Study design and subjects

We designed a nationwide cross-sectional survey of first-year middle (12–13 years old) and high (15–16 years old) school students based on random sampling of single-gender, coeducational, general, and vocational schools in eight metropolitan and eight suburban regions. The sampling protocol was described previously. [11] After a pilot study, which showed 50% agreement on the part of schools to participate in the survey, we listed 248 from 5,587 schools in Korea. The number of participants in each school was 25, based on the average number of students per class. The target class was carefully selected to be representative of the characteristics of the school. Two types of written informed consent were provided individually, to the students and their parents. Only the students their parents, who had both agreed to participate in the survey, were evaluated. Among the participants, those who had conductive hearing loss and/or any abnormalities in an otologic examination were excluded. The study was approved by the Institutional Review Board of Seoul National University Hospital (No. 1604-086-755). It was conducted in accordance with the principles that have their origin in the Declaration of Helsinki.

Data collection

As tinnitus is a subjective symptom, no objective instruments are available for its measurement. Therefore, questionnaires were chosen with reference to previously reported epidemiological research on tinnitus. [12, 13] The questionnaires provided to the participants included two questions regarding tinnitus: (1) the presence or absence of tinnitus and (2) the degree of tinnitus. Those participants who responded that they had ever heard a sound were instructed to answer the second question about the degree of tinnitus, the possible responses to which were: “No problem”, “I am annoyed and bothered”, and “I find it hard to sleep.” The participants were asked to choose only one of the responses. Based on the responses, we defined the following three groups: (1) “normal” for students without tinnitus, (2) “mild tinnitus” for students who had tinnitus, and chose “no problem” as the severity, and (3) “severe tinnitus” for students who were annoyed, bothered, or had sleep problems due to their tinnitus.

A part of the questionnaire inquired about the risk of LNE. The usage of personal listening devices, presence or absence of the experience of conflict with others due to the volume of a listening device, and the time and frequency of usage of leisure facilities (such as local area network (LAN) gaming centers, karaoke rooms, and concert auditoriums).

The participants also responded to four questions about auditory behavior. They were about their hearing in a noisy environment (“Have difficulty hearing or understanding in a noisy environment”), sound localization (“Have trouble finding the sound direction”), verbal working memory (“Have difficulty memorizing by listening to verbal orders”), and fatigue (“Easily feel tired”). The participants responded to the questions by assigning a score of 1 (worst) to 5 (best).

The subjects self-rated their academic performance to explore the association between educational attainment and tinnitus. The subjects responded to the following question: “Please rate your grades for the past year” using a five-point Likert scale (1, highest; 2, high; 3, middle; 4, low; and 5, lowest).

The subjects’ health status, medical history of themselves and their families, alcohol consumption, and smoking were also assessed. The questionnaire was based on Forms 1–2 and 1–3 of the Enforcement Rule of Health Screening at School, according to the School Health Act of South Korea. The health condition questionnaire inquired about general symptoms, respiratory system symptoms, circulatory system symptoms, digestive system symptoms, mental health, hematopoietic system symptoms, and about lifestyle factors such as diet, personal hygiene, exercise, safety, Internet use, home and school life, drug use, and sexuality; the possible responses were yes or no.

Audiological evaluation

PTA was performed by audiometer (AD299b, Interacoustics, Assen, Denmark) in a sound-proof booth in a mobile vehicle. Four experienced audiologists conducted PTA testing at 0.5, 1, 2, 3, 4, 6, and 8 kHz. The hearing threshold was determined as the lowest sound level that was responded to correctly for 50% of stimuli. For frequencies with hearing impairment of ≥ 25 dB, the bone conduction threshold was evaluated to rule out the possibility of conductive hearing loss. High frequencies were defined as 3, 4, 6, and 8 kHz and speech frequencies were defined as 0.5, 1, and 2 kHz. Before the PTA, an otologic examination was performed to remove materials (such as cerumen and foreign bodies) that could affect hearing.

Statistical analyses

We estimated the prevalence of tinnitus among Korean adolescents. The unweighted frequencies and weighted prevalences were used because of the multi-stage complex sampling design of the study. According to characteristics of participants, the prevalence of any tinnitus (mild or tinnitus) and severe tinnitus was calculated, and univariate logistic regression analyses were performed. We also estimated the arithmetic mean and standard error of the pure-tone threshold of each frequency, the pure-tone averages of speech and high frequencies, according to the severity of tinnitus. Differences among means were tested by analysis of variance. A multivariate logistic regression was performed to investigate the association of LNE with severe tinnitus, with adjustment variables for the following potential confounders: age, gender, household income, obesity, alcohol consumption, history of otitis, and history of sinusitis with reference to the results of our analyses and a previous report. [14] Finally, the mean values of HP (hearing in a noisy environment, sound localization, verbal working memory, and fatigue) were compared according to tinnitus severity. Multivariate linear regression analyses were performed to assess the differences in mean values by tinnitus severity with adjustment for potential confounders. Additionally, to investigate the associations between academic performance and hearing status, the mean subjective academic performance scores and 95% confidence intervals were calculated and subjected to multivariate linear regression analysis according to tinnitus status. Moreover, we investigated the prevalence of tinnitus according to the responses to the physical health, mental health, and lifestyle items in the questionnaire by chi-squared test with Bonferroni correction for multiple comparisons (corrected significance level, P < 0.0009). Statistical analyses were performed using the SAS SURVEYFREQ, SAS SURVEYLOGISTIC, SAS SURVEYREG, and SAS FREQ procedures (ver. 9.4; SAS Institute, Inc., Cary, NC), and statistical significance was defined as a two-sided P ≤ 0.05.

Prevalence of tinnitus among adolescents

The prevalence of tinnitus among South Korean adolescents is 45.6% and that of severe tinnitus is 9.1%. It is more common in older adolescents and girls. The prevalence of tinnitus among adolescents differs according to the definition, study population, and method of evaluation. [3] The prevalence of tinnitus among adolescents has been reported to be 6.5% [20] to 43.9% [11]. The prevalence of severe tinnitus ranges from 0.6% [15] to 15.72% [16]. The prevalence of transient tinnitus is 39.7% [21] to 73.5% [22]. Prior studies applied different definitions of tinnitus severity, e.g., bothersome, uncomfortable, worried or concerned, and some used a yes or no response. [23, 24] Here, we used three tinnitus severity categories: no, any, and severe. Severe tinnitus was defined as that which disturbs sleep, a definition that is simple and reliable. [16] Use of the tinnitus handicap inventory or VAS for tinnitus may have yielded more comprehensive data on tinnitus severity. [25]

The prevalence of tinnitus increases with age. [5, 15, 26] However among adolescents, those in their mid-teens have the highest incidence of tinnitus. [23] In this study, older adolescents had a higher prevalence of tinnitus, possibly due to the higher level of stress in high school; however, our use of only two age groups prevents the drawing of any firm conclusions. Among the high school, but not the middle school, students, tinnitus was correlated with academic attainment. The prevalence of tinnitus was higher in females than males, as reported previously. [3, 18, 27] This may be due to the greater tendency of girls to describe symptoms [28] and their more frequent generation of spontaneous otoacoustic emissions [29]. A history of ear infection is reportedly a risk factor for tinnitus and an association between tinnitus and a low household income has been described; [18] our findings were in agreement with these reports and we found that the same factors were associated with hearing loss. [11]

Factors associated with tinnitus among adolescents

Tinnitus is associated with LNE at karaoke facilities and LAN gaming centers, as well as with a poor HP, but not with hearing loss. Hearing loss is often associated with tinnitus, [6] but not in adolescents. [4, 18] Our data also showed no relationship between tinnitus and hearing loss up to 8 kHz. Noise exposure is a risk factor for tinnitus, but does not induce hearing loss. [18, 22, 27] We found that tinnitus was associated with usage of LAN gaming centers and karaoke facilities, which are noisy leisure activities that many Korean adolescents enjoy. [11] Williams et al. also reported that experience of tinnitus is not associated with hearing loss as detected by PTA or otoacoustic emissions (OAEs) in young Australians. [30] However, there was correlation between the experience of tinnitus and the frequency of experience of tinnitus with cumulative life-time noise exposure. [30] An elevated noise level can damage the synapse of the auditory nerve without damaging the outer hair cells. [31] Such damage or hearing loss could induce neuroplasticity in the auditory system and lead to the development of tinnitus. [32] Some patients have a normal hearing threshold by conventional PTA but decreased HP in noisy situations, temporal resolution, and wave I potential in the auditory brain stem response. [9, 32] This is known as hidden hearing loss or cochlear synatopathy. [31, 33] In this study, adolescent patients with tinnitus showed poor HP (poor localization, hearing difficulty in noisy situations, decreased memory for verbal orders, and susceptibility to fatigue). These findings suggest the existence of hidden hearing loss among adolescents. Ultra-high-frequency PTA or other hidden hearing loss tests (e.g., hearing-in-noise test and auditory brain stem response) may enable detection of hearing deterioration in adolescents. [34] Childhood tinnitus patients are more likely report to deterioration in their hearing ability over time. [4] Patients with tinnitus with normal hearing should wear appropriate ear protection in noisy situations. [27, 30]

Comorbid conditions of tinnitus among adolescents

Adolescents with tinnitus have a variety of somatization symptoms and frequently smoke cigarettes and consume alcohol. Adolescent patients with tinnitus more often complain about physical health issues, e.g., severe headache, severe menstrual cramping, throbbing pain in the body, sore throat, digestive problems, and bruising. Attentional mechanism in tinnitus could make this difference. [35] Tinnitus may be associated with somatization and somatoform disorder. [36, 37] Because some conditions, e.g., pain and headache, have similar pathophysiologic mechanisms they could be symptoms of the same disease. [38, 39]

Adolescents are vulnerable to psycho-emotional factors, which can have detrimental effects on emotional development. [40] Adolescent patients with tinnitus often have mental health problems, such as depression, anxiety, suicidal thoughts, and sleep disturbance. [8, 41, 42] The severity of tinnitus is correlated with psychiatric disorders and the severity of depression and anxiety. [8, 43] Tinnitus can induce psychological symptoms [36], but the causality of the relationship has not been established. In this study, a large proportion of the adolescent patients with tinnitus had a history of bullying or isolation and of carrying a weapon (e.g., knife or club). In addition, the adolescent patients with tinnitus have a high prevalence of smoking, alcohol drinking, and substance abuse; [5, 18, 19, 44] this is in agreement with our findings. Adolescent are vulnerable to substance abuse, which is common in those with psychiatric comorbidities. [45] Therefore, adolescents with tinnitus should receive appropriate counselling to improve their mental health and prevent substance abuse.

Limitations

The cross-sectional design of this study prevented assessment of the causality of the associations identified. Smoking and alcohol consumption are reportedly causes of tinnitus. [5, 19, 44] A longitudinal follow up study is needed to assess the causality of the associations and the long-term effect of tinnitus and hearing change in adolescents. This study enrolled students attending middle or high schools. In South Korea, attending middle school is mandated by law and the rate of enrollment in high school was about 93% in 2016. [46] We did not investigate the laterality of tinnitus and objective tinnitus. In addition, testing of tinnitus loudness and pitch matching would have provided further information on tinnitus in adolescents.