Introduction
Sudden sensorineural hearing loss (SNHL) is generally accepted by clinicians as a rapid hearing loss of more than 30 dB in at least 3 contiguous frequencies over 3 days.Reference Stachler, Chandrasekhar, Archer, Rosenfeld, Schwartz and Barrs1 Several different causes of sudden SNHL have been postulated, including vascular disorders, viral infection, autoimmune disorders, neurological disorders, neoplasms and ototoxic drugs.Reference Lin, Hsu, Lin and Weng2 Psychogenic factors are also considered to be one of the causes of sudden SNHL. Sudden SNHL has been reported to be related to psychological strain.Reference Schmitt, Patak and Kroner-Herwig3 According to reports in the literature, lifestyle stress can cause sudden SNHL.Reference Lopez-Gonzalez, Lopez-Lorente, Abrante, Benaixa and Esteban4 Sudden mental stress induced by a gunshot was found to be related to sudden SNHL.Reference Hong and Mun5 Sudden hearing loss can also be caused by panic anxiety attacks.Reference Mori, Fujieda, Yamamoto, Takahashi, Saito and Saito6 In a study of sudden deafness related to psychological factors, major depressive disorder was reported in one of the patients affected.Reference Ban and Jin7 Functional causes of impaired hearing can also include psychogenic hearing impairment, where physical symptoms result as a symbolic expression of long-term psychological problems or conflicts.Reference Ban and Jin7, Reference Vlaski, Dragicević, Dankuc, Kljajić, Lemajić-Komazec and Komazec8
Depressive disorders include major depressive disorders, dysthymic disorder and depressive disorder not otherwise specified. Patients with these disorders may have a state of low mood and may lose interest in activities. Symptoms can last for weeks, months or years. Anxiety disorders include various forms of mental disorders characterised by feelings of anxiety and fear. Previous literature indicates that long-term psychological disease such as depressive disorders or anxiety disorders may increase the risk of sudden SNHL.Reference Stachler, Chandrasekhar, Archer, Rosenfeld, Schwartz and Barrs1–Reference Vlaski, Dragicević, Dankuc, Kljajić, Lemajić-Komazec and Komazec8 However, there are insufficient large-scale studies to support a correlation between sudden SNHL and depressive disorders or anxiety disorders. This retrospective cohort study evaluated the relationship between depressive disorders and sudden SNHL through a population-based cohort study, taking advantage of a large-size dataset available from the National Health Insurance programme in Taiwan.
Materials and methods
Data sources
The National Health Insurance programme in Taiwan was established in 1995 and over 98 per cent of the population are enrolled in the programme. The National Health Insurance Research Database contains encrypted patient identification numbers, International Classification of Diseases (9th revision, Clinical Modification version) codes for diagnoses, procedures, prescriptions, dates of admission, dates of discharge, sex, birth date and other basic sociodemographic information.
The data used in this study came from the Longitudinal Health Insurance Database 2000, a subset of the National Health Insurance database, which contains all the claims data from 1996 to 2011 of 1 million beneficiaries systemically and randomly selected in 2000. There is no significant difference in age, sex and healthcare costs between the sample group and all programme enrollees.
Design
This retrospective cohort study comprised two groups: depressive disorder patients and matched controls. The depressive disorders group consisted of patients with initial diagnoses of depressive disorders (including International Classification of Diseases (9th revision) codes 311, 296.2, 296.3 and 300.4). The control group consisted of subjects without a medical claim for depressive disorders, who were matched (case–control ratio of 1:1) in terms of: age, sex, diabetes mellitus (code 250.xx), hypertension (codes 401–405), chronic kidney disease (codes 582, 583, 585, 586 and 588), hyperlipidaemia (code 272), living area and income. The index date for patients with depressive disorders was the date, between 2001 and 2008, that they were first registered. The index dates for the non-depressive disorders group (control group) were matched to those of the depressive disorders group.
Patients in the control group with anxiety disorders (International Classification of Diseases (9th revision) codes 300.00–300.9 (but not 300.4), and code 300.11 which was conversion disorder) before the index date were excluded. The patients with depressive disorders were subdivided into two groups according to whether or not anxiety disorders (codes 300.00–300.9 (not 300.4) and code 300.11) were diagnosed before the index date.
Subjects in the depressive disorders group and control group were assessed to evaluate the incidence of sudden SNHL (International Classification of Diseases (9th revision) code 388.2) at the end of 2011.
This study was approved by the Institutional Review Board of the Chi-Mei Medical Center in Taiwan. All individual identification numbers in the National Health Insurance Research Database were encrypted.
Statistical analysis
Differences in sociodemographic characteristics and co-morbidities between the depressive disorders group and the control group were compared using Pearson's chi-square test. The incidence rate was calculated as the number of sudden SNHL subjects identified during the follow up, divided by the total person-years for each group by age, sex, co-morbidity and years of follow up. We used Poisson regression to calculate incidence rate ratios of sudden SNHL, thereby enabling comparison of incidence rates in the depressive disorders group and control group. Cox proportional hazard regression and Kaplan–Meier analyses were used to estimate the risk of sudden SNHL in the three study populations, which were: the depressive disorders group with no anxiety disorders before index date, the depressive disorders group with anxiety disorders before the index date and the control group (reference group). All analyses were performed using SAS software, version 9.3 (SAS Institute, Cary, North Carolina, USA). The statistical significance level was set at p < 0.05 (two-tailed).
Results
Hearing loss incidence by group
A total of 27 547 patients with depressive disorders diagnosed between 2001 and 2008 met the criteria. A total of 27 547 subjects were randomly selected for inclusion in the control group. There was no significant difference between the depressive disorders group and the control group in terms of the distribution of age, sex, living area, income, and baseline co-morbidities including diabetes mellitus, hypertension, chronic kidney disease and hyperlipidaemia (Table I).
Demographics and co-morbidities of depressive disorder patients and controls
Data represent numbers (and percentages) unless indicated otherwise. *Total n = 27 547; †total n = 27 547
The incidence of sudden SNHL was found to be higher in the depressive disorders group than in the control group (p = 0.0041) by the end of follow up. The incidence of sudden SNHL was significantly increased for females (p = 0.0187), and was higher for the first to third follow-up years (p = 0.0028) (Table II). The association between cumulative incidence rate and follow-up years is shown in Figure 1.
Association between cumulative incidence rate and follow-up years for depressive disorder patients and controls.
Risk of sudden deafness in depressive disorder patients and controls
*Rate: per 10 000 person-years. IRR = incidence rate ratio; CI = confidence interval
Multivariate analysis using Cox's proportional hazards regression model with age, sex and co-morbidities revealed adjusted hazard ratios of 1.46 (95 per cent confidence interval (CI) = 1.103–1.934) for the depressive disorders group without anxiety disorders before the index date and 1.423 (95 per cent CI = 1.025–1.975) for the depressive disorders group with anxiety disorders before the index date (Table III). The association between cumulative incidence rate and follow-up years in the control group and both depressive disorder subgroups (i.e. depressive disorder patients without anxiety disorders before the index date and depressive disorder patients with anxiety disorders before the index date) is shown in Figure 2.
Association between cumulative incidence rate and follow-up years for the control group and both depressive disorder subgroups (i.e. depressive disorder patients without anxiety disorders before the index date and depressive disorder patients with anxiety disorders before the index date).
Hazard ratios for development of sudden hearing loss*
*Based on Cox proportional hazard regression analyses, adjusted for age, sex, diabetes mellitus, hypertension, chronic kidney disease, hyperlipidaemia, income and residing area. †n = 93 out of 17 726. ‡p < 0.05. **n = 55 out of 9821. §n = 103 out of 27 547. CI = confidence interval
Hearing loss by age, sex and co-morbidities
The multivariate Cox proportional hazard regression analysis of all patients with sudden SNHL in both the depressive disorders group and control group indicated that hearing loss increased with age. Male sex was found to be a significant risk factor for developing sudden SNHL, with hazard ratio of 1.318 (95 per cent CI = 1.026–1.693). Among the co-morbidities, chronic kidney disease had the highest hazard ratio, of 2.177 (95 per cent CI = 1.101–4.305), followed by hyperlipidaemia (hazard ratio of 1.714; 95 per cent CI = 1.156–2.542) and diabetes mellitus (hazard ratio of 1.698; 95 per cent CI = 1.181–2.443) (Table III).
Discussion
A few studies have described sudden SNHL in patients suffering from stress including psychological strain.Reference Schmitt, Patak and Kroner-Herwig3, Reference Lopez-Gonzalez, Lopez-Lorente, Abrante, Benaixa and Esteban4 In addition, panic anxiety attacks and major depressive disorders have been reported in patients with sudden SNHL.Reference Hong and Mun5–Reference Ban and Jin7 Psychogenic hearing impairment where physical symptoms occur as a symbolic expression of long-term psychological problems or conflicts has also been reported.Reference Ban and Jin7, Reference Vlaski, Dragicević, Dankuc, Kljajić, Lemajić-Komazec and Komazec8 Based on such literature, we postulated that long-term psychological diseases such as depressive disorders or anxiety disorders might increase the risk of sudden SNHL.
There were two groups in this study: a depressive disorders group and a matched control group. We attempted to reduce the effect of previous anxiety disorders on developing sudden SNHL by further dividing the depressive disorders group into two subgroups according to whether depressive disorder patients had an anxiety disorder before the index date. Other psychological conditions such as schizophrenia were not taken into consideration because of a lack of evidence indicating that they increased the risk of developing sudden SNHL. It is worth mentioning that anxiety disorders in our study included conversion disorder, which is related to conversion deafness.Reference Wang, Wang, Lin and Lee9
Patients with anxiety disorders (International Classification of Diseases (9th revision) codes 300.00–300.9 (not 300.4) and code 300.11 which is conversion disorder) before the index date were excluded from the control group. In addition, the depressive disorder patients without anxiety disorders had no history of conversion disorder before the index date. Hence, the possibility of sudden SNHL being related to conversion disorder was reduced to a certain extent in the patients in this subgroup.
It was noted that diabetes mellitus, chronic kidney disease and hyperlipidaemia might increase the risk of developing sudden SNHL.Reference Lin, Hsu, Lin and Weng2, Reference Lin, Lin, Weng and Chou10–Reference Aimoni, Bianchini, Borin, Ciorba, Fellin and Martini13 In order to reduce any effect of this association, subjects without claims for depressive disorders in the control group were matched in terms of diabetes mellitus, hypertension, chronic kidney disease and hyperlipidaemia. There was no significant difference in the distribution of the co-morbidities at baseline (Table I and Table II).
As shown in Table II, the incidence of sudden SNHL was higher in the depressive disorders group than in the control group by the end of follow up (p = 0.0041). In addition, as shown in Table III, the depressive disorders subgroup comprising patients without anxiety disorders before the index date had a significant hazard ratio of 1.460 compared to the control group. These findings suggest an increased risk of sudden SNHL in patients with depressive disorders. Table III also shows that the depressive disorders subgroup comprising patients with anxiety disorders before the index date had a significant hazard ratio of 1.423 compared to the control group.
There was no significant difference in hazard ratio between the depressive disorder subgroups, comprising patients with and without anxiety disorders. It is not clear from this study whether or not anxiety disorders increase the risk of developing sudden SNHL; further investigation is needed.
Table II shows that females in the depressive disorders group had a significantly higher incidence of sudden SNHL than females in the control group (p = 0.0187). This might imply that depressive disorders had more influence on the development of sudden SNHL in females than in males. In contrast, as shown in Table III, the males in the depressive disorders group and the control group had more risk of developing sudden SNHL than the females in the depressive disorders group and the control group, with a hazard ratio of 1.318 (95 per cent CI = 1.026–1.693). This might suggest that males are at higher risk than females of developing sudden SNHL. This finding is compatible with research on the association between chronic kidney disease and sudden SNHL conducted in Taiwan, which indicated that males have a higher risk of developing sudden SNHL than females.Reference Lin, Hsu, Lin and Weng2
The incidence of sudden SNHL was significantly higher between the first and third follow-up years (p = 0.0028). This implies that depressive disorders have more impact on the development of sudden SNHL during this period. Stress, including emotional or psychosocial strain, has been related to sudden SNHL by the possible mechanisms of vasoconstriction, decreased blood supply, haemoconcentration and microcirculation occlusion in the inner ear.Reference Schmitt, Patak and Kroner-Herwig3, Reference Lopez-Gonzalez, Lopez-Lorente, Abrante, Benaixa and Esteban4 Decreased regional cerebral blood flow has also been reported in patients with depression.Reference Qiu, Hu, Shi and Wang14–Reference Ota, Noda, Sato, Hattori, Teraishi and Hori16 According to these findings, the possible mechanism by which depressive disorders are related to sudden SNHL might be vasoconstriction or decreased regional cerebral blood flow supply. However, it is possible that other mechanisms are responsible; further investigation is needed.
In the depressive disorders group and control group, patients with chronic kidney disease had a significantly higher hazard ratio (of 2.177; 95 per cent CI = 1.101–4.305) than those without. Patients with diabetes mellitus and hyperlipidaemia also had a significantly higher hazard ratio than those without. These findings are compatible with previous reports.Reference Lin, Hsu, Lin and Weng2, Reference Lin, Lin, Weng and Chou10, Reference Weng, Devine, Xu, Yan and Dong12, Reference Aimoni, Bianchini, Borin, Ciorba, Fellin and Martini13 However, patients with hypertension did not show a significantly higher hazard ratio in our study. According to the literature, there is no strongly increased risk for sudden SNHL in patients with hypertension.Reference Mosnier, Stepanian, Baron, Bodenez, Robier and Meyer11, Reference Aimoni, Bianchini, Borin, Ciorba, Fellin and Martini13
This population-based dataset allows us to demonstrate the relationship between depressive disorders and sudden SNHL in Taiwan. The large sample size increases the statistical power and precision of risk assessment. We reduced the possible effects of other risk factors for sudden SNHL, such as chronic kidney disease, diabetes mellitus and hyperlipidaemia, by matching the control group.
Nonetheless, this population-based study has its limitations. First, the diagnosis of sudden SNHL was based on the diagnostic code registered by the physicians responsible for the treatment of the patients, and was not corroborated by the investigators. Second, other suspected risk factors for sudden SNHL, such as noise exposure, were not available in the insurance data. The lack of access to these factors might result in some degrees of bias. Third, the severity of hearing loss was not provided by the data. In addition, although vasoconstriction or decreased regional cerebral blood flow have been suggested as possible mechanisms by which depressive disorders are related to sudden SNHL, this population-based study was unable to explore this issue.
Non-organic hearing loss is a term used to describe hearing loss that appears to be of a greater magnitude than can be explained by an auditory system lesion and pathophysiology, and in which subjective hearing levels are different from objective levels.Reference Georgescu, Stan, Marinescu and Păun17 Non-organic hearing loss includes hearing loss related to malingering, factitious disorder and conversion disorder.Reference Georgescu, Stan, Marinescu and Păun17 Psychogenic hearing loss or conversion deafness is defined as hearing loss without the intent to deceive, and is differentiated from malingering and factitious disorders. In psychogenic hearing loss, a discrepancy exists between the patient's subjective and objective audiological test results, such as a discrepancy between pure tone audiometry and auditory brainstem response. Although experienced clinicians may easily detect psychogenic hearing loss using conventional audiological procedures, more advanced audiological tests have been used to diagnose psychogenic hearing loss, including auditory brain responses and otoacoustic emissions.Reference Oishi, Kanzaki, Kataoka, Tazoe, Takei and Nagai18–Reference Taiji and Morimoto21 In Taiwan, these subjective and objective audiological tests are widely available, enabling otolaryngologists to diagnose malingering, factitious disorder and psychogenic hearing loss.
• This study indicates that depressive disorder patients have an increased risk of developing sudden sensorineural hearing loss (SNHL)
• Co-morbidities of diabetes mellitus, chronic kidney disease and hyperlipidaemia significantly aggravated this risk
• Depressive disorders might be a risk factor for sudden SNHL
• Control of depressive disorders could potentially decrease sudden SNHL incidence in depressive disorder patients
The findings of this study indicate that depressive disorders might be a risk factor for sudden SNHL. It remains to be seen whether control of depressive disorders can decrease the incidence of sudden SNHL in patients with depressive disorders; further investigation is needed.
