Discussion

In our multinational cohort of 478 patients with chronic otitis media, 58.9 per cent reported suffering from tinnitus of at least moderate severity. This is consistent with other studies utilising the Chronic Otitis Media Questionnaire-12 questionnaire to assess disease-specific health-related QoL, with mean item-7 scores ranging from 1.32 to 2.86 (minor-moderate tinnitus severity).^{Reference Phillips, Haggard and Yung17,Reference Fonseca, Ramos, Balsalobre, Freitas, Phillips and Yung22–Reference Celis-Aguilar, Medina-Cabrera, Torrontegui-Zazueta, Núñez-Millán, García-Valle and Ochoa-Miranda25} In three studies evaluating chronic otitis media patients following middle-ear surgery, a large proportion of patients reported tinnitus pre-operatively (42.7 to 47.6 per cent).^{Reference Kim, Park, Kim, Lee, Park and Yeo13,Reference Baba, Yagi and Fujikura26,Reference Guo, Sun and Wang27} It is evident that tinnitus is a common burden in patients suffering from chronic otitis media alongside hearing impairment and ear discharge. Therefore, in the present study, we sought to determine whether demographic and disease-related factors influence tinnitus severity.

Several studies have shown that subjective tinnitus improves following middle-ear surgery, and this appears to coincide with hearing restoration.^{Reference Kim, Park, Kim, Lee, Park and Yeo13,Reference Guo, Sun and Wang27–Reference Habesoglu, Habesoglu, Karatas, Tosun, Gursel and Oysu30} Guo et al.^{Reference Guo, Sun and Wang27} reported that improved low-frequency air conduction hearing was an important factor in post-operative reduction in tinnitus severity, whereas Kim et al.^{Reference Kim, Jang, Kim, Seong, Kang and Cho28} proposed that the degree of pre-operative air-bone gap may influence tinnitus outcome following tympanoplasty. The existing literature concerning the possible relationship between improvement in tinnitus severity and audiological outcome in patients with chronic otitis media is limited to small cohort studies within single medical centres, often performed retrospectively.

In comparison, based on a cross-sectional evaluation of tinnitus in a large cohort across nine otology centres, the present study found no significant effect of overall hearing disability and worse-ear hearing on tinnitus severity. Instead, the presence of cholesteatoma and ear discharge was found to be inversely associated with tinnitus severity; that is, chronic otitis media patients with cholesteatoma or discharging ear(s) appear to have less severe tinnitus, which has not previously been reported. In chronic otitis media, both cholesteatoma and persistent otorrhoea through a perforated tympanic membrane indicate an ongoing inflammatory response within the middle-ear cleft. Over time, this can cause ossicular erosion and disruption of the ossicular chain.^{Reference Mittal, Lisi, Gerring, Mittal, Mathee and Narasimhan31} Tympanoplasty with or without mastoidectomy is the main curative treatment strategy. The primary objectives of middle-ear surgery are complete removal of disease, provision of a dry ear, and preserved or improved hearing levels. From the patients’ perspective, alongside subjective hearing improvement and relief from otorrhoea, improvement or resolution of tinnitus may influence post-operative patient satisfaction.^{Reference Erkorkmaz, Yilmaz, Guven and Kaymaz32,Reference Phillips, Haggard, Spencer and Yung33} It is possible that when patients with chronic otitis media have a discharging ear, cholesteatoma or both, there is an expectation that the disease and any associated symptoms are treatable. We propose that such expectations prior to definitive treatment may influence patients’ perception of the disease and reduce the burden of individual symptoms such as tinnitus. In contrast, patients presenting with tinnitus as their primary otological complaint are more likely to be annoyed or distressed by it.^{Reference Stouffer and Tyler34}

In large epidemiological studies, tinnitus severity is directly correlated with increasing age.^{Reference Kim, Lee, An, Sim, Park and Kim19,Reference Bhatt, Lin and Bhattacharyya35} In contrast, our data suggest that tinnitus severity may decrease with age, although this association is not significant in the adjusted model. A strong association between poorer disease-specific health-related QoL and tinnitus severity was expected since both variables were determined using component scores of the Chronic Otitis Media Questionnaire-12, which has high internal consistency as demonstrated in the original study and other validation studies.^{Reference Phillips, Yung, Nunney, Doruk, Kara and Kong21,Reference Quaranta, De Robertis, Milella, Pontrelli, Greco and Fiorella23,Reference Doruk, Çelik, Kara, Polat, Güldiken and Orhan24,Reference Otoya-Tono, Pérez-Herrera, Peñaranda, Moreno-López, Sánchez-Pedraza and García36}

This study used a single-item measure within the Chronic Otitis Media Questionnaire-12 to determine tinnitus severity in patients with chronic otitis media, focusing on tinnitus-related inconvenience. However, the consensus is that perceived tinnitus severity has multiple dimensions, not only degree of annoyance but also duration, signal intensity or loudness, and associated disability (e.g. sleep disruption).^{Reference McCormack, Edmondson-Jones, Somerset and Hall1,Reference Hiller and Goebel20} Several validated tinnitus-specific measures exist to assess the level of tinnitus-related distress with varying emphasis on different domains of quality-of-life.^{Reference Kennedy, Wilson and Stephens37} A single-item measure is therefore unlikely to capture the breadth or complexity of individual experience.^{Reference McCormack, Edmondson-Jones, Somerset and Hall1} Nevertheless, a single question can serve as a useful screening tool, identifying at-risk patients for whom further assessment or characterisation of their tinnitus is necessary. An expectation for all patients with chronic otitis media to complete a battery of symptom-specific instruments in addition to a disease-specific health-related QoL measure (such as the Chronic Otitis Media Questionnaire-12) would only lead to a higher response burden and information overload on the part of the clinician.

One limitation of the present study is that we could not assess the relationship between high frequency hearing loss and tinnitus in our cohort, since pure-tone audiometry was performed between 0.5 kHz and 4 kHz for all study participants. Elevation of hearing threshold levels in high and extended high frequencies (more than 8 kHz) is commonly detected in patients with chronic otitis media and has been noted to persist even after resolution of acute otitis media,^{Reference Cordeiro, Monsanto, Kasemodel, Gondra and Penido38–Reference Park, Park, Kim and Park41} presumably because of penetration of inflammatory mediators and bacterial toxins into the inner ear via the round window membrane.^{Reference Schachern, Paparella, Goycoolea, Duvall and Choo42,Reference Engel, Blatz, Kellner, Palmer, Weller and Bhadki43} Of note, Cordeiro et al.^{Reference Cordeiro, Monsanto, Kasemodel, Gondra and Penido38} suggested that the presence of persistent tinnitus six months after an episode of acute otitis media may be associated with worse extended high frequency thresholds. Since pure-tone audiometry was limited to air conduction testing only, we were unable to report the air–bone gap. Chronic otitis media patients with predominantly conductive hearing loss are not distinguishable from patients with underlying and perhaps long-standing sensorineural hearing loss, which may represent distinct cohorts of patients with different experiences of tinnitus.

Due to the brevity of tinnitus assessment, we were unable to characterise whether tinnitus was ipsilateral (localising to the ear affected with chronic otitis media), contralateral or bilateral. It is possible that patients with contralateral or bilateral tinnitus may have chronic otitis media plus sensorineural or mixed hearing loss. Patients with bilateral tinnitus are also more likely to report a greater degree of handicap and psychological burden.^{Reference Yang, Jung, Kim, Byun, Park and Yeo44} Further characterisation of tinnitus (e.g. localisation, loudness, pitch) in chronic otitis media sufferers in relation to disease-specific health-related QoL should be explored in future studies.

Another limitation is that we were unable to investigate a wide array of demographic, behavioural and socioeconomic characteristics or patient co-morbidities since the number of variables included in the logistic regression analyses was restricted to data available in the dataset from the original study.^{Reference Phillips, Yung, Nunney, Doruk, Kara and Kong21} For example, we were unable to explore whether there were gender differences in severity of tinnitus since this characteristic was not reported for UK study participants. Nevertheless, this is the first study to examine potential risk factors influencing tinnitus severity in a large, multinational cohort of chronic otitis media patients with prior unknown tinnitus status. The resulting broad range of tinnitus severity within the study population is reflective of real-world clinical practice, thus increasing the robustness of our findings.