Global Prevalence and Incidence of Tinnitus

Importance
To date, no systematic review has taken a meta-analytic approach to estimating the prevalence and incidence of tinnitus in the general population.


Objective
To provide frequency estimates of tinnitus worldwide.


Data Sources
An umbrella review followed by a traditional systematic review was performed by searching PubMed-MEDLINE and Embase from inception through November 19, 2021.


Study Selection
Research data from the general population were selected, and studies based on patients or on subgroups of the population with selected lifestyle habits were excluded. No restrictions were applied according to date, age, sex, and country.


Data Extraction and Synthesis
Relevant extracted information included type of study, time and location, end point, population characteristics, and tinnitus definition. The study followed the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guideline.


Main Outcomes and Measures
Pooled prevalence estimates of any tinnitus, severe tinnitus, chronic tinnitus, and diagnosed tinnitus as well as incidence of tinnitus were obtained using random-effects meta-analytic models; heterogeneity between studies was controlled using the χ2 test, and inconsistency was measured using the I2 statistic.


Results
Among 767 publications, 113 eligible articles published between 1972 and 2021 were identified, and prevalence estimates from 83 articles and incidence estimates from 12 articles were extracted. The pooled prevalence of any tinnitus among adults was 14.4% (95% CI, 12.6%-16.5%) and ranged from 4.1% (95% CI, 3.7%-4.4%) to 37.2% (95% CI, 34.6%-39.9%). Prevalence estimates did not significantly differ by sex (14.1% [95% CI, 11.6%-17.0%] among male individuals; 13.1% [95% CI, 10.5%-16.2%] among female individuals), but increased prevalence was associated with age (9.7% [95% CI, 7.4%-12.5%] among adults aged 18-44 years; 13.7% [95% CI, 11.0%-17.0%] among those aged 45-64 years; and 23.6% [95% CI, 19.4%-28.5%] among those aged ≥65 years; P < .001 among age groups). The pooled prevalence of severe tinnitus was 2.3% (95% CI, 1.7%-3.1%), ranging from 0.5% (95% CI, 0.3%-0.7%) to 12.6% (95% CI, 11.1%-14.1%). The pooled prevalence of chronic tinnitus was 9.8% (95% CI, 4.7%-19.3%) and the pooled prevalence of diagnosed tinnitus was 3.4% (95% CI, 2.1%-5.5%). The pooled incidence rate of any tinnitus was 1164 per 100 000 person-years (95% CI, 479-2828 per 100 000 person-years).


Conclusions and Relevance
Despite the substantial heterogeneity among studies, this comprehensive systematic review on the prevalence and incidence of tinnitus suggests that tinnitus affects more than 740 million adults globally and is perceived as a major problem by more than 120 million people, mostly aged 65 years or older. Health policy makers should consider the global burden of tinnitus, and greater effort should be devoted to boost research on tinnitus.

T he term tinnitus comes from the Latin word tinnire, which means "to ring."Individuals experiencing tinnitus report an unspecified acoustic sound like ringing, but also buzzing, clicking, pulsations, and other noises. 1 Tinnitus is considered a symptom of an underlying condition, rather than a disease, and it refers to the perception of sounds in the head or ears when no corresponding external sounds are present. 2,3A severe form of tinnitus is associated with hearing loss, thus impairing quality of life. 4,5oday there is no globally accepted categorization of tinnitus, although different attempts have been made. 6Moreover, evidence on the frequency (ie, prevalence and incidence) of tinnitus among the general population is still scant.The difficulties in obtaining adequate data are due to the multifactorial etiology of tinnitus, its associated disorders, the various characteristics of the symptoms, and the subjective nature of any assessment of tinnitus. 7he very few longitudinal studies on tinnitus hamper any accurate estimate of its incidence.Moreover, the prevalence of tinnitus, which is estimated as either point prevalence, period prevalence, or lifetime prevalence, 8 ranges widely, partly because of the lack of standardization in its assessment, illustrated in a systematic review in which McCormack et al 3 attempted to identify and collect data on the global prevalence of tinnitus.
Since that last review on the prevalence of tinnitus, the literature on tinnitus has increased by at least 30%.An update of the evidence, which also includes pediatric tinnitus, is now necessary.We conducted a systematic review to identify the relevant publications in the scientific literature on the frequency of tinnitus at a global level, using an original search method. 9

Methods
This systematic review and meta-analysis is based on 2 subsequent literature searches on the prevalence and incidence of tinnitus.The first search was an umbrella review: a systematic review to identify published meta-analyses, pooled analyses, and systematic reviews providing data on the prevalence or incidence of tinnitus.The second search was a traditional review of original publications: a systematic review of all original articles on the prevalence or incidence of tinnitus to update the results identified in the umbrella review.A review protocol was registered in advance on PROSPERO (registration number: CRD42021283684).The study followed the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guideline.

Umbrella Review
We conducted an umbrella review to systematically collect existing evidence on the prevalence and incidence of tinnitus.We searched in PubMed-MEDLINE and Embase for all systematic reviews or meta-analyses published from inception through November 19, 2021, that had the word tinnitus in the title (eTable 1 in the Supplement).We retrieved 310 reviews from PubMed and 346 from Embase.After checking for du-plicates using EndNote, version X7 (Clarivate), we excluded protocols, scoping reviews, case studies or animal model studies, and articles that were not in English.After applying our inclusion criteria (ie, reporting data on the prevalence or incidence of tinnitus), we excluded 369 reviews as not relevant, ending up with 15 publications.We added 1 study that we were aware of that followed our eligibility criteria but was not identified by our search string because it was not classified as a review.From each of these 16 relevant systematic reviews, we extracted the citations of all the original articles providing data on the prevalence and incidence of tinnitus, collecting 284 original studies in total.
We included only articles in English, based on samples representative of the general population, and with estimates specifically of tinnitus.Reports, letters to the editor, book chapters, conference proceedings, dissertations, and theses were not considered.We excluded studies based on patients or on subgroups of the population with selected lifestyle habits or other characteristics (eg, musicians or people regularly exposed to noise).No restrictions were applied regarding the date of publication, age, sex, and country.Two researchers (C.M.J. and M.S.) independently checked for eligibility.Any disagreement was resolved by discussion; in case of disagreement, a third reviewer (A.L.) helped to reach consensus.The umbrella review yielded 93 eligible original articles.

Traditional Review
We then conducted a traditional review to check any relevant articles in the literature that might not have been identified through the umbrella review.We searched articles in PubMed-MEDLINE and Embase published from inception through November 19, 2021, using a string that included a combination of the words tinnitus, prevalence, and incidence in the title.From 245 publications, we checked for duplicates and retrieved 154 unique references.After excluding articles identified in the umbrella review (n = 45), other duplicates (n = 15), and noneligible articles (n = 78), we obtained 16 new records.To these, we added 2 other references retrieved from other sources that we knew followed our eligibility criteria.All the articles excluded from both the umbrella and the traditional review, as well as the reasons for exclusion, are listed in eTable 2 in the Supplement.

Key Points
Question What is the global prevalence and incidence of tinnitus?Findings This systematic review and meta-analysis estimated that the annual incidence of tinnitus is approximately 1%, with 14% of adults experiencing any tinnitus and 2% experiencing a severe form of it.The prevalence of tinnitus did not differ by sex, but increased prevalence was associated with increasing age, with any tinnitus being present in 10% of young adults, 14% of middle-aged adults, and 24% of older adults.

Meaning
This study suggests that the global burden of tinnitus is large, similar to migraine and pain, and the lack of effective treatment options justifies a major investment in research in this area.

Data Extraction and Management
We used a standardized form in Excel 2016 (Microsoft Corp) to extract data from each article identified.Relevant information included first author, year of publication, journal, type of study, time and location, end point (prevalence and/or incidence), other information (country and sample size), population characteristics (sex and age group), and tinnitus definition.Data were blindly extracted by 2 independent reviewers (C.M.J. and M.S.).Any disagreement was resolved by discussion, or with the help of a third reviewer (A.L.).Each prevalence estimate was extracted and classified by age group: children (≥17 years), young adults (18-44 years), middle-aged adults (45-64 years), older adults (≥65 years), and all adults (≥18 years).
If, while extracting, we came across summary tables that gave additional relevant citations, these were evaluated using the same inclusion and exclusion criteria.This evaluation led to 2 additional eligible articles, yielding a final total of 113 eligible articles.Among these, 24 articles were not included in the extraction for meta-analysis because their results were already included in other more complete or more recent articles (eTable 3 in the Supplement).We extracted prevalence or incidence estimates from 89 articles.

Statistical Analysis
The pooled prevalence and incidence of any tinnitus and severe tinnitus were calculated overall for children, adolescents, and adults and separately by tinnitus definition (eTable 4 in the Supplement).For any tinnitus in adults, we identified 6 possible classes of definitions (A1-A6), and for severe tinnitus, we identified 5 possible classes (S1-S5).For children and adolescents, classes either had the word tinnitus in the question asked (any tinnitus or severe tinnitus) or had a phrase, such as "noises in your ears" (any noises or severe noises).Other possible definitions were chronic tinnitus or diagnosed tinnitus.
Pooled estimates were obtained using random-effects meta-analytic models to take account of the heterogeneity of the estimates.Heterogeneity among studies was controlled using the χ 2 test, and inconsistency was measured using the I 2 statistic, which represents the proportion of total variation associated with between-study variance, with higher values denoting a greater degree of heterogeneity.Stratified analyses by selected individual-level characteristics (eg, sex and age) and country-specific characteristics (eg, continent, gross domestic product [GDP], and latitude of the main city) were performed to detect possible sources of heterogeneity.The quality of the studies was not assessed because it was beyond the scope of meta-analyses on disease frequency.Because most prevalence and incidence estimates were provided without 95% CIs, we recalculated all the 95% CIs from the raw data given in the original articles.All P values were from 2-sided tests and results were deemed statistically significant at P < .05.
All statistical analyses were performed using the R Studio software, version 1.4.1717(R Group for Statistical Computing), particularly the "meta" and "metaphor" packages.To assess publication bias, we examined the funnel plots visually and applied the Egger test for funnel plot asymmetry.

Results
Among 767 publications (384 reviews, 284 identified original publications, 94 articles from the traditional review, and 5 articles known by the authors), 113 eligible articles published between 1972 and 2021 were identified.We extracted prevalence estimates from 83 articles and incidence estimates from 12 articles.eFigure 1 in the Supplement shows the flowchart of study selection.Details on country, age group, and tinnitus definition in the 89 eligible articles included in metaanalyses are summarized in eTable 5 in the Supplement.
Converting our pooled prevalence estimates to absolute numbers, we found that there were 749 million adults (95% A possible publication bias emerged for the prevalence of any tinnitus and for the prevalence of severe tinnitus (P < .001for the Egger test; eFigure 4 in the Supplement).For adults, the pooled prevalence of diagnosed tinnitus was 3.4% (95% CI, 2.1%-5.5%; 3 studies; I 2 = 99%), and the pooled prevalence of chronic tinnitus, defined as tinnitus occurring most or all of the time or persisting for months, was 9.8% (95% CI, 4.7%-19.3%;3 studies; I 2 = 99%).

Discussion
To our knowledge, this systematic review provides the most comprehensive and up-to-date evidence on the prevalence and incidence of tinnitus worldwide among adults and children or adolescents, summarizing estimates from 89 original studies.For the first time, we provide pooled estimates of  In >5 y of follow-up, 161 (75 female, 86 male) of 2970 participants without any tinnitus at baseline (5.4%) reported any new tinnitus; derived annual incidence rate, 1080 per 100 000 person-years.A total of 166 (76 female, 90 male) of 3134 participants without diagnosed tinnitus at baseline (5.3%) reported new diagnosed tinnitus; derived annual incidence rate 1040 per 100 000 person-years.
Gopinath et al, 14 Population-based study (Blue Mountains Hearing Study) (N = 1214), which is part of a larger cohort of people aged 49 y or older, with follow-up after 5 y In >5 y of follow-up, of 811 individuals without tinnitus at baseline, 156 new cases were found (19.2%).Derived annual incidence rate: 3847 per 100 000 person-years.Tinnitus at baseline persisted in 81.6% of participants (n = 346 of 424).A significant trend (P = .005)was observed for the 5-y incidence for age, with incidence of tinnitus decreasing with age.A significant age trend was observed for men only.There was no sex difference in the 5-y incidence of tinnitus.
Lee et al, 53

(Taiwan)
Retrospective population-based cohort study (N = 30 340) of people aged ≥20 y over 10 y There were 530 incident cases of diagnosed tinnitus.Derived annual incidence rate: 302 per 100 000 person-years.
Tinnitus was identified with ICD-9 code 388.3.Incidence rates were provided according to sex (271 per 100 000 person-years for men, 317 per 100 000 person-years for women) and age (195 per 100 000 person-years for those aged ≤49 y, 500 per 100 000 person-years for those aged 50-64 y, and 576 per 100 000 person-years for those aged ≥65 y).
Clinically significant tinnitus was defined as discharge from hospital with a primary diagnosis of tinnitus, or a primary care recording of tinnitus with subsequent related medical follow-up within 28 d.The incidence rate did not depend on sex, but increased with age, peaking at 11.4 per 10 000 person-years in those aged 60-69 y.The study provided the annual incidence rate for every year from 2002 to 2010, and 5-y and 10-y cumulative incidence by age at start of observation.

(US)
Population-based study (Beaver Dam Study) of adults (N = 2513) aged 48-92 y, without tinnitus at baseline, with a follow up after 5 y In >5 y of follow-up, 144 new cases of tinnitus were observed (67 in men and 77 in women), leading to a 5-y cumulative incidence of tinnitus of 5.7% (95% CI, 4.8%-6.6%).Derived annual incidence rate of 1140 per 100 000 person-years.

(continued)
tinnitus; based on our data, approximately 14% of the world population have experienced tinnitus, and more than 2% have severe tinnitus.The prevalence of tinnitus is similar for both sexes, and increases in prevalence are associated with increasing age.Heterogeneous estimates have been reported by the few studies that provide data on the incidence of tinnitus.The pooled annual incidence rate approaches 1%.The various cross-sectional studies providing data on the frequency of tinnitus used a wide variety of assessment methods. 3,58,59We therefore classified the questions about any tinnitus into 6 groups and about severe tinnitus into 5 groups.Despite the substantial heterogeneity of estimates in the classes of any tinnitus, we did not find statistically significant differences in its prevalence across classes.This finding suggests that, at least for any tinnitus, not all the variability is explained by different definitions, and other factors might explain the prevalence of any tinnitus better at the population level.
Concomitantly, our findings suggest that data on tinnitus among children or adolescents are more prone to different interpretations of the question used to assess tinnitus.One possible reason could be that children are more frequently asked about tinnitus without specifically mentioning the name of the symptom.Other researchers have suggested that children might report the presence of noise to please the interviewers. 60espite the increasing number of studies on the subject, tinnitus remains an unrecognized problem that is inadequately assessed in the pediatric population. 61e found differences in terms of any tinnitus and severe tinnitus in association with age, confirming increasing prevalence of the symptom with age. 7,10,11In particular, whereas the prevalence of any tinnitus among older adults was close to 2.5 times higher than among young adults, the prevalence of severe tinnitus among older adults was almost 20 times higher than among young adults.This finding suggests that tinnitus is a particular disorder of older people. 8he literature is not unanimous about whether there is any association between sex and tinnitus.McCormack and colleagues 3 generally reported a higher prevalence of any tinnitus among men than women, whereas Biswas and colleagues 12 found a higher prevalence of bothersome tinnitus among women than men.The latter is consistent with previous findings of an association between severe tinnitus and suicidal attempts among women but not among men. 4 In our comprehensive review, pooling findings from a vast scientific literature, we did not find any significant difference according to sex for either any tinnitus or severe tinnitus.
As previously noted, 8 information is scant on the differences in tinnitus prevalence among countries, and we were only partially able to fill the gap.In fact, Africa, Oceania, and South America are not well represented.We found only 2 studies on the prevalence of any tinnitus and 2 studies on the prevalence of severe tinnitus from Africa and South America combined, covering more than 1.7 billion people.This finding may, to some extent, be due to the fact that, by protocol, we did not include articles that were not in English.Pooled estimates for any tinnitus from the other continents were somehow similar-between 13% and 15%whereas differences were larger for pooled estimates for severe tinnitus-between 1.8% and 3.3%.
The incidence rates increased with age and were highest in the group aged 60-69 y (41.2 per 10 000 person-years; 95% CI, 40.7-41.7 per 10 000 person-years).Rates were similar for men (25.7 per 10 000 person-years; 95% CI, 25.0-26.4per 10 000 person-years) and women (24.7 per 10 000 person-years; 95% CI, 24.0-25.3per 10 000 person-years).Rates were also given by year (2000-2016).ated with broad tinnitus definitions, such as "any" tinnitus, it appears that rare variants are more associated with severe tinnitus. 65,66Thus, differences in population genetics could be associated with the large discrepancies in the prevalence of severe tinnitus, as for instance in South America.However, more efforts are needed to investigate the association of genetics with any tinnitus or severe tinnitus across different countries and continents.In a European survey, Biswas et al 12 found that the prevalence of tinnitus was greater in countries from the eastern European region than in western Europe, with Bulgaria reaching a prevalence for any tinnitus of 28.3% and Romania with a prevalence for severe tinnitus peaking at 4.2%.This finding is consistent with a greater prevalence of hearing loss among individuals in these countries, according to the Global Burden of Disease study. 67It is possible that less active work-related preventive measures against occupational noise exposure or limited access to rehabilitation for hearing loss by means of hearing aids may cause such disparities across Europe.9][70] Other risk factors could also underlie such differences in prevalence.However, only a handful of case-control and longitudinal studies have investigated the potential causal relationship to tinnitus, most of which focus on hearing-related conditions. 14,15Thus, a comprehensive picture of the association of nonauditory etiologies with any tinnitus or severe tinnitus is required.
Our results do indicate that differences arise when using multiple definitions to assess tinnitus.For future research, therefore, we recommend using a standardized questionnaire for assessing the prevalence of tinnitus, to make better comparisons between different surveys, identifying more solid estimates of tinnitus in various countries worldwide.We acknowledge, however, that no single question can address the multidimensional properties of tinnitus that are critical for its assessment (duration [acute or chronic], temporality [intermittent or constant], and severity [negligible or impactful]).Thus, our suggestion is to systematically use the questions given by a consortium of experts available in multiple languages. 58or children specifically, a large difference was clear between questionnaires that mentioned the term tinnitus and those that did not; we conclude that future surveys addressing children and adolescents must state clearly the name of the disorder in their questions-with an explanation-as the high prevalence of tinnitus might be a result of the participants not recognizing the extraordinary nature of the symptom being investigated.
An association has been hypothesized between socioeconomic status and tinnitus. 16,54Although with all the limitations of an ecological analysis, 71,72 we found no association between per-capita GDP and tinnitus prevalence.
Tinnitus has been reported to have a seasonal pattern, where it is worse in the winter than in the summer. 73Thus, the hours of sunlight per day or certain temperatures might be associated with the onset or severity of tinnitus.Countries with their main city at an intermediate latitude (40°-51°) had the lowest prevalence and the lowest severity of tinnitus.Future analytical studies should investigate this issue in more detail.
In this meta-analysis, we defined as eligible only studies based on samples representative of the general population, excluding subgroups of the population exposed to selected risk factors, such as veterans and musicians.In these 2 particular populations, the prevalence of tinnitus was reported with a point estimate of 31% among veterans 74 and 26% among musicians, 75 much higher than among the general adult population.These 2 subpopulations might therefore be targeted for specific interventions to prevent or limit exposure to noise and, consequently, to reduce tinnitus and other hearing conditions.
There is a paucity of articles on the incidence of tinnitus: of 113 eligible articles, only 12 provided data on the incidence of tinnitus, although many cohorts were available with tinnitus assessed at follow-up.The incidence rates differed by up to 2 orders of magnitude in various studies.Although estimates stratified by sex are frequently provided, information is limited on incident cases by age group.

Limitations and Strengths
This study has some limitations, including the classification of tinnitus into 6 groups of questions for any tinnitus (A1-A6) and 5 groups for severe tinnitus (S1-S5).Although inspired by the 8 different categories for tinnitus identified by McCormack et al, 3 our classification has not been validated and is therefore subject to the interpretation of the researchers who used it.Moreover, we cannot exclude a possible publication bias regarding the prevalence of both any tinnitus and severe tinnitus.
The strengths of the study include the original method used to identify relevant articles, which involves an umbrella review as well as a traditional review. 97][78] Thus, we were able to include almost twice the number of articles included in the most comprehensive review of the literature published before the present one, 3 including, in our opinion, at least 11 articles that could have been retrieved by McCormack and colleagues 3 but were not in that review.Thus, to our knowledge, this is the most comprehensive review conducted to date because it considers a larger publication period (between 1972 and 2021) and is not limited to adults but also includes children and adolescents.

Conclusions
To our knowledge, this is the first meta-analysis on the frequency of tinnitus.Generalizing our estimates to the whole global population, one can infer that more than 740 million people experience tinnitus and more than 120 million people worldwide have a severe form of tinnitus.Such estimates place tinnitus at an order of magnitude similar to the leading causes of years lived with disability, namely, hearing loss, followed by migraine, low back pain, and neck pain. 67Health authori-ties and research institutions, such as the Global Burden of Disease, should consider this prevalence and play a leading role in funding, ultimately to boost research on tinnitus and improve the care and the lives of patients with tinnitus.

Figure 1 .
Figure 1.Forest Plot of Any Tinnitus Prevalence Among Adults, by Different Definition Classes of Any Tinnitus 2017 (Taiwan)Population-based cohort study (Taiwan National Health Insurance Research Database) of adults (N = 741 720) aged ≥18 y, followed for 10 y There were 1173 incident cases of diagnosed tinnitus (615 in men and 558 in women).Derived annual incidence rate was 78 per 100 000 person-years.Tinnitus was identified with ICD-9 code 388.3.Incidence rates were provided by sex.Stohler et al, 56 2019 (UK) Two separate study populations: first, retrospective cohort study of adult patients from 2000 and 2016 (N = 109 783); second, case-control study.

Table 1 .
Pooled Prevalence of Any Tinnitus in Adults, Overall and in Strata of Selected Individual-Level or Country-Specific Characteristics jamaneurology.com (Reprinted) JAMA Neurology Published online August 8, 2022 E5 © 2022 American Medical Association.All rights reserved.Downloaded From: https://jamanetwork.com/ by a Maastricht University User on 09/08/2022 CI, 655-858 million adults) worldwide with any tinnitus and 120 million adults (95% CI, 88-177 million adults) with severe Figure 2. Forest Plot of Pooled Prevalence, by Different Definitions of Severe Tinnitus in Adults

Table 2 .
Pooled Prevalence of Severe Tinnitus Among Adults, Overall and in Strata of Selected Individual-Level or Country-Specific Characteristics jamaneurology.com (Reprinted) JAMA Neurology Published online August 8, 2022 E7 © 2022 American Medical Association.All rights reserved.Downloaded From: https://jamanetwork.com/ by a Maastricht University User on 09/08/2022

Table 3 .
Incidence of Any Tinnitus Among Adults in the 12 Studies Included

Table 3 .
Incidence of Any Tinnitus Among Adults in the 12 Studies Included (continued) Abbreviations: GP, general practitioner; ICD-9, International Classification of Diseases, Ninth Revision.