropped out during or after the first part of the study; none from the second part. Reasons for dropping out (in clonazepam first and G biloba first groups respectively) were concern about adverse effects (1 and 0), reconsidered participating (2 and 2), too time consuming (5 and 3), tinnitus no longer annoying (2 and 1), no improvement (1 and 2), drowsiness (1 and 0), and unable to contact (4 and 3). Upon completion of the study, each group consisted of 19 subjects.

The study group was composed of 27 men (71.1%) and 11 women, with a mean age of 58.2 years and SD of 12 years. The median duration of tinnitus was 48 months (range 2–480 months). Fourteen patients (36.8%) had bilateral tinnitus and 19 (50%) were able to modulate their tinnitus with somatic testing. The patients' baseline characteristics are shown in table 1. Mean pure tone thresholds were 35.5 dB for the bilateral tinnitus patients, and 35.0 dB for the symptomatic ear and 35.9 dB for the asymptomatic ear of the unilateral tinnitus subjects (figure 3).

Table 1

Patient's baseline characteristics

Figure 3

Hearing profiles of patients. Data show a mean audiogram. Bilateral and unilateral tinnitus patients had high-frequency hearing loss.

Effects of G biloba and clonazepam

We randomised the order of drug administration to remove the effect of drug order. Our study results are presented as two different datasets: subjects who received G biloba first (n=19) and those who received clonazepam first (n=19). In the G biloba-first group, there were no significant differences in tinnitus pitch, dB SL or VAS loudness, duration per day annoyance or THI scores before and after G biloba administration (figure 4). In contrast, before and after clonazepam, patients significantly improved as shown by analysing the median of measurements with the Wilcoxon signed-rank test: the median decreased from 5 to 1 dB SL for matching of tinnitus loudness (p<0.001; figure 4A), VAS of tinnitus loudness decreased from 5 to 2 (p<0.05; figure 4B), tinnitus awareness per day decreased from 100% to 60% (p<0.005; figure 4C), VAS of annoyance from 6 to 2 (p<0.001; figure 4D), and THI score went from 30 to 26 (p<0.001; figure 4E). After clonazepam treatment, there was no significant difference in median tinnitus pitch; it remained at 8.0 kHz with a range of 0.25–8.0 kHz.

Figure 4

The effect of clonazepam and Ginkgo biloba in the G biloba-first group (n=19). *p<0.001, †p≤0.01, ‡p≤0.05 by the Wilcoxon signed-rank test. SL, sensation level; THI, Tinnitus Handicap Inventory; VAS, visual analogue scale.

In the clonazepam-first group, the effect of clonazepam treatment was the same as in the G biloba-first group: significantly decreased dB SL and VAS loudness, duration, annoyance and THI scores (figure 5A–E), but tinnitus pitch was unchanged. While G biloba administration did not change tinnitus pitch, duration annoyance or THI scores, dB SL loudness actually increased from 3 to 5 dB SL with G biloba treatment (p<0.05, figure 5F) and VAS loudness increased from 5 to 6 (p<0.05, figure 5G).

Figure 5

The effect of clonazepam and Ginkgo biloba in the clonazepam-first group (n=19). *p< 0.001, †p≤ 0.01, ‡p≤ 0.05 by the Wilcoxon signed-rank test. SL, sensation level; THI, Tinnitus Handicap Inventory; VAS, visual analogue scale.

Another striking result is that the clonazepam scores went to zero for seven subjects with respect to dB SL loudness, for three subjects regarding VAS loudness, for four subjects regarding tinnitus duration, and for three subjects regarding annoyance (VAS). In contrast the comparable numbers for G biloba are 0, 0, 0 and 2 (figures 4 and 5).

We also compared the baseline values before the start of the first drug and after the 2 weeks of washout. There was no difference in tinnitus pitch, awareness (duration), annoyance, and THI scores, but dB SL loudness of the clonazepam-first group remained relatively decreased even after the washout period (p=0.029): before any medication 5 dB SL (median, range; 1–10), after 3 weeks of clonazepam 2 dB SL, and after 2 weeks of washout 3 dB SL.

To assess statistically the effects of each medication, we divided the 38 patients into four groups: improvement more than two-thirds; improvement between one-third and two-thirds; some improvement but less than one-third; and no change or aggravation. Overall improvement for G biloba (combining groups A, B, and C) was 21.1% for dB SL loudness, 23.7% for VAS loudness, 13.2% for tinnitus duration, 39.5% for annoyance VAS, and 55.3% for THI. In comparison the statistics for clonazepam revealed 73.7% improvement for dB SL loudness, 65.8% for VAS loudness, 63.2% for awareness, 78.9% for annoyance, and 60.5% for THI (table 2).

Table 2

Improvement by Ginkgo biloba and clonazepam (n=38, patients, %)

At the end of the study, patients were asked, ‘Which medication do you want to continue using for your tinnitus?’ 66% selected clonazepam, 29% G biloba and 5% did not wish to continue either medication.

Factors associated with positive effect of clonazepam upon tinnitus

The factors associated with benefit from clonazepam were investigated using multiple regression analysis (p<0.05), and the following factors were identified: a longer duration of tinnitus in the pretreatment period was associated with a greater reduction in dB SL loudness; a lower THI score in the pretreatment period was associated with a greater decrease in tinnitus duration per day; and lower initial tinnitus dB SL loudness and higher annoyance scores were associated with a greater improvement in annoyance VAS and THI scores (table 3).

Table 3

The factors associated with positive effect of clonazepam identified by multiple regression analysis

The effect of clonazepam did not differ between unilateral and bilateral tinnitus, symmetric and asymmetric hearing in unilateral tinnitus, and somatic-modulation positive and negative groups. The tinnitus-relieving effect of clonazepam did not differ between hearing-loss-related tinnitus and other types of tinnitus.

Additional results and side effects

The median adjusted dose of clonazepam was 1.5 mg (range 0.5–2.0 mg). All patients took four tablets (160 mg) of G biloba.

Side effects from clonazepam were reported in 42.1% of patients. Fourteen patients (36.8%) had drowsiness, and 2 (5.3%) reported dizziness during clonazepam dose adjustment. The side effects of G biloba were not reported.

The effect of clonazepam on tinnitus was not correlated with its adjusted dose. The median adjusted dose of clonazepam did not differ between the effective group (n=25) and the non-effective group (n=13): 1.5 mg (range 0.5–2.0 mg) versus 1.5 mg (range 0.5–2.0 mg) respectively. The frequency of drowsiness as a result of clonazepam treatment also did not differ between the effective group (n=9, 36%) and non-effective group (n=5, 38%).


This study found that that clonazepam but not G biloba had a significant benefit for people with tinnitus. This included a small subgroup of people in our study whose tinnitus was completely suppressed by clonazepam, but none were completely suppressed by G biloba. Neither result was unexpected.

This now third well designed short-term benzodiazepine study replicates two earlier well designed alprazolam studies. Both found alprazolam effective in quieting tinnitus and significantly superior to placebo.5 ,7 A third prior study with bromazepam was said to quiet tinnitus in 78% of subjects; this percentage is about the same as both the present study (74%) and the 1993 alprazolam study (76%). Percentages were not provided in the 2009 alprazolam report.

G biloba had been shown to be ineffective in several prior systematic studies. A meta-analysis of randomised trials11 concluded as follows: ‘G biloba does not benefit patients with tinnitus’. This fact motivated its choice for our study; G biloba acted effectively as a placebo in our study. Comparable to the placebo-controlled studies, which found that 22% of subjects gained benefit from G biloba and 18% from placebo, 29% of our subjects wished to continue G biloba after completing our study.

We used a randomised, crossover, controlled design. The crossover design allowed each patient to serve as his or her own control and diminished the impact of important factors with imprecisely defined conditions, for example, different baseline disease quality, severity and aetiology. One potential problem of a crossover study is that the washout period may not be long enough to remove all the effects of the drug used in the first part of the study. In fact, that appears to be the case in our study of clonazepam, because of its long duration of action (half life of 39 h).16

When we compare the results for G biloba for the first and second parts of our study, there is a difference. When G biloba was used as the first part medication, there were no significant changes in any of the measures of tinnitus (figure 4). In contrast, when G biloba was used as the second part medication there was significant worsening of tinnitus loudness as measured by VAS and dB SL (figure 5). This result is likely because clonazepam was still wearing off from the first part of the study. Tinnitus loudness worsens because the biological effects of clonazepam from the first part of the study were still wearing off during the 3 weeks of the second part. Hence this finding lends further support to our finding that clonazepam benefits tinnitus.

This third benzodiazepine replication clearly establishes benzodiazepines as the one class of medications that can effectively treat tinnitus in the short term. No study has addressed the question of whether all benzodiazepines have the same effect upon a given individual. From clinical experience, it is likely that there are individuals whose tinnitus does not benefit from one benzodiazepine, but does benefit from a different benzodiazepine. One factor is probably the large inter-subject variability in the pharmacokinetics of benzodiazepines which argues for the individualisation of therapy.16 Because each of these three formal studies was performed with a restricted dose range and one benzodiazepine, it is likely that the proportion of people benefitting in clinical practice from benzodiazepines will be even higher with individualisation of dose and trials with more than one benzodiazepine. Likewise with individualisation of benzodiazepine therapy the percentage of people with intolerable side effects will diminish from what was found in this study.

GABA is a major central nervous system (CNS) inhibitory neurotransmitter. It is widely distributed throughout the auditory pathway from cochlear nuclei to auditory cortex and beyond. Benzodiazepines potentiate the inhibition caused by release of GABA. If tinnitus is due to auditory CNS hyperactivity, then it is likely that benzodiazepines in general and clonazepam in particular benefit tinnitus by reducing this hyperactivity through enhancing GABA-mediated inhibition. For example, animal experiments have implicated elevated fusiform activity of the dorsal cochlear nucleus as a source of tinnitus.17 Among the influences upon the fusiform cell is a GABA-mediated synapse from the stellate cell to the fusiform cell's apical dendrite.18 Benzodiazepine enhancement of this synapse would decrease fusiform activity and thereby tinnitus. Similar mechanisms in other parts of the auditory pathway would likewise be expected to reduce tinnitus. Which part or parts of the auditory pathway are involved must await further study.

Support for the GABA hypothesis comes from other pharmacological studies. Aminoxyacetic acid, a GABA potentiator, significantly quieted tinnitus in 21% of subjects in a placebo-controlled study.19 In an open uncontrolled study, primidone, another GABA potentiator, has been reported to provide more than 80% relief to 11 of 41 tinnitus patients.20 In animals, vigabatrin and taurine, both GABA potentiators, reversibly reduces behavioural evidence of tinnitus.21 ,22

A SPECT pilot study implicates GABA mechanisms in the pathophysiology of tinnitus. Daftary et al reported statistically significant asymmetry of benzodiazepine receptor function in a segment of the superior temporal cortex of six patients with tinnitus as compared to twelve controls.23

Because of its widespread effects throughout the CNS, there is a possibility that the effect of clonazepam on tinnitus may be due to sedation in addition to reduction in auditory neuronal activity. Fourteen of our patients (36.8%) did complain of drowsiness while receiving clonazepam. However, the incidence of drowsiness did not differ between the clonazepam responsive subjects and the non-responsive subjects. In addition, sedation would not account for complete suppression of tinnitus, as occurred in about 10% of clonazepam subjects but in no G biloba subjects. These results support the notion that sedation or cognitive blurring from benzodiazepines is not responsible for their tinnitus benefit. In conclusion, like two other short-term studies using alprazolam, this benzodiazepine study using long-acting clonazepam found a significant quieting effect upon the tinnitus percept. The long-term benefit of benzodiazepines for tinnitus, while likely to persist, will require further study.



  • See Editorial commentary, p 765

  • Linked article 302823.

  • Funding This research was supported by the Basic Science Research Program through the National Research Foundation (NRF) of Korea funded by the Ministry of Education, Science and Technology (2011-0006602).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by the review board and the ethics committee of Kangwon National University Hospital, Korea.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles