Acoustic Parameters and Salivary IL-6 Levels in

Overweight and Obese Teachers

*,†Jesenia Acurio, †Cristian Celis, †Jazmin Perez, and *Carlos Escudero, *yChillan, Chile

Summary: Objective. Either obesity or vocal loading task leads to elevation of proinflammatory cytokines such as
interleukin 6 (IL-6). However, it is unknown whether vocal parameters after vocal loading are correlated with body mass
index (BMI) or IL-6. We hypothesize that vocal loading induces an elevation of acoustic parameters of voice and sali-
vary IL-6 in overweight and obese people.
Methods. A total of 33 schoolteachers without any self-reported voice alterations were invited to participate in this
study. Participants were classified according to BMI into normal, overweight, and obese groups. The vocal loading task
consisted of loud speech (60–90 minutes) in their classroom. Salivary and voice samples were taken before and after
vocal loading. Perceptual and self-reported voice alterations and objective voice analyses were investigated. The rela-
tive concentration of salivary IL-6 was estimated by a colorimetric assay.
Results. Obese teachers showed a significant elevation in fundamental frequency value after vocal loading. In addi-
tion, reduction in harmonics-to-noise ratio (HNR) was observed in teachers with normal weight after vocal loading but
not in overweight or obese groups. No significant correlation was observed between BMI and any of the acoustic pa-
rameters analyzed or salivary IL-6 levels. Furthermore, teachers who were overweight showed a significant increase in
the salivary IL-6 levels after vocal loading. Interestingly, salivary IL-6 levels were positively correlated with HNR value
in the overweight group after vocal loading.
Conclusion. Excessive body weight is related to alterations in fundamental frequency and HNR. In addition, HNR,
but not BMI, is associated with salivary IL-6 levels in overweight teachers.
Key Words: Salivary IL-6–Acoustic parameter–Voice alteration–Obesity and overweight–Vocal loading.

INTRODUCTION
Overweight and obesity have become major public health prob-
lems worldwide,1 being major contributors to the burden of the
disease in low-income and middle-income countries.2,3 In
addition, excessive fat deposits compromise virtually all body
systems including the upper airway. Some reports4–6 have
demonstrated that the pattern and amount of local fat deposits
in the pharynx are closely related to the occurrence of
obstructive sleep apnea syndrome (OSAS). In addition, fat
deposits may affect other components of the vocal tract
structures including uvula, soft palate, and the posterior
region of the tongue.7 Even a modest increase in weight can
diminish lung function because of the presence of adipose tis-
sue around the rib cage and abdomen and in the visceral cavity,
which in turn reduces functional residual capacity.8
It has been suggested that excessive fat deposits might lead to
impaired vocalization, but results are contradictory. For instance,
a reduced maximum frequency of harmonics of /i/ and /e/ vowels,
as well as the number of harmonics of /i/ vowel have been
described in OSAS patients compared with controls.9 However,
Accepted for publication March 4, 2014.
The authors J.A. and C.C. contributed equally to this publication.
Financial disclosure: None to report.
Conflict of interest: None to report.
From the *Department of Basic Sciences, Faculty of Sciences, Vascular Physiology
Laboratory, Group of Investigation in Tumor Angiogenesis (GIANT), Group of Research
and Innovation in Vascular Health (GRIVAS Health), University of Bıo-Bıo, Chillan,
Chile; and the ySchool of Phonoaudiology, Universidad Pedro de Valdivia, Chillan, Chile.
Address correspondence and reprint requests to Carlos Escudero, Department of Basic
Sciences, Faculty of Sciences, Vascular Physiology Laboratory, Group of Investigation
in Tumor Angiogenesis (GIANT), Group of Research and Innovation in Vascular Health
(GRIVAS Health), Universidad del Bio-Bio, Chillan, Chile. E-mail: cescudero@ubiobio.cl
Journal of Voice, Vol. 28, No. 5, pp. 574-581
0892-1997/$36.00
Ó 2014 The Voice Foundation
http://dx.doi.org/10.1016/j.jvoice.2014.03.002
these alterations in vocalization were not related to body mass in-
dex (BMI) because obese OSAS patients exhibited similar alter-
ations to nonobese patients. In addition, only a few studies10,11
have directly analyzed the impact of obesity on vocal quality
and function. One study10 described that obese people exhibited
an elevated prevalence of altered laryngoscopic analysis, associ-
ated with enhanced acoustic parameters (jitter, shimmer, and
noise) but reduced maximum phonation time (MPT). Accord-
ingly, this study concludes that voice in obese people can be char-
acterized as hoarse (62%), murmuring (27%), and with phonation
instability (44%). On the other hand, Solomon et al11 reported that
obese patients had similar acoustic parameters (ie, jitter, shimmer,
noise-to-harmonic ratio [NHR], fundamental frequency [F0],
MPT) when compared with nonobese controls. Among several
methodological differences between these two studies, it is inter-
esting to highlight that both include overweight people as nonob-
ese controls. Therefore, we believe that stratification of
participants according to BMI may elucidate whether excessive
body fat deposits impair voice production.
On the other hand, Verdolini Abbott et al12 have described that
the acute response to vocal fold injury in healthy people may
include the secretion of proinflammatory markers such as matrix
metalloproteinase-8, interleukin 1b (IL-1b), and interleukin 6
(IL-6).12 In particular, elevation of IL-6 detected in laryngeal
secretion after 24 hours post initial 60-minute vocal loading ses-
sion was partially reverted by voice resting and completely over-
come by resonant voice therapy.12 Furthermore, this same group13
has described that fibroblasts from rabbit vocal folds exposed
to IL-1b activate proinflammatory gene transcription, an effect
reversed by exposure to various magnitudes of cyclic tensile
strain (0–0.5 Hz), suggesting that vibratory therapy might induce
wound healing in the vocal folds. Despite the promise of interleu-
kins (ILs) measurements, particularly IL-6, as a biomarker for
Jesenia Acurio, et al IL-6 and Acoustic Parameter in Obese and Overweight
detecting vocal fold alterations and its healing process after injury,
one limitation is the accessibility to laryngeal secretion.
IL-6 has been detected in saliva in normal volunteers,14 the
levels being associated with the activity of salivary glands, as
well as oral epithelial cells, periodontal ligament cells, and
gingival fibroblasts among other cells.15,16 Salivary IL-6 may
reflect localized inflammation in the upper airway that is rela-
tively independent of the general systemic level of inflamma-
tion.17 There is no information regarding the association
between salivary IL-6 levels and voice analysis. However, it
has been reported that salivary IL-6 increases in approximately
50% of people after acute psychosocial stress induced by deliv-
ering a speech and performing a mental arithmetic task in front
of two audiences,17 suggesting that salivary IL-6 may increase
after vocal loading. Because obesity has been related with
elevated circulating levels of proinflammatory cytokines
including IL-6,18–20 it is intriguing how BMI and vocal
loading may affect salivary levels of IL-6. We aim to investigate
whether a vocal loading task increases acoustic parameters of
voice and also salivary IL-6 in overweight and obese people
compared with healthy weight controls.
PARTICIPANTS AND METHODS
This study was approved by The Ethical Committee from the
Universidad Pedro de Valdivia, Chillan, Chile. All participants
signed a written informed consent. In this study, a total of 33
self-reported healthy schoolteachers were included. School-
teachers were studied because they have high occupational
voice demands.21 All teachers were subjectively free from up-
per airway infections and allergies at the time of the examina-
tion during an ordinary workweek.
Teachers were classified into normal weight, overweight, and
obese groups following the World Health Organization classifi-
cation. Thus, BMI <25 kg/m2 was considered as normal weight,
BMI 25–30 kg/m2 was classified as overweight, and BMI
>30 kg/m2 as obese. Inclusion criteria were school teacher, vol-
unteers, in good overall health by self-report, and ability to pro-
duce a loud voice (ie, between 85 and 95 dB). Exclusion criteria
were a self-report of a current voice problem or a voice problem
more than once monthly during the preceding year; current use
of any medications that might influence voice (eg, deconges-
tants or antiinflammatory and antihistamines) or signs of
currently active allergic process; known or suspected allergy;
as well as current smoker and report of pregnancy.
All evaluations took place in a silenced classroom between
8:30 and 11.00 AM to minimize the circadian variation of
IL-6.14 Also, previous to experimentation, the participants
were instructed to maintain their regular activities, including
sleep patterns and no medication (in particular, antiinflamma-
tory drugs). They were asked to refrain from eating and drink-
ing at least 1 hour before the experimental session. A single
voice or saliva sample was taken before and after vocal load.
Vocal loading task
After their regular schedule, teachers were asked to give their
first expositive class in the morning for 60 minutes minimum
575
(range, 60–90 minutes), which was defined as vocal loading,
defined by Vilkman22 as a combination of prolonged voice
use and additional factors, such as background noise, acoustics,
and air quality, which affect the F0, mode, and intensity of pho-
nations as well as the external frame of the larynx. Therefore,
teachers gave a loud speech (ie, at 65–90 dB23) in front of 25
primary school children. Acoustic conditions and humidity in
the classroom as well as background noise were standardized
for all participants, using the same style of classrooms at the
school.
Subjective analysis of voice
All participants had a subjective perceptive-audible assessment
of their vocal quality using the Grade, Roughness, Breathiness,
Asthenia, and Strain (GRBAS) scale. This scale defines the
dysphonic voice based on the grade of dysphonia (G) by iden-
tifying four independent parameters: roughness or irregular
vocal fold vibration (R), instability or fluctuation of the vocal
quality, breathlessness or escape of sonorous transglottic air
(B), asthenia or vocal weakness (A), and strain compression
of the vocal fold in the middle line (S). Each parameter is scored
by using a four-point ordinal scale from zero to three (zero,
normal; one, mild; two, moderate; and three, severe).10
Also, the participants self-completed the Voice Handicap In-
dex (VHI). This patient-based self-assessment tool consists of
30 items. These items are equally distributed over three do-
mains: functional, physical, and emotional aspects of voice dis-
orders.24 A scale of five, where zero means never and four
always defined the grade of alteration of each item.
Voice samples and perceptual evaluation
All voice samples were recorded in the same silent classroom in
the school. To analyze MPT and voice acoustic parameters, all
subjects were asked to produce sustained phonation of the
vowel /a/ at a comfortable pitch and volume. A unidirectional
microphone (Shure Incorporated, IL) connected to a laptop
through a Lexicon interface (Lexicon Pro, UT) was used for
collecting voice samples. When the first trial was not suffi-
ciently long or oscillographically too unsteady for further
research, more trials were undertaken until an acceptable
recording was obtained. Voice recording was done for more
than 5 seconds at a distance of 30 cm from the microphone.
Acoustic parameters included MPT, F0 (or pitch), frequency
perturbation (jitter local absolute, jitter local, jitter rap, and
jitter ppq5), amplitude perturbation (shimmer in decibels,
shimmer local, shimmer apq3, shimmer apq5, and shimmer
apq11), harmonic-to-noise ratio (HNR), and NHR. All acoustic
analyses were performed with the Praat software (Phonetic Sci-
ences, University of Amsterdam, Amsterdam, The Netherlands),
using the recording done from second 2 to 3 in the 5-second
trace in a mono sound recording of 44.100 Hz.
Saliva samples
Saliva samples were collected for measuring IL-6 levels as
previously described.17 In brief, the participants were asked
to pool saliva in the mouth for a couple of minutes and to drool
it down into a previously labeled sterile container. Saliva was
576 Journal of Voice, Vol. 28, No. 5, 2014

TABLE 1.
Characteristic of Included Teachers
ANOVA
Characteristic Normal (n ¼ 12) Overweight (n ¼ 8) Obese (n ¼ 7) P value
Male/female 4/8 4/4 1/6
Age (y) 31.1 ± 1.5 38.0 ± 3.1 45.2 ± 6.8 0.067
Weight (kg) 67.8 ± 3.1 76.1 ± 2.7* 87.1 ± 3.3* 0.008
Height (m) 1.7 ± 0.03 1.6 ± 0.03 1.5 ± 0.02 0.197
BMI (kg/m2) 23.2 ± 0.3 28.1 ± 0.5* 34.1 ± 1.0*,y <0.0001
Classes per week (h/wk) 34.6 ± 3.2 34.0 ± 5.0 32.2 ± 4.3 0.641
Experience as teachers (y) 6.5 ± 1.3 8.1 ± 2.0 19.4 ± 6.1 0.055
Notes: Values in mean ± SEM.
* P < 0.05 vs normal weight.
y
P < 0.05 vs overweight.

kept at 4 C. After vocal loading, another sample was collected
as previously mentioned. Both containers were transported on
ice to the laboratory and stored at 80 C until immunoassay
for IL-6 was performed.
Enzyme-linked immunosorbent assay
Semiquantitative IL-6 levels were assayed using a commer-
cially available kit (eBioscience, Inc, San Diego, CA) and
following the manufacturer’s instructions. Briefly, 100 mL saliva
was deposited in a 96-well plate coated previously with the
specific antibody (IL-6) and incubated overnight at 4 C. After
sequential washes, IL-6 was detected by using an avidin-
horseradish peroxidase system. Optical density was measured
at 450 nm. Also 490-nm filter was used for wavelength subtrac-
tion (Autobio Labtec Instruments, Zhengzhou City, PR China).
Intra- and interassay variation errors were 3.1 ± 0.2% and
8.2 ± 1.7%, respectively. Limit of detection was 2 pg/mL.
Statistical analysis
Qualitative data are expressed as a percentage with respect to
their own group. Chi-squared test was used for comparison be-
tween qualitative variables. Quantitative data are expressed as
mean ± standard error of the mean. Kruskal-Wallis nonpara-
metric analysis of variance (ANOVA) followed by the Mann-
Whitney U test was performed for comparing the data from
normal, overweight, and obese teachers. In addition, BMI and
IL-6 values were correlated with the other quantitative vari-
ables, including acoustic parameters, by using a Spearman
test. The statistical software GraphPad Prism 5.00 (GraphPad
Software, Inc, CA) was used. Nominal level of statistical signif-
icance was set at 0.05.
As expected, differences in BMI and weight were statistically
significant between the groups. We also found that obese
people were older and had more years of service than the
other groups; however, this difference was not statistically
significant.
None of the participants exhibited a voice alteration during
the subjective evaluation by the GRBAS scale. Mean value of
the total score in the GRBAS test was 0.6 ± 0.2 arbitrary units
in the control group, which was not statistically different
compared with overweight (0.6 ± 0.4) or obese groups
(0.8 ± 0.3, ANOVA, P ¼ 0.53). Moreover, all the participants
reported no alteration in their voice according to the scores
given to each of the items analyzed by the VHI (Table 2).
Acoustic analysis
In relation to the acoustic parameters of the voice, no differ-
ences were found in the acoustic measures for jitter, shimmer,
or NHR either between groups or when values were compared
before and after vocal loading in each group (Table 3). Never-
theless, ANOVA showed statistically significant differences in
pitch (F0) and HNR (P < 0.05 in both cases) between the groups
when all groups (before and after loading) were compared. Post
hoc analysis (Figure 1A) showed that the F0 value was signifi-
cantly increased after vocal loading only in the obese group
(176 ± 18 vs 224 ± 6.4 Hz, respectively). Furthermore, HNR
(Figure 1B) was significantly reduced in the normal weight
group after vocal loading (10.5 ± 0.9 dB) compared with values
before the loading task (13.6 ± 0.8 dB).
TABLE 2.
Score Reported in the VHI

Normal Overweight Obese ANOVA

RESULTS
Characteristic (n ¼ 12) (n ¼ 8) (n ¼ 7) P value
Participant characteristics
The initial sample size was 33 teachers. Six individuals were Functional 3.9 ± 1.1 3.0 ± 1.3 4.7 ± 2.1 0.73
excluded because of inappropriate sample collections before Physical 10.0 ± 2.5 9.1 ± 1.9 7.5 ± 2.6 0.82
or after vocal loading. Thus, 27 teachers were included in the Emotional 1.0 ± 0.5 1.0 ± 0.5 3.1 ± 1.7 0.61
Total 14.9 ± 3.9 13.1 ± 2.5 15.4 ± 6.2 0.96
final analysis. They were divided into normal weight (controls,
n ¼ 12), overweight (n ¼ 8), and obese (n ¼ 7). The general Notes: Values presented by different levels of analysis studied by the VHI.
Values in mean ± SEM.
characteristics of the participants are presented in Table 1.
Jesenia Acurio, et al IL-6 and Acoustic Parameter in Obese and Overweight 577

TABLE 3.
Acoustic Parameters in the Studied Groups
Normal Overweight Obese

Characteristic Before After Before After Before After

5
Jitter absolute (310 ) 2.99 ± 0.51 3.97 ± 0.82 4.97 ± 1.01 2.48 ± 0.33 6.38 ± 2.29 4.57 ± 0.92
Jitter (local): % 0.49 ± 0.08 0.51 ± 0.06 0.63 ± 0.11 0.43 ± 0.08 0.51 ± 0.09 0.59 ± 0.07
Jitter (rap): % 0.3 ± 0.1 0.3 ± 0.1 0.4 ± 0.1 0.5 ± 0.1 0.4 ± 0.1 0.4 ± 0.1
Jitter (ppq5): % 0.3 ± 0.1 0.3 ± 0.0 0.4 ± 0.1 0.5 ± 0.1 0.3 ± 0.1 0.4 ± 0.1
Shimmer (local): % 9921 ± 1019 11 594 ± 1387 12 876 ± 2369 9549 ± 1456 12 012 ± 1651 14 149 ± 1976
Shimmer (local): dB 0.87 ± 0.08 1.03 ± 0.12 1.14 ± 0.20 0.85 ± 0.13 1.09 ± 0.14 1.25 ± 0.16
Shimmer (apq3): % 5364 ± 579 5330 ± 765 6195 ± 715 6279 ± 865 6694 ± 1309 7045 ± 916
Shimmer (apq5): % 6018 ± 650 5664 ± 817 7186 ± 898 7333 ± 1009 7846 ± 1499 9068 ± 1335
Shimmer (apq11): % 7453 ± 939 6656 ± 965 9063 ± 1315 8736 ± 1482 10 459 ± 1927 10 799 ± 1885
NHR 0.06 ± 0.01 0.11 ± 0.02 0.20 ± 0.08 0.06 ± 0.02 0.10 ± 0.05 0.13 ± 0.04
Abbreviations: rap, relative average perturbation; ppq, period perturbation quotient; apq, amplitude perturbation quotient.
Notes: Acoustic parameters measured before and after vocal loading. Frequency perturbation (jitter) expressed as absolute value. Amplitude perturbation
(shimmer) expressed as local in decibels (dB). Numeral subtypes accompanying analysis indicates the specific calculation performed by the Praat software.
Before and after indicates before and after vocal loading, respectively. Values in mean ± SEM.

The value for the MPT was not statistically significant be-
tween groups or when values were compared before and after
vocal loading (Figure 1C). However, a trend to higher levels
of MPT was observed in the overweight group compared with
the other two groups, before and after vocal loading but which
did not reach statistically significant values (P > 0.05 in all
comparisons).
Taking into account the significant differences in F0 and
HNR, we performed a post hoc analysis considering the
behavior of these two acoustic parameters before and after
loading in the analyzed groups. Thus, Figure 2 describes the
proportion (%) of participants in each group who showed an in-
crease or decrease in the value of these two parameters after
vocal loading. The percentage of overweight teachers who ex-
hibited an increase in F0 (Figure 2A) and HNR (Figure 2B) was
higher than in the normal weight group (c2, 10.4 and 16.2,
respectively, P < 0.05 in both cases). A lower percentage of
obese teachers (n ¼ 1, 14%) exhibited an increase in the
HNR value after vocal loading compared with normal weight
teachers (n ¼ 4, 33%). In the same analysis, a lower percentage
of overweight teachers showed increased jitter or shimmer
compared with normal weight group (data not shown).
Table 4 shows the characteristics of the individual that
showed an increase or decrease in the HNR (Figure 2). No sig-
nificant changes were observed in age, BMI, years of service,
hours of classes per week, or score in the VHI test, when the
subgroups were compared within or across the various BMI
groups.
We performed a nonparametric correlation between BMI and
every single acoustic parameter analyzed in this research. We
did not find significant correlations in any of the analyses
when all teachers were included (see Supplementary Table
S1). However, in the analysis within groups, HNR (r ¼ 0.66;
95% confidence interval, 0.12–0.89; P ¼ 0.01) and F0
(r ¼ 0.72; 95% confidence interval, 0.22–0.91; P ¼ 0.01)
were positively correlated with BMI only in the normal weight
group.
Salivary IL-6, acoustic parameters, and BMI
Relative concentrations (absorbance units in enzyme-linked
immunosorbent assay) of IL-6 in saliva are presented in
Figure 3. IL-6 concentrations were significant (ANOVA,
P ¼ 0.03) when all groups were analyzed before and after vocal
loading. Also, overweight teachers exhibited a significant
decrease in salivary IL-6 after loading compared with before
loading (0.40 ± 0.05 vs 0.30 ± 0.007 absorbance units, respec-
tively, P ¼ 0.0003). In normal or obese groups, the vocal load
did not affect the levels of IL-6 in saliva.
Concentrations of IL-6 in saliva were not correlated with
BMI either before (r ¼ 0.15, P ¼ 0.42) or after
(r ¼ 0.008, P ¼ 0.66) vocal loading when all teachers were
included. Also, no significant correlation was observed between
IL-6, either before or after loading, and BMI within the three
analyzed groups (data not shown). Interestingly, IL-6 in saliva
was negatively correlated with HNR before (r ¼ 0.74,
P ¼ 0.008) but not after (r ¼ 0.54, P ¼ 0.06) vocal loading
in the group of normal weight (Figure 4A). In contrast, in the
overweight group, the values of IL-6 and HNR were not corre-
lated before vocal loading but were significantly correlated af-
ter loading (r ¼ 0.69, P ¼ 0.05) (Figure 4B). There was no other
significant correlation when IL-6 in saliva was compared with
the analyzed acoustic parameters in normal, overweight, or
obese group, either before or after vocal loading (Table S2).
DISCUSSION
Overweight and obesity are well-recognized public health prob-
lems worldwide;25,26 however, its impact on voice alterations is
unclear. Initially, indirect evidence suggested OSAS, a
pathology commonly related with obesity,27,28 might be
associated with reduction in the compliance of the upper
airway or vocal tract that will cause a change in damping,
which in turn will change perceived vocal quality and may be
identified as a resonance disorder.29 Indeed, fat deposits in
the tongue and pharyngeal walls as well as changes in the
578
FIGURE 1. Acoustic parameter of voice in the studied groups.
Acoustic parameters were analyzed in normal (normal, n ¼ 12), over-
weight (n ¼ 8), or obese (n ¼ 7) teachers. The Praat software was used
for measuring (A) fundamental frequency (pitch, F0); (B) harmonic-to-
noise ratio; and (C) maximum phonation time before (,) and after
(-) vocal loading (speech for 60–90 minutes) in the three groups.
In (A) yP < 0.05 versus before vocal loading in the obese group. In
(B) *P < 0.05 versus before vocal loading in normal weight. Values
are mean ± standard error of the mean.
pharyngeal shape have been described in obese patients with
OSAS,7 whereas normal patients with OSAS exhibited reduced
maximum frequency and number of harmonics utterances
compared with healthy subjects.9 Moreover, Solomon et al11
showed no alteration in the vocal parameters including jitter,
shimmer, NHR, semitone sound pressure levels, and MPT in
the obese compared with normal-overweight group. Also, these
authors did not find significant variations in any aeromechanic
variables, including phonation threshold pressure or F0,
concluding that obesity does not affect either perception
of voice quality or acoustic and aeromechanic measures.
Journal of Voice, Vol. 28, No. 5, 2014
FIGURE 2. Subgroups of teachers organized according to changes
in acoustic parameters. The variation between values after vocal
loading minus respective value before vocal loading was calculated,
then the percentage of participant who showed increase (,) or
decrease (-) in (A) fundamental frequency (pitch, F0) or (B)
harmonic-to-noise ratio with respect to their own group of analysis
(ie, normal, overweight, and obesity). *P < 0.05 versus before normal
weight.
Interestingly, these authors also analyzed the effect of weight
loss induced by bariatric surgery in the obese group and sug-
gested that changes in phonation threshold pressure may be
influenced by BMI.
da Cunha et al10 found an increased frequency of altered lar-
yngoscopic analysis in obese versus normal-overweight adults
(26% vs 13%). Also, they found that a higher proportion of
obese participants presented vocal strangulation, as well as
alteration in jitter, shimmer, and noise compared with controls,
but obese people exhibited a reduced MPT. The authors
concluded that voice in obese individuals is characterized by
hoarseness, murmuring, and vocal instability and that excessive
fat accumulation in the larynx might alter MPT, which in turn
would impair myoelastic and aerodynamic forces in the larynx
adjustment required for adequate phonation.
In this report, teachers without any subjective signs of voice
alterations and classified by BMI exhibited nonsignificant dif-
ferences in jitter, shimmer, NHR, and MPT either before or af-
ter vocal loading. Yet the obese group exhibited elevation in the
F0 after vocal loading compared with their own basal condi-
tion, whereas no effect was evident in the other two groups.
In addition, a significant reduction in the HNR was observed
after vocal load in the normal weight group. Also, we found
that more overweight teachers exhibited elevation in F0 and
HNR after loading than the normal weight teachers. At first
Jesenia Acurio, et al IL-6 and Acoustic Parameter in Obese and Overweight 579

TABLE 4.
Characteristics of the Participants According to Delta of HNR Between Measurements Before and After Vocal Load
Normal Overweight Obese

Increase Decrease Increase Decrease Increase Decrease

Characteristic (n ¼ 4) (n ¼ 8) (n ¼ 5) (n ¼ 3) (n ¼ 1) (n ¼ 6)
BMI (kg/m2) 23.5 ± 0.5 23.1 ± 0.4 28.4 ± 0.5 27.7 ± 1.2 36.0 33.8 ± 1.2
Age (y) 33.7 ± 5.8 30.3 ± 0.8 39.4 ± 3.8 34.5 ± 6.5 35.0 47.8 ± 8.3
Classes per week (h/wk) 26.7 ± 11.0 37.6 ± 1.7 32.0 ± 7.1 39.0 ± 1.0 38.0 30.8 ± 5.4
Experience as teacher (y) 10.0 ± 3.6 5.3 ± 1.1 9.4 ± 2.7 5.0 ± 1.0 7.0 22.5 ± 6.8
VHI total 9.8 ± 5.2 17.9 ± 5.4 16.5 ± 3.0 8.7 ± 2.9 10.0 16.3 ± 7.3
Notes: Increase and decrease indicates trend of value after vocal loading, respectively. Values in mean ± SEM.

sight, our results may suggest that obesity may impact voice Contrary to the other studies,10,11 we included participants
generation, via restriction of vocal fold vibration that is evi- without any subjective signs of abnormal voice. At least one-
denced after vocal loading, as suggested by da Cunha et al10 quarter of participants in the study by da Cunha et al and
and partially supported by Solomon et al.11 However, we could one-third in the study by Solomon et al exhibited some degree
not find a significant correlation between any of the acoustic of vocal fold alterations. Data analysis was also different mak-
parameters and BMI. Lack of statistical significance might ing direct comparisons hard to make. Nevertheless, there
be explained by sample size. Alternatively, the site of fat distri- seems to be a general agreement that excessive fat deposits
bution, abdominal versus upper airway, may induce selective may impact F0.
changes in vocal fold vibration that are not revealed just by
measuring BMI.
There are many methodological differences between our
study and those of Solomon et al11 and da Cunha et al.10 We
did not include individuals with morbid obesity
(BMI > 40 kg/m2) and categorized the participants into three
groups, normal, overweight, and obese. Solomon et al11
included only two participants with normal weight (BMI
<25 kg/m2) in the control group, whereas the rest (n ¼ 6)
were overweight people. In the study by da Cunha et al,10
only 51% of participants in the control group had a BMI
<25 kg/m2, whereas 49% were overweight. In our study, we
included teachers, who are trained in the use of their voice.

FIGURE 4. Correlation between salivary IL-6 levels and harmonic-

FIGURE 3. Salivary IL-6 level before and after vocal loading in
the studied groups. Samples of saliva were obtained before (,)
and after (-) vocal loading from normal weight (n ¼ 12), over-
weight (n ¼ 8), or obese (n ¼ 7) teachers. Samples were used for
estimating the relative concentration of IL-6 by colorimetric assay
as detailed in Methods section. yP < 0.05 versus value before vocal
load in the overweight group. Values are mean ± standard error of
the mean.
to-noise ratio in the studied groups. Relative concentration of IL-6 was
correlated with harmonic-to-noise ratio in normal (B), overweight
(,), and obese teachers (6). Comparisons were performed using
the respective values (A) before vocal load or (B) after vocal load. In
(A) significant inverse correlation was observed in normal group
(r2 ¼ 0.44, P < 0.05). In (B) significant positive correlation was
observed in overweight group (r2 ¼ 0.61, P < 0.05). Each dot repre-
sents an individual participant.
580
Salivary IL-6 levels in overweight and obese
Saliva has been considered as a potential source of bio-
markers.17,30 In particular, quantification of proinflammatory
cytokines in saliva has been associated with proinflammatory
conditions such as oral cancer or chronic oral inflammation,31
but there are no studies regarding the potential role of salivary
proinflammatory cytokines during voice alterations. It is not
clear which cells are responsible for IL-6 synthesis and release
into the salivary fluid. Potential candidates include local macro-
phages, epithelial cells, and even adipocytes, among others.
Salivary production of IL-6 exhibits a circadian behavior
during the day in healthy individuals, whose levels are between
10 and 30 pg/mg of total protein.14 Another study reported IL-6
salivary concentration in the range of 1–5 pg/mL.17 In our
study, saliva samples were taken during the morning peak of
IL-6 production, and levels were estimated by using a highly
sensitive technique, whose limit of detection is 2 pg/mL. De-
tected IL-6 levels in our samples were too low, and most
(85%) of the values were below 3 pg/mL. Considering this lim-
itation, we present the results as relative values by using units of
observance rather than values in concentration. Then, we estab-
lished that clinical significance of reported relative value of IL-
6 in our results is limited, and they should be analyzed in the
context of research.
We find that relative concentrations of IL-6 in saliva, either
before or after vocal loading, do not correlate with BMI.
Because circulating IL-6 does not correlate with salivary IL-
6,32 then salivary IL-6 levels may represent local rather than
systemic inflammation.
IL-6 and voice parameters
Contrary to our original hypothesis, our results indicate a
significant reduction in the relative values of IL-6 in over-
weight teachers after vocal loading with a trend to decrease
in normal weight but no change in the obese teachers.
Considering evidences in human and culture models might
be feasible that reduction of IL-6 levels after vocal loading
can be related to an adequate vibration in the vocal fold.
Thus, in subjects exposed to vocal loading it has been re-
ported that vibratory therapy reduces the IL-6 levels in vocal
fold exudate,12 whereas in rabbit fibroblasts isolated from
vocal fold it has been described that cyclic tensile strain re-
duces IL-1b levels.13 Consistent with this, we describe a pos-
itive correlation between salivary IL-6 levels and HNR after
vocal loading, suggesting that the higher the vibration, the
lower the salivary IL-6 levels, at least in overweight teachers.
However, in participants with normal weight, it was found
that salivary IL-6 is negatively correlated with HNR only
before vocal loading. Clearly, more studies are required to
elucidate these findings.
In conclusion, obesity is associated with elevation of F0 after
vocal loading, without modification of the other parameters.
Moreover, we found that the level of IL-6 in saliva was reduced
after vocal loading only in the overweight group. In addition,
salivary IL-6 seems to be associated with HNR but not with
BMI. Further investigations are needed to determine the phys-
iological and therapeutic significance of these changes.
Journal of Voice, Vol. 28, No. 5, 2014
Acknowledgments
We thank teachers in the Sidney School Chillan, Chile, research
staff at Vascular Physiology Laboratory, Group of Investigation
of Tumor Angiogenesis (GIANT), and Group of Research and
Innovation in Vascular Health for their technical support. Also,
we thank Prof Leslie Myatt from Center for Pregnancy and
Newborn Research, University of Texas Health Science Center
San Antonio for his diligent review of this manuscript. We
thank Prof Rodrigo Fuenzalida for helping with sample collec-
tion. Fondecyt Regular 1100684, Conicyt 79112027, DIUBB
122109 GI/EF financed this study.
Individual contribution: This work was carried out in collab-
oration between all authors. C.E. and J.P. defined the research
topic. J.P. and C.E. designed the experiments. C.C. and J.A. per-
formed all experiments. All authors cowrote the manuscript. All
authors approved final version of the manuscript.
Supplementary material
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jvoice.2014.03.002.
REFERENCES
1. Malik VS, Willett WC, Hu FB. Global obesity: trends, risk factors and pol-
icy implications. Nat Rev Endocrinol. 2013;9:13–27.
2. Misra A, Khurana L. Obesity and the metabolic syndrome in developing
countries. J Clin Endocrinol Metab. 2008;93:S9–S30.
3. Rueda-Clausen CF, Silva FA, Lopez-Jaramillo P. Epidemic of overweight
and obesity in Latin America and the Caribbean. Int J Cardiol. 2008;125:
111–112.
4. Shelton KE, Woodson H, Gay S, Suratt PM. Pharyngeal fat in obstructive
sleep apnea. Am Rev Respir Dis. 1993;148:462–466.
5. Mortimore IL, Marshall I, Wraith PK, Sellar RJ, Douglas NJ. Neck and to-
tal body fat deposition in nonobese and obese patients with sleep apnea
compared with that in control subjects. Am J Respir Crit Care Med.
1998;157:280–283.
6. Sakakibara H, Tong M, Matsushita K, Hirata M, Konishi Y, Suetsugu S.
Cephalometric abnormalities in non-obese and obese patients with obstruc-
tive sleep apnoea. Eur Respir J. 1999;13:403–410.
7. Horner RL, Mohiaddin RH, Lowell DG, Shea SA, Burman ED,
Longmore DB, Guz A. Sites and sizes of fat deposits around the pharynx
in obese patients with obstructive sleep apnoea and weight matched con-
trols. Eur Respir J. 1989;2:613–622.
8. Salome CM, King GG, Berend N. Physiology of obesity and effects on lung
function. J Appl Physiol. 2010;108:206–211.
9. Fiz JA, Morera J, Abad J, et al. Acoustic analysis of vowel emission in
obstructive sleep apnea. Chest. 1993;104:1093–1096.
10. da Cunha MG, Passerotti GH, Weber R, Zilberstein B, Cecconello I. Voice
feature characteristic in morbid obese population. Obes Surg. 2011;21:
340–344.
11. Solomon NP, Helou LB, Dietrich-Burns K, Stojadinovic A. Do obesity and
weight loss affect vocal function? Semin Speech Lang. 2011;32:31–42.
12. Verdolini Abbott K, Li NY, Branski RC, Rosen CA, Grillo E, Steinhauer K,
Hebda PA. Vocal exercise may attenuate acute vocal fold inflammation. J
Voice. 2012;26:814.e11–814.e13.
13. Branski RC, Perera P, Verdolini K, Rosen CA, Hebda PA, Agarwal S. Dy-
namic biomechanical strain inhibits IL-1beta-induced inflammation in
vocal fold fibroblasts. J Voice. 2007;21:651–660.
14. Izawa S, Miki K, Liu X, Ogawa N. The diurnal patterns of salivary
interleukin-6 and C-reactive protein in healthy young adults. Brain Behav
Immun. 2013;27:38–41.
15. Okada N, Kobayashi M, Mugikura K, Okamatsu Y, Hanazawa S, Kitano S,
Hasegawa K. Interleukin-6 production in human fibroblasts derived from
Jesenia Acurio, et al IL-6 and Acoustic Parameter in Obese and Overweight
periodontal tissues is differentially regulated by cytokines and a glucocor-
ticoid. J Periodontal Res. 1997;32:559–569.
16. Sun D, Emmert-Buck MR, Fox PC. Differential cytokine mRNA expres-
sion in human labial minor salivary glands in primary Sjogren’s syndrome.
Autoimmunity. 1998;28:125–137.
17. Izawa S, Sugaya N, Kimura K, et al. An increase in salivary interleukin-6
level following acute psychosocial stress and its biological correlates in
healthy young adults. Biol Psychol. 2013;94:249–254.
18. Bastard JP, Jardel C, Bruckert E, et al. Elevated levels of interleukin 6 are
reduced in serum and subcutaneous adipose tissue of obese women after
weight loss. J Clin Endocrinol Metab. 2000;85:3338–3342.
19. Meijer K, de Vries M, Al-Lahham S, et al. Human primary adipocytes
exhibit immune cell function: adipocytes prime inflammation independent
of macrophages. PLoS One. 2011;6:e17154.
20. Wieser V, Moschen AR, Tilg H. Inflammation, cytokines and insulin resis-
tance: a clinical perspective. Arch Immunol Ther Exp (Warsz). 2013;61:
119–125.
21. Sala E, Laine A, Simberg S, Pentti J, Suonpaa J. The prevalence of voice
disorders among day care center teachers compared with nurses: a question-
naire and clinical study. J Voice. 2001;15:413–423.
22. Vilkman E. Occupational safety and health aspects of voice and speech pro-
fessions. Folia Phoniatr Logop. 2004;56:220–253.
23. Eysel-Gosepath K, Daut T, Pinger A, Lehmacher W, Erren T. Sound levels
and their effects on children in a German primary school. Eur Arch Otorhi-
nolaryngol. 2012;269:2475–2483.
581
24. Rosen CA, Lee AS, Osborne J, Zullo T, Murry T. Development and
validation of the voice handicap index-10. Laryngoscope. 2004;114:
1549–1556.
25. Nguyen DM, El-Serag HB. The epidemiology of obesity. Gastroenterol
Clin North Am. 2010;39:1–7.
26. Obesity: preventing and managing the global epidemic. Report of a WHO
consultation. World Health Organ Tech Rep Ser. 2000;894:1–253.
27. Alam I, Lewis K, Stephens JW, Baxter JN. Obesity, metabolic syndrome
and sleep apnoea: all pro-inflammatory states. Obes Rev. 2007;8:119–127.
28. Lam JC, Mak JC, Ip MS. Obesity, obstructive sleep apnoea and metabolic
syndrome. Respirology. 2012;17:223–236.
29. Fox AW, Monoson PK, Morgan CD. Speech dysfunction of obstructive
sleep apnea. A discriminant analysis of its descriptors. Chest. 1989;96:
589–595.
30. Yoshizawa JM, Schafer CA, Schafer JJ, Farrell JJ, Paster BJ, Wong DT.
Salivary biomarkers: toward future clinical and diagnostic utilities. Clin
Microbiol Rev. 2013;26:781–791.
31. Lisa Cheng YS, Jordan L, Gorugantula LM, Schneiderman E, Chen HS,
Rees T. Salivary Interleukins 6 and 8 in Oral Cancer Patients and in Pa-
tients With Chronic Oral Inflammatory Diseases. J Periodontol. 2013
Oct 23. [Epub ahead of print].
32. Fernandez-Botran R, Miller JJ, Burns VE, Newton TL. Correlations among
inflammatory markers in plasma, saliva and oral mucosal transudate in
post-menopausal women with past intimate partner violence. Brain Behav
Immun. 2011;25:314–321.