Tinnitus and suicidal ideation

The pooled data in this study represents over 7000 people with tinnitus across four countries, giving an overall prevalence for suicidal ideation of 20.6 per cent (95 per cent CI, 10.8–30.3 per cent). The cross-national 12-month period prevalence and lifetime prevalence for suicidal ideation have been estimated to be 2 per cent and 9 per cent, respectively, for the general population.^{Reference Nock, Borges, Bromet, Alonso, Angermeyer and Beautrais37,Reference Borges, Nock, Haro Abad, Hwang, Sampson and Alonso38} This would suggest that suicidal ideation may be more prevalent in tinnitus sufferers. The control populations of two included studies also appear to support this.^{Reference Han, Ko, Shin, Lee, Choi and Kwon29,Reference Vogel, Van de Looij-Jansen, Mieloo, Burdorf and de Waart36}

Chronic medical illness is an independent risk factor for suicidal ideation, especially in the elderly population.^{Reference Druss and Pincus39,Reference Harris and Barraclough40} The pooled prevalence of suicidal ideation in tinnitus sufferers (20.6 per cent) is comparable to that of people with traumatic brain injury (3–33 per cent),^{Reference Simpson and Tate41} chronic pain (13.5 per cent)^{Reference Calati, Laglaoui Bakhiyi, Artero, Ilgen and Courtet42,Reference Stubbs43} and general medical conditions including cancer (0.7–46.3 per cent),^{Reference Kolva, Hoffecker and Cox-Martin44} Parkinson's disease (30 per cent),^{Reference Shepard, Perepezko, Broen, Hinkle, Butala and Mills45} diabetes mellitus (16.2 per cent),^{Reference Elamoshy, Bird, Thorpe and Moraros46} rheumatic diseases (26 per cent),^{Reference Li, Yang, Dong, Li, Cui and Zhao47} coronary artery disease (14 per cent)^{Reference Lehmann, Kohlmann, Gierk, Murray and Löwe48} and insomnia (25–32 per cent).^{Reference Drapeau and Nadorff49} A greater disease burden (more than one medical illness) is also shown to increase the odds of suicidal ideation; this relationship persists after adjusting for moderating factors such as heavy alcohol use and depression.^{Reference Druss and Pincus39} Comparatively, we found a higher prevalence of suicidal ideation with increasing tinnitus severity.^{Reference Aazh and Moore32–Reference Aazh, Landgrebe and Danesh34} Furthermore, the analysis in Han et al. indicated that the association between tinnitus and suicidal ideation persists after adjusting for concomitant depression.^{Reference Han, Ko, Shin, Lee, Choi and Kwon29} Further work assessing the relationship between tinnitus severity and suicidal ideation should attempt to control for comorbid depression and other moderating factors.

We are able to comment on the gender profile of suicidal ideation in the tinnitus-affected population in three of six included studies. In the study by Fox-Thomas, it was possible to derive prevalence figures by gender based on the summary data reported in the study: prevalence of suicidal ideation in women with tinnitus was 18.1 per cent compared with 14.8 per cent in men with tinnitus.^{Reference Fox-Thomas33} The adjusted analysis in Han et al. identified female gender as a risk factor for tinnitus, depressive mood and suicidal ideation in Korean adults.^{Reference Han, Ko, Shin, Lee, Choi and Kwon29} However, Aazh and Moore did not report significant gender differences in suicidal ideation.^{Reference Aazh and Moore32} Even in adult general populations, the association between gender and suicidal ideation is not well characterised. One cross-national study demonstrated higher rates of suicidal ideation in women compared with men, but the analyses did not adjust for sociodemographic factors and psychiatric illness.^{Reference Weissman, Bland, Canino, Greenwald, Hwu and Joyce50} In contrast, there is a recognised gender difference in lethality of suicidal behaviour; men have a lower rate of suicide attempts, but a higher rate of suicide completion compared with women.^{Reference Mergl, Koburger, Heinrichs, Székely, Tóth and Coyne51} It is also important to consider the prevalence of suicidal ideation across different age groups. Aazh and Moore found a small negative correlation between suicidal ideation and age in patients recruited from a specialist tinnitus clinic (r = −0.19; p = 0.02),^{Reference Aazh and Moore32} suggesting that suicidal ideation may be more frequent in younger tinnitus sufferers. This concurs with the findings of a large national cross-sectional study of suicidal ideation across different age groups.^{Reference Dennis, Baillon, Brugha, Lindesay, Stewart and Meltzer52}

The variation in reported prevalence of suicidal ideation among included studies is likely because of methodological factors, such as population setting, demographic characteristics and criteria for suicidal ideation assessment.

Firstly, there was a high level of heterogeneity in population characteristics. As an example, Vogel et al.^{Reference Vogel, Van de Looij-Jansen, Mieloo, Burdorf and de Waart36} reported the highest prevalence for suicidal ideation (31 per cent), but participants were limited to Dutch students (mean age of 18 years) from low educational backgrounds. The population also consisted of patients with chronic tinnitus as well as other permanent hearing-related symptoms (muffled sounds, distortion, hyperacusis or hearing loss). The complexity of presentation hinders generalisability of the data because degree of hearing loss is a potential moderator of tinnitus distress,^{Reference Andersson and Westin53} which the original study has not accounted for.^{Reference Vogel, Van de Looij-Jansen, Mieloo, Burdorf and de Waart36} In addition, three included studies recruited participants from specialist tinnitus clinics.^{Reference Aazh and Moore32–Reference Aazh, Landgrebe and Danesh34} This sample may not be representative of tinnitus sufferers in the general population because such participants have sought help for their symptoms, which implies a pre-existing degree of distress. Furthermore, the study sample in Tyler and Baker was self-selected because it consisted of members of a tinnitus self-help association who responded to a mailed questionnaire.^{Reference Tyler and Baker35} This form of sampling is susceptible to voluntary response bias from people with a greater degree of tinnitus-related distress taking part in the study. Thus, the prevalence of suicidal ideation within each of these studies (6.9–16 per cent) may be an overestimation. Interestingly, the study by Han et al. which used data from the Korean National Health and Nutrition Examination Survey dataset to obtain a representative South Korean population sample reported a higher prevalence of 21.6 per cent,^{Reference Han, Ko, Shin, Lee, Choi and Kwon29} although the general prevalence of suicidal ideation in South Korea appears higher than in other developed countries.^{Reference Choi and Lee54–Reference Olfson, Weissman, Leon, Sheehan and Farber57}

Secondly, the variation in prevalence data may be related to differences in methodology of suicidal ideation assessment. The lowest prevalence (6.9 per cent) was reported by Tyler and Baker,^{Reference Tyler and Baker35} who used an open-ended questionnaire to elucidate negative effects of tinnitus. Although closed questioning (as commonly encountered in self-report questionnaires) is potentially leading, individuals are likely to need prompting to report distressing symptoms. For studies that asked participants whether they had experienced thoughts of suicide within the previous year,^{Reference Han, Ko, Shin, Lee, Choi and Kwon29,Reference Vogel, Van de Looij-Jansen, Mieloo, Burdorf and de Waart36} the reported prevalence was greater than that reported in studies specifying a time interval of week(s).^{Reference Aazh and Moore32–Reference Aazh, Landgrebe and Danesh34} A shorter time period may not identify people with less frequent or episodic, and possibly less severe, thoughts of suicide. This highlights the importance of mandating a single, specific period interval when screening patients in clinical and research settings. Another consideration is that simplified screening tools (using a single question as criteria for determining suicidality) are unlikely to capture the complexity and spectrum of suicidal ideation (passive and active), which represents a broad continuum of distress.

Implications for clinical practice

Recent guidelines recommend the use of tinnitus questionnaires, such as the Tinnitus Functional Index and Tinnitus Questionnaire or mini-Tinnitus Questionnaire, to assess the impact of tinnitus and psychological effects.⁵⁸ For adult patients with distressing tinnitus, psychological therapies should be offered in a stepped approach.⁵⁸ Routine use of antidepressant medication in managing tinnitus is not evidenced.^{Reference Baldo, Doree, Molin, McFerran and Cecco59} Although high-risk patients may benefit from antidepressants, this is outside the scope of practice of an otolaryngologist and should only be considered following guidance from a mental health professional.

Our findings suggest that suicidal ideation may be elevated in the tinnitus-affected population, specifically for individuals with a higher degree of tinnitus-related distress. Clinicians should therefore be assessing suicidality as standard. A history of suicidal ideation in the absence of a plan or attempt does not increase the risk of future suicidal behaviour.^{Reference Borges, Angst, Nock, Ruscio and Kessler60} However, if suicidal ideation is present, formal suicide risk assessment and appropriate referral to mental health services is important. Clinicians may feel they have insufficient expertise to make such an enquiry or risk exacerbating negative mood states; however, the general consensus is that asking about suicidality does not increase subsequent suicidal thoughts or behaviour.^{Reference Mathias, Furr, Sheftall, Hill-Kapturczak, Crum and Dougherty61} ENT and audiology services may require additional training and support from mental health services in risk assessment and initial management.

Study limitations and future work

We aimed to conduct a comprehensive review through an inclusive search query. However, the existing literature forming the data for this study has significant issues, meaning that the quantitative analysis of the dataset must be interpreted cautiously. These limitations can be summarised as follows: lack of literature in this topic with one study^{Reference Han, Ko, Shin, Lee, Choi and Kwon29} driving the weighted meta-analysis findings; between-study heterogeneity; sampling and possible response biases; the lack of a control (no tinnitus) population in four of six studies; variable study quality; and the fact that the majority of studies did not control for known moderating factors when assessing the relationship between tinnitus and suicidal ideation. However, by presenting a careful qualitative synthesis, we hope to draw attention to this important topic as well as promote understanding of psychological sequelae of tinnitus among the tinnitus research community and members of the broad multidisciplinary team involved in patient care.

All of the included studies rely on single-item measures for assessing suicidal ideation. Clinical interviews, although time-consuming, are effective in discriminating patients and more reflective of routine clinical practice, whereas single items from validated measures or custom self-report questionnaires are likely to result in inflation of prevalence data.^{Reference Walker, Hansen, Hodges, Thekkumpurath, O'Connor and Sharma62} Further study should include the use of full (multi-item) validated questionnaires, of which there are many, and clinical interviews. Similarly, despite the existence of validated patient-reported tools relating to impact or severity of tinnitus,^{Reference Hall, Haider, Szczepek, Lau, Rabau and Jones-Diette63} only three studies incorporated such instruments in tinnitus evaluation. These tools are helpful in distinguishing clinically significant tinnitus from patients who do not find their tinnitus bothersome. Our findings support a potential association between tinnitus severity and suicidal ideation. Future research within routine clinical practice, possibly evaluating a combination of validated tools or development of a new tool, will therefore be of value.

We are unable to comment further on the underlying association and potential moderating or mediating factors between tinnitus severity and suicidal ideation because of a lack of data on temporality of symptom onset (an inherent weakness of cross-sectional research) and limited data on the co-existence of psychiatric illness. Furthermore, this study only assessed suicidal ideation. There is a growing body of literature examining the factors that may increase the risk of transition from suicidal ideation to suicidal behaviour, including substance misuse, severity of comorbid anxiety or depression, and types of personality disorder.^{Reference Bolton, Pagura, Enns, Grant and Sareen64–Reference May, Klonsky and Klein66} Although suicidal ideation is a key contributing factor and is a very distressing experience, there is a need to further evaluate what else might predict suicidality in tinnitus sufferers before suicidal ideation can be considered an appropriate proxy for studying suicidal behaviour in this population. Longitudinal follow up of tinnitus sufferers with suicidal ideation will be beneficial, with particular reference to socioeconomic factors, psychiatric co-morbidity, life events, and subsequent attempted and completed suicide rates.