Ginkgo biloba for tinnitus

Abstract

Background

Tinnitus is a symptom defined as the perception of sound in the absence of an external source. In England alone there are an estimated ¾ million general practice consultations every year where the primary complaint is tinnitus, equating to a major burden on healthcare services. Clinical management strategies include education and advice, relaxation therapy, tinnitus retraining therapy (TRT), cognitive behavioural therapy (CBT), sound enrichment using ear‐level sound generators or hearing aids, and drug therapies to manage co‐morbid symptoms such as insomnia, anxiety or depression. 

Objectives

To assess the effects of Ginkgo biloba for tinnitus in adults and children.

Search methods

The Cochrane ENT Information Specialist searched the Cochrane ENT Register; CENTRAL (2022, Issue 6); Ovid MEDLINE; Ovid Embase; CINAHL; Web of Science; ClinicalTrials.gov; ICTRP and additional sources for published and unpublished trials. The date of the search was 7 June 2022.

Selection criteria

Randomised controlled trials (RCTs) recruiting adults and children with acute or chronic subjective tinnitus. We included studies where the intervention involved Ginkgo biloba and this was compared to placebo, no intervention, or education and information. Concurrent use of other medication or other treatment was acceptable if used equally in each group. Where an additional intervention was used equally in both groups, we analysed this as a separate comparison. The review included all courses of Ginkgo biloba, regardless of dose regimens or formulations, and for any duration of treatment.

Data collection and analysis

We used standard Cochrane methods. Our primary outcomes were tinnitus symptom severity measured as a global score on a multi‐item tinnitus questionnaire and serious adverse effects (bleeding, seizures). Our secondary outcomes were tinnitus loudness (change in subjective perception), tinnitus intrusiveness, generalised depression, generalised anxiety, health‐related quality of life and other adverse effects (gastrointestinal upset, headache, allergic reaction). We used GRADE to assess the certainty of the evidence for each outcome.

Main results

This review included 12 studies (with a total of  1915 participants). Eleven studies compared the effects of Ginkgo biloba with placebo and one study compared the effects of Ginkgo biloba with hearing aids to hearing aids alone. All included studies were parallel‐group RCTs. In general, risk of bias was high or unclear due to selection bias and poor reporting of allocation concealment and blinding of participants, personnel and outcome assessments. Due to heterogeneity in the outcomes measured and measurement methods used, only limited data pooling was possible. 

Ginkgo biloba versus placebo

When we pooled data from two studies for the primary outcome tinnitus symptom severity, we found that Ginkgo biloba may have little to no effect (Tinnitus Handicap Inventory scores) at three to six months compared to placebo, but the evidence is very uncertain (mean difference (MD) ‐1.35 (scale 0 to 100), 95% confidence interval (CI) ‐8.26 to 5.55; 2 studies; 85 participants) (very low‐certainty). Ginkgo biloba may result in little to no difference in the risk of bleeding or seizures, with no serious adverse effects reported in either group (4 studies; 1154 participants; low‐certainty).

For the secondary outcomes, one study found that there may be little to no difference between the effects of Ginkgo biloba and placebo on tinnitus loudness measured with audiometric loudness matching at 12 weeks, but the evidence is very uncertain (MD ‐4.00 (scale ‐10 to 140 dB), 95% CI ‐13.33 to 5.33; 1 study; 73 participants) (very low‐certainty). One study found that there may be little to no difference between the effects of Ginkgo biloba and placebo on health‐related quality of life measured with the Glasgow Health Status Inventory at three months (MD ‐0.58 (scale 0 to 100), 95% CI ‐4.67 to 3.51; 1 study; 60 participants) (low‐certainty). Ginkgo biloba may not increase the frequency of other adverse effects (gastrointestinal upset, headache, allergic reaction) at three months compared to placebo (risk ratio 0.91, 95% CI 0.52 to 1.60; 4 studies; 1175 participants) (low‐certainty). None of the studies reported the other secondary outcomes of tinnitus intrusiveness or changes in depressive symptoms or depression, anxiety symptoms or generalised anxiety.

Gingko biloba with concurrent intervention versus concurrent intervention only

One study compared Ginkgo biloba with hearing aids to hearing aids only. It assessed the mean difference in the change in Tinnitus Handicap Inventory scores and tinnitus loudness using a 10‐point visual analogue scale (VAS) at three months. The study did not report adverse effects, tinnitus intrusiveness, changes in depressive symptoms or depression, anxiety symptoms or generalised anxiety, or health‐related quality of life. This was a single, very small study (22 participants) and for all outcomes the certainty of the evidence was very low. We were unable to draw meaningful conclusions from the numerical results.

Authors' conclusions

There is uncertainty about the benefits and harms of Ginkgo biloba for the treatment of tinnitus when compared to placebo. We were unable to draw meaningful conclusions regarding the benefits and harms of Ginkgo biloba when used with concurrent intervention (hearing aids). The certainty of the evidence for the reported outcomes, assessed using GRADE, ranged from low to very low. Future research into the effectiveness of Ginkgo biloba in patients with tinnitus should use rigorous methodology. Randomisation and blinding should be of the highest quality, given the subjective nature of tinnitus and the strong likelihood of a placebo response. The CONSORT statement should be used in the design and reporting of future studies. We also recommend the use of validated, patient‐centred outcome measures for research in the field of tinnitus.

Plain language summary

Ginkgo biloba herbal supplement for tinnitus

What is tinnitus?

Tinnitus is a symptom where people have a perception of sound without there being an external source. It is often described as a ringing, hissing, buzzing or whooshing sound. It is common, affecting between 5% and 43% of the general population, and its prevalence increases with age. For some people tinnitus is persistent and troublesome, and it may lead to sleep problems (insomnia), difficulty concentrating, difficulties in communication and social interaction, and anxiety and depression. Management can include education and advice, relaxation therapy, tinnitus retraining therapy (TRT), cognitive behavioural therapy (CBT), sound generators or hearing aids, and drug therapies. The herbal supplement Ginkgo biloba has also been used.

What did we want to find out?

We wanted to find out whether Ginkgo biloba reduces tinnitus severity and whether it has any unwanted or harmful effects.

What did we do? 

We searched for studies that looked at Ginkgo biloba compared to placebo ('dummy' treatment), no treatment or education/information alone in adults and children with tinnitus. We compared and summarised the results of the studies and rated our confidence in the evidence, based on factors such as how the studies had been done and how many people were involved.

What did we find? 

We found 12 studies (with a total of 1915 people who participated). Eleven studies compared the effects of Ginkgo biloba with placebo. One study compared the effects of Ginkgo biloba combined with hearing aids to hearing aids alone.

Main results 

When we combined the results of two studies that measured tinnitus severity in the same way we found that Ginkgo biloba may have little to no effect compared to placebo, but the evidence is very uncertain. We looked at four studies that recorded any serious harmful effects, all of which reported none, so Ginkgo probably does not result in any difference in risk compared to placebo. However, the included studies did not look at the potentially harmful effects of Ginkgo biloba when used alongside other drugs. There may not be any difference between Ginkgo biloba and placebo in the effect on tinnitus loudness, but this is very uncertain. We also found that there may not be any difference in other outcomes (health‐related quality of life and minor unwanted effects such as gastrointestinal upset, headache and allergic reaction). There is no evidence to suggest that Ginkgo biloba has an effect on tinnitus when compared to placebo.

We looked at the study that compared Ginkgo biloba combined with hearing aids to hearing aids alone. It assessed the difference in the change in tinnitus severity and loudness using a scale at three months. The study did not report any of the other outcomes we were interested in. This was a single, very small study (22 people) and the evidence was very uncertain. We were unable to draw meaningful conclusions from the findings of this study.

What are the limitations of the evidence?

Although we found 12 studies, half of them did not report outcomes that we were interested in. We were not able to combine the results from many of the remaining studies. We are not confident in the evidence for the effect on tinnitus severity of Ginkgo biloba compared to placebo. This is because some people dropped out of one study, only people over 60 were included, the studies were small and very few studies reported this important outcome. We have little confidence in the evidence about serious harmful effects because none were reported in either group and the studies may have had some problems in the way they were done. For tinnitus loudness we are not confident in the evidence because the study that measured this was very small, some people dropped out and only this one study reported this important outcome. We have little confidence in the evidence for health‐related quality of life and minor unwanted effects because the studies were small and may have had problems with the way they were done.

We are not confident in the evidence for the effects of Ginkgo in combination with hearing aids because the number of participants in the study was very small.

How up to date is this evidence?

The evidence is up to date to June 2022.

Summary of findings

Summary of findings 1. Ginkgo biloba compared to placebo for tinnitus.

Ginkgo biloba compared to placebo for tinnitus
Patient or population: adults with tinnitus Setting: departments of otorhinolaryngology in Brazil, Germany and Turkey and one study conducted by telephone/email Intervention: Ginkgo biloba Comparison: placebo
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with placebo	Risk with Ginkgo biloba
Tinnitus symptom severity    Assessed with: THI (range: 0 to 100) Follow‐up: 3 to 6 months	The mean tinnitus symptom severity at 3 to 6 months was 4	MD 1.35 lower (8.26 lower to 5.55 higher)	—	85 (2 RCTs)	⊕⊝⊝⊝ very low1,2,3,4	Ginkgo biloba may have little to no effect on tinnitus severity compared to placebo, but the evidence is very uncertain.
Serious adverse effects (bleeding or seizures)   Yes or no Follow‐up: at 3 months	Study population		—	1154 (4 RCTs)	⊕⊕⊝⊝ low5,7	Ginkgo biloba may result in little to no difference in the risk of serious adverse effects (bleeding or seizures), with zero cases reported in either group.
	Zero events in the placebo group	Zero events in the Ginkgo biloba group
Tinnitus loudness   Assessed with: audiometric loudness matching Follow‐up: at 12 weeks	The mean tinnitus loudness was 0.8	MD 4 lower (13.33 lower to 5.33 higher)	—	73 (1 RCT)	⊕⊝⊝⊝ very low3,4,6	Ginkgo biloba may result in little to no difference in tinnitus loudness compared to placebo.
Health‐related quality of life   Assessed with: GHSI (range: 0 to 100) Follow‐up: 3 months	The mean quality of life at 3 months was 2.52	MD 0.58 lower (4.67 lower to 3.51 higher)	—	60 (1 RCT)	⊕⊕⊝⊝ low3,5	Ginkgo biloba may result in little to no difference in quality of life compared to placebo.
Other adverse effects (gastrointestinal upset, headache, allergic reaction)   Yes or no Follow‐up: at 3 months	Study population		RR 0.91 (0.52 to 1.60)	1175 (4 RCTs)	⊕⊕⊝⊝ low3,5	Ginkgo biloba may not increase the frequency of other adverse effects (gastrointestinal upset, headache, allergic reaction) compared to placebo.
	41 per 1000	37 per 1000 (21 to 66)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; GHSI: Glasgow Health Status Inventory; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio; THI: Tinnitus Handicap Inventory
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

¹One of the two studies omitted primary outcome data, and had dropout without reason; downgraded by 0.5 level for study limitations. 
²Only people over the age of 60 are included in this body of evidence; downgraded by 0.5 level for indirectness.
³Wide confidence interval around the point estimate; small sample size; downgraded by 1 level for imprecision. 
⁴Only a small proportion of the included studies reported on this critical outcome: publication bias is suspected; downgraded by 1 level for publication bias. 
⁵Some of the studies have multiple domains of unclear or high risk of bias; downgraded by 1 level for study limitations.
⁶There are no data regarding the primary outcome measure for 14 participants in the intervention group and 12 participants in the control group at 12 weeks. Dropout is not explained. Almost all of the bias domains are rated as either unclear or high risk; downgraded by 1 level for study limitations. 
⁷Zero events in either group; downgraded by 1 level for imprecision.

Summary of findings 2. Gingko biloba with concurrent intervention versus concurrent intervention only.

Gingko biloba with concurrent intervention versus concurrent intervention only
Patient or population: adults with tinnitus Setting: audiology centre in Brazil Intervention: Ginkgo biloba plus hearing aids Comparison: hearing aids
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with hearing aids	Risk with Ginkgo biloba plus hearing aids
Tinnitus symptom severity   Assessed with: THI (range: 0 to 100) Follow‐up: at 3 months	The mean tinnitus symptom severity at 3 months was ‐40	MD 7.5 higher (2.83 higher to 12.17 higher)	—	22 (1 RCT)	⊕⊝⊝⊝ very low1,2,3	When Ginkgo biloba was used in combination with hearing aids, the reduction in tinnitus symptom severity at 3 months was not as great as when hearing aids were used alone, but the evidence is very uncertain.
Serious adverse effects (bleeding or seizures)	The study did not report serious adverse effects.
Tinnitus loudness   Assessed with: 10‐point VAS (range: 0 to 10) Follow‐up: at 3 months	The mean tinnitus loudness at 3 months was ‐3.2	MD 1 higher (0.52 higher to 1.48 higher)	—	22 (1 RCT)	⊕⊝⊝⊝ very low1,2,3	When Ginkgo biloba was used in combination with hearing aids, the reduction in tinnitus loudness at 3 months was not as great as when hearing aids were used alone, but the evidence is very uncertain.
Health‐related quality of life	The study did not include measures of health‐related quality of life.
Other adverse effects (gastrointestinal upset, headache, allergic reaction)	The study did not reported adverse effects.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; SD: standard deviation; THI: Tinnitus Handicap Inventory; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

¹Not downgraded for indirectness, but it is worth noting that the study only included adult participants (mean age 56.3 years, SD 16.8 years).
²Very small sample size, below the requirement for optimal information size; downgraded by two levels for imprecision.
³Downgraded by one level for study limitations (high risk of bias) due to the lack of blinding during the conduct of the trial.

Background

This review supersedes the Cochrane Review 'Ginkgo biloba for tinnitus', which was first published in the Cochrane Library in Issue 3, 2013. The following paragraphs and Description of the condition are based on the Cochrane Review 'Amplification with hearing aids for patients with tinnitus and co‐existing hearing loss' and are reproduced with permission (Hoare 2014). Description of the intervention and How the intervention might work are based on the Cochrane Review 'Ginkgo biloba for tinnitus' and are reproduced with permission (Hilton 2013).

Description of the condition

Tinnitus is defined as the perception of sound in the absence of an external source (Jastreboff 2004). It is typically described by those who experience it as a ringing, hissing, buzzing or whooshing sound and is thought to result from abnormal neural activity at some point or points in the auditory pathway, which is erroneously interpreted by the brain as sound. Tinnitus can be either objective or subjective. Objective tinnitus refers to the perception of sound that can be also heard by the examiner and is usually due to turbulent blood flow or muscular contraction (Roberts 2010). Most commonly, however, tinnitus is subjective; the sound is only heard by the person experiencing it and no source of the sound is identified (Jastreboff 1988). Tinnitus affects between 5% and 43% of the general population and prevalence increases with age (McCormack 2016). It can be experienced acutely, recovering spontaneously within minutes to weeks, but is considered chronic and unlikely to resolve spontaneously when experienced for more than three months (Gallus 2015; Hall 2011). For many people tinnitus is persistent and troublesome, and has disabling effects such as insomnia, difficulty concentrating, difficulties in communication and social interaction, and negative emotional responses such as anxiety and depression (Hall 2018). In approximately 90% of cases, chronic tinnitus is co‐morbid with some degree of measurable hearing loss, which may confound these disabling effects (Fowler 1944; Sanchez 2002). Nevertheless, the association between hearing loss and tinnitus is not simple or straightforward; not all people with hearing loss experience tinnitus, and conversely some people with clinically normal hearing have tinnitus (Baguley 2013). It has been reported that 40% of patients are unable to identify what health condition is associated with their tinnitus onset, i.e. the tinnitus is idiopathic (Henry 2005). An important implication in clinical research is that outcome measures need to distinguish benefits specific to improved hearing from those specific to improvement in the psychological aspects of tinnitus.

Diagnosis and clinical management of tinnitus

There is no standard procedure for the diagnosis or management of tinnitus. Practice guidelines and the approaches described in studies of usual clinical practice typically reflect differences between the clinical specialisms of the authors or differences in the clinical specialisms charged with meeting tinnitus patients' needs (medical, audiology/hearing therapy, clinical psychology, psychiatry), or the available resources of a particular country or region (access to clinicians or devices, for example) (Biesinger 2010; Cima 2012; Department of Health 2009; Hall 2011; Henry 2008; Hoare 2011). Common across all these documents, however, is the use or recommendation of written questionnaires to assess tinnitus and its impact on patients and their families by measuring tinnitus symptom severity (e.g. impact of tinnitus on quality of life, activities of daily living or sleep), and a judgement about patients who are experiencing a degree of psychological distress (depression or anxiety). Assessment of the perceptual characteristics of tinnitus (pitch, loudness, minimum masking level) and residual inhibition are also recommended (Cima 2019). Although these measures do not correlate well with tinnitus symptom severity (Hiller 2006), they can prove useful in patient counselling (Henry 2004), as a baseline before start of treatment (El Refaie 2004), or by demonstrating stability of the tinnitus percept over time (Department of Health 2009).

Clinical management strategies include education and advice, relaxation therapy, tinnitus retraining therapy (TRT), cognitive behavioural therapy (CBT), sound enrichment using ear‐level sound generators or hearing aids, and drug therapies to manage co‐morbid symptoms such as insomnia, anxiety or depression (for example, Department of Health 2009; Tunkel 2014). As yet, no drug has been approved for tinnitus by a regulatory body (e.g. the European Medicines Agency or US Food and Drug Administration).

Pathophysiology

Most people with chronic tinnitus have some degree of measurable hearing loss (Ratnayake 2009), and the prevalence of tinnitus increases with greater hearing loss (Han 2009; Martines 2010). The varying theories of tinnitus generation involve changes in either function or activity of the peripheral (cochlea and auditory nerve) or central auditory nervous systems (Henry 2005). Theories involving the peripheral systems include the discordant damage theory, which predicts that the loss of outer hair cell function, where inner hair cell function is left intact, leads to a release from inhibition of inner hair cells and aberrant activity (typically hyperactivity) in the auditory nerve (Jastreboff 1990). Such aberrant auditory nerve activity can also have a biochemical basis, resulting from excitotoxicity or stress‐induced enhancement of inner hair cell glutamate release with upregulation of N‐methyl‐D‐aspartate (NMDA) receptors (Guitton 2003; Sahley 2001).

In the central auditory system, structures implicated as possible sites of tinnitus generation include the dorsal cochlear nucleus (Middleton 2011; Pilati 2012), the inferior colliculus (Dong 2010; Mulders 2010), and the auditory and non‐auditory cortex (discussed further below). There is a strong rationale that tinnitus is a direct consequence of maladaptive neuroplastic responses to hearing loss (Moller 2000; Muhlnickel 1998). This process is triggered by sensory deafferentation and a release from lateral inhibition in the central auditory system allowing irregular spontaneous hyperactivity within the central neuronal networks involved in sound processing (Eggermont 2004; Rauschecker 1999; Seki 2003). As a consequence of this hyperactivity, a further physiological change noted in tinnitus patients is increased spontaneous synchronous activity occurring at the subcortical and cortical level, measurable using electroencephalography (EEG) or magnetoencephalography (MEG) (Dietrich 2001; Tass 2012; Weisz 2005). Another physiological change thought to be involved in tinnitus generation is a process of functional reorganisation, which amounts to a change in the response properties of neurons within the primary auditory cortex to external sounds. This effect is well demonstrated physiologically in animal models of hearing loss (Engineer 2011; Norena 2005). Evidence in humans, however, is limited to behavioural evidence of cortical reorganisation after hearing loss, demonstrating improved frequency discrimination ability at the audiometric edge (Kluk 2006; McDermott 1998; Moore 2009; Thai‐Van 2002; Thai‐Van 2003), although Buss 1998 did not find this effect. For comprehensive reviews of these physiological models, see Adjamian 2009 and Norena 2011.

It is also proposed that spontaneous hyperactivity could cause an increase in sensitivity or 'gain' at the level of the cortex, whereby neural sensitivity adapts to the reduced sensory inputs, in effect stabilising mean firing and neural coding efficiency (Norena 2011; Schaette 2006; Schaette 2011). Such adaptive changes would be achieved at the cost of amplifying 'neural noise' due to the overall increase in sensitivity, ultimately resulting in the generation of tinnitus.

Increasingly, non‐auditory areas of the brain, particularly areas associated with emotional processing, are also implicated in bothersome tinnitus (Rauschecker 2010; Vanneste 2012). Vanneste 2012 describes tinnitus as "an emergent property of multiple parallel dynamically changing and partially overlapping sub‐networks", implicating the involvement of many structures of the brain more associated with memory and emotional processing in tinnitus generation. However, identification of the structural components of individual neural networks responsible for either tinnitus generation or tinnitus intrusiveness, which are independent of those for hearing loss, remains open to future research (Melcher 2013). One further complication in understanding the pathophysiology of tinnitus is that not all people with hearing loss have tinnitus and not all people with tinnitus have a clinically significant and measurable hearing loss. Other variables, such as the profile of a person's hearing loss, may account for differences in their tinnitus report. For example, König 2006 found that the maximum slope within audiograms was higher in people with tinnitus than in people with hearing loss who do not have tinnitus, despite the 'non‐tinnitus' group having the greater mean hearing loss. This suggests that a contrast in sensory inputs between regions of normal and elevated threshold may be more likely to result in tinnitus. However, this finding is not consistent across the literature (Sereda 2011; Sereda 2015).

Description of the intervention

Extracts of Ginkgo biloba leaves have been used for medicinal purposes for at least 5000 years in China, where they form an important component of the traditional Chinese pharmacopoeia (a book which lists drugs and instructions for their use). More recently Ginkgo biloba extracts have been used in Western countries. In the USA, Canada and the UK extracts are widely available as non‐prescription food supplements (Diamond 2013; Mei 2017; Ude 2013). In France and Germany a standardised dry leaf extract is registered as a drug and is commonly prescribed for tinnitus (Hall 2011; Ude 2013). However, there are several components in the available Gingko biloba preparations. A purified and enriched liquid extract is prepared from dried leaves of the maidenhair plant. The liquid extract is dried to give one part extract from 50 raw leaves. The most important active chemical compounds are flavonoids (ginkgo‐flavone glycosides) and terpenoids (ginkgolides A, B, C, J and bilobalide). Ginkgolides appear to be unique to Ginkgo biloba and have not been isolated from any other plant species. Standardised Ginkgo leaf extracts have been used in clinical trials for tinnitus, and cognitive and cardiovascular disorders, at daily doses of 60 mg to 450 mg (Mei 2017; Yang 2011). These preparations contain standardised amounts of the above compounds. EGb761 (Tebonin, Tanakan, Rökan) contains 24% ginkgo‐flavone glycosides and 6% terpenoids, and LI 1370 (Kaveri) contains 25% ginkgo flavone glycosides and 6% terpenoids (Blumenthal 1998; Mei 2017). Although the quantities are standardised, the manufacturing process is different and the ratio of active ingredients within each sub‐class may be different. There is no standardisation for food supplement preparations (Mei 2017).

The most commonly reported adverse effect of Ginkgo biloba is mild gastrointestinal disturbance (e.g. stomach pain, change in bowel habit). Serious adverse effects are rare, but include bleeding problems, interaction with anticoagulant medication and seizures (Ernst 2002; Mei 2017; Rajarajan 2018).

How the intervention might work

Several mechanisms of action of Ginkgo biloba have been proposed in the light of its many active ingredients. Human, animal and in vitro studies indicate the following effects:

• A vasoregulatory effect (altering the tone of blood vessels) promoting increased blood flow (Diamond 2013; Lichota 2019; Nuhu 2014; Shu 2019; Xia 2007; Zhou 2004). Animal and human studies have shown that Ginkgo biloba can increase skin (Boelsma 2004; Jung 1990; Koltinger 1989), cardiac (Xiao 2019), and cerebral blood flow (Li 2018; Mashayekh 2011).

• Antagonism of platelet activating factor (PAF) (Diamond 2013; Nash 2015; Xia 2007; Zhou 2004). This effect is specific to the ginkgolides (predominantly B). PAF causes platelet (a blood constituent involved in blood clot formation) aggregation, neutrophil degranulation (activation of immune cells within the blood stream) and oxygen radical production. Ginkgolides appear to protect against the effects of hypoxic brain injury from cerebral ischaemia (permanent brain damage caused by insufficient blood and oxygen supply) in laboratory animals (Braquet 1991; DeFeudis 1991; Li 2018; Smith 1996), and humans (Oskouei 2013).

• Antioxidant activity including scavenging of free radicals, indirectly inhibiting formation of free radicals, regulation of oxidative stress and anti‐lipid peroxidation (Lichota 2019; Mahadevan 2008; Singh 2019; Zhou 2004; Zuo 2017).

• Changes in the metabolism of neurons (Blecharz‐Klin 2009; DeFeudis 2000; Eckert 2005), and restoration of age‐related deficiencies in central neurotransmitter systems (Blecharz‐Klin 2009; DeFeudis 2000; Fehske 2009).

• Enhancement of neuronal plasticity including increased long‐term potentiation, spine density, neuritogenesis and neurogenesis, as shown in pre‐clinical reports (Müller 2012).

• Anti‐inflammatory effects and protective actions against brain damage, possibly through its terpenoids and ginkgolides (Cheng 2003; Lichota 2019; Nuhu 2014; Shu 2019; Xia 2007; Zhang 2016). The Ginkgo biloba leaf extract has been shown to reduce the level of cytokines and inflammatory factors such as tumour necrosis factor alpha (TNF‐α), interleukin 6 (Il‐6), interleukin 1 beta (Il 1‐β) and matrix metalloproteinase 9 (Omidkhoda 2019; Zhang 2016).

These mechanisms may treat tinnitus by preventing free‐radical damage to the cochlea, or increasing blood flow, improving the health of the inner ear (Didier 1996; Smith 2013). Tziridis 2014 tested the effectiveness of prophylactic treatment with EGb 761 for noise‐induced hearing loss and development of tinnitus after noise trauma in an animal model. Based on the results they suggested significant neuroplastic effects of EGb 761 on auditory processing at the peripheral and central level of the auditory pathway as measured with behavioural and electrophysiological approaches. They proposed a model of the effects of EGb 761 on auditory processing with two main effects: 1) an increase in auditory brainstem activity leading to an increased thalamic input to the primary auditory cortex; and 2) an asymmetric effect on lateral inhibition in the primary auditory cortex.

A study by Krauss 2016 examined the therapeutic effects of EGb 761 after the formation of permanent noise‐induced hearing loss and tinnitus in an animal model. They found that treatment with EGb 761 led to recovery of auditory thresholds and reduced behavioural signs of tinnitus. Interestingly, while the auditory thresholds were maintained, behavioural signs of tinnitus reappeared after discontinuation of treatment. An analysis of the auditory brainstem responses (ABRs) showed changes in ABR wave amplitude and latency at different levels of the auditory pathway (increase of response to low stimulus intensities and decrease at high intensities) rather than restoration of the auditory processing back to pre‐trauma conditions. Based on that result, the authors suggested a global inhibitory mechanism that counteracts tinnitus. The EGb 761 extract was also shown to protect against noise‐induced hearing loss by inhibiting the expression of proinflammatory cytokines and cyclooxygenase 2 (COX‐2), and increasing values of heat shock proteins in the rat cochlea (Dogan 2018). It was also shown to protect against cisplatin‐ and gentamicin‐induced hair cell loss, and subsequent changes in brain activity in animal models (Huang 2007; Yang 2011).

Why it is important to do this review

In England alone there are an estimated ¾ million general practice consultations every year where the primary complaint is tinnitus (El Shunnar 2011), equating to a major burden on healthcare services. Use of Gingko biloba for tinnitus is currently recommended against in the European tinnitus guideline (Cima 2019) and the American Academy of Otolaryngology Clinical Practice Guideline (Tunkel 2014). Both guidelines conclude that there is no proven efficacy of Gingko biloba and that there is potential for harm. There is evidence that Ginkgo biloba interacts with antithrombotic drugs to cause serious bleeding and increases bleeding risk in clotting disorders (Posadzki 2012). A worldwide survey of dietary supplements used to treat tinnitus reported that Ginkgo biloba was the most cited supplement resulting in adverse effects (diarrhoea, nausea, hearing, dizziness, headache, bleeding, blood pressure changes, chest pain, palpitation and increased urination) (Coelho 2016). Despite this, Ginkgo biloba is commonly used for tinnitus (Hall 2011). A survey of treatment options for subjective tinnitus showed that Ginkgo biloba was a popular first‐line treatment prescribed by general practitioners (GPs) and ENT physicians across Europe, with a proportion of patients prescribed Ginkgo biloba as high as 71% in some countries (Hall 2011). Ginkgo biloba is freely available for purchase in health food stores across Europe and America (Chan 2007). A recent survey showed that 1 in 10 people with tinnitus in the UK use alternative therapies, including Gingko biloba (McFerran 2018).

The previous Cochrane Review on this question concluded that there was no evidence that Gingko biloba was effective in patients with a primary complaint of tinnitus (Hilton 2013). However, the methods and searches used in that review now require updating.

Objectives

To assess the effects of Ginkgo biloba for tinnitus in adults and children.

Methods

Criteria for considering studies for this review

Types of studies

We included studies with following design characteristics:

• randomised controlled trials, including cluster‐randomised (cross‐over trials were eligible if data from before the cross‐over could be extracted, to avoid the potential for a carry‐over phenomenon).

We excluded studies with the following design characteristics:

• quasi‐randomised controlled studies.

We applied no restrictions on language, year of publication or publication status.

Types of participants

Adults and children with acute (≤ 3 months) or chronic (> 3 months) subjective tinnitus.

Types of interventions

The review included all courses of Ginkgo biloba, regardless of dose regimens or formulations, and for any duration of treatment.

The main comparison was:

• Ginkgo biloba versus placebo.

Other possible comparison pairs included:

• Ginkgo biloba versus no intervention;

• Ginkgo biloba versus education and information only.

Concurrent use of other medication or other treatment was acceptable if used equally in each group. For example, Ginkgo biloba with an additional intervention versus placebo with an identical intervention. Where an additional intervention was used equally in both groups, we analysed this as a separate comparison.

Types of outcome measures

We analysed the following outcomes in the review, but did not use them as a basis for including or excluding studies.

Primary outcomes

• Tinnitus symptom severity (such as the impact of tinnitus on quality of life, activities of daily living and sleep), as measured by the global score on a multi‐item tinnitus questionnaire (Table 3). These include:

  • Tinnitus Questionnaire (Hallam 1996; Hiller 1992);

  • Tinnitus Functional Index (Meikle 2012);

  • Tinnitus Handicap Inventory (Newman 1996);

  • Tinnitus Handicap Questionnaire (Kuk 1990);

  • Tinnitus Reaction Questionnaire (Wilson 1991);

  • Tinnitus Severity Scale (Sweetow 1990).

• Serious adverse effects: bleeding, seizures.

1. Examples of questionnaires measuring tinnitus symptom severity.

Measurement instrument (author, year)	Number of items and subscales	Internal consistency (Cronbach’s alpha for global score)
Tinnitus Functional Index (Meikle 2012)	25 items, 8 subscales	a = 0.97
Tinnitus Handicap Inventory (Newman 1996)	25 items, 3 subscales	a = 0.93
Tinnitus Handicap Questionnaire (Kuk 1990)	27 items, 3 subscales	a = 0.94
Tinnitus Questionnaire (Hallam 1996)	52 items, 5 subscales	a = 0.94
Tinnitus Reaction Questionnaire (Wilson 1991)	26 items, 4 subscales	a = 0.96
Tinnitus Severity Scale (Sweetow 1990)	15 items	Not reported

Secondary outcomes

• Tinnitus loudness (a change in subjective perception) measured using either patient‐reported instruments (including visual analogue scales or numerical rating scales) or performance‐based procedures (including tinnitus loudness matching or minimum masking level).

• Tinnitus intrusiveness measured using a self‐report multi‐item questionnaire or validated subscale (Hall 2018a).

• Generalised depression as measured by validated questionnaires, such as the Beck Depression Inventory II (Beck 1996), the depression scale of the Hospital Anxiety and Depression Scale (HADS; Zigmond 1983), and the Hamilton Rating Scale for Depression (Hamilton 1960).

• Generalised anxiety as measured by a validated scale, for example the anxiety scale of the HADS or Beck Anxiety Inventory (Beck 1988) or the Anxiety Sensitivity Index (Reiss 1986).

• Health‐related quality of life as measured by a validated scale, for example the Short‐Form 36 (Hays 1993), WHOQoLBREF (Skevington 2004), and other WHOQoL versions, and the Health Utilities Index (Furlong 2001).

• Other adverse effects: gastrointestinal upset, headache, allergic reaction.

We assessed outcomes as short‐term (less than three months) and long‐term (three to six months). We also considered whether these outcomes are sustained beyond six months.

Search methods for identification of studies

The Cochrane ENT Information Specialist conducted systematic searches for randomised controlled trials and controlled clinical trials. There were no language, publication year or publication status restrictions. The date of the search was 7 June 2022.

Electronic searches

The Information Specialist searched:

• the Cochrane ENT Register (searched via the Cochrane Register of Studies 7 June 2022);

• the Cochrane Central Register of Controlled Trials (CENTRAL 2022, Issue 6) (searched via the Cochrane Register of Studies);

• Ovid MEDLINE(R) Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE(R) (1946 to 7 June 2022);

• Ovid EMBASE (1974 to 7 June 2022);

• Ovid AMED (1985 to 7 June 2022);

• Web of Science, Web of Science (1945 to 7 June 2022);

• EBSCO CINAHL (1982 to 7 June 2022);

• LILACS (Latin American and Caribbean Health Science Information database), lilacs.bvsalud.org (searched 7 June 2022);

• CNKI, www.cnki.com.cn (searched via Google Scholar 1999 to 7 June 2022);

• ClinicalTrials.gov, (searched via clinicaltrials.gov to 7 June 2022);

• World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (search via https://trialsearch.who.int 7 June 2022).

The Information Specialist modelled subject strategies for databases on the search strategy designed for CENTRAL. Search strategies for major databases including CENTRAL are provided in Appendix 1.

Searching other resources

We scanned the reference lists of identified publications for additional trials and contacted trial authors where necessary. In addition, the Information Specialist searched Ovid MEDLINE to retrieve existing systematic reviews relevant to this systematic review, so that we could scan their reference lists for additional trials. The Information Specialist also ran non‐systematic searches of Google Scholar to retrieve grey literature and other sources of potential trials.

We did not perform a separate search for adverse effects. We considered adverse effects described in the included studies only.

Data collection and analysis

Selection of studies

Two authors (MS, MH, AER and DJH) independently reviewed all records retrieved to determine their eligibility for inclusion in the review. Two authors (MS, MH, AER and DJH) reviewed the full‐text articles of the retrieved studies and independently applied the inclusion criteria. Any disagreements were discussed, involving a third author if necessary until a consensus was reached.

Data extraction and management

MS, MH, AER and DJH independently extracted data using a purposefully designed data extraction form. We piloted the data extraction form on a subset of articles and revised it as indicated before formal data extraction began. Where necessary or where insufficient data were provided for the study, we contacted study authors for further information.

Information extracted included: study design, setting, methods or randomisation and blinding, power, inclusion and exclusion criteria, type of intervention and control, treatment duration, treatment fidelity, type and duration of follow‐up, and outcome measures and statistical tests. 

Data extracted included: baseline characteristics of participants (age, sex, duration of tinnitus, tinnitus symptom severity, tinnitus loudness and pitch estimates, details of co‐morbid hearing loss, anxiety or depression) and details of any attrition or exclusion.

Outcome data included: group mean and standard deviation at pre‐ and post‐intervention and follow‐up, and results of any statistical tests of between‐group comparisons. 

Where not reported or provided by the authors we estimated standard deviations in RevMan 5.3 (RevMan 2014) using the available data, such as standard errors, confidence intervals, P values and t values. Where data were only available in graph form, we made and agreed numeric estimates. 

After independent data extraction by MS, MH, AER and DJH, authors reviewed the extracted data for disagreements, and revisited and discussed the relevant studies as required to reach a final consensus. 

Assessment of risk of bias in included studies

MS, MH, AER and DJH independently assessed risk of bias of the included studies, with the following taken into consideration, as guided by the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019):

• sequence generation;

• allocation concealment;

• blinding;

• incomplete outcome data;

• selective outcome reporting; and

• other sources of bias.

We used the Cochrane risk of bias tool in RevMan 5.3 (RevMan 2014), which involves describing each of these domains as reported in the trial and then assigning a judgement about the adequacy of each entry: 'low', 'high' or 'unclear' risk of bias. We resolved differences of opinion by discussion. 

Measures of treatment effect

We analysed dichotomous data as risk ratios (RR) with 95% confidence intervals (CIs). We summarised continuous outcomes as mean differences (MD) with 95% CI, provided that the selected studies used the same scale of measurement. We had planned to use standardised mean differences (SMD) when different scales of measurement had been used to measure the same outcome but the studies all used the same scale.

Unit of analysis issues

For parallel‐group RCTs the unit of analysis was the group mean. We planned to use alternative analyses for cluster‐randomised trials but none were identified. For studies with more than two intervention groups, we planned to either combine groups to create a single pair‐wise comparison or, if this was not appropriate, select the most relevant pair of interventions for comparison. We selected the most relevant pair of interventions for comparison for studies with more than two arms, in line with the comparison of interest for this review.

Dealing with missing data

We planned that where necessary and where sufficient data from the study were not provided, we would contact the authors of the study requesting further details about missing data and reasons for the incompleteness of the data. We were alert to potential mislabelling or non‐identification of standard errors and standard deviations. Our method for imputation would have been according to chapter Chapter 6 of the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019). Other missing data would not have been imputed. The studies identified did not have missing data so we performed complete case analyses and no sensitivity analysis was used.

Assessment of heterogeneity

We determined whether the selected studies suffered from clinical, statistical and methodological heterogeneity. We  quantified statistical heterogeneity using the I² statistic and the Chi² test. With respect to the I² statistic, an approximate guide to interpretation is provided in the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019). If the I² value was 50% or higher, we considered the data to suffer from substantial or considerable heterogeneity. For the Chi² test, we used the indicator that where the Chi² is greater than the degrees of freedom (where the degrees of freedom are the number of studies K minus 1), then heterogeneity is likely to be present. We considered heterogeneity to be statistically significant if the P value was less than 0.10. Subsequently, in the absence of heterogeneity (I² = 0%, P = 0.41) we performed the meta‐analysis using a fixed‐effect model. 

Assessment of reporting biases

We searched for and requested study protocols for the included studies and, where available, we evaluated whether there was evidence of selective reporting. We planned to assess publication bias using a funnel plot and Egger's test if a meta‐analysis contained at least 10 studies. Unfortunately, none of the meta‐analyses contained more than two studies.

Data synthesis

We planned to perform meta‐analyses using RevMan 5.3 if more than one study was identified for a given outcome (RevMan 2014). Two studies were identified for our main outcome of tinnitus symptom severity and four for other adverse effects for the comparison of Ginkgo biloba versus placebo For other comparisons there was not data from more than one study so it was not possible to calculate a pooled estimate.

Subgroup analysis and investigation of heterogeneity

We planned to carry out subgroup analyses to explore the potential effect modifiers of hearing loss, baseline tinnitus symptom severity and baseline anxiety or depression. However, insufficient data were available.

Sensitivity analysis

We planned to conduct a sensitivity analysis by excluding those studies with a high risk of bias, thereby checking the robustness of the conclusion from the studies included in the meta‐analysis. However, only one to four studies were included in the meta‐analyses, all with similar, non‐significant estimates of effect. Moreover, all of the included studies carried a high or unclear risk of bias.

Summary of findings and assessment of the certainty of the evidence

Two authors (MS and JX) used the GRADE approach to rate the overall certainty of evidence using GRADEpro GDT (https://gradepro.org/). The certainty of evidence reflects the extent to which we are confident that an estimate of effect is correct and we applied this in the interpretation of results. There are four possible ratings: high, moderate, low and very low. A rating of high certainty of evidence implies that we are confident in our estimate of effect and that further research is very unlikely to change our confidence in the estimate of effect. A rating of very low certainty implies that any estimate of effect obtained is very uncertain.

The GRADE approach rates evidence from RCTs that do not have serious limitations as high certainty. However, several factors can lead to the downgrading of the evidence to moderate, low or very low. The degree of downgrading is determined by the seriousness of these factors:

• study limitations (risk of bias);

• inconsistency;

• indirectness of evidence;

• imprecision; and

• publication bias.

We planned to include a summary of findings table, constructed according to the recommendations described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019), for the following comparison:

• Ginkgo biloba versus placebo.

In addition we included a summary of findings table for the comparison:

• Gingko biloba with concurrent intervention versus concurrent intervention only.

We included the following outcomes in the summary of findings tables:

• tinnitus symptom severity;

• serious adverse effects (bleeding disorders, seizures);

• tinnitus loudness;

• health‐related quality of life;

• other adverse effects (gastrointestinal upset, headache, allergic reaction).

Results

Description of studies

Results of the search

Our electronic database search on 7 June 2022 identified 1036 records, of which 492 remained after removing duplicates. We discarded 455 records based on title and/or abstract. We retrieved 37 articles for full‐text screening. We excluded 17 studies because they had the wrong study design (n = 11), had the wrong intervention or comparator (n = 4), had the wrong patient population (n = 1) or were a trial that was terminated before obtaining evaluable data (n = 1) (see Characteristics of excluded studies). Two records are awaiting further classification as full texts could not be obtained (Characteristics of studies awaiting classification). Five records supplemented the methodological information that was extracted for three included studies. We identified one ongoing study (Characteristics of ongoing studies).

In total, 12 completed studies met our inclusion criteria (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Radnuz 2019; Rejali 2004; Yarmohammadi 2007). Five of these studies reported quantitative data that we included in quantitative synthesis (Cekkayan 1996; Drew 2001; Morgenstern 1997; Polanski 2016; Rejali 2004).

We identified no additional records from other sources. A flowchart of study retrieval and selection is provided in Figure 1.

1.

Included studies

See Characteristics of included studies.

We included 12 published studies (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Radnuz 2019; Rejali 2004; Yarmohammadi 2007). 

Design

All included studies were parallel‐group RCTs. Four of the included studies had more than two treatment arms. Cekkayan 1996 was a three‐arm trial comparing Ginkgo biloba, betahistine and placebo. Nishad 2019 was a four‐arm trial comparing Ginkgo biloba to placebo and caroverine to placebo. Radnuz 2019 was a three‐arm trial comparing Ginkgo biloba, hearing aids and Ginkgo biloba with hearing aids. Polanski 2016 was a four‐arm trial comparing Gingko biloba, α‐lipoic acid plus vitamin C, papaverine hydrochloride plus vitamin E and placebo.

Sample sizes

The total sample size for all included studies was 1915 (range 22 to 978 participants).

Setting

Seven studies were set in ENT departments in Brazil, France, Germany, India, Iran, Turkey and the UK (Cekkayan 1996; Meyer 1986; Morgenstern 1997; Nishad 2019; Polanski 2016; Rejali 2004; Yarmohammadi 2007), one in an audiology centre in Brazil (Radnuz 2019), one in a neurology practice in Germany (Halama 1988), one in psychiatric or neurological hospitals in Ukraine (Napryeyenko 2009), and one was conducted completely by email and telephone (Drew 2001). One study did not provide setting details (Fucci 1992).

Participants

Eight studies included participants with tinnitus as a main complaint. Two studies included participants with neurological conditions who also had tinnitus; in Napryeyenko 2009 all participants had either Alzheimer's disease or vascular dementia and in Halama 1988 all patients had "mild cerebral insufficiency of vascular origin". The participant groups in these two studies may not fully represent the tinnitus population (Halama 1988; Napryeyenko 2009). Polanski 2016 included participants with sensorineural hearing loss who also complained of tinnitus. 

Eleven studies recruited adult participants (18 years or over). In one study the age range was 14 to 70 years (Cekkayan 1996). Three studies recruited older participants: Halama 1988 recruited participants over 55 years of age, Napryeyenko 2009 50 years or older and Polanski 2016 60 years and over. Nishad 2019 had an upper age cut‐off for participants of 60 years and Drew 2001 of 70 years. The mean age of participants in the included studies ranged from 38.8 to 66.6 years. Mean age was not reported in two studies (Fucci 1992; Nishad 2019).

Three studies did not report the gender of the participants (Fucci 1992; Halama 1988; Morgenstern 1997), and three did not report gender separately for the arms of interest, only for all participants in all study arms (Nishad 2019; Polanski 2016; Radnuz 2019). In the remaining studies 56% of participants were men and 44% of participants were women. Men accounted for between 28% and 69% of participants, depending on the study. One study had a markedly larger proportion of women than men (Napryeyenko 2009; 72%) and two studies had a markedly larger proportion of men than women (Drew 2001; Nishad 2019; 69% and 67% respectively). 

Baseline hearing level was rarely reported in the included studies. Only one study reported baseline hearing thresholds for participants (Rejali 2004), with averaged air‐conduction thresholds at 0.5 kHz, 1 kHz, 2 kHz and 4 kHz for both ears of 26.8 dB in the intervention group and 35.2 dB in the placebo group. Participants in Cekkayan 1996 had various degrees of uni‐ or bilateral sensorineural hearing loss varying from mild to profound. In Morgenstern 1997, all participants had normal hearing at three neighbouring frequencies in the audiogram. Nishad 2019 performed audiometry (0 kHz to 16,000 Hz) for all participants and had an inclusion criterion of "high frequency loss and sensorineural hearing loss is consistent with the diagnosis of cochlear synaptic tinnitus". Participants in Polanski 2016 had a variable degree of sensorineural hearing loss confirmed by previous audiometric testing. Yarmohammadi 2007 recruited participants with normal hearing. Five studies did not report the baseline hearing level (Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Napryeyenko 2009; Radnuz 2019). 

Individual tinnitus duration ranged from four months to over 10 years. Meyer 1986 recruited participants with recent tinnitus, i.e. tinnitus that had appeared within less than one year and Radnuz 2019 with a complaint of tinnitus for at least three months. Drew 2001 excluded patients with a tinnitus duration of less than 12 months. Tinnitus duration was not reported in Cekkayan 1996, Fucci 1992, Halama 1988, Napryeyenko 2009, Nishad 2019 or Polanski 2016. 

The majority of included studies did not specify inclusion criteria based on tinnitus symptom severity, presence of tinnitus being sufficient. Polanski 2016 recruited participants with "clinical complaints of tinnitus", Radnuz 2019 with "complaint of tinnitus" and Rejali 2004 with "tinnitus as only or main presenting complaint". Five studies reported baseline tinnitus symptom severity. In Meyer 1986, the average tinnitus severity measure on the 1 to 4 scale (mild to severe) was 2.46 (standard deviation (SD) 0.9) in the intervention group and 2.47 (SD 0.6) in the control group. In Napryeyenko 2009, baseline tinnitus symptom severity in the whole study sample of participants with dementia measured on the 11‐point box scale (from 0 indicating absence to 10 indicating extreme severity of the symptom) was 2.1 (SD 2.3) in the Ginkgo biloba group and 2.1 (SD 2.2) in the placebo group. From 400 participants enrolled in the study 204 had tinnitus (101 in the Ginkgo biloba group and 103 in the placebo group). Baseline symptom severity for participants with tinnitus was 4.02 in the Ginkgo biloba group and 3.90 in the placebo group. Three studies reported baseline tinnitus symptom severity measured with the THI (Polanski 2016; Radnuz 2019; Rejali 2004). In Polanski 2016, baseline tinnitus symptom severity measured with the THI was 32.8 (SD 19.9; range 12 to 80) in the Ginkgo biloba group and 28.2 (SD 25.1; range 2 to 72) in the placebo group. In Radnuz 2019, the baseline THI scores were 57.5 (SD 7.5) in the Ginkgo biloba plus hearing aids group and 52.5 (SD 5) in the hearing aid group. In Rejali 2004, these were 37.5 (SD 20.5) in the intervention group and 50.7 (SD 23.3) in the placebo group. Baseline tinnitus symptom severity was not reported in seven studies (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Morgenstern 1997; Nishad 2019; Yarmohammadi 2007).

Baseline anxiety and/or depression scores were not reported in any of the included studies. In Napryeyenko 2009, participants with severe depression were excluded by requiring a score below 20 on the 17‐item Hamilton Rating Scale for Depression (HAMD). 

Interventions and comparisons

Eleven studies evaluated the effects of Ginkgo biloba (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Rejali 2004; Yarmohammadi 2007), and one study the effects of Ginkgo biloba with concurrent intervention (digital Beltone hearing aids) (Radnuz 2019). 

The daily dosage of ginkgolides varied from 80 mg daily to 240 mg daily. Daily dosage was not reported in Fucci 1992. Duration of treatment varied from two weeks to six months. 

The comparator in nine studies was a placebo not otherwise specified (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Rejali 2004; Yarmohammadi 2007). In Polanski 2016, starch capsules were used as a placebo. In Radnuz 2019, the comparator was the digital Beltone hearing aid. 

Outcomes

Seven included studies used one of the pre‐specified outcome measures (Types of outcome measures) (Cekkayan 1996; Drew 2001; Fucci 1992; Morgenstern 1997; Polanski 2016; Radnuz 2019; Rejali 2004).

Primary outcomes

Three studies reported changes in tinnitus symptom severity before and after treatment, as measured by the THI (Polanski 2016; Radnuz 2019; Rejali 2004). Four studies assessed the frequency of serious adverse events (Cekkayan 1996; Drew 2001; Morgenstern 1997; Rejali 2004). Outcomes were measured at three months (Cekkayan 1996; Drew 2001; Morgenstern 1997; Radnuz 2019; Rejali 2004), and six months (Polanski 2016). 

Secondary outcomes

Three studies assessed tinnitus loudness before and after treatment (Fucci 1992; Morgenstern 1997; Radnuz 2019). Fucci 1992 assessed tinnitus loudness measured with loudness matching and minimum masking levels, however it did not provide before and after data for those measures. Morgenstern 1997 assessed loudness measured with audiometric loudness matching. Radnuz 2019 assessed loudness using a visual analogue scale (VAS) (range 0 to 10). One study reported change in quality of life before and after treatment, as measured by the Glasgow Health Status Inventory (GHSI) (Rejali 2004). Four studies assessed other adverse effects (Cekkayan 1996; Drew 2001; Morgenstern 1997; Rejali 2004). 

Outcomes were measured at three months (Cekkayan 1996; Drew 2001; Morgenstern 1997; Radnuz 2019; Rejali 2004).  

None of the studies reported changes in depressive symptoms or depression, anxiety symptoms or generalised anxiety. None of the studies included measures of tinnitus intrusiveness.

Other outcomes

Cekkayan 1996 reported change in tinnitus symptom severity using a single‐item five‐point Likert scale: 0 = the tinnitus disappeared completely; 1 = great relief, but the complaint was still ongoing; 2 = relieved by 50%; 3 = relief was very small; 4 = no changes were noticed. Drew 2001 reported change in loudness measured with a categorical rating scale (‐6 to 6) and change in troublesome nature of tinnitus measured with a categorical rating scale (‐4 to 4). Halama 1988 reported change in the severity of tinnitus assessed on a four‐point rating scale. Meyer 1986 reported change in tinnitus severity on a scale from 1 to 7 (1 to 2 = deterioration; 3 = no change; 4 = slight improvement; 5 = improvement; 6 = major improvement; 7 = discontinued due to intolerance). Napryeyenko 2009 reported change in tinnitus severity, using an 11‐point box scale, 0 representing absence and 10 indicating extreme severity of a symptom. Yarmohammadi 2007 reported change in tinnitus severity measured with an undefined questionnaire. Outcomes were collected at three months. 

Excluded studies

We excluded 16 studies after reviewing the full‐text paper. We excluded 11 studies because they were not RCTs (Abascal 2012; Ahsan 2017; Coles 1988; Dau 2000; Feinberg 2003; Hajna 1999; Holgers 1994; Novotny 2000; Sadner 2017; Schneider 2000; von Wedel 1995). We excluded four studies because of the intervention or control they used (Claussen 1988; Kiefer 2019; Plath 1995; Walger 1993), and one study because of the patient population (Bruchert 1991). See Characteristics of excluded studies for details.

Two studies are awaiting classification as full texts could not be obtained (Fandriantika 2017; Rogowski 2001).

We identified one ongoing study comparing oral administration of Ginkgo biloba extract tablets with medication combined with music in people with tinnitus (Characteristics of ongoing studies). 

Risk of bias in included studies

We assessed risk of bias based on the information provided in the published reports. See Figure 2 and Figure 3 for a graph and summary of risk of bias across studies.

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

3.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Random sequence generation

We considered the risk of selection bias due to inadequate sequence generation to be high in two studies (Fucci 1992; Yarmohammadi 2007), unclear in three studies (Cekkayan 1996; Meyer 1986; Morgenstern 1997), and low in seven studies (Drew 2001; Halama 1988; Napryeyenko 2009; Nishad 2019; Polanski 2016; Radnuz 2019; Rejali 2004). In Fucci 1992, an author confirmed the study was randomised but disclosed that there was a potential bias in the selection of participants. In Yarmohammadi 2007, Ginkgo and placebo tablets were placed in an equal number of paper bags and the doctor randomly administered the content of the bags to the patient. Four studies reported that participants had been randomised to the treatment arms, but did not provide further information on the method of sequence generation (Cekkayan 1996; Meyer 1986; Morgenstern 1997). Drew 2001 paired participants according to pre‐defined criteria and each pair was then allocated two numbers from a randomly arranged code; one number corresponded to placebo treatment and one to active treatment. In Halama 1988, intervention or placebo were randomly assigned by a computer program to participant enrolment numbers. In Napryeyenko 2009, centre‐stratified randomisation (drug–placebo ratio 1:1) in blocks of four was performed by the sponsor's biometrics unit using a validated computer program that linked ascending drug numbers to active drug or placebo, respectively. Nishad 2019 used computer‐generated block randomisation to randomise participants into four groups. Polanski 2016 used http://www.randomization.com for distribution and randomisation of participants. In Radnuz 2019, the random allocation process consisted of generating a random sequence using an Excel file and used the random allocation rule, which chooses at random one of the possible balanced assignments of the given number of participants per treatment. In Rejali 2004, randomisation was carried out by an independent third party using a card from bag system.

Allocation concealment

We judged three studies to have low risk of bias for allocation concealment (Drew 2001; Napryeyenko 2009; Rejali 2004). In Drew 2001, placebo tablets of identical size and colour were dispensed in coded bottles, and treatment allocation was masked. The allocation procedure ensured that all matched participants received active or placebo tablets without the code being identified. In Napryeyenko 2009, the randomisation list was sealed and stored safely at the sponsor's biometrics unit, and block length was not disclosed to investigators. In Rejali 2004, randomisation was carried out by an independent third party using a card from bag system. For the remaining nine studies, risk of bias due to allocation concealment was unclear as the information was not reported (Cekkayan 1996; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Nishad 2019; Polanski 2016; Radnuz 2019; Yarmohammadi 2007).

Blinding

Blinding of participants and personnel 

We judged the risk of performance bias to be low in six studies (Drew 2001; Halama 1988; Napryeyenko 2009; Polanski 2016; Rejali 2004; Yarmohammadi 2007). In Drew 2001, placebo tablets were identical to the active tablets in shape, size, colour and packaging, and provided by the third party (Lichtwer Pharma). The allocation procedure ensured that all matched participants received active or placebo tablets without the code being identified. Halama 1988 used identical tablets for intervention and placebo; neither the patient nor the examining physician knew the assignment and the code was only opened after the data collection had been completed at the beginning of the evaluation. In Napryeyenko 2009, drug and placebo tablets were indistinguishable by appearance, packaging and labelling. Polanski 2016 stated that the substances were not identified by name in the containers into which they were packed, but rather through symbols defined by a professional who did not participate in the research, as a way of blinding investigators and participants. In Rejali 2004, all tablets were provided by third party (Lambert’s Health Care, Tunbridge Wells, UK). Yarmohammadi 2007 stated that the doctor and patients were unaware of the type of medication, with only the second co‐researcher in charge of numbering and packaging the bags being aware. We judged the risk of performance bias to be unclear in five of the included studies, as the information regarding blinding was not provided in the records (Cekkayan 1996; Fucci 1992; Meyer 1986; Morgenstern 1997; Nishad 2019), and high in one study as due to the nature of the intervention (hearing aids) participants were not blinded (Radnuz 2019). However, the examiners responsible for applying the questionnaires during the study were blinded to the intervention (Radnuz 2019).

Blinding of outcome assessment 

The risk of detection bias was low in three studies (Drew 2001; Halama 1988; Radnuz 2019). In Drew 2001, data entry and initial analyses were carried out by a researcher blinded to the participant's allocation. In Halama 1988, neither the patient nor the examining physician knew the assignment; the code was only opened after the data collection had been completed at the beginning of the evaluation. In Radnuz 2019, the examiners responsible for applying the questionnaires during the study were blinded to the intervention. In addition, an employee outside the research team inserted data into the computer in separate data sheets so that the researchers could analyse data without having access to information about the allocation. The risk of detection bias was unclear in the remaining nine studies as the information was not provided (Cekkayan 1996; Fucci 1992; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Rejali 2004; Yarmohammadi 2007). 

Incomplete outcome data

We judged four studies to have high risk of bias due to incomplete outcome data (Cekkayan 1996; Fucci 1992; Morgenstern 1997; Polanski 2016).  Cekkayan 1996 reported data on changes in tinnitus symptom severity only for 8 out of 15 participants, without explaining reasons for dropout; all missing data were for participants in the placebo group. For Fucci 1992 only a brief abstract was available and we were not able to obtain additional information from the authors. No information was provided on the number of participants who completed the study and completeness of outcome measures. In Morgenstern 1997, there were missing data for the primary outcome measure for 14 (out of 49) participants in the intervention group and 12 (out of 50) in the control group at 12 weeks; this dropout was not explained. In Polanski 2016, 15 placebo and 14 test participants who had tinnitus participated in the trial, however data for only 12 and 13 participants respectively were included in the report without reasons for dropout/exclusion given. We judged Drew 2001 to have an unclear risk of bias. The study paired participants according to pre‐defined criteria, however only 956 out of 1121 participants were paired. The remaining 165 participants were also randomised and analysed, however analyses on unmatched data were not reported, except the statement that: "Unmatched analyses did not provide any additional information and have therefore been excluded from this paper". We judged seven studies to have low risk of bias due to incomplete outcome data as all participant data were reported or reasons for dropout were explained (Halama 1988; Meyer 1986; Napryeyenko 2009; Nishad 2019; Radnuz 2019; Rejali 2004; Yarmohammadi 2007).

Selective reporting

We judged seven studies to have high risk of bias due to selective reporting (Cekkayan 1996; Fucci 1992; Halama 1988; Napryeyenko 2009; Nishad 2019; Polanski 2016; Yarmohammadi 2007). In Cekkayan 1996, routine and audiological examinations were performed every 15 days, however those data were not reported. For Fucci 1992 it was not clear whether all primary outcomes were reported. In addition, no data on secondary outcomes were reported. In Halama 1988, the data for most secondary outcome measures were not reported. For Napryeyenko 2009 there were two reports from the study, one reporting the tinnitus score in the whole study population (from 0 ‐ absence of tinnitus to 10 ‐ extreme severity) and one brief report reporting tinnitus scores only in those participants with tinnitus. No additional data were reported specifically for participants with tinnitus. Nishad 2019 specified quality of life as a primary outcome measure but no results were reported. The manuscript mentioned that "tinnitus matching" improved but this was not pre‐defined as an outcome measure. The trial registration for Polanski 2016 lists hearing improvement in an audiometric test as a primary outcome, however this outcome was not described. In Yarmohammadi 2007, a questionnaire of tinnitus severity was used as an outcome measure but not described; the authors interpreted the result as no change, improvement and exacerbation but it is unclear what were the criteria for such classifications. We judged two studies to have an unclear risk of bias (Meyer 1986; Morgenstern 1997). Meyer 1986 did not specify outcomes. In Morgenstern 1997, no values for secondary outcome measures were reported, only brief statements regarding whether those improved or not were included. Drew 2001, Radnuz 2019 and Rejali 2004 reported all pre‐specified outcome measures, therefore we judged the risk of bias to be low. 

Other potential sources of bias

Conflict of interest was not reported in eight studies (Cekkayan 1996; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Rejali 2004; Yarmohammadi 2007), and funding was not disclosed in seven studies (Cekkayan 1996; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Nishad 2019; Polanski 2016). For nine studies there was no prospective protocol or trial registration available (Cekkayan 1996; Fucci 1992; Halama 1988; Meyer 1986; Napryeyenko 2009; Nishad 2019; Radnuz 2019; Rejali 2004; Yarmohammadi 2007). 

Cekkayan 1996 did not report the statistical tests used, therefore it was not possible to judge whether the analysis of the results was appropriate. In Drew 2001, there was limited reporting of baseline data regarding tinnitus characteristics and severity. Participants were matched for age, sex and duration of tinnitus and stated cause of tinnitus but not severity. No data regarding 'unmatched' participant characteristics (i.e. was a specific group excluded from the analysis) were reported. In Fucci 1992, when contacted, one of the authors expressed concerns regarding the quality of the study, however no additional details were provided. Halama 1988 was a trial with patients with "mild cerebral insufficiency of vascular origin" and tinnitus was a secondary outcome measure. No study registration or protocol was available. For Meyer 1986 only an abbreviated version of the full study report was available. Napryeyenko 2009 was a study of the effects of Ginkgo biloba on Alzheimer's disease or vascular dementia and tinnitus was a secondary outcome measure. In Nishad 2019, there were no criteria for defining tinnitus improvement, no statistical analyses were described and the results did not justify the conclusions. Audiological selection criteria were complex and of uncertain significance in selecting the stated aetiological group. Polanski 2016 was a trial concentrating on hearing loss, with tinnitus as a secondary outcome. In Rejali 2004, baseline tinnitus severity was significantly higher in the control group than in the intervention group. No other sources of bias were identified for Radnuz 2019.

Effects of interventions

See: Table 1; Table 2

See: Table 1; Table 2.

Ginkgo biloba versus placebo

Primary outcomes

Tinnitus symptom severity 

Two studies measured tinnitus symptom severity using a multi‐item questionnaire (Polanski 2016; Rejali 2004). Both studies used the Tinnitus Handicap Inventory (THI). Outcomes were measured at 12 weeks (Rejali 2004) and six months (Polanski 2016). Ginkgo biloba may have little to no effect on THI scores compared to placebo, but the evidence is very uncertain. The pooled mean difference (MD) was ‐1.35 (scale 0 to 100; 95% confidence interval (CI) ‐8.26 to 5.55; 2 studies; 85 participants) (Analysis 1.1) (GRADE: very low‐certainty).  

1.1. Analysis.

Comparison 1: Ginkgo biloba versus placebo, Outcome 1: Tinnitus symptom severity (THI) at 3 to 6 months

Serious adverse effects

Four studies reported adverse effects at three months follow‐up (Cekkayan 1996; Drew 2001; Morgenstern 1997; Rejali 2004). Ginkgo biloba may results in little to no difference in the risk of bleeding or seizures, with no serious adverse effects reported in either the Ginkgo biloba or placebo group (4 studies; 1154 participants; low‐certainty) (Analysis 1.2). 

1.2. Analysis.

Comparison 1: Ginkgo biloba versus placebo, Outcome 2: Serious adverse effects at 3 months (yes or no)

Secondary outcomes

Tinnitus loudness

Morgenstern 1997 found that there may be little to no difference between the effect of Ginkgo biloba or placebo on tinnitus loudness measured with audiometric loudness matching at 12 weeks, but the evidence is very uncertain (MD ‐4.00 (scale ‐10 to 140 dB), 95% CI ‐13.33 to 5.33; 1 study; 73 participants) (Analysis 1.3) (GRADE: very low‐certainty). 

1.3. Analysis.

Comparison 1: Ginkgo biloba versus placebo, Outcome 3: Tinnitus loudness measured by audiometric loudness matching at 12 weeks

Tinnitus intrusiveness

None of the studies included measures of tinnitus intrusiveness.

Generalised depression

None of the studies included measures of generalised depression.

Generalised anxiety

None of the studies included measures of generalised anxiety.

Health‐related quality of life

Rejali 2004 found that there may be little to no difference in health‐related quality of life measured with the Glasgow Health Status Inventory (GHSI) at three months (MD ‐0.58 (scale 0 to 100), 95% CI ‐4.67 to 3.51; 1 study; 60 participants) (Analysis 1.4) (GRADE: low‐certainty). 

1.4. Analysis.

Comparison 1: Ginkgo biloba versus placebo, Outcome 4: Health‐related quality of life at 3 months

Other adverse effects

Four studies reported adverse effects at three months follow‐up (Cekkayan 1996; Drew 2001; Morgenstern 1997; Rejali 2004). None of the participants reported adverse effects in Cekkayan 1996 and Morgenstern 1997. In Drew 2001, participants reported gastrointestinal upset (14 in the intervention group and 14 in the placebo group), headache (4 and 4 respectively) and allergic skin reaction (2 and 3 respectively) and in Rejali 2004 participants reported gastrointestinal upset (2 in the intervention group and 1 in the placebo group) and headache (1 in each group). When we pooled the data we found that Ginkgo biloba may not increase the frequency of other adverse effects compared to placebo (risk ratio 0.91, 95% CI 0.52 to 1.60; 4 studies; 1175 participants) (Analysis 1.5) (GRADE: low‐certainty).

1.5. Analysis.

Comparison 1: Ginkgo biloba versus placebo, Outcome 5: Other adverse effects at 3 months (yes or no)

Gingko biloba with concurrent intervention versus concurrent intervention only

Primary outcomes

Tinnitus symptom severity 

Radnuz 2019 measured tinnitus symptom severity with the THI at three months. We found that when Ginkgo biloba was used alongside hearing aids, the reduction in tinnitus symptom severity (as measured by the THI) at three months was not as great as when hearing aids were used alone, but the evidence is very uncertain (MD 7.50 (scale 0 to 100), 95% CI 2.83 to 12.17; 1 study; 22 participants) (Analysis 2.1) (GRADE: very low‐certainty). 

2.1. Analysis.

Comparison 2: Gingko biloba with concurrent intervention versus concurrent intervention only, Outcome 1: Tinnitus symptom severity (THI) at 3 months

Serious adverse effects

The study did not report serious adverse effects.

Secondary outcomes

Tinnitus loudness

Radnuz 2019 measured tinnitus loudness using a visual analogue scale (VAS) at three months. We found that when Ginkgo biloba was used alongside hearing aids, the reduction in tinnitus loudness (as measured by VAS) at three months was not as great as when hearing aids were used alone, but the evidence is very uncertain (MD 1.00 (scale 0 to 10), 95% CI 0.52 to 1.48; 1 study; 22 participants) (Analysis 2.2) (GRADE: very low‐certainty). 

2.2. Analysis.

Comparison 2: Gingko biloba with concurrent intervention versus concurrent intervention only, Outcome 2: Tinnitus loudness (VAS) at 3 months (range 0 to 10)

Tinnitus intrusiveness

The study did not include measures of tinnitus intrusiveness.

Generalised depression

The study did not include measures of generalised depression.

Generalised anxiety

The study did not include measures of generalised anxiety.

Health‐related quality of life

The study did not include measures of health‐related quality of life.

Other adverse effects

The study did not report other adverse effects.

Discussion

Summary of main results

The objective of this review was to assess the effects of Ginkgo biloba for subjective tinnitus in adults and children. This review included 12 studies (1915 participants). Eleven studies investigated the effects of Ginkgo biloba compared to placebo (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Rejali 2004; Yarmohammadi 2007), and one study investigated the effects of Ginkgo biloba plus hearing aids compared to hearing aids alone (Radnuz 2019). Only six studies reported outcomes of interest to this review. 

Ginkgo biloba versus placebo

At three to six months, Ginkgo biloba may have little to no effect on Tinnitus Handicap Inventory (THI) scores but the evidence is very uncertain. No participants in either the Ginkgo biloba or placebo groups reported serious adverse effects (bleeding, seizures) in the included studies. One study found that there may be little to no difference between Ginkgo biloba and placebo in tinnitus loudness at 12 weeks. One study found that there may be little to no difference between Ginkgo biloba and placebo in health‐related quality of life at three months. Ginkgo biloba may not increase the frequency of other adverse effects (such as gastrointestinal upset, headache, allergic reaction) at three months compared to placebo.

None of the studies reported the other secondary outcomes of changes in depressive symptoms or depression, anxiety symptoms or generalised anxiety, or tinnitus intrusiveness.

There is insufficient evidence to support the superiority or inferiority of Ginkgo biloba over placebo. See Table 1.

Gingko biloba with concurrent intervention versus concurrent intervention only

The use of Ginkgo biloba alongside hearing aids may result in a smaller reduction in tinnitus symptom severity (measured with the THI) and tinnitus loudness (measured with a visual analogue scale) at three months compared to when hearing aids are used alone but the evidence is very uncertain.

The study did not report tinnitus intrusiveness, or changes in depressive symptoms or depression, anxiety symptoms or generalised anxiety as measured by a validated instrument.

The evidence is very uncertain about the effect of Ginkgo biloba on the reduction in tinnitus symptom severity when used alongside hearing aids. 

See Table 2.

Overall completeness and applicability of evidence

Only six studies reported outcomes of interest to this review and contributed to the meta‐analyses (Cekkayan 1996; Drew 2001; Morgenstern 1997; Polanski 2016; Radnuz 2019; Rejali 2004). Nine studies included participants with subjective idiopathic tinnitus as a main complaint. Two studies included participants with neurological conditions who also had tinnitus; in Napryeyenko 2009 all participants had either Alzheimer's disease or vascular dementia and in Halama 1988 all patients had "mild cerebral insufficiency of vascular origin". The participant groups in these two studies may not fully represent the tinnitus population. 

Only one study recruited participants with recent onset of tinnitus (< 12 months; Meyer 1986), therefore we were not able to compare the effects of Ginkgo biloba for acute (≤ 3 months) versus chronic (> 3 months) tinnitus. Eleven out of 12 included studies included adult participants, with only one study recruiting patients from 14 years. Three studies recruited older participants (Halama 1988; Napryeyenko 2009; Polanski 2016), while two studies had an upper age cut‐off of 60 or 70 years (Drew 2001; Nishad 2019). There were no studies assessing the effects of Ginkgo biloba for tinnitus in children as a standalone group. One study had a markedly larger proportion of women than men (Napryeyenko 2009; 72%) and two studies had a markedly larger proportion of men than women (Drew 2001; Nishad 2019; 69% and 67% respectively). Baseline hearing level was rarely reported in the included studies. Only one study reported baseline hearing thresholds for participants (Rejali 2004). Yarmohammadi 2007 recruited participants with normal hearing. Most included studies did not specify inclusion criteria based on tinnitus symptom severity, presence of tinnitus being sufficient. Therefore, it was not possible to judge whether different study populations were representative of the whole tinnitus population with respect to hearing loss and tinnitus symptom severity. One study excluded participants with severe depression (Napryeyenko 2009). 

Eleven studies evaluated the effects of Ginkgo biloba (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Polanski 2016; Rejali 2004; Yarmohammadi 2007), and one study evaluated the effects of Ginkgo biloba with digital Beltone hearing aids (Radnuz 2019). The daily dosage varied from 80 mg daily to 240 mg daily and the duration of treatment varied from two weeks to six months. 

The comparator in nine studies was a placebo not otherwise specified (Cekkayan 1996; Drew 2001; Fucci 1992; Halama 1988; Meyer 1986; Morgenstern 1997; Napryeyenko 2009; Nishad 2019; Rejali 2004; Yarmohammadi 2007). In Polanski 2016, starch capsules were used as a placebo. In the analyses we pooled the different dosages and types of placebo together. In Radnuz 2019, the comparator was a digital Beltone hearing aid.

Quality of the evidence

The certainty of the evidence ranged from very low to low. We are uncertain about the estimates for change in tinnitus symptom severity, tinnitus loudness assessed with audiometric matching or visual analogue scale, quality of life and other adverse effects. The certainty of the evidence was affected by several issues, including multiple domains of high or unclear risk of bias, indirectness of evidence, imprecision and possible publication bias. Allocation concealment and blinding of participants, personnel and outcome assessments were poorly reported. Most studies were characterised by high or unclear risk of selection bias and multiple other sources of bias were identified. Only three of the included studies had a pre‐published protocol available.

Potential biases in the review process

Our searches of the electronic databases were comprehensive. We also searched the reference lists of the included studies and previous Cochrane Review (Hilton 2013). Language was not a barrier to inclusion and we obtained translations of the articles in French, German, Turkish and Persian. All author roles were pre‐defined in the review process. Two authors selected studies for inclusion and judged risk of bias independently, with recourse to a third author for resolution of any disagreement or uncertainty. Two authors independently extracted data to minimise personal bias. We adhered to a pre‐published protocol and no post hoc decisions or changes were made (Sereda 2019). 

Agreements and disagreements with other studies or reviews

This is a new review, superseding the previous Cochrane Review on Ginkgo biloba for tinnitus (Hilton 2013). Hilton 2013 included three studies, all considering patients with tinnitus as a primary complaint. All three trials included in Hilton 2013 were also included in the current review. We also included two studies that had been excluded from Hilton 2013 due to lack of preferred outcomes (Halama 1988) and differences in baseline tinnitus severity between intervention and placebo groups (Meyer 1986), as those reasons are no longer considered sufficient for exclusion (Handbook 2019). In the current review we identified and included an additional seven studies with both participants with a primary complaint of tinnitus as well as tinnitus associated with neurological conditions. Similarly to the current review Hilton 2013 concluded that there was no evidence that Ginkgo biloba was effective for tinnitus, with a relatively small incidence of adverse effects that was similar between the intervention and placebo groups. 

A search for systematic reviews of the effectiveness of Ginkgo biloba for tinnitus identified three reviews (Ernst 1999; Hilton 2013; von Boetticher 2011) referring to four randomised trials (Drew 2001; Meyer 1986; Morgenstern 1997; Rejali 2004), which included 1246 patients (Kramer 2018). The authors concluded that Ginkgo biloba probably did not decrease the severity of tinnitus, did not improve quality of life and did not reduce the intensity of tinnitus. Additionally, Ginkgo biloba was not associated with adverse effects in patients with tinnitus, namely there was no difference in the type and frequency of adverse effects between the intervention and placebo groups (Kramer 2018).  

A review looking at the effects of Ginkgo biloba on neurosensory symptoms (tinnitus and dizziness) in patients with dementia concluded that Ginkgo biloba was effective in alleviating tinnitus in those patients (Spiegel 2018). It is interesting to hypothesise why there might be a favourable response to Ginkgo biloba when tinnitus is associated with cognitive insufficiency. The aetiology of tinnitus in cognitive insufficiency is different from primary tinnitus. For example, tinnitus in cognitive insufficiency may be caused by central vascular insufficiency or a neural metabolic disorder, whereas in primary tinnitus the initiating pathology is a cochlear disorder. Changes in vascular perfusion and neuronal metabolism are well‐documented effects of Ginkgo biloba in animal and human studies. Improved cognitive functioning with Ginkgo biloba may also allow habituation to tinnitus. Habituation to repetitive presentation of a sensory stimulus or perception is the norm. There is a strong psychological component in appropriate habituation to tinnitus. If Ginkgo biloba causes a significant improvement in overall cognitive functioning then a positive effect on tinnitus may be real, but non‐specific. A Cochrane Review, however, concluded that the evidence that Ginkgo biloba has a predictable and clinically significant benefit for people with dementia or cognitive impairment is inconsistent and unreliable (Birks 2009). In the current review we included the studies Napryeyenko 2009, where all participants had either Alzheimer's disease or vascular dementia, and Halama 1988, where all patients had "mild cerebral insufficiency of vascular origin". However, neither of the studies included outcomes of interest to the current review. 

Several reviews reported side effects of Ginkgo biloba such as gastrointestinal problems, headache, nausea and vomiting, which were usually mild and reversible (Ernst 2002; Mei 2017; Rajarajan 2018). Serious adverse effects were rare but these can include bleeding problems and seizures. The most serious adverse effects reported relate to the interactions of Ginkgo biloba with other drugs, such as anticoagulant medications, which results in the reports of haemorrhage, haematoma, apraxia, permanent neurological deficit and death (Posadzki 2012). Only four studies in the current review reported adverse effects (Cekkayan 1996; Drew 2001; Morgenstern 1997; Rejali 2004). They included gastrointestinal upset, headache and allergic skin reaction. None of the studies reported any serious adverse effects, however it is worth noting that six studies excluded patients using concurrent anticoagulant medications (Drew 2001; Halama 1988; Napryeyenko 2009; Nishad 2019; Polanski 2016; Rejali 2004). The frequency of adverse effects was relatively small and similar in the intervention and placebo groups.

Authors' conclusions

Implications for practice.

Ginkgo biloba is commonly used as a food supplement for tinnitus and is also a popular first‐line treatment prescribed by general practitioners (GPs) and ENT physicians across Europe (Hall 2011). However, we did not find any evidence to support or refute the prescription of Ginkgo biloba for subjective tinnitus. There is uncertainty about the benefits and harms of Ginkgo biloba for the treatment of tinnitus when compared to placebo. The included studies did not investigate the potentially harmful interactions of Ginkgo biloba with other drugs (Posadzki 2012). 

In line with the lack of evidence for the effectiveness of Ginkgo biloba and potentially harmful interactions with other drugs, current tinnitus management guidelines recommend against its use for tinnitus treatment (Cima 2019; Tunkel 2014). The Multidisciplinary European Guideline for Tinnitus (Cima 2019) based their recommendation on the results of several systematic reviews, which either concluded that Ginkgo biloba was not effective (Hilton 2013; Rejali 2004), or highlighted the low methodological rigour of the included trials (von Boetticher 2011). The authors also highlighted that there was evidence that Ginkgo biloba can interact with other anticoagulant drugs to cause serious bleeding and worsen bleeding risk in patients with underlying clotting disorders (Posadzki 2012). Similarly, the American Academy of Audiology Clinical Practice Guideline recommended against the use of Ginkgo biloba for treating patients with persistent, bothersome tinnitus (Tunkel 2014). The above recommendation was based on variation in conclusions, methodological limitations and heterogeneity in the study protocols of the randomised controlled trials (RCTs) in addition to conflicting conclusions from systematic reviews. The guideline also referenced the evidence for serious side effects, which involve platelet inhibitory actions of Ginkgo biloba, particularly if taken along with other medications that impair coagulation (Posadzki 2012), with the recommendation to avoid use of Ginkgo biloba in older adults, in which the use of anticoagulants and analgesics is widespread. Other mentioned herb‐drug interactions included thiazide diuretics, which resulted in increased blood pressure, and trazodone, which results in increased sedation (Tunkel 2014).

It is unlikely that the aetiology of tinnitus is the same for every tinnitus sufferer. Ginkgo biloba has been shown to affect vascular permeability and neuronal metabolism. If a greater level of understanding and diagnostic accuracy can be reached about the different aetiologies of tinnitus, this may naturally highlight subgroups of tinnitus patients in whom further controlled trials of Ginkgo biloba are worth considering.

Implications for research.

Future research into the effectiveness of Ginkgo biloba in patients with tinnitus should use rigorous methodology. Randomisation and blinding should be of the highest quality, given the subjective nature of tinnitus and the strong likelihood of a placebo response. The CONSORT statement should be used in the design and reporting of future studies (CONSORT 2010).

We also recommend the development of validated, patient‐centred outcome measures for research in the field of tinnitus. Visual analogue scales have limited value in this regard because quantifying change using only a single item has inadequate measurement properties (e.g. internal consistency cannot be established and test‐retest scores are at greater risk of instability). Only three out of 12 studies included in this review used multi‐item questionnaires of tinnitus symptom severity, and only one used a measure of heath‐related quality of life. Other outcomes such as depression and anxiety were not measured. Only four of the studies reported adverse effects. In future trials, multi‐item questionnaires of tinnitus symptom severity, validated instruments measuring depressive symptoms or depression, anxiety symptoms or generalised anxiety and health‐related quality of life should also be used. 

Core outcome measures for adults with subjective tinnitus have been identified (Hall 2018a). For pharmacological interventions, these are tinnitus intrusiveness and loudness. None of the included studies assessed tinnitus intrusiveness, and while loudness was assessed in two out of the five included studies, the measurement methods varied. There is no standard test for tinnitus loudness, and psychoacoustic loudness matching and subjective rating methods are equally common. These measurement methods are generally applied, and findings interpreted, with the assumption that they measure the same underlying construct (i.e. that they have convergent validity). However, retrospective analysis of one randomised placebo‐controlled trial in 91 participants with subjective idiopathic tinnitus indicates otherwise (Hall 2017). Use of the core outcome set as a minimum standard for what should be assessed and reported in RCTs will facilitate comparison between studies and meta‐analyses (Tunis 2016).

Given the heterogeneity of tinnitus patients, future trials should assess and report baseline characteristics so that the risk of potential confounding factors can be better understood. Examples include tinnitus duration, tinnitus symptom severity, age, hearing loss and co‐morbidities since these might reasonably modify treatment success. Future trials might also consider, as a subgroup analysis, the differential effect of Ginkgo biloba on acute (i.e. less than three months duration) versus chronic (more than three months duration) subjective idiopathic tinnitus. Individual participant characteristics were poorly described in the included studies. All studies included in the current review recruited chronic tinnitus patients (more than three months). Only three included studies performed a sample size estimation (Drew 2001; Halama 1988; Rejali 2004). Future studies should seek to recruit an adequate sample size based on an appropriate power calculation for the primary outcome.

It was not possible to address the question of the effectiveness of Ginkgo biloba for tinnitus associated with cognitive insufficiency as only two studies included such groups of patients (Halama 1988; Napryeyenko 2009), with neither study including outcomes of interest to the current review, but it is a key area to consider for future development. Therefore, any future research into the effect of Ginkgo biloba on tinnitus must be careful to define its patient population with respect to cerebral function. It would be beneficial if further trials in cognitive insufficiency explicitly report the results for the subgroup of patients who have tinnitus at the outset of the study. 

What's new

Date	Event	Description
9 January 2023	Amended	Award number for University of Maryland School of Medicine/Cochrane Complementary Medicine Field bursary added (Sources of support and Acknowledgements).

History

Protocol first published: Issue 12, 2019
Review first published: Issue 11, 2022

Appendices

Appendix 1. Search strategies

CENTRAL and Cochrane ENT Register (CRS)	MEDLINE (Ovid)	Embase (Ovid)	AMED (Ovid)
1 MESH DESCRIPTOR Tinnitus EXPLODE ALL  2 (zumbido or tinnit*):AB,EH,KW,KY,MC,MH,TI,TO  3 MESH DESCRIPTOR Ginkgo biloba EXPLODE ALL  4 (Ginkgo* or Gingko* or Ginko* or Maidenhair):AB,EH,KW,KY,MC,MH,TI,TO  5(Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761):AB,EH,KW,KY,MC,MH,TI,TO  6 (GBE50 or EGB50):AB,EH,KW,KY,MC,MH,TI,TO  7 (rokan or tanakan or tebofortran or tebokan or tebonin or Kavari):AB,EH,KW,KY,MC,MH,TI,TO  8 (LI 1370 or LI1370 or LI‐1370):AB,EH,KW,KY,MC,MH,TI,TO  9 (flavanoid* or terpenoid* or bioflavanoid*):AB,EH,KW,KY,MC,MH,TI,TO  10 (gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*):AB,EH,KW,KY,MC,MH,TI,TO  11 (biloba):AB,EH,KW,KY,MC,MH,TI,TO  12 (Eun‐haeng* OR Fossil Tree* OR Ginkyo* OR Icho* OR Ityo* OR Japanese Apricot* OR Kew Tree* OR Salisburia* OR Silver Apricot* OR Pterophyllus salisburiensis OR Yinxingye* OR Prunus Ume OR Prunus mume):AB,EH,KW,KY,MC,MH,TI,TO  13 #1 OR #2  14 #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12  15 #13 AND #14	1 exp Tinnitus/  2 (tinnit* or zumbido).ti,ab.  3 1 or 2  4 exp Ginkgo biloba/  5 (Ginkgo* or Gingko* or Ginko* or Maidenhair).ab,ti.  6 (Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761).ab,ti.  7 (GBE50 or EGB50).ab,ti.  8 (rokan or tanakan or tebofortran or tebokan or tebonin or Kavari).ab,ti. 9 (LI 1370 or LI1370 or LI‐1370).ab,ti.  10 (flavanoid* or terpenoid* or bioflavanoid*).ab,ti.  11 (gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*).ab,ti.  12 (ginkgo* or "LI 1370").nm.  13 biloba.ab,ti.  14 (Eun‐haeng* or Fossil Tree* or Ginkyo* or Icho* or Ityo* or Japanese Apricot* or Kew Tree* or Salisburia* or Silver Apricot* or Pterophyllus salisburiensis or Yinxingye* or Prunus Ume or Prunus mume).ab,ti.  15 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14  16 3 and 15	1 exp tinnitus/  2 (tinnit* or zumbido).ab,ti.  3 1 or 2  4 exp Ginkgo biloba/  5 (Ginkgo* or Gingko* or Ginko* or Maidenhair).ab,ti.  6 (Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761).ab,ti.  7 (GBE50 or EGB50).ab,ti.  8 (rokan or tanakan or tebofortran or tebokan or tebonin or Kavari).ab,ti.  9 (LI 1370 or LI1370 or LI‐1370).ab,ti.  10 (flavanoid* or terpenoid* or bioflavanoid*).ab,ti.  11 (gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*).ab,ti.  12 biloba.ab,ti.  13 (Eun‐haeng* or Fossil Tree* or Ginkyo* or Icho* or Ityo* or Japanese Apricot* or Kew Tree* or Salisburia* or Silver Apricot* or Pterophyllus salisburiensis or Yinxingye* or Prunus Ume or Prunus mume).ab,ti.  14 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13  15 3 and 14	1 exp Tinnitus/           2 (tinnit* or zumbido).ti,ab. 3 1 or 2            4  exp Ginkgo biloba/            5 (Ginkgo* or Gingko* or Ginko* or Maidenhair).ab,ti.            6 (Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761).ab,ti. 7 (GBE50 or EGB50).ab,ti.            8 (rokan or tanakan or tebofortran or tebokan or tebonin or Kavari).ab,ti.            9 (LI 1370 or LI1370 or LI‐1370).ab,ti.            10 (flavanoid* or terpenoid* or bioflavanoid*).ab,ti.            11 (gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*).ab,ti.            12 biloba.ab,ti.            13 (Eun‐haeng* or Fossil Tree* or Ginkyo* or Icho* or Ityo* or Japanese Apricot* or Kew Tree* or Salisburia* or Silver Apricot* or Pterophyllus salisburiensis or Yinxingye* or Prunus Ume or Prunus mume).ab,ti.            14 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13            15 3 and 14
CINAHL (EBSCO)	Web of Science (Web of Knowledge)	Trial registries	Other
S15      S3 AND S14                   S14      S4 OR S5 OR S6 OR S7 OR S8 OR S9 OR S10 OR S11 OR S12 OR S13            S13      TX Eun‐haeng* or Fossil Tree* or Ginkyo* or Icho* or Ityo* or Japanese Apricot* or Kew Tree* or Salisburia* or Silver Apricot* or Pterophyllus salisburiensis or Yinxingye* or Prunus Ume or Prunus mume            S12      TX biloba            S11      TX gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*  S10      TX flavanoid* or terpenoid* or bioflavanoid*  S9        TX LI 1370 or LI1370 or LI‐1370  S8        TX rokan or tanakan or tebofortran or tebokan or tebonin or Kavari  S7        TX GBE50 or EGB50  S6        TX Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761  S5        TX Ginkgo* or Gingko* or Ginko* or Maidenhair  S4        (MH "Ginkgo Biloba")  S3        S1 OR S2  S2        TX tinnit* or zumbido  S1        (MH "Tinnitus")	# 12 #11 AND #1 Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years # 11 #10 OR #9 OR #8 OR #7 OR #6 OR #5 OR #4 OR #3 OR #2 Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years # 10 TOPIC: (Eun‐haeng* or Fossil Tree* or Ginkyo* or Icho* or Ityo* or Japanese Apricot* or Kew Tree* or Salisburia* or Silver Apricot* or Pterophyllus salisburiensis or Yinxingye* or Prunus Ume or Prunus mume) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years         # 9 TOPIC: (biloba)  Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years           # 8 TOPIC: (gingkco* or gingho* or ginosan* or bilobalid* or tanakene or supergin*) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years  # 7 TOPIC: (flavanoid* or terpenoid* or bioflavanoid*) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years # 6 TOPIC: (LI 1370 or LI1370 or LI‐1370)            Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years           # 5 TOPIC: (rokan or tanakan or tebofortran or tebokan or tebonin or Kavari) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years           # 4 TOPIC: (GBE50 or EGB50) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years            # 3 TOPIC: (Egb 761 or Egb761 or EGb‐761 or GBE 761 or GBE761 or GBE‐761) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years #2 TOPIC: (Ginkgo* or Gingko* or Ginko* or Maidenhair) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years            # 1 TOPIC: (tinnit* or zumbido) Indexes=SCI‐EXPANDED, CPCI‐S Timespan=All years	ClinicalTrials.gov tinnitus | ( Ginkgo OR Gingko OR Ginko OR gingkco OR gingho OR maidenhair )           ICTRP tinnitus AND ginkgo OR tinnitus AND gingko OR tinnitus AND ginko OR tinnitus AND gingkco OR tinnitus AND gingho OR tinnitus AND maidenhair	LILACS tw:(tinnit* OR zumbido) AND ( db:("LILACS") AND type_of_study:("clinical_trials")) CNKI via Google Scholar site:en.cnki.com.cn tinnitus ( Ginkgo OR Gingko OR Ginko OR gingkco OR gingho OR maidenhair ) Google Scholar allintitle: tinnitus ( Ginkgo OR Gingko OR Ginko OR gingkco OR gingho OR maidenhair )

Data and analyses

Comparison 1. Ginkgo biloba versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Tinnitus symptom severity (THI) at 3 to 6 months	2	85	Mean Difference (IV, Random, 95% CI)	‐1.35 [‐8.26, 5.55]
1.2 Serious adverse effects at 3 months (yes or no)	4	1154	Risk Ratio (M‐H, Fixed, 95% CI)	Not estimable
1.3 Tinnitus loudness measured by audiometric loudness matching at 12 weeks	1	73	Mean Difference (IV, Fixed, 95% CI)	‐4.00 [‐13.33, 5.33]
1.4 Health‐related quality of life at 3 months	1	60	Mean Difference (IV, Fixed, 95% CI)	‐0.58 [‐4.67, 3.51]
1.5 Other adverse effects at 3 months (yes or no)	4	1175	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.52, 1.60]

Comparison 2. Gingko biloba with concurrent intervention versus concurrent intervention only.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Tinnitus symptom severity (THI) at 3 months	1	22	Mean Difference (IV, Fixed, 95% CI)	7.50 [2.83, 12.17]
2.2 Tinnitus loudness (VAS) at 3 months (range 0 to 10)	1	22	Mean Difference (IV, Fixed, 95% CI)	1.00 [0.52, 1.48]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Cekkayan 1996.

Study characteristics
Methods	A 3‐armed (betahistine, Gingko biloba and placebo), parallel‐group randomised controlled trial with 3 months duration of treatment and 3 months duration of follow‐up
Participants	Location: Malatya, Turkey Setting: single‐centre study, Department of Otorhinolaryngology, Inönü University, from September 1993 to April 1994 Sample size: Number randomised: 26 in Ginkgo biloba group and 13 in placebo group Number completed: 26 in Ginkgo biloba group and unclear how many (8 or 13) in placebo group Participant baseline characteristics: Age (mean (range) years): 38.8 (14 to 63) in Gingko biloba group versus 50 (26 to 70) in placebo group Gender (male/female, n): 14/12 in Gingko biloba group versus 4/9 in placebo group Other characteristics: group data for degree of hearing loss are not available. Participants had degrees of uni‐ or bilateral sensorineural hearing loss varying from mild to profound. Group data for tinnitus laterality are not available. Participants with bilateral and unilateral tinnitus were included. Data for baseline tinnitus severity, duration tinnitus, baseline tinnitus loudness and characteristics, baseline depression and anxiety were not provided. Inclusion criteria: patients with subjective tinnitus Exclusion criteria: none
Interventions	Intervention group: Ginkgo biloba tablets (9.6 mg ginkgo glycoside/tablet; Tebokan Fort tbl./Rökan/Tebonin), 1 tablet 3 times a day, 3 months Comparator group: placebo capsules, 1 capsule 3 times a day, 3 months Use of additional interventions: none reported
Outcomes	Primary outcome: change in tinnitus severity scale (range of 0 to 4) at 3 months Secondary outcome: adverse effects at 3 months
Funding sources	No information provided
Declarations of interest	No information provided
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The authors stated that participants were randomly allocated to groups but did not provide any details on methods Quote: "Randomised groups were created (...)"
Allocation concealment (selection bias)	Unclear risk	Information not reported in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	High risk	Data on changes in tinnitus severity categories are reported only for 8 out of 13 participants (Table 3) but dropout and reasons are not mentioned. It seems that data for 5 participants in the placebo group are missing.
Selective reporting (reporting bias)	High risk	Routine and audiological examinations were performed every 15 days, however those data are not reported. Only data regarding tinnitus severity categories are reported at 3 months.
Other bias	Unclear risk	No prospective protocol available. No funding sources and declarations of interest reported. Statistical tests not reported ‐ not clear whether they were appropriate.

Drew 2001.

Study characteristics
Methods	A 2‐armed (Gingko biloba and placebo), parallel‐group randomised controlled trial with 12 weeks duration of treatment and 4, 12 and 14 weeks duration of follow‐up
Participants	Location: Birmingham, UK Setting: carried out entirely by mail and telephone Sample size: Number randomised: 489 in Ginkgo biloba group and 489 in placebo group Number completed: 478 in Ginkgo biloba group and 478 in placebo group Participant baseline characteristics: Age (mean (SD) years): 52.9 (9.3) in Gingko biloba group versus 53 (9.3) in placebo group Gender (male/female, n): 338/151 in Gingko biloba group versus 338/151 in placebo group Other characteristics: average duration of tinnitus was 10 years in the intervention group and 10.1 years in the control group. Data for baseline tinnitus severity, baseline tinnitus loudness, laterality and characteristics, baseline depression and anxiety, baseline degree and characteristics of hearing loss were not provided. Inclusion criteria: participants with tinnitus Exclusion criteria: age < 18 years or > 70 years old, pregnant or trying to get pregnant, previously taken Ginkgo biloba extract, had had tinnitus for < 12 months, reported that their tinnitus had varied greatly in the 6 months before the screening questionnaire, tried any treatment for tinnitus in the 6 months before completing the screening questionnaire (for example, acupuncture, homoeopathy, hypnotherapy, etc.), not generally in good health, taking anticoagulant drugs or antidepressants, abnormal blood pressure
Interventions	Intervention group: Ginkgo biloba tablets (50 mg of Ginkgo biloba standardised extract LI 1370 containing 25% flavonoids, 3% ginkgolides and 5% bilobalides), 3 tablets daily for 12 weeks Comparator group: placebo tablets, 3 tablets daily for 12 weeks Use of additional interventions: none reported
Outcomes	Primary outcomes: change in loudness measured with a categorical rating scale (‐6 to 6) and change in troublesome nature of tinnitus measured with a categorical rating scale (‐4 to 4) at 4, 12 and 14 weeks Secondary outcomes: non‐standard questionnaire created for this study: loudness questions, awareness/ability to ignore tinnitus, impact questions, tinnitus variability, cerebral insufficiency, compliance and side effects at 4, 12 and 14 weeks
Funding sources	This work was funded by the British Tinnitus Association in conjunction with Lichtwer Pharma UK, manufacturer of the extract used in the study
Declarations of interest	The study was financed (2 years' salary for SD and running costs) by a contract between the British Tinnitus Association and Lichtwer Pharma GmbH, Berlin, who also supplied the Ginkgo biloba extract and placebo tablets
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were paired according to the criteria described. Each pair was then allocated two numbers from a randomly arranged code. One number corresponded to placebo treatment and one to active treatment."
Allocation concealment (selection bias)	Low risk	Quote: "Placebo tablets were identical to the active tablets in shape, size, colour, and packaging. The tablets were dispensed in coded bottles, and treatment allocation was masked. The allocation procedure ensured that all matched participants received active or placebo tablets without the code being identified."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "Placebo tablets were identical to the active tablets in shape, size, colour, and packaging. The tablets were dispensed in coded bottles, and treatment allocation was masked. The allocation procedure ensured that all matched participants received active or placebo tablets without the code being identified."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Data entry and initial analyses were carried out by a researcher blinded to the participant’s allocation."
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Analyses on unmatched data were performed but not reported, except statement that: Quote: "Unmatched analyses did not provide any additional information and have therefore been excluded from this paper."
Selective reporting (reporting bias)	Low risk	All stated outcome measures reported.
Other bias	Unclear risk	Limited reporting of baseline data regarding tinnitus characteristics and severity. Participants were matched for age, sex and duration of tinnitus and stated cause of tinnitus but not severity. No data regarding ‘unmatched’ participants characteristics (i.e. was a specific group excluded from the analysis).

Fucci 1992.

Study characteristics
Methods	A 2‐armed (Gingko biloba and placebo), parallel‐group randomised controlled trial with 6 months duration of treatment and 3 and 6 months duration of follow‐up
Participants	Location: Philadelphia, USA Setting: Department of Otorhinolaryngology, Temple University Medical School, Philadelphia, USA Sample size: Number randomised: 40 in Ginkgo biloba and placebo group Number completed: not reported Participant baseline characteristics: Age (mean (SD) years): not reported Gender (male/female, n): not reported Other characteristics: data for baseline tinnitus severity, baseline tinnitus duration, loudness, laterality and characteristics, baseline depression and anxiety, baseline degree and characteristics of hearing loss were not provided. Inclusion criteria: participants with tinnitus Exclusion criteria: not reported
Interventions	Intervention group: Ginkgo biloba extract (EGb, bioflavonoid) for 6 months Comparator group: placebo for 6 months Use of additional interventions: none reported
Outcomes	Primary outcomes: tinnitus loudness and annoyance measured with numerical rating scales (1 to 10) at 3 and 6 months Secondary outcomes: pure tone air and bone conduction thresholds, speech discrimination scores, speech reception thresholds, impedance audiometry, tinnitus loudness matching levels, tinnitus minimum masking levels and tinnitus pitch matching frequencies at 3 and 6 months
Funding sources	No information provided
Declarations of interest	No information provided
Notes	Only conference abstract was available for the study. Author was contacted but did not provide additional data.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Author confirmed that study was randomised but did not provide further details. Methods not described. Author disclosed that there was a potential bias in the selection of participants.
Allocation concealment (selection bias)	Unclear risk	Only brief abstract available. Information not provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Only brief abstract available. Information not provided.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Only brief abstract available. Information not provided.
Incomplete outcome data (attrition bias) All outcomes	High risk	Only brief abstract available. No information provided regarding number of participants who completed the study and completeness of outcome measures.
Selective reporting (reporting bias)	High risk	Only brief abstract available. Not clear if all outcomes were reported. No data for secondary outcomes reported.
Other bias	High risk	Author expressed concerns regarding quality of the study; no additional details provided. No prospective protocol available. No funding sources and declarations of interest reported.

Halama 1988.

Study characteristics
Methods	A 2‐armed (Gingko biloba and placebo), parallel‐group randomised controlled trial with 12 weeks duration of treatment and 4, 12 and 14 weeks duration of follow‐up
Participants	Location: Hamburg, Germany Setting: outpatients in the neurology practice Sample size: Number randomised: 20 in Ginkgo biloba group and 20 in placebo group Number completed: 20 in Ginkgo biloba group and 20 in placebo group Participant baseline characteristics: Age (mean (SD) years): 65.2 (6.96) in Gingko biloba group versus 66.6 (9.43) in placebo group Gender (male/female, n): not reported. Unequal gender composition of the treatment groups was mentioned in the manuscript. Other characteristics: baseline data regarding tinnitus severity, characteristics, hearing loss, depression and anxiety were not reported in the manuscript. All patients had "mild cerebral insufficiency of vascular origin". Inclusion criteria: outpatients over 55 years of age from a neurology practice with diagnosis of "mild to mass cerebral insufficiency of vascular origin", the diagnosis was based on the clinical findings and the results of the psychometric tests performed and scale ratings (Hachinski: > 7, Crichton: levels 1 to 3). Exclusion criteria: patients with psychoses or neuroses, with primary degenerative dementia, with secondary brain impairment (treatable underlying disease, e.g. intoxication, metabolic disorder, alcoholism, etc.), epilepsy or other serious illnesses (e.g. myocardial infarction in the last 6 months, severe kidney failure, severe, life‐threatening) were excluded cardiac arrhythmias, malignancy) and those who were not sufficiently capable of co‐operating. In addition, clonidine, reserpine, antihistamines, anti‐Parkinson drugs, anticoagulants, platelet aggregation inhibitors, psychotropic drugs, centrally acting stimulants or drugs that influence the hormone metabolism or blood flow were not allowed to be taken.
Interventions	Intervention group: Gingko biloba extract (Tebonin forte film tablets) 3 x 1 per day, corresponding to a daily dose of 120 mg Gingko biloba extract for 12 weeks Comparator group: placebo tablets, 3 x 1 per day for 12 weeks Use of additional interventions: none reported
Outcomes	Primary outcomes: Sandoz Clinical Assessment Geriatric (SCAG) scale at 12 weeks Secondary outcomes: Crichton scale (CRBRS), background interference method (HIV), short syndrome test (SKT), craniocorpography (CCG). The severity of dizziness symptoms, noises in the ears, headaches and reduced hormones was assessed with the help of a 4‐point rating scale at 12 weeks.
Funding sources	Not reported
Declarations of interest	Not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Intervention or placebo randomly assigned by computer program to participant enrolment numbers.
Allocation concealment (selection bias)	Unclear risk	Information not provided in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Neither the patient nor the examining physician knew the assignment; the code was only opened after the data collection had been completed at the beginning of the evaluation. Identical tablets.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Neither the patient nor the examining physician knew this assignment; the code was only opened after the data collection had been completed at the beginning of the evaluation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All patients who were randomised were assessed.
Selective reporting (reporting bias)	High risk	Data for the majority of secondary outcome measures were not reported.
Other bias	High risk	This was a trial with patients with "mild cerebral insufficiency of vascular origin"; tinnitus was a secondary outcome measure. No prospective protocol available. No funding sources and declarations of interest reported.

Meyer 1986.

Study characteristics
Methods	A 2‐armed (Ginkgo biloba and placebo), parallel‐group randomised controlled trial with 3 months duration of treatment and 3 months duration of follow‐up
Participants	Location: Paris, France Setting: single‐centre study, ENT Department, Saint‐Antoine Hospital, Paris Sample size: Number randomised: 58 in Ginkgo biloba group and 45 in placebo group Number completed: 56 in Ginkgo biloba group and 44 in placebo group Participant baseline characteristics: Age: 51 years in Ginkgo biloba group versus 50 years in placebo group Gender (male/female, n): 29/29 in Gingko biloba group versus 25/20 in placebo group Other characteristics: mean tinnitus symptom severity measure on a 1 to 4 scale (mild to severe) was 2.46 (SD 0.9) in the Ginkgo biloba group and 2.47 (SD 0.6) in the placebo group. Mean baseline tinnitus discomfort measured on the scale from 0 to 3 was 1.83 (SD 0.73) in the intervention group and 1.71 (SD 0.79) in the control group. Mean baseline tinnitus intensity measured on a scale from 0 to 3 was 2 (SD 0.58) in the intervention group and 1.9 (SD 0.54) in the placebo group. Mean tinnitus duration was 127 days in the Ginkgo biloba group and 143 days in the placebo group. Participants with intermittent or constant and uni‐ and bilateral tinnitus were included. Baseline data for hearing, anxiety and depression were not reported. Inclusion criteria: recent tinnitus, i.e. tinnitus that had appeared within less than 1 year Exclusion criteria: patients undergoing surgical or anti‐infectious treatment, those affected by acute outer or middle ear illnesses (otitis, tubal catarrh etc.), or associated pathological issues that could potentially skew the analysis of the results. The study also excluded patients who were undergoing treatment and for whom development was favourable as well as those requiring medical treatment that could interfere with the assessment of the results.
Interventions	Intervention group: Ginkgo biloba extract (GbE 761 Tanakan – I.P.S.E.N Institute), 4 mL daily, taken in 2 doses for 3 months Comparator group: placebo, 4 mL daily, taken in 2 doses for 3 months Use of additional interventions: none reported
Outcomes	Primary outcome: change in tinnitus on a scale from 1 to 7 (1 to 2 = deterioration; 3 = no change; 4 = slight improvement; 5 = improvement; 6 = major improvement; 7 = discontinued due to intolerance) at 3 months Secondary outcomes: time to tinnitus disappearance or clear improvement, change of tinnitus intensity, change in tinnitus discomfort at 3 months
Funding sources	Not reported
Declarations of interest	Not reported
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Information on methods of random sequence generation not reported in the manuscript.
Allocation concealment (selection bias)	Unclear risk	Information not reported in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blinding stated but no details provided.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Three participants dropped out due to intolerance.
Selective reporting (reporting bias)	Unclear risk	Outcomes not specified in the methods section
Other bias	High risk	No prospective protocol available. No funding sources and declarations of interest reported. Scales for intensity, discomfort, time to disappear or clear improvement are not defined; unclear what numbers 0 to 3 mean. 1 participant dropped out due to intolerance in the placebo group – unclear whether placebo was a real placebo.

Morgenstern 1997.

Study characteristics
Methods	A 2‐armed (Gingko biloba and placebo), parallel‐group randomised controlled trial with 12 weeks duration of treatment and 4, 8 and 12 weeks duration of follow‐up
Participants	Location: Hamburg, Germany Setting: single‐centre study, St. Georg Hospital, Hamburg, Germany Sample size: Number randomised: 49 in Ginkgo biloba group and 50 in placebo group Number completed: 35 in Ginkgo biloba group and 38 in placebo group Participant baseline characteristics: Age: over 18 years Gender: not reported Other characteristics: baseline tinnitus loudness was 42.2 dB HL in the intervention group and 44.3 dB HL in the control group. Mean tinnitus duration in all participants was 4.5 years. All participants had normal hearing at 3 neighbouring frequencies in the audiogram. Baseline data for tinnitus symptom severity, tinnitus characteristics, anxiety and depression were not reported. Inclusion criteria: over 18 years old and chronic subjective tinnitus, normal hearing at 3 neighbouring frequencies in the audiogram Exclusion criteria: objective tinnitus, middle ear pathology, reproducibility of the transient evoked otoacoustic emissions at the intake examination < 85%, latency extensions or pathological wave models of earlier brain stem auditory evoked potentials (BAEPs), severe organic diseases, alcohol or drug abuse, pregnancy
Interventions	Intervention group: Ginkgo biloba coated tablets (40 mg of Gingko special extract EGb 761, a dry extract made from gingko biloba leaves (35 to 67:1), set to 9.6 mg gingko flavone glycoside and 2.4 mg of terpene lactone), 1 tablet 3 x day (120 mg), 12 weeks Comparator group: placebo coated tablets, 1 tablet 3 x day, 12 weeks Use of additional interventions: none reported
Outcomes	Primary outcome: change in tinnitus loudness measured with audiometric methods in the ear with louder tinnitus at 4, 8 and 12 weeks Secondary outcomes: change in tinnitus intensity measured with a scale of Subjective Evaluation of Tinnitus Intensity (0 to 5), click‐evoked otoacoustic emissions (OAEs) at 4, 8 and 12 weeks
Funding sources	No information provided
Declarations of interest	No information provided
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Information on methods of random sequence generation not reported in the manuscript.
Allocation concealment (selection bias)	Unclear risk	Information not reported in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Double‐blinding stated, methods not described.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	High risk	There are no data regarding the primary outcome measure for 14 patients in the intervention group and 12 in the control group at 12 weeks. Dropout is not explained.
Selective reporting (reporting bias)	Unclear risk	No values for secondary outcome measures are reported. Only brief statements regarding whether those improved or not.
Other bias	Unclear risk	This is an abbreviated version of the full study report, which is not available. No prospective protocol available. No funding sources and declarations of interest reported.

Napryeyenko 2009.

Study characteristics
Methods	A 2‐armed (Ginkgo biloba and placebo), parallel‐group randomised controlled trial with 22 weeks duration of treatment and 12 and 22 weeks duration of follow‐up
Participants	Location: Kyiv, Ukraine Setting: multi‐centre study, outpatient clinics of 16 psychiatric or neurological hospitals Sample size: Number randomised: 200 in Ginkgo biloba group and 200 in placebo group Number completed: 196 in Ginkgo biloba group and 195 in placebo group Participant baseline characteristics: Age (mean (SD) years): 65 (8) in Ginkgo biloba group versus 63 (8) in placebo group Gender (male/female, n): 55/143 in Ginkgo biloba group versus 55/142 in placebo group Other characteristics: all participants had either Alzheimer's disease (AD; n = 214) or vascular dementia (VaD; n = 181). Baseline tinnitus symptom severity in the whole study sample measured on the 11‐point box scale (from 0 indicating absence to 10 indicating extreme severity of the symptom) was 2.1 (SD 2.3) in the Ginkgo biloba group and 2.1 (SD 2.2) in the placebo group. From 400 participants enrolled in the study 204 had tinnitus (101 in the Ginkgo biloba group and 103 in the placebo group). Baseline tinnitus symptom severity for participants with tinnitus was 4.02 in the Ginkgo biloba group and 3.90 in the placebo group. Baseline data for tinnitus characteristics and hearing were not reported. Inclusion criteria: participants were eligible for the study if they were at least 50 years of age and diagnosed with AD (probable AD or possible AD with cerebrovascular disease) or VaD. Diagnoses were established employing the criteria specified by the National Institute for Neurological and Communicative Disorders and Stroke (NINCDS) together with the Alzheimer's Disease and Related Disorders Association (ADRDA) and by the National Institute of Neurological Disorders and Stroke (NINDS) together with the Association Internationale pour la Recherche et l'Enseignement en Neurosciences (AIREN) as appropriate. A CT or MRI scan, no more than 1 year old, had to be consistent with the inclusion diagnosis. The Test for Early Detection of Dementia with Discrimination from Depression (TE4D) was used as a screening instrument and to verify the presence of cognitive impairment in at least 2 domains. A total cognitive score of no more than 35 was required for inclusion. Patients had to have mild to moderate dementia as evidenced by a total score from 9 to 23 (both inclusive) on the SKT test battery, which roughly corresponds to a range from 14 to 25 on the MMSE or 17 to 35 on the cognitive subscale of the Alzheimer's Disease Assessment Scale (ADAS‐cog). The Clock‐Drawing Test (CDT) after Sunderland et al was employed as a second screening instrument, the score of which had to be below 6. Patients had to score at least 5 on the 12‐item Neuropsychiatric Inventory (NPI), with at least one item score (other than delusions or hallucinations) being 3 or higher. Severe depression was excluded by requiring a score below 20 on the 17‐item Hamilton Rating Scale for Depression (HAMD). The presence of a caregiver was required who was able and willing to provide information on the patient's behaviour and competence to perform activities of daily living. Exclusion criteria: patients were excluded from the study if they had any other type of dementia or neurological disorder, current or recent major depression or other psychiatric disorder, severe or insufficiently controlled cardiovascular, renal or hepatic disorder, diabetes, anaemia or thyroid dysfunction. Active malignant disease, HIV or lues infection and gastrointestinal diseases with uncertain absorption were not accepted. Treatment with other anti‐dementia drugs, cognitive enhancers, cholinergic, anti‐cholinergic or haemo‐rheologically active drugs, anti‐Parkinson drugs or Ginkgo supplements was prohibited during the study and at least 8 weeks preceding randomisation.
Interventions	Intervention group: EGb 761 dry extract from Ginkgo biloba leaves (35 to 67:1), extraction solvent: acetone 60% (w/w). The extract was adjusted to 22.0% to 27.0% Ginkgo flavonoids, calculated as Ginkgo flavone glycosides, and 5.0% to 7.0% terpene lactones consisting of 2.8% to 3.4% ginkgolides A, B, C and 2.6% to 3.2% bilobalide, with a content of ginkgolic acids below 5 ppm. Daily dose of 240 mg (2 × 120 mg), 22 weeks. Comparator group: placebo, 22 weeks Use of additional interventions: none reported
Outcomes	Primary outcome: the SKT, a 9‐item cognitive test battery with scores ranging from 0 to 27 at 12 and 22 weeks Secondary outcomes: 12‐item Neuropsychiatric Inventory (NPI), which assesses frequency and severity of neuropsychiatric symptoms (composite score range 0 to 144) and caregiver distress caused by such symptoms (score range 0 to 60); the activities‐of‐daily living (ADL) subscale of the Gottfries–Bråne–Steen (GBS) overall geriatric assessment scale; the total score of the GBS (comprised by the intellectual “I”, the emotional “E”, and ADL subscores, but not the behavioural “S” subscore which was not documented to avoid redundancy); the Verbal Fluency Test (animal fluency); the Clock‐Drawing Test and the Hamilton Rating Scale for Depression. Patient self‐ratings of presence and severity of dizziness and tinnitus, symptoms often associated with old age, were documented using 11‐point box scales, 0 representing absence and 10 indicating extreme severity of a symptom. All measures were collected at baseline, 12 and 22 weeks. Safety was assessed by documentation of adverse events at 6, 12, 17 and 22 weeks and physical examination, electrocardiography and laboratory tests at 22 weeks.
Funding sources	The clinical trial was sponsored by Dr. Willmar Schwabe GmbH & Co. KG, Karlsruhe, Germany
Declarations of interest	No information provided
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Centre‐stratified randomisation (drug–placebo ratio 1:1) in blocks of four was performed by the sponsor's biometrics unit using a validated computer program that linked ascending drug numbers to active drug or placebo, respectively."
Allocation concealment (selection bias)	Low risk	Quote: "The randomisation list was sealed and stored safely at the sponsor's biometrics unit, and block length was not disclosed to investigators."
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The multi‐centre trial with two parallel treatment arms was carried out in a double‐blind manner. Drug and placebo tablets were indistinguishable by appearance, packaging and labelling."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	Low risk	214 of 218 patients with Alzheimer's and 181 of 182 patients with vascular dementia with full data for analysis.
Selective reporting (reporting bias)	High risk	There are 2 reports from this study, one reporting tinnitus score in the whole study population (from 0 ‐ absence of tinnitus to 10 ‐ extreme severity) and one brief report reporting tinnitus scores only in those participants with tinnitus.
Other bias	High risk	No prospective protocol available. No declarations of interest reported. This was a study of effects of Ginkgo biloba on Alzheimer's disease or vascular dementia. Tinnitus was secondary outcome measure.

Nishad 2019.

Study characteristics
Methods	A 4‐armed (Gingko biloba and placebo; caroverine and placebo), parallel‐group randomised controlled trial with 3 months duration of treatment and 4 weeks, 12 weeks and 6 months duration of follow‐up
Participants	Location: India Setting: single‐centre study, Otolaryngology Department, India Sample size: Number randomised: 22 in Ginkgo biloba group and 21 in placebo group Number completed: 22 in Ginkgo biloba group and 21 in placebo group Participant baseline characteristics: Age: range 18 to 60; for all 4 arms most of the patients were in the 5th decade of life (40.69%) followed by 4th decade of life (22.09%) Gender (male/female, %): 67/33 for all 4 arms  Other characteristics: group level data for baseline tinnitus, hearing loss, depression and anxiety were not provided. All patients had cochlear synaptic tinnitus diagnosed with extensive test battery (Audiometry (0 to 16,000 Hz); impedance audiometry; Eustachian tube function test; acoustic reflex (Ipsi and Contra); clinching reflex test; reflex decay; tinnitus matching; tinnitus grading; short increment sensitivity index; brainstem evoked response audiometry). Inclusion criteria: age between 18 and 60 years, diagnosis of tinnitus, unilateral or bilateral, cochlear hearing defect, reflex audiometry confirming cochlear‐synaptic tinnitus and excluding middle ear tinnitus, written consent after explanation by consultant in charge Exclusion criteria: patients not willing to give written consent after receiving due instructions, reasonable doubts as to the patient’s co‐operation, severe secondary disorders (i.e. any acute/chronic illness), contra‐indication for caroverine therapy (e.g. severe hypertension), concomitant symptoms, pregnancy or plan for having children, hydrops cochlea or Ménière's disease, retrocochlear hearing defect, blast injury, status post psychiatrist therapy, status post operation of the middle ear, pulse‐synchronous tinnitus, excessive consumption of alcohol, drug or nicotine, known to intolerance to caroverine, concomitant medication, medicative tinnitus therapy within 1 week prior to enrolment, masticatory movements influencing subjective tinnitus sensation
Interventions	Intervention group: Ginkgo biloba (60 mg of Ginkgo biloba extract) 12‐hourly for a period of 3 months Comparator group: placebo (60 mg of placebo) 12‐hourly for a period of 3 months Use of additional interventions: none reported
Outcomes	Primary outcome: change in tinnitus grading measured with unspecified non‐standard questionnaire at 4 weeks, 12 weeks and 6 months Secondary outcome: quality of life (method not described) and tinnitus matching (method not described) at 4 weeks, 12 weeks and 6 months
Funding sources	None specified
Declarations of interest	The authors declare that they have no conflict of interest
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	By using a computer‐generated block randomisation method participants were randomised into 4 groups.
Allocation concealment (selection bias)	Unclear risk	Information not reported in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants seemed to complete the study, no dropout reported.
Selective reporting (reporting bias)	High risk	Quality of life stated as primary outcome but no results were reported. Report mentions "tinnitus matching" improved but this was not pre‐defined as an outcome measure.
Other bias	High risk	Lack of criteria for defining improvement, no statistical analyses, results do not justify conclusions. Audiological selection criteria complex and of uncertain significance in selecting stated aetiological group. No prospective protocol available. No funding sources reported.

Polanski 2016.

Study characteristics
Methods	A 4‐armed (Gingko biloba; α‐lipoic acid plus vitamin C; papaverine hydrochloride plus vitamin E; placebo (starch capsules)), parallel‐group randomised controlled trial with 6 months duration of treatment and 6 months duration of follow‐up
Participants	Location: São Paulo, Brazil Setting: single‐centre study, Department of Otorhinolaryngology and Head and Neck Surgery, Universidade Federal de São Paulo (UNIFESP), São Paulo, Brazil Sample size: Number randomised: 12 in Ginkgo biloba group and 13 in placebo group Number completed: 12 in Ginkgo biloba group and 13 in placebo group Participant baseline characteristics: Age: over 60 years, data reported only for all participants in 4 arms of the trial (72.6 (SD 6.6)) Gender (male/female, %): 26/32 for all 4 arms. Other characteristics: baseline tinnitus symptom severity measured with the THI was 32.8 (SD 19.9; range 12 to 80) in the Ginkgo biloba and 28.2 (SD 25.1; range 2 to 72) in the placebo group. Participants had variable degrees of sensorineural hearing loss confirmed by previous audiometric testing. 65.5% participants had hypertension. Data for baseline tinnitus characteristics, depression and anxiety were not reported. Inclusion criteria: aged 60 years or older with clinical complaints of tinnitus associated with a variable degree of sensorineural hearing loss confirmed by previous audiometric testing Exclusion criteria: known allergy to any substance to be tested or with clinical contraindications to the use of these substances; anticoagulant users or subjects with coagulopathy, as well as diabetics
Interventions	Intervention group: Ginkgo biloba dry extract (120 mg/day) for a period of 6 months Comparator group: placebo (starch capsules) for a period of 6 months Use of additional interventions: none reported
Outcomes	Primary outcome: hearing improvement measured with audiometry at 250, 500, 1000, 2000, 3000, 4000, 6000 and 8000 Hz at 6 months Secondary outcome: change in tinnitus symptom severity measured with the THI at 6 months
Funding sources	None specified
Declarations of interest	The authors declare no conflict of interest
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "For distribution and randomization of participants, the resources available at http://www.randomization.com were used."
Allocation concealment (selection bias)	Unclear risk	Information not reported in the manuscript.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The substances were not identified by name in the containers into which they were packed, but rather through symbols defined by a professional who did not participate in the research, as a way of blinding investigators and patients."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "In the original trial paper there were 15 placebo and 14 test participants who had tinnitus. Data for only 12 and 13 participants are included in this report. So some tinnitus participants were excluded without reason given."
Selective reporting (reporting bias)	High risk	Trial registration lists primary outcome: hearing improvement in an audiometric test as primary outcome, which is not described in this report.
Other bias	High risk	Funding not disclosed. This was a hearing loss trial with tinnitus as secondary outcome.

Radnuz 2019.

Study characteristics
Methods	A 3‐armed (Ginkgo biloba, hearing aid, Ginkgo biloba plus hearing aid), parallel‐group randomised controlled trial with 90 days duration of treatment and 90 days duration of follow‐up
Participants	Location: São Paulo, Brazil Setting: single‐centre study, outpatient clinic of the audiology centre in Rondonópolis (MT), Brazil Sample size: Number randomised: 11 in Ginkgo biloba plus hearing aid group and 11 in hearing aid group Number completed: 11 in Ginkgo biloba plus hearing aid group and 11 in hearing aid group The sample size (n = 35) was determined by the number of individuals submitted to medical consultation between July and August 2015 at an audiological centre in Rondonópolis (MT), Brazil Participant baseline characteristics: Age (mean, (SD)): group level data for age were not provided; mean age for participants in all 3 trial arms was 56.3 (16.8) years Gender (female/male, n): group level data for gender were not provided; gender distribution in all 3 trial arms was 18/15 Other characteristics: baseline tinnitus severity measured with the THI was 57.5 (SD 7.5) in the Ginkgo biloba plus hearing aids group and 52.5 (SD 5) in the hearing aids group. The 3 groups in the trial were homogeneous in the pre‐treatment period, when THI scores were compared within the confidence intervals using the ANOVA test with the application of the Tukey Average Comparison Post Test (P = 0.4708). Baseline tinnitus loudness measured with visual analogue scale (range of 0 to 10) was 6 (SD 0.6) in the Ginkgo biloba plus hearing aids group and 5.6 (SD 0.4) in the hearing aid group. Group level data for tinnitus duration were not provided. Duration of tinnitus in all 3 trial arms was 58.9 (SD 17.7) months. Baseline tinnitus quality, baseline hearing level and baseline anxiety/depression were not reported. Inclusion criteria: individuals over 18 years of age; complaint of tinnitus (uni‐ or bilateral) for at least 3 months; (single or bilateral) sensorineural or mixed hearing loss independent of degree and configuration Exclusion criteria: having used aspirin or acetylsalicylic acid in the last month, having used Ginkgo biloba in the last 3 months, on antidepressants, diagnosed with compromised middle ear (otitis or tubal dysfunction) at the time of evaluation
Interventions	Intervention group: Ginkgo biloba (EGb 761 Equitam leaf extract, patented/trademarked by Eurofarma Laboratorios S.A.), 240 mg/day plus digital Beltone hearing aid for a period of 3 months Comparator group: digital Beltone hearing aid for a period of 3 months Use of additional interventions: none
Outcomes	Primary outcome: change in tinnitus symptom severity measured using the THI and change in tinnitus loudness measured using a visual analogue scale (range of 0 to 10) at 3 months Secondary outcomes: percentage of variation of the THI score before and after treatment in relation to the tinnitus onset time at 3 months
Funding sources	Quote: "S.N.D is grateful to Anhanguera University of São Paulo(UNIAN) for providing financial support and Momenta Farmacêutica Ltd for provide the EGb 761 Equitam® Ginkgo biloba extract."
Declarations of interest	The authors declared no conflicts of interest
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The random allocation process has consisted in generating a random sequence using an Excel file for randomization and used the random allocation rule, which chooses at random one of possible balanced assignments of the given number of subjects per treatment."
Allocation concealment (selection bias)	Unclear risk	Methods of allocation concealment not reported.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "The examiners responsible for applying the questionnaires during the study were blinded to the intervention. Due to the nature of the intervention participants were not blinded."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The examiners responsible for applying the questionnaires during the study were blinded to the intervention. In addition, an employee outside the research team inserted data into the computer in separate data sheets so that the researchers can analyze data without having access to information about the allocation."
Incomplete outcome data (attrition bias) All outcomes	Low risk	All randomised patients' data were analysed.
Selective reporting (reporting bias)	Low risk	All stated outcomes reported.
Other bias	Low risk	No prospective protocol available.

Rejali 2004.

Study characteristics
Methods	A 2‐armed (Gingko biloba and placebo), parallel‐group randomised controlled trial with 12 weeks duration of treatment and 12 weeks duration of follow‐up
Participants	Location: Airdrie, UK Setting: single‐centre study, Department of Otolaryngology, Monklands Hospital Sample size: Number randomised: 33 in Ginkgo biloba group and 33 in placebo group Number completed: 31 in Ginkgo biloba group and 29 in placebo group Participant baseline characteristics: Age (mean (SD) years): 60 (11.4) in Ginkgo biloba group versus 59 (10.4) in placebo group Gender (male/female, n): 17/14 in Ginkgo biloba group versus 17/12 in placebo group Other characteristics: baseline tinnitus symptom severity measured with THI was 37.5 (SD 20.5) in the intervention group and 50.7 (SD 23.3) in the placebo group. Mean tinnitus duration was 4.4 years (SD 4.6) in the intervention group and 6 years (SD 4.9) in the placebo group. Tinnitus was constant in 87% of participants in the intervention group and 90% in the placebo group and intermittent in 13% and 10% of participants, respectively. Tinnitus was unilateral in 39% of participants in the intervention group and 31% of participants in the placebo group. Averaged air‐conduction threshold at 0.5 kHz, 1 kHz, 2 kHz and 4 kHz for both ears was 26.8 dB in the intervention group and 35.2 dB in the placebo group. Inclusion criteria: adult patients recruited from general otolaryngology outpatient clinics with tinnitus as their only or main presenting complaint Exclusion criteria: patients with active middle or external ear disease, known hypersensitivity to Ginkgo biloba and pregnant patients, and patients already on Ginkgo biloba, warfarin or with bleeding diatheses were excluded from the study
Interventions	Intervention group: Ginkgo biloba tablets, 120 mg once a day, 12 weeks Comparator group: placebo tablets, 1 tablet a day, 12 weeks Use of additional interventions: none reported
Outcomes	Primary outcome: change in tinnitus symptom severity measured with THI at 12 weeks Secondary outcomes: change in handicap caused by otolaryngological disease measured with Glasgow Health Status Inventory (GHSI) at 12 weeks; change in hearing determined by averaging air conduction thresholds for both ears at 0.5 kHz, 1 kHz, 2 kHz and 4 kHz at 12 weeks; adverse effects
Funding sources	Quote: "We acknowledge Lamberts Healthcare for the supply of G. biloba and placebo medication."
Declarations of interest	No information provided
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was carried out by an independent third party using a card from bag system.
Allocation concealment (selection bias)	Low risk	Randomisation was carried out by an independent third party using a card from bag system.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	The study was double‐blind. The patients received either 120 mg once daily sustained release formulation of Ginkgo biloba or placebo (all tablets provided by Lambert’s Health Care, TunbridgeWells, UK) for 12 weeks.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Small dropout (3 participants: 1 in intervention and 2 in placebo group); reasons for dropout explained.
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	Unclear risk	No prospective protocol available. No declarations of interest reported. Baseline tinnitus severity higher in the control group.

Yarmohammadi 2007.

Study characteristics
Methods	A 2‐armed (Ginkgo biloba and placebo), parallel‐group randomised controlled trial with 2 weeks duration of treatment and 2 weeks duration of follow‐up
Participants	Location: Teheran, Iran Setting: single‐centre study, otolaryngology clinic Sample size: Number randomised: 30 in Ginkgo biloba group and 30 in placebo group Number completed: 30 in Ginkgo biloba group and 30 in placebo group Participant baseline characteristics: Age (mean (SD) years): 51.2 (SD 1.9) in Ginkgo biloba group versus 51.5 (SD 1.7) years in placebo group Gender (male/female, n): 14/16 in Ginkgo biloba group versus 12/18 in placebo group Other characteristics: baseline tinnitus symptom severity was not reported. Mean tinnitus duration was 51.2 (SD 9.1) months in the intervention group and 56.9 (SD 7.9) months in the placebo group. Tinnitus was unilateral in 56.7% of participants in the intervention group and 66.7% participants in the placebo group and bilateral in the rest of the participants. All participants had normal hearing. Inclusion criteria: patients diagnosed with tinnitus and with no reduction of hearing who had a normal MRI were included in the study Exclusion criteria: none
Interventions	Intervention group: Ginkgo T.D. tablets (Tolid Darou Pharmaceutical Company), 40 mg, twice daily for 2 weeks Comparator group: placebo tablets, 40 mg, twice daily for 2 weeks Use of additional interventions: none reported
Outcomes	Primary outcome: improvement in severity of tinnitus measured with undefined questionnaire at 2 weeks Secondary outcomes: none
Funding sources	Quote: "The authors would like to thank Tolid Darou Pharmaceutical Company for providing the medication and the placebo."
Declarations of interest	No information provided
Notes	—
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: “The ginkgo T.D. tablets and the placebos were placed in an equal number of paper bags—30 bags containing ginkgo T.D. and 30 bags containing the placebo, labelling the bags with numbers. The doctor randomly administered the contents of the bags to the patients.”
Allocation concealment (selection bias)	Unclear risk	Bags with tablets but concealment methods not described (doctor randomly administered the contents of the bags to the patients).
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "In this study, the doctor and patients were unaware of the type of medication, with only the second co‐researcher in charge of numbering and packaging the bags being aware."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Information not reported in the manuscript.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data for all participants reported.
Selective reporting (reporting bias)	High risk	Questionnaire for tinnitus severity was not described; authors interpreted the result as no change, improvement and exacerbation but it is unclear what the criteria were for such classification.
Other bias	Unclear risk	No prospective protocol available. No declarations of interest reported.

AD: Alzheimer's disease; CT: computerised tomography; MRI: magnetic resonance imaging; SD: standard deviation; SKT: Short Cognitive Performance Test (Syndrom‐Kurztest); THI: Tinnitus Handicap Inventory; VaD: vascular dementia

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abascal 2012	ALLOCATION: not randomised; review
Ahsan 2017	ALLOCATION: not randomised; review
Bruchert 1991	ALLOCATION: randomised controlled trial POPULATION: patients with cognitive insufficiency not with tinnitus
Claussen 1988	ALLOCATION: randomised trial  POPULATION: patients with vertigo, nausea and tinnitus INTERVENTION: combined therapy, Ginkgo biloba and other medications
Coles 1988	ALLOCATION: not randomised; uncontrolled before‐and‐after study
Dau 2000	ALLOCATION: not randomised; uncontrolled before and after study
Feinberg 2003	ALLOCATION: not randomised; Q&A
Hajna 1999	ALLOCATION: not randomised; uncontrolled before‐and‐after study
Holgers 1994	ALLOCATION: the study was divided into two steps: an open phase where all participants were offered Ginkgo biloba extract, followed by a randomised cross‐over study in patients who reported benefit of the extract for tinnitus
ISRCTN38408464	The trial stopped on 11 September 2015 due to participant recruitment issues. No evaluable data available.
Kiefer 2019	ALLOCATION: randomised controlled trial POPULATION: adults with tinnitus INTERVENTION: the study compared Ginkgo biloba extract with pentoxifylline; there was no placebo group
Novotny 2000	ALLOCATION: non‐randomised controlled trial
Plath 1995	ALLOCATION: randomised controlled trial POPULATION: adults with chronic tinnitus INTERVENTION: the study compared a low‐power laser and Ginkgo biloba extract with sham laser and Ginkgo biloba extract
Sadner 2017	ALLOCATION: not randomised; review
Schneider 2000	ALLOCATION: not randomised; uncontrolled before‐and‐after study
von Wedel 1995	ALLOCATION: non‐randomised controlled study
Walger 1993	ALLOCATION: randomised controlled trial POPULATION: adults with chronic tinnitus INTERVENTION: the study compared a low‐power laser with Ginkgo biloba

Characteristics of studies awaiting classification [ordered by study ID]

Fandriantika 2017.

Methods	—
Participants	—
Interventions	—
Outcomes	—
Notes	Thesis under embargo, request for accessing full text refused.

Rogowski 2001.

Methods	—
Participants	—
Interventions	—
Outcomes	—
Notes	No full text available: contacted authors, no response.

Characteristics of ongoing studies [ordered by study ID]

ChiCTR2000038850.

Study name	'A randomized controlled clinical study of sound therapy combined with drug therapy for tinnitus'
Methods	RCT
Participants	n = 44 Inclusion criteria: 1) patients with tinnitus frequency between 125 Hz and 8000 Hz; 2) patients with chronic tinnitus (Otolaryngology ‐ Head and Neck Surgery 2014 tinnitus clinical guidelines: chronic tinnitus is defined as continuous tinnitus lasting more than 6 months at the time of examination); 3) patients without tinnitus treatment; 4) patients who did not take stimulants and ototoxic drugs; 5) patients who are willing to follow up Exclusion criteria: 1) pulsatile tinnitus; 2) patients with a history of auditory trauma; 3) patients with middle or inner ear diseases (such as otosclerosis, chronic suppurative otitis media or endolymphatic hydrops); 4) severe hearing asymmetry between ears, patients with post cochlear lesions, tinnitus induced by previous ear surgery and other diseases; 5) patients with head and neck discomfort or tinnitus caused by temporomandibular or cranial neck; 6) have used or are using other oral drugs to improve microcirculation within 3 months; 7) women who are pregnant or prepare for pregnancy within 12 months; 8) patients with severe chronic diseases not under control
Interventions	Oral administration of Ginkgo biloba extract tablets vs music combined with medication
Outcomes	Tinnitus Handicap Inventory, Zung Self‐Rating Anxiety Scale (SAS), electroencephalogram (EEG)
Starting date	1 December 2020
Contact information	Jianlong Wu
Notes	Status: recruiting

RCT: randomised controlled trial

Differences between protocol and review

In the protocol, under Electronic searches we planned to search ClinicalTrials.gov and ICTRP via the Cochrane Register of Studies. The Information Specialist determined that these searches were not necessary because they would be entirely covered by the searches run in CENTRAL and the Cochrane ENT Register.

We have changed the wording of the outcome 'Significant adverse effects' to 'Serious adverse effects' to avoid any confusion with statistical significance.

Contributions of authors

MS and DJH conceived and all authors contributed to the design of the study. MS drafted the protocol. All authors critically revised the protocol for important intellectual content.

Author contributions to the full review:

• The Cochrane ENT Information Specialist developed and ran the search strategy.

• MS obtained copies of studies with the assistance of the University of Nottingham library.

• MS, MH, AER and DJH were responsible for selection of studies.

• MS, MH, AER and DJH were responsible for data extraction.

• MS, MH, AER, DJH and JX were responsible for assessing risk of bias.

• MS entered data into RevMan.

• MS and PS conducted the analysis.

• MS, PS and JX interpreted the analysis.

• MS drafted the final review.

• MS and DJH will be responsible for updating the review.

Sources of support

Internal sources

• No sources of support provided

External sources

• National Institute for Health Research, UK

  Infrastructure funding for Cochrane ENT

• University of Maryland School of Medicine/Cochrane Complementary Medicine Field Bursary, USA

  Cochrane Complementary Medicine Field Bursary; award number: R24 AT001293

Declarations of interest

Magdalena Sereda: MS is a member of the Steering Committee for the British Society of Audiology Tinnitus and Hyperacusis Special Interest Group and Associate Editor for the International Journal of Audiology and BMC Health Services Research. She is funded by the NIHR Nottingham Biomedical Research Centre. She has received tinnitus research funding from the British Tinnitus Association, the British Society of Audiology and the NIHR.

Jun Xia: none known.

Polly Scutt: none known.

Malcolm Hilton: none known.

Amr El Refaie: none known.

Derek J Hoare: DJH is Editor‐at‐large for Ear and Hearing and Chair of the British Society of Audiology. He is funded by the NIHR and research lead for hearing at the NIHR Nottingham Biomedical Research Centre. He has received tinnitus research funding from the British Society of Audiology, the British Tinnitus Association, RNID, Horizon 2020, Royal British Legion, Help Musicians UK and the NIHR.

Edited (no change to conclusions)