Introduction
Depression is a common and significant health problem that negatively affects the quality of life of patients and their family members (Kessler and Bromet, Reference Kessler and Bromet2013). A 2009–2012 report from the National Health and Nutrition Examination Survey (NHANES) found that 7.6% of Americans aged ⩾12 years had moderate or severe depression (Pratt and Brody, Reference Pratt and Brody2014). Depression is among the leading causes of decreased work productivity (Stewart et al., Reference Stewart, Ricci, Chee, Hahn and Morganstein2003). In fact, a Gallup survey conducted during 2011–2012 estimated an annual loss of $23 billion in lost productivity as a result of workers suffering from depression (Witters et al., Reference Witters, Liu and Agrawal2013).
Hearing loss is the third most common self-reported chronic physical condition in the USA (Blackwell et al., Reference Blackwell, Lucas and Clarke2014), The estimated prevalence of hearing loss in US adults ages 20–69 years in 2012 was 31.1% for high-frequency hearing loss (HFHL) and 14.1% for speech-frequency hearing loss (SFHL) (Hoffman et al., Reference Hoffman, Dobie, Losonczy, Themann and Flamme2017). The diagnosis of hearing loss often is associated with poor quality of life, depression, and social isolation and its management starts after the occurrence of these negative consequences.
A recent study reported that among the risk factors for dementia, 35% of them were potentially modifiable: hearing loss was the main potentially modifiable factor (9.1%) with depression accounting for an additional 4% (Livingston et al., Reference Livingston, Sommerlad, Orgeta, Costafreda, Huntley, Ames, Ballard, Banerjee, Burns, Cohen-Mansfield, Cooper, Fox, Gitlin, Howard, Kales, Larson, Ritchie, Rockwood, Sampson, Samus, Schneider, Selbæk, Teri and Mukadam2017).
Several studies reported a significant positive association between hearing impairment measured by self-reported hearing impairment or pure tone audiometry and depression among older persons (Blay et al., Reference Blay, Andreoli, Fillenbaum and Gastal2007; Huang et al., Reference Huang, Dong, Lu, Yue and Liu2010; Morikawa et al., Reference Morikawa, Okamoto, Kiuchi, Tomioka, Iwamoto, Harano, Saeki, Fukusumi, Hashimoto, Amano, Hazaki, Yanagi, Iki, Yamada, Kishimoto and Kurumatani2013), although this is not a universal finding (Chou and Chi Reference Chou and Chi2004; Bergdahl et al., Reference Bergdahl, Gustavsson, Kallin, von Heideken Wågert, Lundman, Bucht and Gustafson2005). NHANES cycle 2005–2010 measured pure tone audiometry only in a subset of the population: adolescents (12–19 years old) and in adults 70 years and older, whereas self-reported hearing problem was assessed in all age group. Using data from NHANES 2005–2010, Li et al. (Reference Li, Zhang, Hoffman, Cotch, Themann and Wilson2014) reported a significant positive association between moderate hearing loss, defined by pure tone average, and depression in females aged ⩾70 years. The researchers found no significant association between self-reported hearing impairment and depression in participants of both sexes aged ⩾70 years.
NHANES wave 2011–2012 measured pure tone audiometry only in a subpopulation of adults aged 20–69 years. This cycle allowed us to evaluate the potential association between hearing loss (defined by pure tone audiometry) and depression among adults aged 20–69 years. This is the first study reporting the association of hearing loss based on thresholds of pure tone audiometry with depression in US adult (ages 20–69 years) from the most recently available NHANES data set.
Methods
Study population
NHANES is a cross-sectional, nationally representative survey of the non-institutionalized civilian population in the USA conducted annually by the National Center for Health Statistics (NCHS) (Johnson et al., Reference Johnson, Paulose-Ram, Ogden, Carroll, Kruszon-Moran, Dohrmann and Curtin2013). The survey uses a multistage stratified probability sample based on selected counties, blocks, households, and persons within households. NCHS-trained professionals conducted interviews in participants’ homes; extensive physical examinations, including blood and urine collection, were conducted at mobile exam centers. For our analyses, we included adult participants (ages ⩾20 years) of NHANES 2011–2012 who answered the 9-question Patient Health Questionnaire (PHQ-9) depression screening module and had complete audiometric results. Participants with missing co-variables included in the multivariable-adjusted models were excluded, for a final sample size of 3316 participants.
Outcome measure
The outcome was the presence or absence of depressive symptoms as determined by a participant's score on the PHQ-9, a depression screening module of the Primary Care Evaluation of Mental Disorders Questionnaire (PRIME-MD). The Depression Screener questions were asked by trained interviewers using the Computer-Assisted Personal Interviewing (CAPI) system as part of the Mobile Examination Center (MEC) private interview in Spanish or English. No proxies or interpreters were permitted for these questions PHQ-9 contains nine questions that are used as a depression screener in NHANES 2011–2012. The questions are based on the 9 signs and symptoms for depression listed in the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) (APA, 1994). Responses to these nine questions were scored on a four-point Likert scale of 0–3, indicating that the participant experienced the symptom ‘not at all,’ ‘on several days,’ ‘on more than half the days,’ or ‘nearly every day’ during the past 2 weeks, for a total score ranging from 0 to 27. A prior validation study found that a score ⩾10 achieved 88% sensitivity and 88% specificity for major depression (Kroenke et al., Reference Kroenke, Spitzer and Williams2001). Therefore, participants who scored ⩾10 were defined as having current depressive symptoms.
Audiometric measurements and definition of hearing loss
National Institute for Occupational Safety and Health trained examiners followed a standardized NCHS protocol to obtain audiometric measurements in specially designed and equipped mobile examination center sound-treated rooms. Audiometers were calibrated according to NCHS specifications. Air conduction thresholds were measured for each ear at 0.5, 1, 2, 3, 4, 6, and 8 kHz across an intensity range of −10 to 120 dB hearing level. Participant's test–retest response reliability was measured at 1 kHz frequency in each ear and the average was used in the analyses. Results were rejected if the test–retest thresholds differed by ⩾10 dB. Further details of the measurement techniques are available on the NCHS website (https://www.ncdc.gov/nchs/data/nhanes/2011-2012/manuals/audiometry_procedures_manual.pdf). Four audiometric frequencies (0.5, 1, 2, and 4 kHz) were averaged to define a SFHL and three audiometric frequencies (3, 4, and 6 kHz) were averaged to define a HFHL. Degrees of hearing impairment for SFHL and HFHL were categorized according to the following World Health Organization classifications: normal ⩽25 dB, slight/mild = 26–40 dB, and moderate/worse ⩾41 dB (see http://www.who.int/pbd/deafness/hearing_impairment_grades/en/).
Covariates
The a priori covariates selected for inclusion in the models are associated with hearing impairment or depressive symptoms. These covariates include age (categorized in weighted tertiles), sex, race/ethnicity, education, poverty income ratio, marital status, obesity, alcohol consumption, cigarette smoking, diabetes, vision impairment, and chronic disease (cardiovascular disease or cancer).
We obtained information about age (years), sex, race/ethnicity, and education from the household interview. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Hispanic, and other. Poverty income ratio is a measure of socioeconomic status and represents the calculated ratio of household income to the poverty threshold after accounting for inflation and family size. Marital status was categorized as ‘married or living with a partner’ and ‘single, or divorced, widowed.’
Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). The adult population was categorized as either normal/underweight (BMI <25 kg/m2) or overweight/obese (BMI ⩾25 kg/m2).
Alcohol consumption (amount consumed per week) and smoking information were obtained from the associated questionnaire. Smoking status was defined as neversmoker (smoked <100 cigarettes ever), former smoker (not currently smoking, but has smoked ⩾100 cigarettes ever), and current smoker (⩾100 cigarettes in lifetime and currently smoking every day or some days).
Diabetes was defined as glycated hemoglobin (A1C) ⩾6.5%, or self-reported current use of insulin or diabetes medication. Cardiovascular disease was defined as having a positive response to any of the following statements: ‘ever told you had congestive heart failure,’ ‘ever told you had coronary heart disease,’ ‘ever told you had angina/angina pectoris,’ ‘ever told you had a heart attack,’ or ‘ever told you had a stroke.’ Cancer was defined as having a positive response to ‘ever told you had cancer or malignancy.’ Vision impairment was defined as self-reported ‘trouble seeing even with glass/contacts.’
Statistical methods
We used mobile examination center sample weights for analyses to account for the complex sampling design and non-response in NHANES. We used logistic regression to calculate adjusted odds ratios for having depressive symptoms. Sex and age group analyses were stratified (ages 20–35, 36–51, and 52–69 years); age group stratification was based on the weighted tertiles of the study population SAS 9·3 software (SAS Institute, Cary, NC) was used for all statistical analyses and SAS-Callable SUDAAN 10 (Research Triangle Institute, Research Triangle Park, NC) was used to account for the NHANES complex sample design. p values from Satterthwaite statistics for regression analyses were presented at the significance level <0.05.
Results
Both audiometric and depression data were available for 3316 of the NHANES 2011–2012 adult participants ages 20–69 years. Table 1 presents the weighted characteristics of the study population. The geometric mean age of all participants was approximately 41 years. Among the participants, 68.2% were non-Hispanic whites, 49.4% were females, and 16.8% were living in a family with income at or below poverty. The weighted prevalence of slight/mild and moderate/worse SFHL grades were 10.4% and 3.9%, respectively. Men had a statistically significant higher prevalence of SFHL than did women for slight/mild (14.3% v. 6.3%, χ2 p = 0.001) and moderate/worse (5.1% v. 2.7%, χ2 p = 0.001). The weighted prevalence of slight/mild and moderate/worse HFHL was 14.7% and 16.1%, respectively. Men had a statistically significant higher prevalence of HFHL than did women for slight/mild (16.7% v. 12.6%, χ2 p < 0.001) and moderate/worse (25.5% v. 6.6%, χ2 p < 0.001). Among the adults in this subpopulation, 8% were categorized as having depressive symptoms (PHQ-9 ⩾ 10), with a statistically significant higher prevalence in women (10.2%) than in men (5.8%) (Table1).
Sample size and weighted characteristics of the National Health and Nutrition Examination Survey (NHANES) 2011–2012 adult participants
GM, geometric mean; BMI, body mass index; SE, standard estimate; PHQ-9, Patient Health Questionnaire; HFHL, high-frequency hearing loss; SFHL, speech frequency hearing loss.
Multivariate logistic regression analyses found that among all adults, moderate/worse HFHL (⩾41 dB) was statistically significantly associated with depressive symptoms (OR 1.54, 95% CL 1.04–2.27) (Table 2). After stratification by sex, this statistically significant association was found in adult females (OR 2.39, 95% CL 1.06–5.39), but not in males (OR 1.23, 95% CL 0.63–2.42) (Table 2). Slight/mild HFHL was not associated with depression. Among all adults, SFHL was not associated with depressive symptoms (slight/mild OR 1.08, 95% CL 0.59–2.00; moderate/worse OR 1.88, 95% CL 0.92–3.86) (Table 2). After stratification by sex, a statistically significant association was found in adult female participants between moderate/worse SFHL (⩾41 dB) and depression (OR 3.00, 95% CL 1.20–9.31) (Table 2).
Multivariate logistic regressiona OR (95% CL) of having depression in the National Health and Nutrition Examination Survey (NHANES) 2011–2012 adult participants ages 20–69 years
a Adjusted for age, race/ethnicity, body weight status, smoking, education, alcohol consumption, poverty income ratio, diabetes, marital status, cardiovascular disease, cancer, vision impairment, and sex (except when stratified by sexes).
Further stratification by age group showed found that moderate/worse HFHL and moderate/worse SFHL were associated with depression in women ⩾52 years (Table 3). Moreover, in men, we found no associations between hearing loss and depression in the age groups 36–51 and 52–69 years (Table 3).
Multivariate logistic regressiona OR (95% CL) of having depression in the National Health and Nutrition Examination Survey (NHANES) 2011–2012 participants, by age group and sex
a Adjusted for age, race/ethnicity, body weight status, smoking, education, alcohol consumption, poverty income ratio, diabetes, marital status, cardiovascular disease, cancer, vision impairment, and sex (except when stratified by sexes).
Discussion
Our study of a US national population-based sample of adults aged ⩾20 years found that moderate/worse SFHL and HFHL (⩾41 dB) are a significant predictor of depressive symptoms and that the person's sex is a potential moderator for these associations. When further stratified by age group, hearing loss was positively associated with depression only in female aged 52–69 years. Similarly, Li et al. (Reference Li, Zhang, Hoffman, Cotch, Themann and Wilson2014) analyzed NHANES 2005–2010 data and found a positive association between moderate SFHL (range 35–49 dB) and depression (PHQ-9 ⩾10) in females, but not in males aged ⩾70 years. Data from the Blue Mountains Eye Study, also showed sex differences in the association between hearing loss and depression, with mild SFHL (>25–40 dB hearing loss) positively associated with depressive symptoms in women (OR 1.95, 95% CI 1.15–3.31), but not in men (Gopinath et al., Reference Gopinath, Wang, Schneider, Burlutsky, Snowdon, McMahon, Leeder and Mitchell2009). In contrast, Shin and Hwang (Reference Shin and Hwang2017), using data from the Korean National Health and Nutrition Examination Survey (KNANES) 2010–2013 on 18 563 participants aged ⩾20 years, reported a positive association between hearing loss (speech frequency pure tone average ⩾25 dB) and depression only in males aged ⩾60 years. Harada et al. (Reference Harada, Nishiwaki, Michikawa, Kikuchi, Iwasawa, Nakano, Ishigami, Saito and Takebayashi2008) also found that hearing loss (defined as failing to hear a 30 dB signal at 1 kHz bilaterally) was positively associated with depression in Japanese males aged ⩾65 years, but not in females.
Several studies addresses the relationship between hearing and depression based on perceived self-reported hearing impairment, rather than thresholds of pure tone audiometry, with inconsistent results In a cross-sectional study of 3796 Japanese adults aged 65 and older, Morikawa et al. (Reference Morikawa, Okamoto, Kiuchi, Tomioka, Iwamoto, Harano, Saeki, Fukusumi, Hashimoto, Amano, Hazaki, Yanagi, Iki, Yamada, Kishimoto and Kurumatani2013) reported that self-reported hearing or visual impairment were independent risk factors for depressive symptoms. Blay et al. (Reference Blay, Andreoli, Fillenbaum and Gastal2007) reported associations of hearing impairment (OR 1.26, 95% CI 1.12–1.42) and visual impairment (OR 1.32, 95% CI 1.16–1.50) with depression among Brazilian adults aged 60 years and older (n = 7040). Conversely, Lee et al. (Reference Lee, Tong, Yuen, Tang and Vanhasselt2010) reported and association of hearing loss measured by audiometry, but not self-reported hearing impairment, with depressive symptoms in a Chinese population aged 60 years and older. Li et al. (Reference Li, Zhang, Hoffman, Cotch, Themann and Wilson2014) found that self-reported hearing impairment was significantly associated with an increased risk of depression in adults aged 18–69 years, but not in those 70 years and older, which audiometric hearing loss was associated with depression in older female.
Hearing impairment was also positively associated with depression in longitudinal studies. In the National Longitudinal Study on Hearing conducted in the Netherlands, reduced hearing ability was significantly associated with higher levels of depression in adults aged 40–49 years (Nachtegaal et al., Reference Nachtegaal, Smit, Smits, Bezemer, van Beek, Festen and Kramer2009). Saito et al. (Reference Saito, Nishiwaki, Michikawa, Kikuchi, Mizutari, Takebayashi and Ogawa2010) studied 580 adults aged ⩾65 years (261 men and 319 women) without depressive symptoms at the baseline. At the 3 years follow-up, they found that hearing impairment was associated with an elevated risk for depressive symptoms (OR 2.45, 95% CI 1.26–4.77).
Our results demonstrate that moderate/worse hearing loss is associated with depressive symptoms in women ages 52–69 years in the USA, independent of important covariates. Hearing loss has been shown to negatively affect the person's quality of life. For instance, hearing loss alters the ability of communicate with others, thereby affecting social relationships (Kamil and Lin, Reference Kamil and Lin2015), as well as employment opportunities (Hogan et al., Reference Hogan, O'Loughlin, Davis and Kendig2009). Our study found that women compared with men were more likely to experience depression related to hearing loss, one reason for this finding may be that women form deeper social relationship based on communicating feelings (Vigil, Reference Vigil2007), whereas men form social relationships based on shared activities (Spain, Reference Spain and Nardi1992) which may not be affected by hearing loss. Additionally, hearing loss has a profound effect on self-esteem and physical functioning, contributing to loss of independence and diminished quality of life (Strawbridge et al., Reference Strawbridge, Wallhagen, Shema and Kaplan2000; Dalton et al., Reference Dalton, Cruickshanks, Klein, Klein, Wiley and Nondahl2003; Chia et al., Reference Chia, Wang, Rochtchina, Cumming, Newall and Mitchell2007; Hawkins et al., Reference Hawkins, Bottone, Ozminkowski, Musich, Bai, Migliori and Yeh2012; Chen et al., Reference Chen, Genther, Betz and Lin2014). Hearing impairment also negatively affects cognitive performance (Lin, Reference Lin2011; Lin et al., Reference Lin, Ferrucci, Metter, An, Zonderman and Resnick2011, Reference Lin, Yaffe, Xia, Xue, Harris, Purchase-Helzner, Satterfield, Ayonayon, Ferrucci and Simonsick2013), increases mortality risk (Yamada et al., Reference Yamada, Nishiwaki, Michikawa and Takebayashi2011) and has been associated with Alzheimer disease (Gates et al., Reference Gates, Anderson, McCurry, Feeney and Larson2011) and particularly with dementia (Davies et al., Reference Davies, Cadar, Herbert, Orrell and Steptoe2017; Golub et al., Reference Golub, Luchsinger, Manly, Stern, Mayeux and Schupf2017).
A meta-analysis commissioned by The Lancet found that the relative risk for dementia in persons with hearing loss was almost twice the risk for dementia in those without hearing loss (pooled RR 1.94, 95% CI 1.38–2.73) (Livingston et al., Reference Livingston, Sommerlad, Orgeta, Costafreda, Huntley, Ames, Ballard, Banerjee, Burns, Cohen-Mansfield, Cooper, Fox, Gitlin, Howard, Kales, Larson, Ritchie, Rockwood, Sampson, Samus, Schneider, Selbæk, Teri and Mukadam2017). The authors reported that 35% of the risk factors for dementia were potentially modifiable, with hearing loss representing the highest percentage (9.1%) and depression an additional 4% (Livingston et al., Reference Livingston, Sommerlad, Orgeta, Costafreda, Huntley, Ames, Ballard, Banerjee, Burns, Cohen-Mansfield, Cooper, Fox, Gitlin, Howard, Kales, Larson, Ritchie, Rockwood, Sampson, Samus, Schneider, Selbæk, Teri and Mukadam2017). Management of hearing loss and depression might contribute to the prevention or delay of dementia (Livingston et al., Reference Livingston, Sommerlad, Orgeta, Costafreda, Huntley, Ames, Ballard, Banerjee, Burns, Cohen-Mansfield, Cooper, Fox, Gitlin, Howard, Kales, Larson, Ritchie, Rockwood, Sampson, Samus, Schneider, Selbæk, Teri and Mukadam2017).
The use of the structured self-report assessment PHQ-9, although, per se, not a diagnostic instrument, is a strength of the present analysis. The PHQ-9 is widely used in psychiatric research and has a high degree of correlation with clinical interviews (Martin et al., Reference Martin, Rief, Klaiberg and Braehler2006). However, the present study also has several limitations, the most important being its cross-sectional design. The depressive symptoms status in our study is limited to answers in the PHQ-9 about the experience of the participants during the past 2 weeks, so the depressive symptoms status might reflect a short-term health condition, and not necessarily a chronic condition. Another limitation to conclusive evidence is the single point in time measurement of audiometric thresholds. Several studies have raised concerns of potential false-positive rates of high frequency audiogram notches, such as 6 and 8 kHz, which might be the result of high subject variability and calibration error for TDH-type headphones (Schlauch and Carney, Reference Schlauch and Carney2011).
Our analysis found that hearing loss is a risk factor for depression. Many studies have shown that hearing loss can affect the quality of life through diminished social interaction, increased loneliness, and cognitive performance. The National Academy of Sciences recommends using patient visits ‘to assess and discuss potential hearing difficulties that could affect doctor–patient communication and overall patient well-being, to encourage individuals and their family members and caregivers to discuss hearing concerns, to raise awareness among older adults about age-related hearing loss, and to encourage referral when appropriate.’ (NAS, Reference Blazer, Domnitz and Liverman2016). Health professionals should be aware of signs and symptoms of depression among persons with hearing loss and refer patients for mental health services.
Acknowledgements
We thank our colleagues at the CDC National Center for Health Statistics who conducted the NHANES surveys and our colleagues at the National Institute for Occupational Safety and Health who analyzed the NHANES audiograms.
Financial support
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflict of interest
None.
Disclaimer
The findings and conclusions in this report are those of the author(s) and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC) or the Agency for Toxic Substances and Disease Registry (ATSDR).
IRB approval
CDC/ATSDR has determined that our study meets the criteria for research not involving identifiable human subjects per federal regulation 45 CFR 46; therefore, it does not require institutional review board (IRB) approval.
