A Cross-sectional Study

Humberto Yévenes-Briones, MS¹; Francisco Félix Caballero, PhD¹; Ellen A. Struijk, PhD¹; et al Jorge Rey-Martinez, MD, PhD²; Lourdes Montes-Jovellar, MD, PhD³; Auxiliadora Graciani, MD, PhD^1,4; Fernando Rodríguez-Artalejo, MD, PhD^1,4; Esther Lopez-Garcia, PhD^1,4Author Affiliations Article Information JAMA Otolaryngol Head Neck Surg. Published online September 23, 2021. doi:10.1001/jamaoto.2021.2399Key Points

Question  Is hearing loss associated with impaired physical function, frailty, and disability in older adults?

Findings  In this cross-sectional study of 1644 older adults from the Seniors-ENRICA-2 cohort study, hearing loss at speech frequency pure-tone average (0.5, 1, 2, and 4 kHz) was associated with impaired physical function, frailty, and disability independently of lifestyles, comorbidities, cognitive status, and social isolation. The results were similar for hearing loss at standard frequency pure-tone average (0.5, 1, and 2 kHz).

Meaning  These results may contribute to better understanding of the association between hearing loss and functional capacity.Abstract

Importance  Several studies have examined the association between hearing loss (HL) and physical function, with inconsistent results. Few used pure-tone thresholds and considered possible confounders in the association.

Objective  To examine the association between hearing loss and impaired lower extremity function, frailty syndrome, and disability in older adults.

Design, Setting, and Participants  The Seniors-ENRICA-2 is a cohort study that was established in 2015-2017 in Spain. The present study was conducted from December 2015 to June 2017. Data were analyzed from January 2, 2021, to March 10, 2021. This cross-sectional analysis included 1644 community-dwelling individuals aged 65 years or older (range, 66-91 years) of both sexes. Participants had hearing threshold measurements and data on impaired lower extremity function, frailty syndrome, and disability.

Exposures  Hearing loss defined as pure-tone average greater than 40 dB HL in the better ear for standard frequency (0.5, 1, and 2 kHz), speech frequency (0.5, 1, 2, and 4 kHz), and high frequency (3, 4, and 8 kHz).

Main Outcomes and Measures  Impaired lower extremity function was defined with the Short Physical Performance Battery; the frailty syndrome was defined as having at least 3 of the 5 criteria, including weakness, slow walking speed, low physical activity, exhaustion, and weight loss; and disability in instrumental activities of daily living (IADL) was evaluated with the Lawton and Brody scale.

Results  Of the 1644 participants, 831 were men (50.5%); mean (SD) age was 73.8 (4.3) years. The prevalence of moderate HL was 13.6% at speech frequency. After adjustment for age, sex, lifestyle, comorbidities, impaired cognition, and social isolation, HL in standard frequency was associated with impaired lower extremity function, with an odds ratio (OR) of 2.20 (95% CI, 1.25-3.88); the corresponding estimate for the frailty syndrome was 1.85 (95% CI, 0.98-3.49) and for the IADL disability, was 2.25 (95% CI, 1.29-3.94). When considering speech frequency pure-tone average, HL was also associated with impaired function (OR, 2.59; 95% CI, 1.57-4.28), frailty syndrome (OR, 1.85; 95% CI, 1.06-3.22), and IADL disability (OR, 2.18; 95% CI, 1.32-3.60).

Conclusions and Relevance  In this cross-sectional analysis of data from the Seniors-ENRICA-2 cohort study, hearing loss was associated with impaired lower extremity function, frailty syndrome, and IADL disability. It may be useful to examine this association with the mechanisms elucidated in further studies.Introduction

Increased life expectancy has been accompanied by a burden of late-life morbidity. Aging is associated with a greater prevalence of impaired sensory, motor, and cognitive function, which lowers quality of life and increases dependency.¹ One of the most prevalent sensory impairments in the older population is hearing loss (HL),² which is the fifth leading cause of disability overall worldwide³ and entails high economic costs for society.⁴ Hearing difficulties in older adults are undertreated, a situation that results in several adverse consequences, such as increased risk of depression⁵ and lower quality of life.⁶

The association between HL and physical function limitation, frailty, and disability is unclear. Some chronic diseases, including hypertension, type 2 diabetes, and cardiovascular disease, have been associated with HL^7–9and are also associated with disability.^10,11 In addition, sensorineural HL is associated with impaired cognition and probably shares several age-related degenerative alterations¹² so that cognitive impairment may lead to physical function limitation.¹³ Hearing loss in older adults may also restrict social participation, which is a definition of disability.¹⁴ In fact, hearing impairment might be an early physiologic marker of physical function limitation.¹⁵

Previous studies have examined the association between HL and physical function, frailty syndrome, and disability, with inconsistent results.^16–24Research to better characterize the association between hearing capacity and functional capacity is needed, as well as to better understand the role of comorbidities, cognitive impairment, and social isolation in this association. Therefore, our objective was to examine the association between HL using the pure-tone average (PTA) of air conduction hearing thresholds in a wide frequency range in association with impaired lower extremity function, frailty, and disability in a well-characterized population of older adults.MethodsStudy Design and Participants

In this cross-sectional study, we analyzed data from the Seniors-ENRICA-2 study, a cohort study of 3273 community-dwelling individuals aged 65 years or older. Study participants were residents of Madrid, Spain, and 4 large surrounding cities holding a national health card. Participants were recruited in 2015-2017 using a random sampling stratified by sex and district. The present study was conducted from December 2015 to June 2017. In 2019, a new data collection was conducted, updating baseline information and adding new measures, including the assessment of hearing function. All study participants provided written informed consent, and the clinical research ethics committee of La Paz University Hospital in Madrid approved the study.

At baseline and follow-up, data were collected in 3 stages: (1) telephone interview for lifestyles, morbidity, health status, and use of health care information; (2) a first home visit to perform a physical examination, including an audiogram, and collect blood and urine samples; and (3) a second home visit to collect information on habitual diet and place an accelerometer on the wrist of the participants. The procedures, instruments, and questionnaires were like those used in the Seniors-ENRICA I cohort.²⁵ In particular, for the performance of the audiograms, interviewers were provided with a specific protocol for an optimal realization, and training sessions were programmed.

A total of 1894 participants provided data in 2019. We selected those who had a hearing assessment that followed the specified protocol and information for the covariates of interest so that analyses were performed with 1644 persons. Participants who refused the audiologic examination were older, had more comorbidities, and had a greater prevalence of social isolation than participants who agreed to be examined.Hearing Assessment

Hearing was assessed by measuring air conduction thresholds using a hearing test at frequencies 0.5, 1, 2, 3, 4, and 8 kHz in both ears. The hearing test was performed with AudCal, an application for iPhone and iPad. The evaluation was carried out face to face with the evaluator in a quiet environment, with the mobile screen visible only to the evaluator. The headphones used were wired in-ear headphones, distributed in iPhone packages (EarPods). The earbuds were fitted under the standard headphones. We began evaluating the frequency of 1 kHz at 0 dB HL, increasing the sound in 5-dB HL intervals until the participant started hearing the stimulus. After the hearing threshold was identified at that frequency, the other frequencies were evaluated for that ear. The same procedure was carried out for the other ear. This application has shown high sensitivity and specificity in both ears with respect to the standard test (tonal audiometry in a soundproof booth) and has shown a high intraclass correlation (r = 0.93) with the standard evaluation using an International Organization for Standardization standard audiometer and standard headphones in the Spanish population.²⁶

To determine hearing capacity, we calculated 3 PTAs according to different frequency ranges, with the first using the standard PTA definition (0.5, 1, and 2 kHz), the second considering the speech frequency (0.5, 1, 2, and 4 kHz), and the third considering the high frequency (3, 4, and 8 kHz). We defined HL in relation to 3 cutoff points according to the American Speech-Language-Hearing Association: greater than 15 dB HL indicating slight to profound HL; greater than 25 dB HL reflecting mild to profound HL; and greater than 40 dB HL indicating moderate to profound HL.²⁷ We considered the hearing threshold of the better ear following the World Health Organization recommendations.²⁸ We used the cutoff point greater than 40 dB HL in the better ear for all frequency ranges in our analyses.Physical Function, Frailty Syndrome, and Disability

Physical function was measured using the Short Physical Performance Battery (SPPB), which assesses gait speed, the degree of ability to rise from a chair, and balance assessment. Gait speed was calculated as the shortest time, in seconds, to complete a walking distance of 2.44 m at a normal pace twice. The ability to rise from a chair was evaluated by asking the participants to stand up and sit down 5 consecutive times without using their hands. For the standing balance test, participants were asked to stand in 3 progressively challenging positions. Each component was scored on a 4-point scale, and the total SPPB score was calculated by the sum of the components ranging from 0 (worst) to 12 (best). Impaired lower extremity function was defined as a total score less than or equal to 6 points.²⁹

The frailty syndrome was assessed according to the Fried criteria,³⁰which defines frailty as the presence of at least 3 of the following factors: (1) unintentional weight loss greater than or equal to 4.5 kg in the preceding year; (2) exhaustion based on an affirmative response to any of the following questions from the Centre for Epidemiologic Studies Depression Scale: I felt that anything I did was a big effort or I felt that I could not get going at least 3 or 4 days a week; (3) low physical activity, defined as walking less than or equal to 2.5 hours/week for men and less than or equal to 2.0 hours/week for women; (4) slow walking speed, defined as the lowest cohort-specific quintile of gait speed over 2.44 m, adjusted for sex and height; and (5) muscle weakness, set as the cohort-specific lowest quintile of grip strength measured with a Jamar dynamometer in the dominant hand, adjusted for sex and body mass index.

To evaluate disability in instrumental activities of daily living (IADL), we used the Lawton and Brody scale, which evaluates complex everyday functional competencies, such as shopping, doing housework, using the telephone, doing laundry, preparing meals, using public transportation, managing money, and taking medications.³¹ Each domain was rated dichotomously (0, capable in some degree; 1, incapable). Owing to the idiosyncrasy of this population, the score for men was calculated without considering the housework, laundry, and preparing meals tasks, with the total ranging from 0 to 5, whereas for women, the score included all of the tasks and ranged from 0 to 8. The score was directly proportional to the degree of dependence; disability was defined when there was a need of assistance for performing 2 or more IADL.³²Other Variables

We collected self-reported information on age, educational level, and tobacco and alcohol consumption. For the measurement of physical activity (metabolic-equivalent tasks per hour/week), an accelerometer (ActiGraph GT9X; ActiGraph LLC) was used, and participants were asked to use the accelerometer for 7 consecutive days.³³ We measured body mass index (calculated as weight in kilograms divided by height in meters squared) under standardized conditions. Diet quality was evaluated according to adherence to the Mediterranean Diet using the Mediterranean Diet Adherence Screener; the score ranges from 0 to 14, and a higher score reflects greater adherence to this diet.³⁴ The hours of sleep were evaluated with the question: Can you tell me, approximately, how long you usually sleep? To assess sedentary behaviors, we considered information about television viewing (hours/week). In addition, we collected the number of habitual drug treatments currently used. Because the consumption of ototoxic medication was low (<1.1% reported consumption of aspirin, acetaminophen, or ibuprofen), we added these drugs to the total number. We defined hypertension as systolic blood pressure greater than or equal to 140 mm Hg, diastolic blood pressure greater than or equal to 90 mm Hg, or receiving antihypertensive medication. In addition, diabetes was defined as fasting glucose levels greater than or equal to 126 mg/dL (to convert to millimoles per liter, multiply by 0.0555) or use of antidiabetic medication.

Participants reported if they had received a physician-based diagnosis of cancer, cardiovascular diseases (myocardial infarction, stroke, heart failure, or atrial fibrillation), and musculoskeletal diseases (arthritis, osteoarthritis, or hip fracture). We also evaluated the cognitive status of the participants with the Mini-Mental State Examination,³⁵ in which orientation, memory, fixation, calculation, and language construction were measured; impaired cognition was defined as a score lower than 23. In addition, we evaluated social participation through 5 questions: How often do you see or talk on the telephone with family members (other than those who live with you)? How often do you see or talk on the telephone with friends or neighbors? How much time do you usually spend alone at home? How often do you attend church or other religious services? How often do you attend senior club meetings, centers, or associations to which you belong? Each response was scored between 1 and 5. We then added the score of each question and created a scale with a range from 5 to 25, with a higher score indicating greater social participation.³⁶ Social isolation was defined as a score below the median (<17).Statistical Analysis

We assessed differences in sociodemographic characteristics, lifestyles, comorbidities, cognitive function, and social participation between the categories of hearing status. The unpaired t test or χ² test was used to compare continuous or categorical variables as appropriate across the categories of HL. Then, we used logistic regression to examine the association of moderate to profound HL at standard PTA, speech frequency, and high-frequency PTA with impaired lower extremity function, frailty syndrome, and IADL disability. The estimates of the associations are expressed as odds ratios (ORs) and 95% CIs. We built 3 logistic regression models: (1) adjusted for age and sex; (2) additionally adjusted for educational level (primary or less, secondary, and university), smoking status (current, former, and never), current alcohol consumption, physical activity (tertiles of metabolic-equivalent tasks per hour per week), body mass index (tertiles), Mediterranean Diet Adherence Screener (tertiles of the score), hours of daily sleep (tertiles), time viewing television (tertiles of hours per week), and number of habitual drug treatments; and (3) further adjusted for hypertension, cancer, diabetes, cardiovascular diseases, musculoskeletal diseases, impaired cognitive function, and social isolation. We also modeled the association of the continuous PTA (per 5-dB HL increment) in association with the studied outcomes. An analysis between HL and individual subscales of the SPPB was performed; the cutoff point to define difficulty to rise from a chair, slow gait, and balance impairment was a score less than or equal to 3 in each scale.

As a sensitivity analysis, we replicated the analyses defining HL as PTA less than 25 dB HL in the better ear in the 3 frequency ranges to understand whether the association varied for milder degrees of severity of HL. The analyses were performed with Stata, version 15.0 (StataCorp LLC).Results

Among 1644 study participants, 831 were men (50.5%) and 813 were women (49.5%); the mean (SD) age was 73.8 (4.3) years (range, 66-91 years). The frequency of HL according to the different definitions and cutoff points is presented in Table 1. We observed a high frequency of slight and mild HL in the 3 types of frequencies (1308 [79.6%] to 1498 [91.1%] of the participants had some degree of HL). For moderate hearing loss, the prevalence at standard PTA was 9.2% for the total population (9.9% for men and 8.6% for women); at speech frequency PTA, the prevalence was 13.6% (15.2% for men and 12.0% for women); and at high-frequency PTA, the prevalence was 45.1% (50.5% for men and 39.7% for women).

The participants’ characteristics according to hearing status are presented in Table 2. In comparison with participants with normal hearing, those with moderate to profound HL were older, reported lower levels of physical activity, had lower adherence to the Mediterranean diet, and spent more time watching television. In addition, they were receiving more drugs and showed a higher prevalence of diabetes, cardiovascular diseases, and impaired cognitive function.

The association of HL with impaired lower extremity function, frailty syndrome, and IADL disability is presented in Table 3. For standard PTA, HL was associated with all 3 outcomes considered in the models adjusted for age and sex. The associations were slightly modified after further adjustment for sociodemographic and lifestyle characteristics, comorbidities, impaired cognitive function, and social isolation: for impaired lower extremity function, the OR was 2.20 (95% CI, 1.25-3.88); for the frailty syndrome, 1.85 (95% CI, 0.98-3.49); and for the IADL disability, 2.25 (95% CI, 1.29-3.94). With use of speech frequency PTA as the variable, HL was associated with the outcomes, and the multivariable adjustment slightly modified the estimates for impaired lower extremity function (OR, 2.59; 95% CI, 1.57-4.28), frailty syndrome (OR, 1.85; 95% CI, 1.06-3.22), and IADL disability (OR, 2.18; 95% CI, 1.32-3.60). In addition, when we used high-frequency PTA, associations were found between HL and frailty and IADL disability in the less-adjusted models; however, when we adjusted for additional confounders, no association was observed.

Continuous PTA in association with the outcomes showed a direct association (eTable 1 in the Supplement). More specific analyses examining the association for the subscales in the SPPB showed that difficulty to rise from a chair and balance impairment were associated with HL after adjustment for all confounders (eTable 2 in the Supplement). In addition, when we focused on mild HL, the associations were similar compared with those found for moderate HL (eTable 3 in the Supplement).Discussion

In this study of community-dwelling older adults, moderate to profound HL at standard and speech frequency PTA was associated with impaired lower extremity function and IADL disability independently of sociodemographic and lifestyle characteristics, comorbidities, cognitive impairment, and social isolation. Moderate to profound HL at speech frequency PTA was also associated with the frailty syndrome after adjustment for these confounders. Similar results were found for milder hearing loss.

In this population, 1308 (79.6%) to 1498 (91.1%) (depending on the PTA used) of the participants had some degree of HL. These figures are higher than the most recent data from the Spanish National Health Survey 2017,³⁷ where 34% of people aged 65 to 74 years, 49% of people aged 75 to 84 years, and 70% of people aged 85 years or older reported some hearing impairment. This discrepancy reflects the fact that self-reported hearing impairment may underestimate the degree of HL, particularly for milder loss. On the other hand, the prevalence of moderate to profound HL in our study (9.2% for standard PTA, 13.6% for speech frequency, and 45.1% for high-frequency PTA) was lower than in the US among people aged 70 years or older (16.5% for standard PTA, 26.5% for speech frequency PTA, and 74.1% for high-frequency PTA).³⁸ Reasons for the lower prevalence in the Spanish population are unknown.

Regarding the association between HL and impaired lower extremity function, our results were consistent with those found by Bang et al.¹⁶ In a cross-sectional study defining PTA at 0.5, 1, 2, and 3 kHz, the investigators found that HL (with a hearing threshold >40 dB HL) was linked to postural instability, which is a key component of the SPPB score. Furthermore, our results are consistent with the Health, Aging and Body Composition study, a prospective cohort of 2190 participants, in which moderate HL was associated with lower SPPB.³⁸ In addition, using data from the US National Health and Nutrition Examination Survey, Lin and Ferrucci¹⁷ observed an association between HL and higher risk of falls. Martinez-Amezcua et al^19,20 analyzed data from the recent Baltimore Longitudinal Study of Aging¹⁹ and the Atherosclerosis Risk in Communities²⁰ studies. The longitudinal analyses noted that participants with HL presented faster decreases in physical function over time compared with those with normal hearing. The Atherosclerosis Risk in Communities study found that gait speed and balance were independently associated with hearing impairment.²⁰ This result is different from our findings, in which the chair stand test showed a significant association, but gait speed did not.

We also observed an association between HL and the frailty syndrome, as did a meta-analysis by Tan et al.²¹ However, although the meta-analysis included studies at low and moderate risk of bias, more than half of the studies used self-reported measures of HL, which could be susceptible to recall bias. Also, only 2 studies used audiometric measures to determine HL and none considered lifestyles, comorbidities, and social isolation simultaneously in their analyses. Regarding the association between HL and IADL disability, our results are consistent with the systematic review carried out by Lin et al.²³ We have extended the results of these other studies by including comorbidities, cognitive function, and social isolation in our analyses of this association.

Hearing loss in older adults can be due to multiple causes that affect the peripheral auditory system. Among the most common causes are degenerative processes associated with age.³⁹ For example, adults with chronic vestibular loss have shown gait deficits.⁴⁰ Vestibular function is responsible for balance, a key component of physical function. Thus, the age-related deterioration of the hair cells that participate in both the vestibular system and auditory system may imply that HL is a surrogate of vestibular dysfunction,⁴¹ and the association observed in this study may not be causal. Other mechanisms include the close relationship between movement and coordination with acoustic input from the environment; hearing difficulty would impede an appropriate physical response.¹⁵ The presence of comorbidities may also indicate that several underlying common mechanisms are damaged, such as the modulation of response to stress, impaired immune response, and impaired cardiometabolic function.⁴² Further noncausal explanations may include the social isolation related to HL, which can lead to a reduction in physical activity and a subsequent deterioration of physical function.¹⁵ We adjusted our analyses for the presence of comorbidities as well as social isolation in an effort to exclude these mechanisms.Limitations

The main limitation of our study was the cross-sectional design, so we could not attribute directionality to the observed associations. In addition, although we used PTA to assess hearing function, which has been proposed as the optimal metric in both clinical settings and epidemiologic studies,⁴³ PTA was performed with a portable device in the at-home environment. Therefore, bias related to environmental noise and interviewer performance, as well as bias related to the instrument, cannot be eliminated. Moreover, the number of cases of HL for standard PTA was small. In addition, we did not measure vestibular function, and it would have been relevant to determine its association with the study findings. On the other hand, well-established measurement tools, such as the SPPB and the IADL scale, were used, as well as the Fried criteria to define the frailty syndrome. Also, the analyses were adjusted for main potential confounders, including comorbidities, cognitive impairment, and social isolation, which suggests that hearing impairment may be a predictor of deterioration of physical function. Furthermore, given that participants who agreed to perform audiometry were different from those who refused the test, the results obtained cannot be extrapolated to a general population of community-dwelling older adults.Conclusions

In this cross-sectional study, HL was associated with impaired lower extremity function, the frailty syndrome, and IADL disability, which are all common conditions in older adults. Longitudinal studies are necessary to further support associations between HL and these outcomes.