Abstract
Objectives
There is a strikingly high prevalence of sensorineural hearing loss among patients with chronic kidney disease, with estimates ranging from 36% to 77%; however, longitudinal data are limited. We assessed whether lower baseline estimated glomerular filtration rate calculated using creatinine (eGFRCr), as well as decline in eGFRCr over time, were associated with incident hearing loss.
Methods
Serum creatinine was measured in 1,843 individuals aged 48 to 80 years without hearing loss at the start of the Epidemiology of Hearing Loss Study in 1993. Follow-up creatinine assessments were conducted at 5 (n=1,526) and 10 (n=1,095) years. Hearing tests were conducted at baseline, and at 5-, 10-, and 15-year follow-up visits. The risk of hearing loss was assessed as a function of baseline eGFRCr, as well as a function of a 20% decline in eGFRCr, between baseline and 5 years, and between 5 and 10 years. Cox proportional hazards regression was used to examine the risk of incident speech frequency hearing loss, defined as pure-tone average >25 dB HL for thresholds at 0.5, 1, 2, and 4 kHz (PTA0.5,1,2,4) in either ear.
Results
During 15,676 person-years of follow up, there were 802 cases of incident hearing loss. There was no statistically significant association between lower baseline eGFRCr and risk of incident hearing loss. Decline in eGFRCr was also not associated with incident hearing loss at speech frequencies.
Conclusions
Overall, there was no significant association between eGFRCr or decline in eGFRCr using the serum creatinine-based equation and risk of incident hearing loss.
Keywords: chronic kidney disease, hearing loss, creatinine, estimated glomerular filtration rate
INTRODUCTION
Hearing loss is an exceedingly common condition, affecting nearly two thirds of adults aged 70 years and older in the United States.1 Since the 1970s, small cross-sectional studies have suggested there may be a causal association between chronic kidney disease (CKD) and hearing loss, with studies reporting a frequency of 36% to 77%.2,3 Animal models show an inverse correlation between serum creatinine level (SCr) and Na+/K+ ATPase activity in the inner ear; as this enzyme is important for maintaining the electrochemical gradient in the ear, its inhibition may be a primary contributor to inner ear dysfunction in individuals with chronic kidney disease.4 CKD is also associated with inflammation and endothelial dysfunction, which may promote hearing loss due to disturbances in the cochlear microcirculation.5
Data on the temporal association between renal function and hearing loss are limited. Investigators from the Epidemiology of Hearing Loss Study (EHLS) previously showed that lower estimated glomerular filtration rate calculated using cystatin C (eGFRCysC) was significantly associated with a higher cumulative incidence of hearing loss (multivariable-adjusted HR, 1.50; 95% CI, 1.02–2.22; <60 vs ≥60 mL/min/1.73 m2).6 In that study, hearing loss was defined as pure-tone average >25 decibels hearing loss (dB HL) for thresholds at frequencies of 0.5, 1, 2, and 4 kHz (PTA0.5,1,2,4) in either ear. Though Cystatin C (CysC) has been shown to be a good predictor of adverse outcomes, including cardiovascular disease, heart failure, and mortality,7–9 it may be a marker of adverse outcomes independent of renal function. CysC is influenced by a number of factors other than renal function, including age, smoking, inflammation, and medical conditions, such as liver disease and thyroid disease.10–13 SCr is more widely available than CysC, and it has been validated across several GFR estimating equations and patient populations.
While there are data on CysC and its relation with hearing loss, there are no longitudinal studies on the association between eGFR using serum creatinine (eGFRCr), and hearing loss. Furthermore, no studies have assessed whether change in renal function is associated with incident hearing loss. We therefore examined the relation between baseline eGFRCr and the risk of hearing loss among EHLS participants. Secondary aims were to assess whether eGFRCr was associated with high frequency or low frequency hearing loss; if baseline estimated glomerular filtration rate calculated using SCr and CysC (eGFRCr-CysC) was associated with incident hearing loss; and whether a steeper decline in eGFRCr over time was associated with a higher risk of hearing loss.
MATERIALS AND METHODS
Study Participants
Starting in 1988, the population-based Beaver Dam Eye Study (BDES) enrolled residents of the city or township who were aged 43–84 years old to collect information related to the incidence of common eye diseases causing loss of vision. The EHLS subsequently invited participants from the BDES to participate in a longitudinal study to examine the incidence and progression of hearing loss. Of the 4,541 eligible individuals from BDES, 3,753 (82.6%) participated in EHLS-1 in 1993. The mean age of the participants was 65.8 years at baseline in EHLS-1. Hearing tests were conducted at baseline, and follow-up studies were conducted at 5 years (EHLS-2), 10 years (EHLS-3), and 15 years (EHLS-4). In the primary analysis, participants with baseline hearing loss at speech frequencies, defined as PTA0.5,1,2,4 > 25 dB HL in either ear, were excluded from the analysis (see Figure 1). Of the remaining 1,925 participants, serum creatinine was measured in 1,847 individuals at EHLS-1. Participants with a history of kidney transplant (n=1) or a history of dialysis (n=3) were excluded, with 1,843 participants remaining. Follow-up creatinine assessments were conducted at 5 (n=1,526) and 10 (n=1,095) years. CysC levels were measured in 863 participants at EHLS-1.
Figure 1:
Flow Diagram of EHLS Participants Included in Each Stage of the Analysis
Abbreviations: EHLS= Epidemiology of Hearing Loss Study; HL= hearing loss; SCr= serum creatinine; Hx=history
When examining the risk of incident high frequency hearing loss as a secondary outcome, those participants with high frequency hearing loss at baseline (defined as PTA4,6,8> 25 dB HL in either ear) were excluded (n=1,210). Similarly, for the outcome of low frequency hearing loss, those with low frequency hearing loss at baseline (defined as PTA0.5,1,2> 25 dB HL in either ear) were excluded (n= 19).
In this population-based study, participants were not excluded based on any underlying health conditions, including cancer.
Estimating Glomerular Filtration Rate
SCr was measured in blood samples collected at baseline examinations in EHLS-1 (1993–1995) and again at EHLS-2 in 1998. Plasma creatinine (rather than SCr) was measured in samples collected in 2003 at EHLS-3. Serum and plasma creatinine levels were used to estimate GFR using the CKD-EPI equation, which has been found to have greater precision and accuracy than other GFRCr -estimating equations, particularly at higher levels of eGFR.14 The eGFRCr at each of the three time points—baseline, EHLS-2, and EHLS-3— were then categorized into clinically relevant cut-points: 0 to <60, 60 to <90, and ≥90 mL/min/1.73 m2. The percent change in the eGFRCr was also calculated for each 5-year follow-up period.
Studies suggest that estimating GFR with an equation that includes both SCr and CysC may more accurately and precisely reflect measured GFR than either marker by itself.15–17 We therefore also examined whether eGFRCr-CysC was associated with incident hearing loss among those participants who had CysC levels measured at baseline.
Hearing Evaluation
Hearing was tested at baseline, and at 5-, 10-, and 15-year follow-up examinations using a standardized protocol.18–21 Participants underwent detailed audiometric testing, including otoscopy, screening tympanogram, and pure tone audiometry at multiple frequencies. Hearing tests were conducted according to the guidelines of the American Speech-Language-Hearing Association.22 Participants who were unable to travel to the clinic site (nursing home residents, home-bound participants, and those living in remote areas), were tested in their own homes or facilities using a portable audiometer. Audiometers were calibrated every 6 months in accordance with American National Standards Institute recommendations.23,24
Pure tone air-conduction thresholds were obtained at 500, 1000, 2000, 3000, 4000, 6000, and 8000 Hz in each ear. Bone-conduction thresholds were measured at two frequencies at baseline (500 and 4000 Hz) and three at follow-up (500, 2000, and 4000 Hz). Masking was used as necessary. Conductive hearing loss was defined as an air-bone of 15dB or greater at any frequency. Sensitivity analyses excluded participants who would have normal pure tone average if the conductive loss resolved (n=99). Of the 1,843 participants at baseline, 20 had a history of Meniere’s, and 3 had a history of otosclerosis. The primary outcome was incident hearing loss at EHLS-2, EHLS-3, or EHLS-4, defined as PTA0.5,1,2,4 kHz > 25 dB HL in either ear. Secondary outcomes included a) incident high-frequency hearing loss, defined as PTA4,6,8 > 25 dB HL in either ear, and b) incident low-frequency hearing loss, defined as PTA0.5,1,2 >25 dB HL in either ear. We also examined whether baseline eGFRCr or a 20% decline in eGFRCr was associated with a difference of the change in PTA at speech frequencies, high frequencies, and low frequencies.
Covariates
Information on demographics and medical history was obtained from baseline questionnaires administered to EHLS-1 participants in 1993, and from follow-up questionnaires at EHLS-2 and EHLS-3. We considered demographic factors, including age, race, sex, income, education, employment status, and presence or absence of health insurance. We considered overall health status, calculated based on results from the Short Form 36 Health Survey, where participants’ responses were scored from 0 to 100 (higher scores indicate better reported health). We also considered regular exercise, defined as self-reported exercise more than once a week (yes/no); smoking (categorized into never smokers who reported smoking fewer than 100 cigarettes in their lifetime, past smokers who smoked at least 100 cigarettes but were no longer smoking at the time of exam, and current smokers); and alcohol consumption (self-reported based on amount of beer, wine, and liquor consumed over the course of a week). We considered medication use (yes/no), including self-reported use of acetaminophen, aspirin, and non-steroidal anti-inflammatory drugs (NSAIDs), post-menopausal hormone therapy (both oral and non-oral agents), loop diuretics, and statins; and medical conditions, including history of cardiovascular disease, epilepsy, head injury, diabetes mellitus, or hypertension. Diabetes classification (yes/no) incorporated information from self-reported physician diagnosis, measured glycated hemoglobin A1c (GHB/A1c) level, and medication history. Hypertension classification (yes/no) was based on information from measured blood pressure and self-reported use of anti-hypertensive medications. Finally, we considered measured body mass index (BMI), waist circumference, and cholesterol levels.
Statistical Analysis
Prospective analyses were performed. In the primary analysis, eGFRCr at baseline was categorized as <60 mL/min/1.73 m2, 60 to <90 mL/min/1.73 m2 (referent group), and ≥ 90 mL/min/1.73 m2. The risk of incident hearing loss was compared among participants in each category of eGFRCr. Participants were censored at the reported onset of hearing loss. We used Cox proportional hazards regression to examine whether baseline eGFRCr was associated with incident hearing loss at speech frequencies (PTA0.5,1,2,4 > 25 dB HL), high frequencies (PTA4,6,8> 25 dB), and low frequencies (PTA0.5,1,2 > 25 dB). We also assessed whether baseline eGFRCr-CysC was associated with incident hearing loss over 15-year follow-up.
In secondary analyses, we examined whether having a 20% decline in eGFRCr between EHLS-1 and EHLS-2; between EHLS-2 and EHLS-3; or both, was associated with an increased risk of hearing loss at speech frequencies, high frequencies, and low frequencies. We also used multivariable-adjusted linear regression to examine whether baseline eGFRCr and a 20% decline in eGFRCr were associated with differences of the change in PTA at speech frequencies, high frequencies, and low frequencies.
In the primary analysis, we considered covariates that could be potential confounders based on clinical knowledge and removed those that were not found to be statistically significant using backward selection, and setting a p-value threshold of <0.10. For consistency, we retained the same covariates in all models, including the linear regression models: age, sex, total cholesterol, smoking, waist circumference, education, NSAID use, loop diuretic use, hypertension, and diabetes mellitus. We calculated 95% confidence intervals for all hazard ratios. P values were all two-sided. SAS software, version 9.4 (SAS Institute, Inc., Cary, North Carolina) was used for all statistical analyses.
RESULTS
The characteristics of EHLS participants at baseline according to category of baseline eGFRCr are shown in Tables 1. Those with lower eGFRCr tended to be older, with lower educational attainment, and were more likely to have higher BMI, waist circumference, and cholesterol levels. They were more likely to have hypertension, diabetes, cardiovascular disease, and to use loop diuretics, aspirin, and NSAIDs.
Table 1:
Baseline Characteristics of EHLS Participants in 1993 According to Categories of eGFRCr Among Those at Risk for Hearing Loss
| 1 eGFRCr <60 (n=182) |
2 60≤ eGFRCr <90 (n=1052) |
3 eGFRCr ≥90 (n=609) |
|
|---|---|---|---|
| Age, years | 66.6 (8.9) | 62.0 (8.4) | 56.0 (5.8) |
| Male | 39 (21.4) | 365 (34.7) | 213 (35.0) |
| BMI (kg/m2) | 30.9 (5.8) | 29.6 (5.4) | 29.2 (5.9) |
| Waist Circumference (cm) | 94.6 (13.8) | 93.0 (15.6) | 92.3 (16.4) |
| Total Cholesterol (mg/dL) | 254.7 (48.0) | 239.7 (42.0) | 239.9 (46.7) |
| SCr (mg/dL) | 1.2 (0.2) | 0.9 (0.1) | 0.8 (0.1) |
| Cystatin C (mg/L) | 1.2 (0.3) | 0.9 (0.2) | 0.8 (0.1) |
| Alcohol (gm/week) | 23.5 (55.1) | 43.7 (78.3) | 52.2 (118.9) |
| Education (>12 years) | 46 (25.3) | 397 (37.7) | 238 (39.1) |
| White Race | 182 (100) | 1045 (99.3) | 601 (98.9) |
| Health Insurance | 176 (97.2) | 1025 (97.4) | 576 (94.7) |
| History of CVD | 34 (18.7) | 89 (8.5) | 28 (4.6) |
| History of HTN | 121 (66.5) | 481 (45.7) | 236 (38.8) |
| History of DM | 27 (14.8) | 75 (7.1) | 50 (8.2) |
| History of Kidney Disease | 1 (0.6) | 22 (2.1) | 9 (1.5) |
| Regular Exercise | 57 (31.3) | 470 (44.7) | 251 (41.2) |
| Smoking | |||
| Never | 104 (57.1) | 530 (50.5) | 252 (41.5) |
| Past | 63 (34.6) | 393 (37.4) | 220 (36.2) |
| Current | 15 (8.2) | 127 (12.1) | 136 (22.4) |
| Excellent or Very Good General Health | 142 (78.0) | 935 (88.9) | 545 (89.4) |
| *Acetaminophen Use | 73 (40.1) | 369 (35.1) | 245 (40.2) |
| *Aspirin Use | 80 (44.0) | 412 (39.2) | 220 (36.1) |
| *NSAID Use | 119 (65.4) | 610 (58.0) | 374 (61.4) |
| *Post-Menopausal Hormone Use (oral and non-oral) | 24 (13.2) | 135 (12.8) | 102 (16.8) |
| *Loop Diuretic Use | 19 (10.4) | 21 (2.0) | 15 (2.5) |
| *Statin Use | 11 (6.0) | 45 (4.3) | 19 (3.1) |
Note: Values represent baseline characteristics in first time period. Values are mean (SD) for continuous variables or n (%) for categorical variables.
Abbreviations: EHLS=Epidemiology of Hearing Loss Study; eGFRCr =estimated glomerular filtration rate from creatinine; BMI= body mass index; SCr= serum creatinine; CVD= cardiovascular disease; HTN= hypertension; DM= diabetes mellitus; NSAID=non-steroidal anti-inflammatory drugs; cm=centimeters; mg=milligrams; dL=deciliters; L=liters; gm= grams
Units for eGFRCr = mL/min/1.73 m2
Self-reported use of medications (yes/no)
Baseline eGFRCr and Incident Hearing Loss
During 15,676 years of person-years of follow-up, there were 802 incident cases of speech frequency hearing loss, defined as PTA0.5,1,2,4 >25 dB HL in either ear. The age-adjusted and multivariable-adjusted HRs for hearing loss according to baseline eGFRCr are shown in Table 2. There was no significant association between baseline eGFRCr and 15-year risk of hearing loss at speech frequencies. In the Cox regression analyses for high frequency and low frequency hearing loss, there was also no significant association between baseline eGFRCr and incident hearing loss. When eGFRCr was analyzed as a continuous predictor, there was similarly no association between lower eGFRCr and incident hearing loss (HR 1.00, 95% CI 0.99–1.01; p-value 0.73). Age was the most important confounder of the relation between eGFRCr and hearing loss. When conducting sensitivity analyses excluding participants who would have normal pure tone average if the conductive loss resolved (n=99), the HR were minimally changed (<10%) and remained non-significant.
Table 2:
Age-Adjusted and Multivariable-Adjusted Risk of Hearing Loss by Baseline eGFRCr Among EHLS Participants, 1993 to 2010
| PTA>25 dB HL at Speech Frequencies (0.5,1,2,4 kHz) | |||
|---|---|---|---|
| eGFRCr <60 (n=248) |
60≤ eGFRCr <90 (n=1,914) |
eGFRCr ≥90 (n=1,243) |
|
| Cases (n) | 86 | 495 | 221 |
| Person-years | 1,060 | 8,674 | 5,942 |
| Age-adjusted HR, CI | 1.03 [0.81–1.30] |
1.00 (ref) |
1.04 [0.87–1.24] |
| *Multivariable-Adjusted HR, CI | 1.04 [0.81–1.32] |
1.00 (ref) |
1.07 [0.90–1.28] |
| PTA >25 dB HL at High Frequencies (4,6,8 kHz) | |||
| eGFRCr <60 (n=52) |
60≤ eGFRCr <90 (n=485) |
eGFRCr ≥90 (n=414) |
|
| Cases (n) | 25 | 247 | 177 |
| Person-years | 201 | 1,846 | 1,666 |
| Age-adjusted HR, CI | 0.90 [0.57–1.41] |
1.00 (ref) |
0.98 [0.79–1.21] |
| *Multivariable-Adjusted HR, CI | 0.90 [0.57–1.42] |
1.00 (ref) |
1.02 [0.81–1.27] |
| PTA >25 dB HL at Low Frequencies (0.5,1,2 kHz) | |||
| eGFRCr <60 (n=260) |
60≤ eGFRCr <90 (n=1,931) |
eGFRCr ≥90 (n=1,120) |
|
| Cases (n) | 47 | 249 | 114 |
| Person-years | 1,224 | 9,387 | 6,148 |
| Age-adjusted HR, CI | 0.97 [0.70–1.35] |
1.00 (ref) |
1.23 [0.96–1.57] |
| *Multivariable-Adjusted HR, CI | 0.88 [0.63–1.22] |
1.00 (ref) |
1.19 [0.92–1.53] |
Abbreviations: EHLS=Epidemiology of Hearing Loss Study; eGFR= estimated glomerular filtration rate; HR= hazard ratio; CI= confidence interval; dB = decibels
Units: mL/min/1.73 m2
Outcome: incident hearing loss, defined as PTA >25 decibels at 0.5, 1, 2 and 4 kilohertz
Covariates in MV-adjusted model: sex, total cholesterol, smoking, waist circumference, education, NSAID use, loop diuretic use, hypertension, diabetes mellitus
In secondary analyses, we evaluated the association between baseline eGFRCr and differences of the change in PTA at speech frequencies, high frequencies, and low frequencies from EHLS-1 to EHLS-4 (Table 3). Overall, there was a mean increase of 10.0 dB (SD 7.3) in PTA at speech frequencies over 15 years. There was no significant difference among the categories of baseline eGFRCr and change in mean PTA at speech frequencies, high frequencies, or low frequencies after multivariable adjustment.
Table 3:
Multivariable-Adjusted Difference of the Change in PTAs, 1993–2010
| Change in PTA at Speech Frequencies (0.5,1,2,4 kHz) | ||
|---|---|---|
| eGFRCr <60 | eGFRCr ≥90 | |
| Change | −1.4 | 0.4 |
| 95% CI | −3.6 to 0.8 | −0.7 to 1.4 |
| P-value | 0.21 | 0.51 |
| Change in PTA at High Frequencies (4,6,8 kHz) | ||
| eGFRCr <60 | eGFRCr ≥90 | |
| Change | −4.4 | 1.5 |
| 95% CI | −15.6 to 6.7 | −2.4 to 5.3 |
| P-value | 0.43 | 0.74 |
| Change in PTA at Low Frequencies (0.5,1,2 kHz) | ||
| eGFRCr <60 | eGFRCr ≥90 | |
| Change | −1.3 | 0.7 |
| 95% CI | −3.7 to 1.0 | −0.5 to 1.8 |
| P-value | 0.26 | 0.28 |
Reference= eGFRCr 60 to <90
Abbreviations: PTA= pure tone average; eGFR= estimated glomerular filtration rate; CI= confidence interval
Units: mL/min/1.73 m2 (eGFRCr), decibels (PTA)
Covariates in MV-adjusted model: sex, total cholesterol, smoking, waist circumference, education, NSAID use, loop diuretic use, hypertension, diabetes mellitus.
Baseline eGFRCr-Cys
In multivariable-adjusted models, baseline eGFRCr-CysC <60 ml/min/1.73 m2 was associated with a higher risk of hearing loss at speech frequencies (HR 1.36, 95% CI 1.01–1.83; p-value 0.05), compared with an eGFRCr-CysC between 60–89 (Table 4). There was no significant association between eGFRCr-CysC ≥ 90 and risk of hearing loss.
Table 4:
Age-Adjusted and Multivariable-Adjusted Risk of Hearing Loss by Baseline eGFRCr-CysC Among EHLS Participants, 1993 to 2010
| Baseline eGFRCr-CysC (ml/min/1.73 m2) | |||
|---|---|---|---|
| <60 | 60 to <90 | ≥90 | |
| Cases (n) | 59 | 265 | 130 |
| Person-years | 468 | 4,330 | 3,323 |
| Age-adjusted HR, CI | 1.39 [1.04–1.87] |
1.00 (ref) |
1.02 [0.82–1.27] |
| *Multivariable-Adjusted HR, CI | 1.36 [1.01–1.83] |
1.00 (ref) |
0.99 [0.79–1.23] |
Abbreviations: EHLS=Epidemiology of Hearing Loss Study; eGFRCr-CysC = estimated glomerular filtration rate measured by creatinine-cystatin C; HR= hazard ratio; CI= confidence interval
Units: mL/min/1.73 m2
Outcome: incident hearing loss, defined as PTA >25 decibels at 0.5, 1, 2 and 4 kilohertz
Covariates in MV-adjusted model: sex, total cholesterol, smoking, waist circumference, education, NSAID use, loop diuretic use, hypertension, diabetes mellitus
Decline in eGFRCr
During 7,707 years of person-years of follow-up, there were 380 cases of speech frequency hearing loss (Supplementary Table S1). There was no overall significant association between 20% decline in eGFRCr and risk of incident hearing loss at speech frequencies, high frequencies, or low frequencies. There was similarly no association between 20% decline in eGFRCr and difference of the change in PTA (Supplementary Table S2).
The association between baseline eGFRCr and 15-year risk of hearing loss did not vary by sex (p-interaction 0.29) or by age <65 versus ≥ 65 (p-interaction 0.43) (Supplementary Tables S3 and S4).
DISCUSSION
In this study of 1,843 EHLS participants, we found no significant association between baseline eGFRCr or decline in eGFRCr and incident hearing loss at speech frequencies. There was also no overall relation between baseline eGFRCr and decline in eGFRCr with the risk of high frequency or low frequency hearing loss. We found that baseline eGFRCr-CysC <60 ml/min/1.73 m2 was associated with a higher risk of hearing loss at speech frequencies.
Our study results contrast with prior longitudinal and cross-sectional studies, which have suggested that individuals with lower eGFRCr may be predisposed to sensorineural hearing loss. Small observational studies have consistently demonstrated a higher prevalence of hearing loss among patients with CKD; however, these studies are limited by sample size, and most have included participants who are already on dialysis.2,3,25 One cross-sectional study of 2,564 individuals in Australia found that those with CKD, defined as an eGFRCr < 60 mL/min/1.73 m2, were 46% more likely to have prevalent mild hearing loss compared with those with an eGFR ≥ 60 mL/min/1.73 m2.26 There was also a significant association between CKD and prevalent severe hearing loss, defined as >40 dB HL (multivariable-adjusted HR 1.50, 95% CI 1.04–2.18; p-value 0.03). Those with an eGFR < 45 mL/min/1.73 m2 had the highest odds of prevalent hearing loss.
Schubert et al. examined the longitudinal association between CysC concentration and eGFRCysC with hearing loss (n=863).6 While a higher concentration of CysC was associated with hearing loss in a model adjusted for age and sex, there was no significant association after adjusting for additional potential confounders. The investigators found that participants with lower eGFRCysC were at higher risk of developing hearing loss, compared with those with better baseline renal function estimated using CysC (multivariable-adjusted HR, 1.50; 95% CI, 1.02–2.22). While CysC has been found in some studies to be a better predictor of adverse outcomes when compared with SCr,15,27–29 it is not clear that it is a better marker of renal function than eGFRCr, particularly when compared with the gold standard of measured GFR.
Several studies have demonstrated that CysC levels are affected by factors other than GFR, suggesting that unmeasured non-GFR determinants may be similar in magnitude for both CysC and SCr.16 For instance, CysC is associated with older age, hemoglobin A1c, smoking, past cardiovascular events, uric acid levels, and hypertriglyceridemia.30 Inflammation is also thought to modulate CysC levels.31 Inflammation has been implicated in animal models of hearing loss; after ototoxic exposures, there is in vivo production of tumor necrosis factor-alpha, interleukin-1β, and interleukin-6, which subsequently induce leukocyte infiltration into the inner ear.32–34 Nash et al. used EHLS data to show that there was a statistically significant association between C-reactive protein (CRP) levels and incident hearing loss in participants <60 years at baseline, but not in adults 60 and above.35 Many of the same conditions associated with hearing loss are also associated with higher circulating levels of CysC, including hypertension, dyslipidemia, and higher BMI.31,36–40 Given our disparate findings, it is likely that CysC is a marker of adverse outcomes including hearing loss, but through mechanisms other than renal function.
Estimation of GFR from CysC alone has been demonstrated to be no more accurate than SCr-based estimates. In a study of renal transplant patients, the eGFRCysC equation was inferior to the performance of eGFRCr when compared to the gold standard of measured GFR, with more bias and less accuracy.30 Stevens et al. found that after adjusting for measured GFR, CysC was 4.3% lower for every 20 years of age and 9.2% lower for female sex.41,42 CysC-based equations were shown to slightly overestimate measured GFR at an eGFR greater than 90 ml/min/1.73 m2, when compared with creatinine-based equations. Furthermore, the precision for CysC-based equations was lower when SCr was not included. Prior studies have demonstrated a wide variability in eGFR for the same level of CysC, which may be related to population differences, or to lack of standardization and calibration of CysC.42,43
When using the combined eGFRCr-CysC equation, we found a significant association between lower baseline eGFRCr-CysC and incident hearing loss at speech frequencies, but not at high frequencies. Some studies suggest that the equation combining SCr and CysC yields a slightly more precise and accurate estimate of GFR, across a range of demographic and clinical characteristics.16 However, using both biomarkers in the eGFR equations separately may provide useful insights, since both CysC and SCr are affected by different non-eGFR determinants.
In our study, age was the most important confounder of the relation between eGFRCr and incident hearing loss. It is well-known that even in healthy persons, eGFR declines with age, though the significance of this has been debated.44 Furthermore, the risk of hearing loss increases with age, with over 2/3 of adults aged 70 or older having some degree of hearing impairment.45 Aging is a shared risk factor for both CKD and hearing loss; it is therefore critical that studies examining the relation between eGFR and hearing loss finely adjust for age.
This is the only prospective study to date that has examined the risk of incident hearing loss in relation to eGFRCr. It is a population-based study with standardized audiometric data measured longitudinally, with >80% follow-up at each time point. There were multiple measurements of creatinine, allowing for assessment of the effect of change in eGFRCr over time. However, this study has limitations. We did not have information on genetic factors, which have been increasingly shown to play an important role in age-related hearing loss.46 While detailed audiograms and tympanograms were performed, bone conduction thresholds were only assessed at two frequencies at baseline and three frequencies at follow-up examinations. We adjusted for diabetes mellitus, which has been associated with hearing loss;47 however, we did not adjust for degree of diabetes control with hemoglobin A1c. Most of the participants were non-Hispanic whites, and we therefore cannot extrapolate the results to other racial groups. We also did not have information on antibiotic use, which could confound the relation between CKD and hearing loss. Few participants had eGFRCr below 45 mL/min/1.73 m2, thereby limiting the power to detect the association between more severe stages of CKD and hearing loss. Furthermore, GFR was estimated rather than directly measured.
Based on the findings from this large longitudinal study, there was no significant association between either lower baseline eGFRCr or decline in eGFRCr and incident hearing loss. Additional studies comprising a greater number of participants with lower eGFRCr are warranted, particularly among those with more severe CKD.
Supplementary Material
Acknowledgements:
The authors would like to thank the participants of the Epidemiology of Hearing Loss Study for their ongoing contributions.
Sources of Support:
This work was supported by the National Institute of Health: 1F32DC017342, DK091417, DC010811, U10EY06594, R37AG021917 and an unrestricted grant from Research to Prevent Blindness.
Footnotes
Conflicts of Interest:
SG: consultant for AstraZeneca
SGC: consultant for Decibel Therapeutics
GCC: grant support from Shoebox Audiometry; consultant for Allena Pharmaceuticals (urine oxalate); consultant for Shire (hypoparathyroidism); consultant for AstraZeneca (hyperkalemia)
The remaining authors have no disclosures or conflicts of interest.
Sponsor’s Role:
The sponsor, GCC, provided guidance for statistical analysis, interpretation of results, manuscript preparation, and presentation of results, and also provided critical revisions.
References
- 1.Lin FR Hearing Loss Prevalence in the United States. Arch. Intern. Med 171, 1851 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bazzi C, Venturini CT, Pagani CT, Arrigo G & D’ Amico G Hearing loss in short- and long-term haemodialysed patients. Nephrol. Dial. Transplant 10, 1865–1868 (1995). [PubMed] [Google Scholar]
- 3.Ozturan O & Lam S The effect of hemodialysis on hearing using pure-tone audiometry and distortion-product otoacoustic emissions. ORL 60, 306–313 (1998). [DOI] [PubMed] [Google Scholar]
- 4.Adler D, Fiehn W & Ritz E Inhibition of Na+, K+-stimulated ATPase in the cochlea of the guinea pig. A potential cause of disturbed inner ear function in terminal renal failure. Acta Otolaryngol 90, 55–60 (1980). [DOI] [PubMed] [Google Scholar]
- 5.Cuna V et al. Hypoacusia and chronic renal dysfunction: New etiopathogenetic prospective. Therapeutic Apheresis and Dialysis 19, 111–118 (2015). [DOI] [PubMed] [Google Scholar]
- 6.Schubert CR, Paulsen AJ, Nondahl DM & et al. Association Between Cystatin C and 20-Year Cumulative Incidence of Hearing Impairment. JAMA Otolaryngol. Neck Surg. doi:doi: 10.1001/jamaoto.2018.0041 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Koenig W, Twardella D, Brenner H & Rothenbacher D Plasma concentrations of cystatin C in patients with coronary heart disease and risk for secondary cardiovascular events: More than simply a marker of glomerular filtration rate. Clin. Chem (2005). doi: 10.1373/clinchem.2004.041889 [DOI] [PubMed] [Google Scholar]
- 8.Shlipak MG et al. Cystatin-C and mortality in elderly persons with heart failure. J. Am. Coll. Cardiol (2005). doi: 10.1016/j.jacc.2004.09.061 [DOI] [PubMed] [Google Scholar]
- 9.Fried LF Kidney Function as a Predictor of Noncardiovascular Mortality. J. Am. Soc. Nephrol (2005). doi: 10.1681/ASN.2005040384 [DOI] [PubMed] [Google Scholar]
- 10.Curhan G Cystatin C: A marker of renal function or something more? Clinical Chemistry (2005). doi: 10.1373/clinchem.2004.044388 [DOI] [PubMed] [Google Scholar]
- 11.Chu SC et al. Increased cystatin C serum concentrations in patients with hepatic diseases of various severities. Clin. Chim. Acta (2004). doi: 10.1016/j.cccn.2003.11.011 [DOI] [PubMed] [Google Scholar]
- 12.Wiesli P, Schwegler B, Spinas GA & Schmid C Serum cystatin C is sensitive to small changes in thyroid function. Clin. Chim. Acta (2003). doi: 10.1016/j.cccn.2003.07.022 [DOI] [PubMed] [Google Scholar]
- 13.Knight EL et al. Factors influencing serum cystatin C levels other than renal function and the impact on renal function measurement. Kidney Int. (2004). doi: 10.1111/j.1523-1755.2004.00517.x [DOI] [PubMed] [Google Scholar]
- 14.Levey AS & Stevens LA Estimating GFR Using the CKD Epidemiology Collaboration (CKD-EPI) Creatinine Equation: More Accurate GFR Estimates, Lower CKD Prevalence Estimates, and Better Risk Predictions. American Journal of Kidney Diseases 55, 622–627 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Shlipak MG et al. Cystatin C versus Creatinine in Determining Risk Based on Kidney Function. N. Engl. J. Med 369, 932–943 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Inker LA et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N. Engl. J. Med (2012). doi: 10.1056/NEJMoa1114248 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schaeffner ES et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann. Intern. Med (2012). doi: 10.7326/0003-4819-157-7-201210020-00003 [DOI] [PubMed] [Google Scholar]
- 18.Cruickshanks KJ et al. The 5-Year Incidence and Progression of Hearing Loss. Arch. Otolaryngol. Neck Surg. 129, 1041 (2003). [DOI] [PubMed] [Google Scholar]
- 19.Cruickshanks KJ et al. Smoking, central adiposity, and poor glycemic control increase risk of hearing impairment. J. Am. Geriatr. Soc 63, 918–924 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Cruickshanks KJ et al. Prevalence of hearing loss in older adults in Beaver Dam, Wisconsin. The Epidemiology of Hearing Loss Study. Am. J. Epidemiol 148, 879–86 (1998). [DOI] [PubMed] [Google Scholar]
- 21.Cruickshanks KJ et al. Education, occupation, noise exposure history and the 10-yr cumulative incidence of hearing impairment in older adults. Hear. Res 264, 3–9 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.American Speech-Language-Hearing Association (ASHA). Guidlines for manual pure-tone threshold audiometry. ASHA 20, 297–301 (1978). [PubMed] [Google Scholar]
- 23.Institute, A. N. S. Maximum Permis-sible Ambient Noise Levels for Audiometric Test Rooms. ANSI S3. 1–1991. New York Am. Natl. {StandardsInstitute}; (2008). doi:https://www.scribd.com/document/356781184/ANSI-ASA-S3-1-1999-R2008 [Google Scholar]
- 24.American National Standards Institute. ANSI/ASA S12.60–2009/Part 2. Acoustical performance criteria, design requirements, and guidelines for schools, part 2: Relocatable classroom factors. America: (2009). doi: 10.1063/1.3056837 [DOI] [Google Scholar]
- 25.Antonelli AR et al. Audiological findings in elderly patients with chronic renal failure. Acta Otolaryngol. Suppl 476, 54–68 (1990). [DOI] [PubMed] [Google Scholar]
- 26.Vilayur E et al. The association between reduced GFR and hearing loss: A cross-sectional population-based study. Am. J. Kidney Dis. 56, 661–669 (2010). [DOI] [PubMed] [Google Scholar]
- 27.Peralta CA et al. Detection of chronic kidney disease with creatinine, cystatin c, and urine albumin-to-creatinine ratio and association with progression to end-stage renal disease and mortality. JAMA - J. Am. Med. Assoc 305, 1545–1552 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Peralta C. a et al. Cystatin C identifies chronic kidney disease patients at higher risk for complications. J. Am. Soc. Nephrol 22, 147–155 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shlipak MG et al. Cystatin C and prognosis for cardiovascular and kidney outcomes in elderly persons without chronic kidney disease. Ann. Intern. Med 145, 237–246 (2006). [DOI] [PubMed] [Google Scholar]
- 30.Keddis MT et al. Creatinine-based and cystatin C-based GFR estimating equations and their non-GFR determinants in kidney transplant recipients. Clin. J. Am. Soc. Nephrol 11, 1640–1649 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Taglieri N, Koenig W & Kaski JC Cystatin C and cardiovascular risk. Clinical Chemistry 55, 1932–1943 (2009). [DOI] [PubMed] [Google Scholar]
- 32.Hirose K, Discolo CM, Keasler JR & Ransohoff R Mononuclear phagocytes migrate into the murine cochlea after acoustic trauma. J. Comp. Neurol 489, 180–194 (2005). [DOI] [PubMed] [Google Scholar]
- 33.Fujioka M, Okano H & Ogawa K Inflammatory and immune responses in the cochlea: Potential therapeutic targets for sensorineural hearing loss. Frontiers in Pharmacology 5, (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Fujioka M et al. Proinflammatory cytokines expression in noise-induced damaged cochlea. J. Neurosci. Res 83, 575–583 (2006). [DOI] [PubMed] [Google Scholar]
- 35.Nash SD et al. Long-term assessment of systemic inflammation and the cumulative incidence of age-related hearing impairment in the epidemiology of hearing loss study. J. Gerontol. A. Biol. Sci. Med. Sci 69, 207–14 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ix JH, Shlipak MG, Chertow GM & Whooley MA Association of cystatin C with mortality, cardiovascular events, and incident heart failure among persons with coronary heart disease: Data from the Heart and Soul Study. Circulation 115, 173–179 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Shlipak MG et al. Cystatin C and the risk of death and cardiovascular events among elderly persons. N. Engl. J. Med 352, 2049–60 (2005). [DOI] [PubMed] [Google Scholar]
- 38.Lin BM et al. Hypertension, Diuretic Use, and Risk of Hearing Loss. Am. J. Med 129, 416–422 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Gates GA, Cobb JL, D’agostino RB & Wolf PA The Relation of Hearing in the Elderly to the Presence of Cardiovascular Disease and Cardiovascular Risk Factors. Arch. Otolaryngol. Neck Surg. 119, 156–161 (1993). [DOI] [PubMed] [Google Scholar]
- 40.Curhan SG, Eavey R, Wang M, Stampfer MJ & Curhan GC Body mass index, waist circumference, physical activity, and risk of hearing loss in women. Am. J. Med 126, 1142.e1–8 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Stevens LA et al. Factors other than glomerular filtration rate affect serum cystatin C levels. Kidney Int. (2009). doi: 10.1038/ki.2008.638 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Stevens LA et al. Estimating GFR Using Serum Cystatin C Alone and in Combination With Serum Creatinine: A Pooled Analysis of 3,418 Individuals With CKD. Am. J. Kidney Dis. 51, 395–406 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Flodin M, Hansson LO & Larsson A Variations in assay protocol for the Dako cystatin C method may change patient results by 50% without changing the results for controls. Clin. Chem. Lab. Med 44, 1481–1485 (2006). [DOI] [PubMed] [Google Scholar]
- 44.Wetzels JFM, Kiemeney LALM, Swinkels DW, Willems HL & Heijer M Den. Age- and gender-specific reference values of estimated GFR in Caucasians: The Nijmegen Biomedical Study. Kidney Int. (2007). doi: 10.1038/sj.ki.5002374 [DOI] [PubMed] [Google Scholar]
- 45.Bainbridge KE & Wallhagen MI Hearing Loss in an Aging American Population: Extent, Impact, and Management. Annu. Rev. Public Health (2014). doi: 10.1146/annurev-publhealth-032013-182510 [DOI] [PubMed] [Google Scholar]
- 46.Dror AA & Avraham KB Hearing Loss: Mechanisms Revealed by Genetics and Cell Biology. Annu. Rev. Genet (2009). doi: 10.1146/annurev-genet-102108-134135 [DOI] [PubMed] [Google Scholar]
- 47.Gupta S, Eavey RD, Wang M, Curhan SG & Curhan GC Type 2 diabetes and the risk of incident hearing loss. Diabetologia [Epub ahea, 1–5 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.