Abstract
Background
Caffeine is a commonly consumed substance that has been thought to play a role in the development of tinnitus, but prospective data are lacking. We prospectively evaluated the association between caffeine intake and self-reported tinnitus in a female cohort.
Methods
Participants were 65,085 women in the Nurses' Health Study II, aged 30 to 44 years and without tinnitus at baseline in 1991, who completed questionnaires about lifestyle and medical history every 2 years and food frequency questionnaires every 4 years. Information on self-reported tinnitus and date of onset was obtained from the 2009 questionnaire, with cases defined as those reporting experiencing symptoms “a few days/week” or “daily.” Multivariable adjusted hazard ratios were calculated using Cox proportional hazards regression models.
Results
At baseline, the mean age of the cohort was 36.3 years and the mean caffeine intake was 242.3 mg/d. After 18 years of follow-up, 5289 incident cases of tinnitus were reported. There was a significant inverse association between caffeine intake and the incidence of tinnitus. Compared with women with caffeine intake less than 150 mg/d (150 mg corresponds to ∼ one 8-ounce cup of coffee), the multivariable adjusted hazard ratios were 0.85 (95% confidence interval, 0.76-0.95) for those who consumed 450 to 599 mg/d and 0.79 (0.68-0.91) for those who consumed 600 mg/d or more.
Conclusions
In this prospective study, higher caffeine intake was associated with a lower risk of incident tinnitus in women.
Keywords
Tinnitus is a highly prevalent condition estimated to affect 50 million Americans, severely disabling 3 million.
1
, 2
In the majority of cases, the precise pathophysiology underlying tinnitus remains unknown, and in the absence of a reversible underlying condition, treatment is generally not highly effective.3
, 4
Preventing the development of the condition may prove to be the best way to reduce its burden on the individual and society. Although some risk factors such as hearing loss and trauma are well established, evidence for other risk factors is scarce.Caffeine is the most commonly used psychoactive substance, frequently consumed in coffee.
5
More than 50% of Americans drink coffee, and the per capita intake is 2 cups per day.6
Although caffeine has been implicated in the development of tinnitus,7
, 8
evidence to support caffeine as a risk factor or exacerbating factor for tinnitus is lacking in the medical literature. However, caffeine cessation or reduction in intake to reduce tinnitus symptoms has been recommended by national primary care and specialty organizations.Mayo Clinic Staff. Tinnitus. 2010 [cited 2012 July 28]; Available at: http://www.mayoclinic.com/health/tinnitus/DS00365/DSECTION=causes. Accessed July 28, 2012.
9
, Insight into Causes and Treatments for Tinnitus. 2010 December 2010 [cited 2012 July 8]. Available at: http://www.entnet.org/HealthInformation/tinnitus.cfm. Accessed July 8, 2012.
10
Of note, a recent randomized controlled crossover trial failed to demonstrate an improvement in tinnitus symptoms with cessation of caffeine intake.11
With the high prevalence of caffeine intake and tinnitus, prospective evaluation of this relation is important. Therefore, we prospectively examined the association between caffeine intake and the risk of incident tinnitus in 65,085 women in the Nurses' Health Study II.
Methods
Study Population
The Nurses' Health Study II cohort began in 1989 with questionnaires completed by 116,430 female registered nurses aged 25 to 42 years. Questionnaires have been mailed on a biennial basis to update the exposure information and the incidence of a wide variety of medical conditions. The follow-up rate for this cohort exceeds 90% of eligible person-time.
We excluded individuals if they reported onset of tinnitus before 1991 (the first year that caffeine intake was assessed) or if they had a history of cancer (excluding nonmelanoma skin cancer). Participants were eligible to contribute person-time to the study as of their 30th birthday, which is the youngest age for which we were able to estimate the onset of tinnitus.
Assessment of Caffeine Intake
Caffeine intake was assessed using detailed, extensively validated semiquantitative food frequency questionnaires that inquired about the average intake of more than 130 foods and beverages over the preceding year in 1991 and at 4-year intervals thereafter.
12
, 13
, 14
The relevant beverages on the questionnaire included low-calorie cola (eg, Diet Coke or Diet Pepsi with caffeine), regular cola (eg, Coke, Pepsi, or other cola beverages with sugar), tea with caffeine, tea without caffeine, coffee with caffeine, and decaffeinated coffee. In 2003 and 2007, a question about the intake of dairy coffee drinks—hot or cold (eg, cappuccino) also was included. The relevant foods included candy bars (eg, Snickers, Milky Way, Reeses), dark chocolate (eg, Hershey's Dark or Dove Dark), milk chocolate—bar or pack (eg, Hershey's, M&M's), and brownies. The response options for specified serving sizes were never or less than once per month; 1 to 3 per month; 1 per week; 2 to 4 per week; 5 to 6 per week; 1 per day; 2 to 3 per day; 4 to 5 per day; and 6 or more per day.
Total caffeine intake was calculated using US Department of Agriculture food composition sources. For example, the caffeine content of the specific items was considered to be 137 mg per cup of coffee, 47 mg per cup of tea, 46 mg per can or bottle of cola beverage, and 7 mg per serving of chocolate candy.
15
This method of determining caffeine intake was shown to be valid in prior studies.12
, 13
, 14
Ascertainment of Tinnitus
Participants were asked on the 2009 questionnaire “In the past 12 months, have you had ringing, roaring, or buzzing in your ears?” The response categories were never, once per week or less, a few days per week, or daily. To determine the date of onset, participants were asked “At what age did this first begin?” (<30, 30-39, 40-49, 50-59, and ≥60 years). Onset of tinnitus was defined to be the midpoint of an age interval (eg, 35 years for the age category of onset of 30-39 years). We defined a case as a participant who reported tinnitus occurring a few days per week or daily.
Assessment of Covariates
Data on body mass index (BMI) and intake of tobacco were collected on each questionnaire. On the 2009 questionnaire, each participant was asked if he/she had a hearing problem and the age when a change was first noticed. Data on other health conditions were ascertained on each questionnaire. Conditions of interest included diabetes, hypertension, multiple sclerosis, and depression, because these have been associated with tinnitus.
2
, 7
, 16
, 17
Data regarding use of medications proposed to be associated with tinnitus were obtained from the biennial questionnaires (analgesics [acetaminophen, aspirin, ibuprofen, other nonsteroidal anti-inflammatory drugs, and cyclooxygenase-2 inhibitors], thiazide diuretics, furosemide, and selective serotonin reuptake inhibitors/other antidepressants). The Crown Crisp Questionnaire, used to assess phobic anxiety, was administered in 1993 and 2005 and was used to adjust for phobic anxiety based on score categories.18
, 19
Statistical Methods
The time of observation was between 1991 and 2009. Participants who did not return a questionnaire for one time period and who remained eligible contributed person-time for later time intervals. Participants were censored after the onset of tinnitus or diagnosis of cancer.
Cox proportional hazards regression models were used to estimate multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) on the basis of categories of cumulative average caffeine intake (in 150 mg/d increments, which is ∼1 cup of coffee). We used the Anderson-Gill data structure and stratified by age and calendar year. Information on caffeine intake on each subsequent questionnaire was used to update the cumulative average caffeine intake up to that point. Covariates considered in multivariable-adjusted models included age (continuous), hearing loss, smoking status (current, past, or never), comorbid conditions (diabetes, hypertension, multiple sclerosis, and depression), medications used at least 1 day per week (analgesics, diuretics, and antidepressants), BMI, and smoking status. We used stratified models to examine whether the relation between caffeine intake and incident tinnitus varied by age (<50 or ≥50 years) and self-reported hearing loss. We also evaluated the association between decaffeinated coffee and incident tinnitus, after excluding those who consumed caffeinated coffee, to explore whether any observed relations were due to coffee (the major source of caffeine in this cohort) or caffeine.
Covariate information was updated at the beginning of each time period using the most recently available responses. Participants with missing covariate data were assigned to a missing category for that specific time period.
Level of significance (α) for 2-tailed P values was <.05. We used SAS version 9.3 for UNIX statistical software package (SAS Institute Inc, Cary, NC).
Results
There were 65,085 women included in the analysis. The baseline characteristics of the cohort according to category of caffeine intake are shown in Table 1. At baseline, the mean age of the cohort was 36.3 years, the mean BMI was 24.5 kg/m2, and the mean caffeine intake was 242.3 mg/d. Comorbidities present at baseline included hearing problems (prevalence = 3.3%), history of depression (13.7%), diabetes (0.8%), hypertension (6.1%), and multiple sclerosis (0.3%). With increasing caffeine intake, there were higher percentages of participants who reported smoking, depression, aspirin use, and ibuprofen use. The incidence rate of tinnitus increased substantially with age, ranging from 104 per 100,000 person-years for those aged <40 years to 1273 per 100,000 person-years for those aged ≥50 years (Table 2).
Table 1Baseline Characteristics of Women in 1991 by Category of Caffeine Intake
| Energy-Adjusted Caffeine Intake, mg/d | |||||
|---|---|---|---|---|---|
| 0-149 (n = 29,037) | 150-299 (n = 13,049) | 300-449 (n = 13,188) | 450-599 (n = 4453) | 600+ (n = 5358) | |
| Age, y | 35.7 (4.7) | 36.2 (4.7) | 37.0 (4.3) | 37.1 (4.4) | 37.7 (4.2) |
| BMI, kg/m² | 24.4 (5.3) | 24.7 (5.4) | 24.2 (4.8) | 24.6 (5.1) | 24.7 (4.9) |
| Smoking Status | |||||
| Never, % | 77.8 | 68.0 | 55.0 | 52.5 | 39.5 |
| Past, % | 16.9 | 22.3 | 30.8 | 27.6 | 27.4 |
| Current, % | 5.2 | 9.5 | 14.0 | 19.7 | 33.0 |
| Hearing problem, % | 3.3 | 3.1 | 3.5 | 3.6 | 3.4 |
| History of hypertension, % | 5.9 | 6.6 | 5.6 | 6.4 | 6.4 |
| History of diabetes, % | 0.7 | 0.9 | 0.6 | 0.8 | 1.1 |
| History of depression, 2003, % | 12.9 | 13.7 | 14.2 | 14.8 | 16.0 |
| History of multiple sclerosis, % | 0.3 | 0.4 | 0.3 | 0.4 | 0.4 |
| Thiazide diuretic use, % | 1.6 | 1.8 | 1.7 | 1.9 | 1.8 |
| Furosemide use, 1989, % | 0.4 | 0.4 | 0.3 | 0.5 | 0.7 |
| Aspirin use, % | 3.6 | 4.8 | 4.5 | 5.0 | 6.2 |
| Ibuprofen use, % | 7.3 | 10.4 | 10.1 | 10.6 | 12.2 |
| Acetaminophen use, % | 6.7 | 8.2 | 7.6 | 8.2 | 8.1 |
| COX-2 inhibitors use, 2001, % | 8.6 | 9.4 | 9.3 | 10.3 | 10.6 |
| SSRI use, 1997, % | 7.2 | 7.9 | 8.3 | 9.0 | 9.1 |
| Other antidepressant use, 1993, % | 9.0 | 10.0 | 9.5 | 10.9 | 11.6 |
| Anxiety Crown-Crisp Score | |||||
| ≤2, % | 58.1 | 56.1 | 56.2 | 55.7 | 53.2 |
| ≥3, % | 31.8 | 32.7 | 32.2 | 32.1 | 35.3 |
The values in the table reflect responses from 1991 unless otherwise indicated. Use of medication refers to use at least 1 day per week. Values are means (standard deviation) or percentages and are standardized to the age distribution of the study population.
BMI = body mass index; COX = cyclooxygenase; SSRI = selective serotonin reuptake inhibitor.
∗ The value is not age adjusted.
Table 2Age-Specific Tinnitus Incidence Rates
| Age (y) | Person-y | Cases | Incidence Rate per 100,000 Person-y |
|---|---|---|---|
| <40 | 348,676 | 361 | 104 |
| 40-49 | 545,612 | 2286 | 419 |
| ≥50 | 207,464 | 2642 | 1273 |
| Total | 1,101,753 | 5289 | 475 |
∗ Total differs slightly because of rounding.
The age and multivariable-adjusted HRs of the association between caffeine intake and incident tinnitus are shown in Table 3. After adjusting for age, there was a significant inverse association between caffeine intake and the incidence of tinnitus at higher levels of caffeine intake (HR, 0.89; 95% CI, 0.80-0.99 for 450-599 mg/d and HR, 0.83; 95% CI, 0.73-0.95 for ≥600 mg/d, P for trend <.001). After adjusting for other potential confounders, a slightly greater inverse association was observed between caffeine intake and tinnitus at higher levels of caffeine intake: HR, 0.85 (95% CI, 0.76-0.95) for 450 to 599 mg/d and HR 0.79 (95% CI, 0.68-0.91) for ≥600 mg/d (P for trend <.001). The association between caffeine intake and tinnitus did not vary by age (P for interaction, .5) or hearing loss (P for interaction, .8).
Table 3Age and Multivariable-Adjusted Hazard Ratios for the Association Between Caffeine Intake and Incident Tinnitus
| Caffeine Intake in mg/d | No. of Cases | Person-y | Age-adjusted HR (95% CI) | Multivariate Model ∗ Covariates in model included caffeine, age, hearing loss, medications, medical conditions, phobic anxiety, BMI, and smoking status. Medications included acetaminophen, aspirin, ibuprofen, other nonsteroidal anti-inflammatory drugs, cyclooxygenase-2 inhibitors, thiazide diuretics, furosemide, and selective serotonin reuptake inhibitors/other antidepressants. Medical conditions included diabetes mellitus, hypertension, multiple sclerosis, and depression. HR (95% CI) |
|---|---|---|---|---|
| 0-149 | 2268 | 470,228 | 1.00 (referent) | 1.00 (referent) |
| 150-299 | 1394 | 274,909 | 0.96 (0.89-1.02) | 0.94 (0.88-1.00) |
| 300-449 | 1029 | 214,588 | 0.93 (0.87-1.01) | 0.91 (0.84-0.98) |
| 450-599 | 364 | 80,433 | 0.89 (0.80-0.99) | 0.85 (0.76-0.95) |
| ≥600 | 234 | 61,595 | 0.83 (0.73-0.95) | 0.79 (0.68-0.91) |
| Ptrend < .001 | P trend < .001 |
CI = confidence interval; HR = hazard ratio.
∗ Covariates in model included caffeine, age, hearing loss, medications, medical conditions, phobic anxiety, BMI, and smoking status. Medications included acetaminophen, aspirin, ibuprofen, other nonsteroidal anti-inflammatory drugs, cyclooxygenase-2 inhibitors, thiazide diuretics, furosemide, and selective serotonin reuptake inhibitors/other antidepressants. Medical conditions included diabetes mellitus, hypertension, multiple sclerosis, and depression.
To exclude further the possibility that tinnitus onset might have led to decreased caffeine consumption, we compared caffeine intake in participants before and after the time they indicated they developed tinnitus, but we found no differences. For example, between 1991 and 1999, there was no significant change in caffeine consumption category among participants who indicated their onset of tinnitus was between 1995 and 1999 (P = .53 using Bowker's test of symmetry). We also examined changes in mean caffeine intake over time among those who did and did not develop tinnitus to account for potential temporal trends. There was a similar slight decrease in total caffeine consumption between 1991 and 1999 among those who did develop tinnitus (19.1 mg, 95% CI, 8.5-29.7) and those who did not develop tinnitus during follow-up (14.3 mg, 95% CI, 12.7-15.9). The decreased consumption over time was not significantly different between groups.
There was no association between decaffeinated coffee intake and incident tinnitus. Compared with those who drank decaffeinated coffee less than once per month, the multivariable adjusted HR for tinnitus was 1.00 (95% CI, 0.89-1.13) for those who drank 2 to 3 cups per day and 0.85 (95% CI, 0.62-1.16) for those who drank 4 or more cups per day (P for trend = .9).
Discussion
Main Findings
In this large, prospective study among women, we observed an inverse association between caffeine intake and incident tinnitus. The majority (∼70%) of caffeine intake in these individuals was from coffee consumption. We did not find an association between decaffeinated coffee intake and incident tinnitus. This suggests it is caffeine rather than another component of coffee that was associated with a decreased incidence of tinnitus.
Comparison with Other Studies
A recent randomized controlled crossover trial failed to demonstrate an improvement in tinnitus symptoms with cessation of caffeine intake while demonstrating harm in the way of withdrawal symptoms among participants.
11
The investigation included men and women with a mean age of 59.2 years. A mean difference in tinnitus severity score of 0.04 (95% CI, 1.99-1.93) was observed between patients during caffeine-consuming and caffeine-abstaining states. Our study examined the risk of incident tinnitus, whereas the aforementioned study11
evaluated the effect of caffeine on existing tinnitus. On the basis of the results of these studies, there is not persuasive evidence that reducing caffeine intake would be beneficial with respect to tinnitus. We are not aware of other prospective studies evaluating the association between caffeine and incident tinnitus.The precise pathophysiology underlying subjective tinnitus remains unknown. Purported mechanisms include hyperactivity of the ascending auditory pathways or a reduced suppressive influence of the central nervous system.
20
Caffeine is a central nervous system stimulant, and this biologic action may have led to the idea that it may play a role in the development of tinnitus. Noting a paucity of evidence for such an association, Bagueley et al21
suggested that if such an association exists, the stimulatory effect of caffeine may lead to increased detection of tinnitus through increased arousal or anxiety.In addition to general stimulatory effects of caffeine on the central nervous system, caffeine has been shown to have a direct effect on the inner ear in both in vitro and animal studies. Caffeine is thought to result in shortening of outer hair cells, and this effect may be mediated through the release of calcium from ryanodine receptors.
22
At this time, it is unclear whether this pathway is involved in the development or protection from tinnitus, and therefore this may be a topic for future investigation.Study Limitations
The presence of tinnitus was established on the basis of self-report; however, there is no diagnostic test for tinnitus, so it is not possible to validate a self-report of tinnitus. Hearing loss also was self-reported, so there was likely some misclassification of this covariate. There does remain the possibility of misclassification due to recall with respect to the date of onset of tinnitus. Furthermore, the questionnaire does not distinguish between subjective and objective tinnitus, but the majority of cases of tinnitus are subjective. Finally, our cohort was composed of women who were predominantly Caucasian. Because there are no previous studies that have examined the association between caffeine and incidence tinnitus in men, it is unknown whether the findings apply to men or other racial groups.
Conclusions
In this prospective study, higher caffeine intake was associated with a reduced risk of tinnitus. At present, the mechanism underlying this association remains unknown. Given that our investigation evaluated the incidence of tinnitus and not the effect of caffeine among participants with existing tinnitus, we are unable to make a recommendation as to whether the addition of caffeine would improve symptoms.
Acknowledgments
The authors thank Elaine Coughlan-Gifford for the programming assistance she provided to our research team.
References
- Current estimates from the National Health Interview Survey, 1996.Vital Health Stat 10. 1999; : 1-203
- Prevalence and characteristics of tinnitus among US adults.Am J Med. 2010; 123: 711-718
- Classification and epidemiology of tinnitus.Otolaryngol Clin North Am. 2003; 36 (Review): 239-248
- Systematic review and meta-analyses of randomized controlled trials examining tinnitus management.Laryngoscope. 2011; 121: 1555-1564
- Caffeine intake, toxicity and dependence and lifetime risk for psychiatric and substance use disorders: an epidemiologic and co-twin control analysis.Psychol Med. 2006; 36: 1717-1725
- Caffeine consumption.in: Spiller G. Caffeine. CRC Press, Boca Raton, FL1998 (199)
- Tinnitus.in: Bailey B.J. Johnson J.T. Newlands S.D. Head and Neck Surgery—Otolaryngology. Lippincott Williams & Wilkins, Philadelphia, PA2006: 2241
Mayo Clinic Staff. Tinnitus. 2010 [cited 2012 July 28]; Available at: http://www.mayoclinic.com/health/tinnitus/DS00365/DSECTION=causes. Accessed July 28, 2012.
Insight into Causes and Treatments for Tinnitus. 2010 December 2010 [cited 2012 July 8]. Available at: http://www.entnet.org/HealthInformation/tinnitus.cfm. Accessed July 8, 2012.
- Diagnostic approach to tinnitus.Am Fam Physician. 2004; 69 (Review): 120-126
- Caffeine abstinence: an ineffective and potentially distressing tinnitus therapy.Int J Audiol. 2010; 49: 24-29
- Reproducibility and validity of a semiquantitative food frequency questionnaire.Am J Epidemiol. 1985; 122: 51-65
- Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption.Int J Epidemiol. 1989; 18: 858-867
- Coffee, caffeine, and cardiovascular disease in men.N Engl J Med. 1990; 323: 1026-1032
- Habitual caffeine intake and the risk of hypertension in women.JAMA. 2005; 294: 2330-2335
- Association between depression and tinnitus in a nationally representative sample of US older adults.Aging Ment Health. 2013; 17: 714-717
- Hearing loss in multiple sclerosis.Arch Neurol. 1983; 40: 33-35
- A short clinical diagnostic self-rating scale for psychoneurotic patients. The Middlesex Hospital Questionnaire (M.H.Q.).Br J Psychiatry. 1966; 112: 917-923
- Discriminant validity of the Crown-Crisp Experiential Index.Br J Med Psychol. 1987; 60: 61-69
- Caffeine and the central nervous system: mechanisms of action, biochemical, metabolic and psychostimulant effects.Brain Res Brain Res Rev. 1992; 17 (Review): 139-170
- Consequences and moderating factors.in: Tinnitus: A Multidisciplinary Approach (Second Edition). Wiley-Blackwell, London2013: 100
- The role of caffeine in otorhinolaryngology: guilty as charged?.Eur Arch Otorhinolaryngol. 2013 Aug 11; ([Epub ahead of print])
Article info
Publication history
Published online: March 10, 2014
Footnotes
Funding: Grants UM1 CA176726 and DK91417 from the National Institutes of Health.
Conflict of Interest: None.
Authorship: All authors had access to the data and played a role in writing this manuscript.
Identification
Copyright
© 2014 Elsevier Inc. Published by Elsevier Inc. All rights reserved.
