International Journal of Obesity (2008) 32, 788–794
& 2008 Nature Publishing Group All rights reserved 0307-0565/08 $30.00
www.nature.com/ijo
ORIGINAL ARTICLE
Epidemiological evidence of altered cardiac autonomic
function in overweight but not underweight subjects
J-S Wu1,2, FH Lu1,2, Y-C Yang1,2, T-S Lin3, Y-H Huang2, C-H Wu2, J-J Chen4 and C-J Chang1,2
1
Department of Family Medicine, College of Medicine, National Cheng Kung University, Tainan, Taiwan, Republic of China;
Department of Family Medicine, National Cheng Kung University Hospital, Tainan, Taiwan, Republic of China;
3
Department of Neurology, College of Medicine, National Cheng Kung University, Tainan, Taiwan, Republic of China and
4
Institute of Biomedical Engineering, National Cheng Kung University, Tainan, Taiwan, Republic of China
2
Background: Little is known about the altered cardiac autonomic function (CAF) across different levels of body mass index
(BMI), including underweight, normal weight, overweight and obesity. This study provides a thorough analysis to clarify the CAF
change in subjects with underweight, overweight and obesity.
Methods: According to the World Health Organization (WHO) Asia-Pacific BMI cutoffs, a total of 1437 participants were
classified as underweight (n ¼ 74), normal weight (n ¼ 588), overweight (n ¼ 313), obesity I (n ¼ 390) and obesity II (n ¼ 72).
CAF was determined by standard deviation of normal-to-normal (SDNN) intervals or RR intervals, power spectrum in low (LF)
and high frequency (HF) (LF, 0.04–0.15 Hz; HF, 0.15–0.40 Hz), and LF/HF ratio at supine for 5 min, the ratio between the 30th
and the 15th RR interval after standing from the supine position (30/15 ratio) and the average heart-rate change while taking six
deep breaths in 1 min (HRDB).
Results: There were significant differences in age, gender, socioeconomic status, blood pressure, HOMA insulin resistance index,
fasting glucose, cholesterol, triglyceride and high-density lipoprotein (HDL)-C, and the prevalence of hypertension, ischemic/left
bundle branch block (LBBB) electrocardiography (EKG) pattern, current smoking and alcohol use among subjects with
underweight, normal weight, overweight, obesity I and II. Univariate analysis showed that SDNN, HRDB, HF power and the
square root of the LF/HF ratio differed among these five groups. Multivariate analysis showed that obesity I and II were inverse
correlates of HRDB and HF power. Overweight, obesity I and II were positively associated with the square root of the LF/HF ratio.
No BMI status was related to SDNN, 30/15 ratio or LF power. Underweight was not the independent correlate of any CAF
indices.
Conclusions: The risk for altered CAF is significant in overweight and obese subjects, independent of cardiovascular risk factors.
Underweight is not apparently associated with CAF change.
International Journal of Obesity (2008) 32, 788–794; doi:10.1038/sj.ijo.0803791; published online 29 January 2008
Keywords: epidemiological study; cardiac autonomic function; heart-rate variability; underweight; overweight
Introduction
The World Health Organization (WHO) cutoff points of body
mass index (BMI) are based on the risk of chronic disease.1
BMI is useful in monitoring not only the levels of overweight
and obesity but also the degree of undernutrition.2 Recent
studies have shown that underweight (BMI o18.5 kg m2)
and obesity (BMIX30 kg m2) are associated with increased
Correspondence: Professor C-J Chang, Department of Family Medicine,
College of Medicine, National Cheng Kung University Hospital, 138 ShengLi Road, Tainan, Taiwan 70403, Republic of China.
E-mail: [email protected]
Received 29 May 2007; revised 20 November 2007; accepted 27 November
2007; published online 29 January 2008
total or cardiovascular mortality relative to the normal
weight category.3–5 Thus, not only obesity but also a low
BMI may be a risk factor for cardiovascular disease.4,6 It is
conceivable that an alteration of the autonomic nervous
system may be associated with a change in BMI, because the
autonomic function is involved in energy metabolism7 and
the regulation of the cardiovascular system.8
A change in heart rate during deep breathing (HRDB) and
the ratio between the 30th and the 15th RR interval after
standing from the supine position (30/15 ratio) are traditional autonomic tests.9 Heart rate variability (HRV), an
indirect measure of cardiac autonomic function (CAF), is a
useful tool for evaluating sympathetic and parasympathetic
modulation of the heart (Table 1).8,9 The vagal activity is the
major contributor to the high-frequency (HF) component
Autonomic function, overweight and underweight
J-S Wu et al
789
Table 1 Indices of cardiac autonomic function tests based on their
physiological meaning
1. Parasympathetic
HRDB (beat min1)
30/15 ratio
SDNN (ms)
HF power (ms2)
The average of six heart-rate changes, defined by the
maximum minus minimum heart rate during each
breathing cycle, while subjects were sitting and
breathing deeply at a rate of six breaths a minute
The ratio between the 30th and the 15th RR interval
after standing from a supine position
The standard deviation of normal-to-normal (SDNN)
intervals or RR intervals during a time interval
High-frequency (0.15–0.40 Hz) power from the
spectral analysis on successive RR intervals of a time
interval
2. Predominantly sympathetic with parasympathetic modulation
LF power (msec2)
Low-frequency (0.04–0.15 Hz) power from the
spectral analysis on successive RR intervals of a time
interval
3. Sympathovagal balance
LF/HF ratio
The ratio between LF power and HF power
Abbreviations: HF, high frequency; HRDB, heart rate during deep breathing;
LF, low frequency; SDNN, standard deviation of normal-to-normal.
and the low-frequency (LF) component is predominantly
under sympathetic control with vagal modulation. LF/HF
ratio is considered an index of sympathovagal balance.8 The
frequency components of HRV are of similar diagnostic value
as the Ewing battery concerning the presence of cardiovascular autonomic neuropathy.10 Furthermore, HF spectral
indices were found to correlate very strongly with conventional methods.11
Studies on the relationship between CAF and BMI have
focused largely on obesity11–21 and only one study has
assessed HRV in underweight subjects.22 However, the results
on the effect of BMI on CAF are inconsistent.11–22 Most of the
above studies were hospital-based. Only one study was
population-based, but the number of study subjects
was limited (n ¼ 93).16 Little is known about CAF change
across the spectrum of BMI, including underweight, normal
weight, overweight and obesity. Therefore, the aim of this
study was to clarify whether CAF is altered in subjects with
obesity, overweight and even underweight after adjustment
for other confounders using population-based data from
Taiwan.
Methods
Participants
The subjects were recruited in Tainan city, which is the oldest
city in southern Taiwan with a population of 700 000. The
selection procedure used a three-stage sampling scheme to
generate a stratified systematic cluster sample of households
throughout Tainan. First, the city was classified formally
into seven administrative districts. In each of the districts,
one area was selected from each stratum by adopting a
probability proportional to the size of the areas within that
specific stratum. Second, every fifth household within each
of the seven selected areas was identified systematically.
Third, all the members of the selected households aged 20
years or older were invited to participate in the study. There
were 2416 eligible subjects for systematic sampling from
seven administrative districts. A total of 1638 subjects,
representing a response rate of 68%, were included. The
non-responders were slightly younger and consisted of more
men compared with the responders, but the differences were
not statistically significant. All the subjects were Chinese and
none of them had any other racial background. Details of the
study sampling have been described elsewhere.23 Written
informed consent was obtained from all the subjects. The
research committee of National Cheng Kung University
Hospital, Taiwan, approved this study. We excluded 201
subjects who had taken medications known to influence
CAF, such as antihypertensive, antiparkinsonism, narcotics,
sedatives, antipsychotics or antidepressants within 2 weeks
of the study. Therefore, 1437 subjects were included for the
final analysis.
Clinical measurements
The participants were instructed not to consume alcohol,
coffee, tea or cigarettes on the day of the examination.
They were interviewed by a well-trained assistant using a
structured questionnaire that included demographic
characteristics, medical history and use of medications,
cigarettes, alcohol and dietary habits as well as physical
activity during the past year. For the modified Hollingshead
two-factor index of social position, occupation was coded
from 1 to 5 and educational level was graded on a 1–5 scale.
The Hollingshead index was then calculated by multiplying
the occupation scale value by a weight of 7, and the
education scale value by 4 and summing the products. Five
levels of socioeconomic status were created: lower (score:
11–18), lower-middle (score: 19–29), middle (score: 30–40),
upper-middle (score: 41–51) and upper (score: 52–55).24
Total physical activity, including work, walking and leisure
time, was assessed in metabolic equivalent-hours per week
for the past year.25
Body weight and height measurements, and fasting blood
samples were collected between 0800 and 1000 hours. All the
subjects had body weight and height measured by welltrained nurses while fasting and wearing only light indoor
clothes and without shoes. Each subject also received a
complete physical examination by physicians. According to
WHO Asia-Pacific BMI cutoffs proposed by the International
Obesity Taskforce,26 the subjects were classified into underweight (BMIo18.5 kg m2), normal weight (BMI 18.5–
22.9 kg m2), overweight (BMI 23–24.9 kg m2), obesity I
(BMI 25–29.9 kg m2) and obesity II (BMI Z30 kg m2) groups.
The laboratory tests included blood biochemistry, hemogram, insulin, urine examination and electrocardiography
(EKG) after an overnight fast of at least 10 h. The subjects
International Journal of Obesity
Autonomic function, overweight and underweight
J-S Wu et al
790
without a history of diabetes received a 75 g oral glucose
tolerance test after completing the measurement of their
blood pressure and HRV. Blood samples were obtained 2 h
after the subjects began to drink the glucose solution. Insulin
resistance (IR) was calculated using HOMA, serum insulin
(mU ml1) fasting glucose (mmol l1)/22.5.27 Ischemic/
LBBB EKG pattern included Q-QS abnormalities, ST segment
depression, T-wave changes and left bundle branch block
(LBBB) EKG pattern included Q-QS abnormalities, ST
segment depression, T wave changes and LBBB according
to the Minnesota code (1.1–3; 4.1–3; 5.1–3 and 7.1).28
Measurements of blood pressure and HRV
Before measurements, all the participants were informed
about the purpose and procedures of the test. Blood pressure
was measured with a DINAMAP TM vital sign monitor
(Model 1846SX; Critikon Inc, Irvine, CA, USA) in a quiet
location. Measurements were obtained in the fasting state
and an appropriate-sized cuff was wrapped around the right
upper arm after the subject had rested in a seated position for
at least 15 min between 0800 and 1000 hours. Two seated
blood pressure readings were separated by at least 5-min
intervals. Hypertension was defined as the average of the two
seated readings of systolic/diastolic blood pressure X140/
90 mm Hg or a positive history of hypertension.29
Before the start of the HRV assessment, the subjects rested
in the supine position for at least 15 min. The RR interval was
recorded by continuous measurement of beat-to-beat cardiac
cycle duration with an EKG monitor (CardiSuny a 800;
Fukuda M-E Kogyo Inc., Tokyo, Japan). A personal computerbased data acquisition system was used to receive the signal
from the EKG monitor. The HRV measurements were
obtained in the following sequence: (1) normal breathing
for 5 min in the supine position, (2) an active change from
lying to standing position and then (3) six deep breaths over
1 min while sitting after a 10-min rest period. The analog
signals from the EKG monitor were immediately sent to the
signal-acquiring and -processing system, which converted
the analog signals to digital signals (DAQPad-6020E and
SCB-68; National Instrument, NI) and then stored them on a
personal computer using a universal serial bus connection.
The EKG signals were processed for R-peak detection with
the LabView 6.1 software program (National Instruments,
NI). A power spectral analysis was used to define the
temporal fluctuations of HRV. The LF (0.04–0.15 Hz) and
HF (0.15–0.4 Hz) components were identified for each
subject.8 In summary, the measures of CAF used in this
study were the following: (1) standard deviation of normalto-normal (SDNN), LF power, HF power and LF/HF ratio in
the supine position for 5 min; (2) 30/15 ratio, and (3) the
average of six HRDB in 1 min while sitting after a 10-min rest.
Statistical analyses
Data analyses were performed using the Statistical Package of
Social Science 10.0 for Windows software. A square root
International Journal of Obesity
transformation was used to make the values follow a normal
distribution for non-normally distributed variables including,
plasma triglyceride, physical activity level, HOMA IR and
LF/HF ratio. In univariate analysis, analysis of variance
(ANOVA) was used for the comparisons of continuous
variables among subjects with underweight, normal weight,
overweight, obesity I and II. Comparisons of categorical
variables were analyzed by the w2 test or Fisher’s exact test for
cells less than 5.
The relationship between CAF and the different BMI levels
was examined in the following way: first without adjustment, then adjusting for age and finally adjusting for age
and other confounders. The outcome variables were HRV,
indicated by SDNN, 30/15 ratio, HRDB, LF power, HF power
and square root of LF/HF ratio, respectively. In the final
model, the predictor variables were underweight vs normal
weight, overweight vs normal weight, obesity I vs normal
weight, obesity II vs normal weight, age, gender, socioeconomic status (lower vs upper/upper-middle, lower-middle vs upper/upper-middle, middle vs upper/upper-middle),
HOMA IR index, plasma cholesterol, triglyceride and highdensity lipoprotein (HDL)-C, hypertension, ischemic/LBBB
EKG pattern, physical activity, current smoking and alcohol
use. The P-value of o0.05 was considered significant.
Results
A total of 1437 participants were included and classified
into underweight (BMIo18.5 kg m2, n ¼ 74), normal
weight (BMI 18.5–22.9 kg m2, n ¼ 588), overweight (BMI
23–24.9 kg m2, n ¼ 313), obesity I (BMI 25–29.9 kg m2,
n ¼ 390) and obesity II (BMI X30 kg m2, n ¼ 72) groups.
For the univariate analysis, Table 2 presents the comparisons of clinical variables among subjects with underweight,
normal weight, overweight, obesity I and obesity II.
There were significant differences in age (Po0.001), gender
(Po0.001), socioeconomic status (Po0.001), BMI (Po0.001),
systolic (Po0.001) and diastolic (Po0.001) blood pressures,
square root of HOMA IR index (Po0.001), fasting plasma
glucose (Po0.001), cholesterol (Po0.001) and HDL-C
(Po0.001), square root of plasma triglyceride (Po0.001)
and the prevalence of hypertension (Po0.001), ischemic/
LBBB EKG pattern (P ¼ 0.002), current smoking (P ¼ 0.05)
and alcohol use (Po0.001) among these five groups. After
adjusting for age, the differences in the above variables
except age were still significant among groups. However, the
levels of physical activity (P ¼ 0.20) and heart rates (P ¼ 0.44)
were not apparently different, even after adjusting for age
(physical activity, P ¼ 0.07; heart rates, P ¼ 0.15), among
these five groups.
Table 3 shows the comparisons of CAF among subjects
with underweight, normal weight, overweight, obesity I and
II. There were significant differences in SDNN (Po0.001),
HRDB (Po0.001), HF power (Po0.001) and square root of
Autonomic function, overweight and underweight
J-S Wu et al
791
Table 2
Comparison of clinical characteristics among subjects with underweight, normal weight, overweight, obesity I and II
Variables
Age (years)
Male gender (%)
Socioeconomic status
Upper/upper-middle (%)
Middle (%)
Lower-middle (%)
Lower (%)
Body mass index (kg m2)
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
Heart rate (beat min1)
Physical activity (m-h week1)a
Homa insulin resistance indexa
Fasting glucose (mmol l1)
Cholesterol (mmol l1)
Triglyceride (mmol l1)a
HDL-C (mmol l1)
Ischemic/LBBB EKG pattern (%)
Hypertension (%)
Current alcohol use (%)
Current smoking (%)
Underweight
(n ¼ 74)
Normal weight
(n ¼ 588)
Overweight
(n ¼ 313)
Obesity I
(n ¼ 390)
Obesity II
(n ¼ 72)
P-value
P-value age-adjusted
35±15
28
38±15
39
45±14
51
46±13
60
48±15
46
o0.001
o0.001
o0.001
3
38
44
15
12
28
48
12
9
24
43
24
13
21
42
24
7
17
41
36
o0.001
o0.001
17.7±0.7
107±13
66±8
68±12
52±121
1.42±0.79
5.0±0.7
4.4±1.0
0.93±0.35
1.5±0.3
3
5
5
19
21.0±1.2
111±15
68±9
66±12
57±80
1.52±1.07
5.1±1.3
4.8±1.0
1.05±0.55
1.4±0.4
10
6
9
19
23.9±0.5
118±17
71±9
66±11
59±58
1.95±1.71
5.4±1.5
5.1±1.1
1.44±1.00
1.3±0.4
11
12
16
20
26.8±1.3
122±18
75±10
66±11
68±106
2.71±3.22
5.7±1.8
5.1±1.1
1.87±2.09
1.2±0.3
14
17
18
27
32.4±2.7
130±18
76±12
67±11
51±52
3.78±3.18
5.5±1.3
5.2±1.1
2.06±1.39
1.2±0.4
21
29
15
18
o0.001
o0.001
o0.001
0.44
0.20
o0.001
o0.001
o0.001
o0.001
o0.001
0.002
o0.001
o0.001
0.05
o0.001
o0.001
o0.001
0.15
0.07
o0.001
o0.001
o0.001
o0.001
o0.001
0.04
o0.001
o0.001
0.03
Abbreviations: ANOVA, analysis of variance; HDL, high-density lipoprotein. aANOVA on square root-transformed data.
Table 3
Comparison of cardiac autonomic function among subjects with underweight, normal weight, overweight, obesity I and II
Variables
Underweight
(n ¼ 74)
Normal weight
(n ¼ 588)
Overweight
(n ¼ 313)
Obesity I
(n ¼ 390)
Obesity II
(n ¼ 72)
P-value
P-value age-adjusted
SDNN (ms)
30/15 ratio
HRDB (beat min1)
LF power (ms2)
HF power (ms2)
Square root of LF/HF ratio
37±20
1.10±0.11
20.6±8.9
762±455
406±225
1.53±0.86
41±25
1.09±0.13
19.0±8.1
779±420
379±200
1.60±0.79
35±21
1.09±0.13
16.6±8.3
822±465
350±199
1.78±1.00
32±22
1.08±0.11
15.7±8.1
782±429
328±205
1.84±1.11
29±15
1.09±0.10
13.2±7.5
859±443
275±188
2.21±1.40
o0.001
0.44
o0.001
0.38
o0.001
o0.001
0.002
0.20
o0.001
0.14
0.004
o0.001
Abbreviations: HF, high frequency; HRDB, heart rate during deep breathing; LF, low frequency; SDNN, standard deviation of normal-to-normal.
LF/HF ratio (Po0.001) but not the 30/15 ratio (P ¼ 0.44) and
LF power (P ¼ 0.38) among these five groups. With adjustment for age, there was still a difference in SDNN (P ¼ 0.002),
HRDB (Po0.001), HF power (P ¼ 0.004) and square root of
LF/HF ratio (Po0.001) among these five groups.
Figure 1 shows the association of different BMI levels with
CAF from multivariate analysis. Obesity I and II were
inversely associated with HRDB (obesity I, P ¼ 0.01; obesity
II, P ¼ 0.01) and HF power (obesity I, P ¼ 0.04; obesity II,
P ¼ 0.02), although underweight and overweight were not
independently related to HRDB (underweight, P ¼ 0.86; overweight, P ¼ 0.15) and HF power (underweight, P ¼ 0.39;
overweight, P ¼ 0.67) after adjusting for other factors.
Overweight (P ¼ 0.04), obesity I (P ¼ 0.03) and obesity II
(P ¼ 0.01), but not underweight (P ¼ 0.85), were positively
associated with the square root of the LF/HF ratio. However,
no BMI status was the independent correlate of SDNN
(underweight, P ¼ 0.41; overweight, P ¼ 0.80; obesity I,
P ¼ 0.43; obesity II, P ¼ 0.45), 30/15 ratio (underweight,
P ¼ 0.76; overweight, P ¼ 0.47; obesity I, P ¼ 0.74; obesity II,
P ¼ 0.18) or LF power (underweight, P ¼ 0.55; overweight,
P ¼ 0.94; obesity I, P ¼ 0.11; obesity II, P ¼ 0.28) in the
multivariate analysis. In summary, underweight was not a
correlate of any indices of CAF. Overweight and obesity were
independently associated factors of altered CAF.
Discussion
To our knowledge, this is the first population-based study to
reveal the change of HRV across the spectrum of BMI,
including underweight, normal weight, overweight and
obesity. For the change of CAF in obesity, our report
International Journal of Obesity
Autonomic function, overweight and underweight
J-S Wu et al
792
LF power
SDNN
Beta ( 95% CI)
-10
0
6
Beta
(95% CI)
-110
Underweight vs. NW
-2.47
(-8.83 ~ 3.36)
37.44
(-85.66 ~ 159.53)
Overweight vs. NW
-0.46
(-3.92 ~ 3.01)
-3.02
(-76.43 ~
70.40)
Obesity I vs. NW
-1.41
(-4.89 ~ 2.08)
19.05
(-52.14 ~
91.25)
Obesity II vs. NW
-2.32
(-8.32 ~ 3.69)
69.82
(-57.85 ~ 197.47)
Beta
(95% CI)
-10
0
6
-110
Underweight vs. NW
-0.18
(-1.77 ~ 2.12)
24.78
(-30.57 ~ 80.13)
Overweight vs. NW
-0.87
(-2.04 ~ 0.31)
-17.15
(-51.43 ~ 12.33)
Obesity I vs. NW
-0.97
(-1.95 ~ -0.21)∗
-40.52
(-70.29 ~ -5.15)∗
Obesity II vs. NW
-2.54
(-4.60 ~ -0.62)∗
-65.59
(-103.44 ~ -25.14)∗
30/15 ratio
Beta
(95% CI)
-0.10
200
0
200
HF power
HRDB
Beta ( 95% CI)
0
Square root of LF/HF ratio
0.0
0.06
Beta
(95% CI)
Underweight vs. NW
-0.01
(-0.04 ~ 0.03)
0.03
(-0.25 ~ 0.30)
Overweight vs. NW
-0.01
(-0.02 ~ 0.01)
0.15
( 0.03 ~ 0.24)∗
Obesity I vs. NW
-0.003 (-0.02 ~ 0.02)
0.16
( 0.06 ~ 0.25)∗
Obesity II vs. NW
-0.03
0.35
( 0.08 ~ 0.62) R
(-0.06 ~ 0.01)
-1.10
0.00
2.00
Figure 1 The association of different BMI levels with cardiac autonomic function from multiple linear regression analysis. b, b coefficient; BMI, body mass index; CI,
confidence interval; NW, normal weight. Dependent variables: HRV (heart rate variability) indicated by SDNN (standard deviation of normal-to-normal), 30/15 ratio,
HRDB, LF power, HF power, and square root of LF/HF ratio, respectively. Independent variables: underweight vs normal weight, overweight vs normal weight, obesity
I vs normal weight, obesity II vs normal weight, age, gender, socioeconomic status (lower vs upper/upper-middle, lower-middle vs upper/upper-middle, middle vs
upper/upper-middle), HOMA IR index, plasma cholesterol, triglyceride, and HDL-C (high-density lipoprotein), hypertension, ischemic/LBBB EKG pattern, physical
activity, current smoking, and alcohol use. *Po0.05, wP ¼ 0.01.
and some hospital-based studies showed that obese subjects
(BMIX30 kg m2) had an altered CAF,11–15,21 but some
reported no apparent association between CAF and
obesity.16–20 The discrepancy may be related to selection
bias and the limited number of study subjects in some
studies. Our results revealed a decrease in parasympathetic
tone (shown by a decrease in HF power and HRDB) and
autonomic imbalance (shown by an increase in LF/HF ratio)
in obese II subjects with BMIX30 kg m2 after carefully
adjusting for other confounding factors. The CAF change in
our obese II subjects is compatible with a pharmacological
intervention study of atropine that showed that weight gain
of 10% above initial body weight resulted in a decrease in
parasympathetic tone in seven lean subjects.30 In contrast,
losing 10% of initial body weight resulted in an increase in
parasympathetic activity in nine lean and seven obese
subjects.30
Our study also revealed that obese I subjects (BMI
25–29.9 kg m2) had an independently higher risk of reduced
CAF with an autonomic imbalance. Although a report for
CAF change in a specific population with BMI
25–29.9 kg m2 is not available, two studies have reported
that changes of HRV to a 5- to 10-min rest and HRDB were
not significantly associated with BMI in subjects with BMI
levels of 24±4 and 25.5±6.0 kg m2, respectively.18,20 Racial
difference may be a partial explanation for the discrepancy
between our study and the other two reports, because the
International Journal of Obesity
Asian population has always been characterized as ‘small’
and with a relatively high percentage of body fat at a given
BMI.26Although the Asia-Pacific population has adopted a
lower BMI cutoff for obesity (X25 kg m2) than the WHO
BMI cutoff (X30 kg m2),26 our study still shows that a
BMI of 25–29.9 kg m2 has a significant association with
alterations in CAF.
Because the BMI cutoff of 23–24.9 kg m2 is considered as
normal weight according to the WHO criterion, the assessment of CAF change in subjects with BMI levels of
23–24.9 kg m2 is ignored. However, overweight for the
Asia-Pacific population is defined as the BMI cutoff of
23–24.9 kg m2, not the WHO cutoff of 25–29.9 kg m2.26
Our overweight subjects (BMI 23–24.9 kg m2) had a significant CAF change in only the LF/HF ratio, not a decrease
in both HF power and HRDB that was seen in obese subjects.
This suggests that the pure impairment of cardiac parasympathetic tone in overweight subjects is not as apparent as
that in obese subjects. In addition to the CAF change in our
obese subjects with BMIX25 kg m2, our study also provides
epidemiological evidence that the impact of mild elevation
of BMI, that is, being overweight (BMI 23–24.9 kg m2), also
has an influence on CAF.
For CAF change in subjects with BMIo18.5 kg m2, there
were no differences in any index of CAF between subjects
with normal weight and underweight after adjusting for
other factors in our study. The results were similar to those of
Autonomic function, overweight and underweight
J-S Wu et al
793
another report from India that showed no difference in HRV
between normal control subjects with BMI levels of
21.4±1.6 kg m2 and well-nourished subjects with a lower
BMIo18.5 kg m2.22 Although the Indian report also
revealed a significant influence of low socioeconomic status
on HRV in underweight subjects,22 our study showed that
underweight was not independently related to any CAF
indices after adjusting for socioeconomic status and other
factors. However, the magnitude of the influence (b value) of
underweight on SDNN is larger than those of the other BMI
categories in multivariate analysis, and the non-significant
results may be due to the small number of underweight
subjects in our study. In contrast, the b value of underweight
for other indices of parasympathetic tone is not so
prominent when compared to other BMI categories, and
the b value of underweight for HF power was positive. Thus,
the parasympathetic modulation of the heart in underweight subjects does not seem apparently reduced.
A reduced 5-min rest HRV has been shown to be associated
with higher HbA1c, triglycerides, systolic blood pressure,
BMI and albumin excretion rate, which are key factors in
IR.31 Abnormal cardiac sympathovagal balance has been
found not only with obesity but also with hyperinsulinemia.32 Obesity has been considered a potential
link between autonomic nervous activity and IR.33
However, our study showed that overweight and obesity
were independently associated with the alteration of CAF
after adjusting for the HOMA IR index and other confounding factors in multivariate analysis. Thus, the
mechanism underlying the change of CAF with overweight and obesity is independent of IR. Although an
attenuated reflex due to impaired lung expansion may be
related to the change of CAF with overweight and obesity,34
the exact mechanism is still unknown. The involvement
of neural, hormonal and metabolic factors deserves to be
investigated.
Because the clinical relevance of CAF measures in overweight and obesity has to be determined, we tried to map the
CAF change across different BMI levels, from underweight,
normal weight to overweight and finally obesity. Our study
shows no significant change in CAF between subjects with
underweight and normal weight. The CAF change in overweight subjects is only an autonomic imbalance shifting
toward augmented sympathetic tone, shown only by an
increased LF/HF ratio but not by a decreased parasympathetic modulation of the heart. The alteration of CAF occurs
in obese subjects with not only an increased LF/HF ratio but
also a decreased parasympathetic tone. Thus, the parasympathetic tone declined with an autonomic imbalance
shifting toward augmented sympathetic tone during the
development from normal weight to obesity. In conclusion,
our results provide epidemiological evidence that the risk for
altered CAF is significant in both obese and overweight
subjects independent of cardiovascular risk factors. In
contrast, underweight is apparently not associated with
CAF change.
Acknowledgements
This study was supported by grants from the National
Science Council, Taiwan, Republic of China (NSC 882314-B-006-096, NSC 89-2314-B-006-043 and NSC 92-2314B-006-117).
References
1 WHO. Obesity: Preventing and Managing the Global Epidemic. WHO:
Geneva, 2000. Report of a WHO Consultation. WHO Technical
Report Series 894.
2 Shetty PS, James WPT. Body mass index: a measure of chronic
energy deficiency in adults. FAO Food Nutr Pap 1994; 56: 1–57.
3 Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deaths
associated with underweight, overweight, and obesity. JAMA
2005; 293: 1861–1867.
4 Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath Jr CW.
Body mass index and mortality in a prospective cohort of U.S.
adults. N Engl J Med 1999; 341: 1097–1105.
5 Higashi Y, Sasaki S, Nakagawa K, Kimura M, Noma K, Sasaki S et al.
Low body mass index is a risk factor for impaired endotheliumdependent vasodilation in humans: role of nitric oxide and
oxidative stress. J Am Coll Cardiol 2003; 42: 256–263.
6 Boschmann M, Schroeder C, Christensen NJ, Tank J, Krupp G,
Biaggioni I et al. Norepinephrine transporter function and
autonomic control of metabolism. J Clin Endocrinol Metab 2002;
87: 5130–5137.
7 Dart AM, Du XJ, Kingwell BA. Gender, sex hormones and
autonomic nervous control of the cardiovascular system.
Cardiovasc Res 2002; 53: 678–687.
8 Task force of the European Society of Cardiology and the North
American Society of Pacing and Electrophysiology. Heart rate
variability standards of measurement, physiological interpretation, and clinical use. Circulation 1996; 93: 1043–1065.
9 Howorka K, Pumprla J, Schabmann A. Optimal parameters for
short-term heart rate spectrogram for routine evaluation of
diabetic cardiovascular autonomic neuropathy. J Auton Nerv Syst
1998; 69: 164–172.
10 Freeman R, Saul P, Roberts M, Berger RD, Broadbridge C, Cohen R.
Spectral analysis of heart rate in diabetic autonomic neuropathy.
Arch Neurol 1991; 48: 185–190.
11 Rossi M, Marti G, Ricordi L, Fornasari G, Finardi G, Fratino P et al.
Cardiac autonomic dysfunction in obese subjects. Clin Sci 1989;
76: 567–572.
12 Zahorska-Markiewicz B, Kuagowska E, Kucio C, Klin M. Heart rate
variability in obesity. Int J Obes Relat Metab Disord 1993; 17:
21–23.
13 Arrone LJ, Mackintosh R, Rosenbaum M, Leibel RL, Hirsch J.
Cardiac autonomic nervous system activity in obese and neverobese young men. Obes Res 1997; 5: 354–359.
14 Freeman R, Weiss ST, Roberts M, Zbikowski SM, Sparrow D. The
relationship between heart rate variability and measures of body
habitus. Clin Auton Res 1995; 5: 261–266.
15 Valensi P, Paries J, Attali JR. French Group for Research and Study
of Diabetic Neuropathy. Cardiac autonomic neuropathy in
diabetic patients: influence of diabetes duration, obesity, and
microangiopathic complicationsFThe French multicenter study.
Metabolism 2003; 52: 815–820.
16 Antelmi I, de Paula RS, Shinzato AR, Peres CA, Mansur AJ, Grupi
CJ. Influence of age, gender, body mass index, and functional
capacity on heart rate variability in a cohort of subjects without
heart disease. Am J Cardiol 2004; 93: 381–385.
17 Emdin M, Gastaldelli A, Muscelli E, Macerata A, Natali A,
Camastra S et al. Hyperinsulinemia and autonomic nervous
International Journal of Obesity
Autonomic function, overweight and underweight
J-S Wu et al
794
18
19
20
21
22
23
24
25
26
system dysfunction in obesity: effects of weight loss. Circulation
2001; 103: 513–519.
Kupari M, Virolainen J, Koskinen P, Tikkanen MJ. Short-term heart
rate variability and factors modifying the risk of coronary artery
disease in a population sample. Am J Cardiol 1993; 72: 897–903.
Andersson B, Wikstrand J, Ljung T, Bjork S, Wennmalm A,
Bjorntorp P. Urinary albumin excretion and heart rate variability
in obese women. Int J Obes Relat Metab Disord 1998; 22: 399–405.
Gelber DA, Pfeifer M, Dawson B, Schumer M. Cardiovascular
autonomic nervous system tests: determination of normative
values and effect of confounding variables. J Auton Nerv Syst 1997;
62: 40–44.
Richter WO, Geiss HC, Aleksic S, Schwandt P. Cardiac autonomic
nerve function and insulin sensitivity in obese subjects. Int J Obes
Relat Metab Disord 1996; 20: 966–969.
Vaz M, Bharathi AV, Sucharita S, Nazareth D. Heart rate variability
and baroreflex sensitivity are reduced in chronically undernourished, but otherwise healthy, human subjects. Clin Sci
2003; 104: 295–302.
Lu FH, Yang YC, Wu JS, Wu CH, Chang CJ. A population-based
study of the prevalence and associated factors of diabetes mellitus
in southern Taiwan. Diabet Med 1998; 15: 564–572.
Hollingshead AB. Hollingshead two factor index of social
position. In: Miller DC (ed). Handbook of Research Design and
Social Measurement, 5th edn. Sage Publications: Newbury Park
California, 1991. pp 351–359.
Paffenbarger Jr RS, Blair SN, Lee IM, Hyde RT. Measurement of
physical activity to assess health effects in free-living populations. Med Sci Sports Exerc 1993; 25: 60–70.
Weisell RC. Body mass index as an indicator of obesity. Asia Pac
J Clin Nutr 2002; 11 (Suppl): S681–S684.
International Journal of Obesity
27 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,
Turner RC. Homeostasis model assessment: insulin resistance and
b-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–419.
28 Blackburn H, Keys A. The electrocardiogram in population
studies: a classification system. Circulation 1960; 21: 1160–1175.
29 Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA,
Izzo Jr JL et al. The seventh report of the Joint National
Committee on Prevention, Detection, Evaluation, and Treatment
of High Blood Pressure: the JNC 7 report. JAMA 2003; 289:
2560–2572.
30 Arone LJ, Mackintosh R, Rosenbaum M, Leibel RL, Hirsch J.
Autonomic nervous system activity in weight gain and weight
loss. Am J Physiol 1995; 269: R222–R225.
31 Colhoun HM, Francis DP, Rubens MB, Underwood SR, Fuller JH.
The association of heart-rate variability with cardiovascular risk
factors and coronary artery calcification: a study in type 1
diabetic patients and the general population. Diabetes Care 2001;
24: 1108–1114.
32 Bergholm R, Westerbacka J, Vehkavaara S, Seppala-Lindroos A,
Goto T, Yki-Jarvinen H. Insulin sensitivity regulates autonomic control of heart rate variation independent of body
weight in normal subjects. J Clin Endocrinol Metab 2001; 86:
1403–1409.
33 Spraul M, Ravussin E, Fontvieille AM, Rising R, Larson DE,
Anderson EA. Reduced sympathetic nervous activity. A potential
mechanism predisposing to body weight gain. J Clin Invest 1993;
92: 1730–1735.
34 Straub RH, Thum M, Hollerbach C, Palitzsch KD, Scholmerich J.
Impact of obesity on neuropathic late complications in NIDDM.
Diabetes Care 1994; 17: 1290–1294.