Investigating the Effects of Caffeine on Phonation
Elizabeth Erickson-Levendoski and Mahalakshmi Sivasankar, West Lafayette, Indiana
Summary: Objective. A core component of vocal hygiene programs is the avoidance of agents that may dry the
vocal folds. Clinicians commonly recommend that individuals reduce caffeine intake because of its presumed dehydrating effects on the voice. However, there is little evidence that ingestion of caffeine is detrimental to voice production.
The first objective of this study was to evaluate whether caffeine adversely affects voice production. The second objective was to evaluate if caffeine exacerbates the adverse phonatory effects of vocal loading.
Study Design. Prospective, double-blinded, sham-controlled study.
Methods. Sixteen healthy adults participated in two sessions where they consumed caffeine (caffeine concentration ¼ 480 mg) or sham (caffeine concentration ¼ 24 mg) beverages. Voice measures (phonation threshold pressure
and perceived phonatory effort) were collected. Subjects then completed a vocal loading challenge and voice measures
were obtained again.
Results. There were no significant differences in voice measures between the caffeine and sham conditions. Ingestion
of caffeine did not adversely affect voice production (P > 0.05) or exacerbate the detrimental phonatory effects of vocal
loading (P > 0.05).
Conclusions. Our findings contribute to emerging knowledge on the effects of caffeine on voice production. Recommendations to completely eliminate caffeine from the diet, as a component of a vocal hygiene program, should be evaluated on an individual basis.
Key Words: Caffeine–Voice production–Dehydration.
INTRODUCTION
Avoiding agents that dehydrate the vocal folds is an integral
component of vocal hygiene education.1,2 Systemic
dehydration is detrimental to voice production.3,4 Because of
the presumed systemic drying action of caffeine,2,5 voice
clinicians recommend avoidance of caffeinated beverages.6,7
This remains common clinical adage even with little
supporting evidence that caffeine consumption induces
negative changes to voice. Consequently, the first objective of
this study was to investigate whether caffeine is detrimental
to phonation. Voice measures were compared with subjects,
who each consumed beverages that contained large amounts
of caffeine (caffeinated beverages) and negligible amount of
caffeine (sham beverages). Voice measures included
phonation threshold pressure (PTP) and perceived phonatory
effort (PPE). PTP is an appropriate voice measure to evaluate
the phonatory effects of caffeine because PTP estimates the
lung pressure that is required for maintaining voicing at
threshold loudness8 and is sensitive to systemic dehydration.4,9
Another useful indicator of the phonatory effects of caffeine
may be participant ratings of effort required for voice
production (PPE).10 To determine if caffeine changes voice,
PTP and PPE ratings were obtained after consuming caffeinated and sham beverages. An increase in PTP and/or PPE following caffeine but not sham beverages would suggest caffeine
is detrimental to voice production.
Accepted for publication February 22, 2011.
A portion of this study was presented at the Annual Meeting of the American Academy
of Otolaryngology; September 26–29, 2010; Boston, MA.
From the Department of Speech, Language and Hearing Sciences, Purdue University,
West Lafayette, Indiana.
Address correspondence and reprint requests to Mahalakshmi Sivasankar, Speech,
Language and Hearing Sciences, Purdue University, 500 Oval Drive, West Lafayette,
IN 47907. E-mail: msivasankar@purdue.edu
Journal of Voice, Vol. -, No. -, pp. 1-5
0892-1997/$36.00
Ó 2011 The Voice Foundation
doi:10.1016/j.jvoice.2011.02.009
The second objective of this investigation was to evaluate
whether caffeine exacerbates the detrimental phonatory effects
of vocal loading. This question is clinically relevant because
individuals often work and socialize in environments where
they engage in vocal loading activity such as prolonged speaking
in background noise.11 It is well known that vocal loading alone
adversely affects the voice,12 and there is emerging evidence
that systemic dehydration may exacerbate the negative phonatory effects of vocal loading.3 To examine the relationship
between dehydration and vocal loading in the context of the purported systemic drying action of caffeine, we examined the
interaction between caffeine consumption and vocal loading
on PTP and PPE. Findings from this investigation are important
for the development of evidence-based recommendations
regarding the effects of caffeine consumption on voice
production.
METHODS
Participants
All procedures used here were approved by the Purdue
University Institutional Review Board. Sixteen healthy
adults (eight males and eight females; mean age ¼ 23 years)
with perceptually normal speech and voice participated in
this investigation. Participants reported a negative history
for hearing problems, respiratory disease, smoking, refluxing, and prescription medication (except oral contraceptives). All female subjects participated during the follicular
phase of their menstrual cycle. Eight individuals had
received vocal training (Table 1), either through group
and/or solo singing (seven subjects) or theatrical speaking
(1 subject). Trained subjects were purposely included
because they frequently engage in vocal loading-type activities and are also instructed to avoid caffeinated beverages
to maintain a healthy voice.
2
Journal of Voice, Vol. -, No. -, 2011
TABLE 1.
Subject Characteristics
Subject
Number
Gender
Age (Years)
Training
(Years)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Male
Male
Male
Male
Male
Male
Male
Male
Female
Female
Female
Female
Female
Female
Female
Female
18
21
20
20
26
27
20
25
24
20
27
32
19
20
20
23
6
10
7
10
None
None
None
None
10
5
9
10
None
None
None
None
Procedure
This investigation utilized a double-blinded, sham-controlled,
repeated-measures design. Participants were blinded to the
objectives of this study and consented to an investigation examining the ‘‘effects of taste and temperature of ingested beverages
on voice production.’’ Likewise, the investigators collecting and
analyzing the data were blinded to whether the consumed beverages were caffeinated or sham. Each subject participated in two
experimental sessions scheduled at the same time on two consecutive days. In each session, subjects ingested either caffeinated beverages (caffeine content 480 mg) or sham beverages
(caffeine content 24 mg). The sessions were counterbalanced
across subjects. The caffeinated beverages consisted of two
12-ounce caffeinated Starbucks coffees, whereas the sham beverages consisted of two 12-ounce Starbucks decaffeinated coffees. During both the caffeine and sham sessions, participants
left the laboratory following consumption of the first beverage
and returned for the second beverage and voice testing within
2.5–3 hours. Total duration of each experimental session was
5 hours beginning with consumption of the first beverage.
This timeline is within the elimination half-life of caffeine.13
Participants were instructed not to consume caffeine before
the experimental sessions. Only one participant reportedly consumed a 20-ounce soda (caffeine content 63 mg) before the
caffeinated session.
Data collection procedures were identical for both sessions
and completed in a humidity-controlled environment
(70% ± 6%). Following consumption of the second beverage,
voice measures were collected. These voice measures are hereafter referred to as preloading measures. Participants then completed a vocal loading challenge. Specifically, the vocal loading
challenge for the nontrained subjects, and the one subject
trained in theatrical speaking consisted of 35 minutes of reading
aloud in background noise (cafeteria noise: 65 dB sound pressure level). The vocal loading challenge for the subjects trained
in group and/or solo singing consisted of 35 minutes of singing
in identical background noise. Voice measures were reassessed
and this time point is hereafter referred to as Loading35. Subjects then completed an additional 35 minute vocal loading
challenge as described above, and voice measures were reobtained, and this time point is hereafter referred to as Loading70.
The voice measures included PTP and PPE.
Initially, subjects were trained on the PTP task.14 The training
consisted of subjects repeatedly practicing the PTP task (enunciating a five to seven /pi/ syllable string in a soft voice, on a single breath at a rate of 1.5 syllables/second guided with
a metronome). Participants were provided visual feedback and
investigator modeling during the training. Participants were
trained on the PTP task at two different pitches. These pitches
were the 10th and 80th percent pitch of each subjects’ maximum
pitch range.15 Extreme pitches were purposely selected based on
prior research that the effects of systemic dehydration are most
apparent at PTP productions at high pitch.4 Participants were
deemed trained when the following criteria were met at both
the 10th and 80th percent of the pitch range: /pi/ pressure peaks
of equal height, consistent quiet voicing, and oral flows approaching 0 mL/s with nostrils occluded.14 Once trained to the
above criteria, PTP measures were collected at the 10th and
80th pitches (hereafter referred to as PTP10 and PTP80, respectively). At least five repetitions of the PTP task were collected
at PTP10 and PTP80. The instrumentation for PTP collection included a circumferentially vented pneumotachograph face mask
coupled to transducers (Glottal Enterprises, Syracuse, NY) for
the measurement of oral flow (sampling rate ¼ 1000 Hz) and
oral pressure (sampling rate ¼ 100 Hz). Within each /pi/ syllable string, the middle three /p/ peaks were selected, and adjacent
peaks were measured to estimate lung pressure during the intervening /i/. Estimated lung pressures from each string were averaged for the final PTP at each pitch.15 PTP analysis was
completed by investigators blinded to the identity of the experimental session. To ensure PTP measurement accuracy between
investigators, 10% of the /pi/ syllable strings were remeasured,
and interrater reliability was computed. To ensure PTP measurement accuracy within an investigator, 10% of the /pi/ syllable
strings were remeasured and intrarater reliability was also computed. Inter- (r ¼ 0.98) and intrarater reliability (r ¼ 0.99)
revealed strong correlations.
To obtain PPE ratings, subjects sang ‘‘Happy Birthday’’ in
a soft voice starting at the 50th percent pitch of the maximum
phonation range.16 Subjects rated the effort for singing on
a 1000 visual analog scale with anchors of ‘‘no vocal effort’’ and
‘‘maximum vocal effort.’’15 Subjects rated PPE at three time
points (Preloading, Loading35, and Loading70). Analysis of
PPE was completed by measuring the distance in inches, from
the ‘‘no vocal effort’’ anchor to the subject’s PPE rating, by investigators blinded to the identity of the experimental session.13 To
ensure PPE measurement accuracy between investigators, 10%
of PPE ratings were remeasured and interrater reliability was
computed. To ensure PPE measurement accuracy within an investigator, 10% of PPE ratings were remeasured and intrarater
reliability was computed. Inter- (r ¼ 0.99) and intrarater reliability (r ¼ 0.99) revealed strong correlations.
Elizabeth Erickson-Levendoski and Mahalakshmi Sivasankar
Caffeine Effects on Voice
3
Statistical analysis
Data were analyzed using a mixed general linear model analysis
of variance (ANOVA; alpha level ¼ .05). The t tests were used
for post hoc analyses. The independent variables were condition (caffeine/sham) and vocal loading (Preloading, Loading35,
and Loading70). The dependent variables were PTP10, PTP80,
and PPE. Because the core objectives of this study were to
investigate the effects of caffeine on phonation and the interaction between vocal loading and caffeine, and because visual
inspection of the data did not reveal group differences because
of vocal training, data for trained and nontrained subjects were
merged in statistical analysis.
RESULTS
This investigation evaluated the effects of caffeine and vocal
loading on voice production. Caffeine consumption did not
adversely affect voice production (Figures 1 and 2). There
were no significant differences between the caffeine and
sham conditions for PTP10, PTP80, or PPE (P > 0.05).
Ingestion of caffeine or sham beverages did not worsen the
effects of vocal loading for PTP10, PTP80, or PPE (P > 0.05).
FIGURE 1. The nonsignificant interaction effect of caffeine consumption (caffeinated beverage/sham beverage) and vocal loading
(Preloading, Loading35, Loading70) on PTP10 and PTP80. Ingestion
of caffeine or sham beverages did not exacerbate the effects of vocal
loading. However, vocal loading significantly increased PTP10 and
PTP80. The * denotes the statistically significant increase in PTP10
and PTP80 after 35 and 70 minutes of vocal loading (P < 0.05). Error
bars represent standard errors.
FIGURE 2. The nonsignificant interaction effect of caffeine consumption (caffeinated beverage/sham beverage) and vocal loading
(Preloading, Loading35, Loading70) on PPE. Ingestion of caffeine or
sham beverages did not exacerbate the effects of vocal loading. In
addition, vocal loading did not significantly increase PPE. Error bars
represent standard errors.
However vocal loading, alone, significantly increased PTP10
and PTP80 (P < 0.01, Figure 1). Post hoc testing revealed that
PTP10 and PTP80 significantly increased after 35 minutes of
vocal loading and after 70 minutes of vocal loading (P < 0.01,
Figure 1). Conversely, PPE did not increase with vocal loading
(P > 0.05, Figure 2). Details from the statistical analyses are
presented in Table 2.
DISCUSSION
Over the past 20 years, caffeine consumption has increased
substantially.17 Caffeine is the world’s most widely used dietary stimulant18 with the average adult consuming approximately 227 mg caffeine per day17,19 in its various forms.
Following oral ingestion, caffeine is rapidly absorbed by the
gastrointestinal tract and elicits numerous physiological and
biochemical effects.13,20 For example, caffeine consumption
increases excretion of water and electrolytes.13 This diuretic
action of caffeine is presumed to induce systemic dehydration.2,5 Given that systemic dehydration negatively affects
voice production,4,9 voice clinicians strongly caution against
excessive caffeine use.1,21,22
To the best of our knowledge, only one pilot investigation has
examined the effects of caffeine consumption (250 mg caffeine
tablets) on voice.5 The results of this pilot investigation were
nonsignificant, but no definite conclusions could be drawn secondary to small sample size and the lack of the following: double
blinding, within-subject controls, and temporal controls. The
present study incorporated a double-blinded, sham-controlled,
repeated-measures design, where each participant consumed
both caffeinated and sham beverages in counterbalanced order,
on different days, to investigate whether caffeine consumption is
detrimental to voice production. Although the caffeine concentrations utilized here are greater than that consumed by the average adult,17 this dosage was purposely chosen. Past research
suggests that the diuretic response to caffeine occurs at doses
exceeding 300 mg.23 It is of note that despite the high dose of
4
Journal of Voice, Vol. -, No. -, 2011
TABLE 2.
Mixed General Linear Model ANOVA Results for PTP and
PPE
Variable
F
DF
P
Nonsignificant effect of caffeine consumption on PTP
and PPE
PTP10
2.83
1, 15
0.11
PTP80
1.67
1, 15
0.22
PPE
0.01
1, 15
0.97
Significant effect of vocal loading on PTP
PTP10
23.64
2, 30
PTP80
12.24
2, 30
PPE
0.20
2, 30
<0.01
<0.01
0.82
Nonsignificant interaction effect of caffeine and vocal
loading on PTP and PPE
PTP10
0.89
2, 30
0.42
PTP80
0.19
2, 30
0.83
PPE
0.28
2, 30
0.76
Abbreviation: DF, degrees of freedom.
caffeine intake in this study (480 mg) no changes in PTP and
PPE were observed.
The absence of a significant effect in PTP and PPE following
consumption of caffeinated beverages is surprising because
these measures are thought to be sensitive to changes in vocal
fold hydration status.4,9 It is possible that caffeine does not
produce a diuretic effect of sufficient magnitude to elicit
negative voice changes in healthy subjects who maintain
a regular balanced diet.20 Although the extent of fluid loss postcaffeine ingestion was not quantified here, previous voice
research demonstrates that adverse voice changes (as measured
by PTP) occur only after substantial systemic fluid loss induced
by dialysis9 or the loop diuretic, Lasix.4 The effects of caffeine
intake on systemic physiology and biochemistry may provide
further insight into the nonsignificant changes in PTP and
PPE observed in this study. Healthy individuals quickly develop
a tolerance to the diuretic effects of caffeine, a tolerance
strengthened by regular caffeine consumption.20,23 In this
investigation, all but two participants reported using caffeine,
although the frequency of caffeine use ranged from once per
month to twice per day. Further study is required on how
tolerance to caffeine affects voice production. Future studies
should also include a caffeine withdrawal period, because the
diuretic action of caffeine is most noted in individuals
deprived of caffeine for a period of days.24 We did not impose
a withdrawal period in this study to mimic realistic conditions
for participants.
Systemic dehydration has been previously shown to exacerbate the negative effects of vocal loading.3 In the current investigation, ingestion of caffeine did not have a significant effect
on voice production in either trained or nontrained speakers
exposed to vocal loading. One possible interpretation of this
finding is that the dose of caffeine used here did not induce
the extent of systemic dehydration needed to exacerbate the
ill effects of vocal loading. However, the vocal loading task,
alone, significantly increased PTP independent of caffeine consumption. Therefore, it is also possible that the detrimental
effects of vocal loading were of sufficient magnitude to override
the negative phonatory effects, if any, associated with caffeine
intake. Surprisingly, PPE did not increase following vocal
loading. The rationale for this finding is unclear, but could be
related to the nature of the singing task used to elicit PPE
ratings.
CONCLUSIONS
Findings from the present study suggest that a high dose of caffeine does not adversely affect PTP or PPE measures within the
timeline examined. Additional measures may be more sensitive
to caffeine-induced perturbations in vocal fold physiology. We
selected PTP and PPE measures as these techniques are commonly used to detect changes in vocal fold function. This study
lays the groundwork for further investigations that quantify the
specific amount of dehydration induced by caffeine, in larger
sample sizes, to develop evidence-based recommendations
regarding caffeine consumption and voice.
Acknowledgments
We acknowledge the contributions of Kaleena Cruz to data collection and analysis. We also thank Grace Scott, Juli Hoon, and
Dakota Robinson.
REFERENCES
1. Broaddus-Lawrence P, Treole K, McCabe R, et al. The effects of preventative vocal hygiene education on the vocal hygiene habits and perceptual
vocal characteristics of training singers. J Voice. 2000;14:58–71.
2. Fletcher H, Drinnan M, Carding P. Voice care knowledge among clinicians
and people with healthy voices or dysphonia. J Voice. 2007;21:80–91.
3. Solomon N, DiMattia M. Effects of a vocally fatiguing task and systemic
hydration on phonation threshold pressure. J Voice. 2000;14:341–362.
4. Verdolini K, Min Y, Titze I, et al. Biological mechanisms underlying voice
changes due to dehydration. J Speech Lang Hear Res. 2002;45:268–281.
5. Akhtar S, Wood G, Rubin J, et al. Effect of caffeine on the vocal folds: a pilot study. J Laryngol Otol. 1999;113:341–345.
6. Chan R. Does the voice improve with vocal hygiene education? A study of
some instrumental voice measures in a group of kindergarten teachers.
J Voice. 1994;8:279–291.
7. Benninger M, Jacobson B, Johnson A. Vocal Arts Medicine: The Care and
Prevention of Professional Voice Disorders. New York, NY: Thieme
Medical Publishers, Inc.; 1994.
8. Titze I. Phonation threshold pressure: a missing link in glottal aerodynamics. J Acoust Soc Am. 1992;91:2926–2935.
9. Fisher K, Ligon J, Sobecks J, et al. Phonatory effects of body fluid removal.
J Speech Lang Hear Res. 2001;44:354–367.
10. Chang A, Karnell M. Perceived phonatory effort and phonation threshold
pressure across a prolonged voice loading task: a study of vocal fatigue.
J Voice. 2004;18:454–466.
11. Vilkman E. Voice problems at work: a challenge for occupational safety and
health arrangement. Folio Phoniatr Logop. 2000;52:120–125.
12. Sodersten M, Ternstrom S, Bohman M. Loud speech in realistic environmental noise: phonetogram data, perceptual voice quality, subjective ratings, and gender differences in healthy speakers. J Voice. 2005;19:29–46.
13. Carrillo J, Benitez J. Clinically significant pharmacokinetic interactions
between dietary caffeine and medications. Clin Pharmacokinet. 2001;39:
127–153.
14. Fisher K, Swank P. Estimating phonation threshold pressure. J Speech Lang
Hear Res. 1997;40:1122–1129.
Elizabeth Erickson-Levendoski and Mahalakshmi Sivasankar
15. Erickson E, Sivasankar M. Evidence for adverse phonatory change following an inhaled combination treatments. J Speech Lang Hear Res. 2010;53:
75–83.
16. Sivasankar M, Erickson E, Schneider S, et al. Phonatory effects of airway
dehydration: preliminary evidence for impaired compensation to oral
breathing in individuals with vocal fatigue. J Speech Lang Hear Res.
2008;51:1494–1506.
17. Frary C, Johnson R, Wang M. Food sources and intakes of caffeine in the
diets of persons in the United States. J Am Diet Assoc. 2005;105:110–113.
18. Lamarine R. Selected health and behavioral effects related to the use of caffeine. J Community Health. 1994;19:449–466.
19. Nehlig A. Are we dependent upon coffee and caffeine? A review on human
and animal data. Neurosci Biobehav Rev. 1999;23:563–576.
Caffeine Effects on Voice
5
20. Armstrong L. Caffeine, body fluid-electrolyte balance, and exercise performance. Int J Sport Nutr Exerc Metab. 2002;12:189–206.
21. Bovo R, Galceran M, Pertuccelli J, et al. Vocal problems among
teachers: evaluation of a preventive voice program. J Voice. 2007;12:
705–722.
22. Simberg S, Sala E, Tuomainen J, et al. The effectiveness of group therapy
for students with mild voice disorders: a controlled clinical trial. J Voice.
2006;20:97–109.
23. Maughan R, Griffin J. Caffeine ingestion and fluid balance: a review. J Hum
Nutr Diet. 2003;16:411–420.
24. Neuhauser-Berthold B, Verwied S, Luhrmann P. Coffee consumption and
total body water homeostasis as measured by fluid balance and bioelectrical
impedance analysis. Ann Nutr Metab. 1997;41:29–36.