Académique Documents
Professionnel Documents
Culture Documents
005361
Running title: Soares-Miranda et al.; Physical activity and heart rate variability
Luisa Soares-Miranda, PhD1,2; Jacob Sattelmair, PhD1; Paulo Chaves, MD, PhD3;
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
Glen Duncan, PhD4; David S. Siscovick, MD, MPH4,5; Phyllis K. Stein, PhD6;
Dariush Mozaffarian, MD, DrPH1,7
1
DOI: 10.1161/CIRCULATIONAHA.113.005361
Abstract
Background Cardiac mortality and electrophysiologic dysfunction both increase with age.
Heart rate variability (HRV) provides indices of autonomic function and electrophysiology that
are associated with cardiac risk. How habitual physical activity (PA) among older adults
prospectively relates to HRV, including nonlinear indices of erratic sinus patterns, is not
established. We hypothesized that increasing levels of both total leisure-time activity and
walking would be prospectively associated with more favorable time-domain, frequency-domain,
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
DOI: 10.1161/CIRCULATIONAHA.113.005361
Introduction
Cardiac diseases and arrhythmias are common with aging and at least partly linked to increasing
cardiac electrophysiological and autonomic dysfunction.1,2-5 Heart rate variability (HRV)
provides indices related to cardiac electrophysiology and autonomic regulation, including
respiratory, baroreflex, and circadian fluctuations, that are indicative of healthier responses.
With advancing age, increased HRV can also reflect abnormal (erratic) sinus patterns that are
associated with increased risk.6 In middle-aged populations, habitual physical activity (PA) has
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
been associated with more favorable HRV indices, especially those reflecting increased vagal
modulation and reduced sympathetic activity.7, 8 However, several key issues remain to be
addressed in older adults 9-24, who are at much higher risk for both abnormal HRV
HR
RV and
a d cardiac
an
caard
card
rdia
iacc
ia
events than middle-aged populations. First, whether habitual PA relates to HRV later in life is
not
no
ot well
well established.
est
stab
ablish
ab
shhed.
ed Most studies in older adults
adultts were
were quite small
lll (e.g.,
(e.g.
g.,, <100
<100 subjects)9, 10, 14-21,
23-25
3-25
5
10,
0, 12
0,
12,, 14,
14, 17, 20,
20,, 21, 23-25
23-2
2325
25
aand
nd utilized
utilize
zedd on
only
nly
y short-term
sho
hort
rt-t
rt
-ter
-t
erm
er
m EC
ECG
ECGs
Gs10
tthat
haat assess
assess only
onlly short-term
shoortsh
ort-te
term
te
rm
m HRV
HRV
indices,
ndi
dice
cees, as
as opposed
oppo
op
possedd to 24-hour
244 ho
houur measurements
mea
eassur
sureme
mennts
nt that
thatt assess
ass
s ess both
bo h short-term
short
rt-t
rt
-ter
-t
errm and
an
nd long-term
long
lo
ng-t
ng
-tterm
m HRV.
HRV.
HRV
V.
Additionally,
Additionally
y, wi
with
th
h aaging,
ging
gi
ng,, wa
ng
walk
walking
l in
lk
ng pr
pred
predominates
ed
dom
omin
in
nat
ates
ess aass th
thee ma
majo
major
jorr PA
jo
PA,, bu
butt di
diff
differences
ffer
ff
eren
er
e ce
en
cess in
n eeffects
f ects of
ff
walking versus general leisure-time activities versus exercise intensity on HRV late in life are
also not established. Furthermore, few prior studies of PA26 assessed nonlinear (erratic) HRV,
which provides important information about abnormal sinus firing that predicts higher mortality
and is especially common in older adults.1, 27 Finally, with only one exception,11 no large, longterm prospective studies have assessed longitudinal (rather than only cross-sectional)
associations between PA and 24-hour HRV among older individuals. To address these key gaps
in knowledge, we prospectively investigated the associations of PA with HRV indices in older
adults. We hypothesized that increasing levels of both total leisure-time activity and walking
DOI: 10.1161/CIRCULATIONAHA.113.005361
Methods
Population
The Cardiovascular Health Study (CHS) design and recruitment have been described.28, 29 Briefly,
5,201 ambulatory, non-institutionalized men and women 65 years of age were randomly
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
selected and enrolled from Medicare eligibility lists in 4 US communities in 1989-1990; and an
additional 687 black participants were similarly recruited and enrolled in 1992, but were not
ncluded in this analysis due to absence of repeated activity measures. The instit
ituutio
utio
iona
naal review
revi
re
view
vi
ew
included
institutional
committee at each center approved the study, and all subjects provided informed consent.
Base
Ba
seli
se
linne
li
ne eevaluation
v luat
va
lu tio
ionn included standardized physic
ical
ic
al examination, di
ddiagnostic
ag
gno
nosstic
st testing, laboratory
Baseline
physical
ev
vallua
u tion, an
nd qu
ques
esti
tion
ionnnaaires
ires oon
n he
heal
alth
th stat
tus, m
tus
eddicaal hi
his
stoory,
ory, an
nd cardiovascular
nd
caard
dio
ova
vasc
scul
ullar
a rrisk
iskk
is
evaluation,
and
questionnaires
health
status,
medical
history,
and
28-30
28
30
fa
act
ctor
orrs.28U
Updated
pddateed
ed iinformation
nffor
o maati
t onn w
was
as ggathered
as
athe
at
herred at an
he
annual
nnu
uall sstudy
tudy
tu
dy vvisit
isit
is
it tthrough
hrrou
ugh
h 11999
9 9 an
99
and 66-mo
6-month
m nthh
mo
factors.
elephone co
ont
ntac
a tss tthereafter.
ac
here
he
reaffte
re
ter.
r. Inn a su
ubsset
e ooff pa
part
r iccip
rt
ipan
a tss ((n=1,361),
an
n 1,
n=
1 36
361)
1 , 22-ch
chan
ch
anne
an
n l 24
ne
24-h
-h
hou
ourr Holter
telephone
contacts
subset
participants
2-channel
24-hour
monitor recordings were obtained at baseline (Del Mar Medical Systems, Irvine, California) and
then 5 years later in the same subjects (n=1,199). We excluded participants with markedly
irregular cardiac rhythms (n=106); insufficient N-N interbeat intervals (n=45 for time-domain;
n=121 for frequency-domain), as described below; or incomplete data on leisure-time activity,
exercise intensity, or walking habits (n=63). In total, 985 participants for longitudinal analyses
of time-domain HRV, and 909 for frequency-domain and nonlinear HRV. had serial
longitudinal measures of both PA and HRV over 5 years and were included in our main analysis.
As supplementary analyses, we have also evaluated cross-sectional relations between PA and 24-
DOI: 10.1161/CIRCULATIONAHA.113.005361
with risk of cardiovascular disease, atrial fibrillation, body composition, venous thrombosis, and
inflammation in expected directions.34-38 The questionnaire evaluated frequency and duration of
15 different activities during the prior 2 weeks, including gardening, mowing, raking,
rak
ak
kin
ng,
g swimming,
swi
wimm
mmin
mm
ingg
hiking, aerobics, tennis, jogging, racquetball, walking, golfing, bicycling, dancing, calisthenics,
and
an
nd ri
ridi
riding
ding
di
ng
g aan
n ex
exercise
xer
erccise
ci cycle (Figure 1).37 Each ac
acti
activity
ivity was pre-d
pre-designated
des
e ig
gna
natted
te as having an intensity
vvalue
alu
ue in units ooff m
metabolic
eta
tabo
boli
bo
licc eq
li
equi
equivalent
uiva
vaale
lent
nt ttask
ask (M
(MET)
MET)) sscore.
corre.32 P
Participant
articcippant
arti
pant rresponses
esspo
onsses rregarding
egar
eg
ardding
ng ttypes
ypes
yp
off activity,
acttiv
ivit
ity,
y, frequency,
frequ
requ
uen
ncy
y, and
an
nd duration
du
ura
rati
tion
ti
on were
were
ere used
used to
t calculate
calcu
ulaate weekly
wee
eekl
klly energy
ener
en
errgyy expenditure
exp
xpen
e di
en
ditu
tu
uree from
fro
ro
om
leisure-time
eisure-time aactivity,
ctiv
ct
i it
iv
ity,
y, expressed
exp
xpreess
ssed
e as
ed
as kc
kcal
kcal/wk.
l/w
/wk.
k U
k.
Usual
sual
su
al eexercise
x rccis
xe
isee in
intensity
nteens
nsityy wa
wass al
also
s aassessed
so
sses
ss
esse
es
sedd separately
se
y
at baseline and at 1992/93 (Figure 1), with responses including no exercise or low, medium, or
high intensity of exercise.37 Usual walking habits, including average walking pace (gait speed),
and distance walked, were assessed annually at each follow-up visit (Figure 1). We evaluated
these metrics in pre-specified categories, including: for leisure-time activity (quintiles), exercise
intensity (none/low, medium/high), blocks walked (quintiles), and usual pace walked (<2, 2
mph).
Covariates
Information on a wide range of covariates was obtained during study visits, including
DOI: 10.1161/CIRCULATIONAHA.113.005361
demographics, education, income, detailed smoking habits, alcohol use, usual dietary habits,
body mass index (BMI), resting heart rate (HR), blood pressure, medication use, and presence or
absence of coronary heart disease (CHD), congestive heart failure (CHF), hypertension, diabetes,
and ECG-defined left ventricular hypertrophy (LVH).39
Assessment of HRV
HRV indices include time-domain, frequency-domain, and nonlinear measures (Table 1).40 41-43
Long-term (e.g., 24-hour Holter) measures evaluate longer-term circadian differences in HRV as
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
well as daytime and nighttime baroreceptor and respiratory autonomic variation. In CHS, HRV
was assessed using two-channel 24-hour Holter recordings (Del Mar Medical Systems, Irvine,
California) at baseline in 1989-90 and again in 1994-95.44 Recordings were analy
analyzed
ly
yzeed at tthe
he
Washington University School of Medicine HRV Laboratory (GE Marquette Mars 8000 Holter
analyzer,
an
nal
alyz
yzeer,
yz
er Mi
Mil
Milwaukee,
lwau
lwau
ukee,
ke Wisconsin). Beat onset de
detection
etec
tection and classi
classification
sifi
f caatiion were reviewed and
edited
ed
ditted by traine
trained
nedd te
technicians
ech
hni
n ci
cian
anss and
an
and ov
over
overread
erre
reaad iin
n detail.
detaaill. From
Frrom
m 1,199
1,1
199 participants,
par
arti
tici
c paant
ci
ntss,
s, w
wee ex
eexcluded
clud
cl
u ed
subjects
irregularity
ubj
bjec
ects
ec
t with
ts
wit
ithh markedly
maark
ked
edlly
ly irregular
irr
rreg
egul
u ar cardiac
ul
car
ardi
diac
a rhythms,
ac
rhy
hyth
hms
m , defined
defiineed as
defi
as extent
extten
nt of irr
rrreg
egul
u ar
ul
arit
ityy of
of tthe
he rrhythm
he
hyyth
hm
or p-waves that
thaat was
was too
to
oo high
hiigh for
forr trained
tra
rain
i ed
in
d personnel
per
erso
sonn
so
n el
nn
e tto
o ac
accurately
ccu
cura
rate
ra
tely
te
ly llabel
ab
bel w
which
hich
hi
ch bbeats
eats
ea
ts w
were
e e normal
er
sinus beats (n=106). We also excluded recordings that provided fewer than 18 hours of usable
data (216 of 288 5-minute segments), requiring for time-domain analyses that at least 50% of
each segment consisted of N-N interbeat intervals (n=45) and, for frequency-domain and
nonlinear analyses, which are more sensitive to missing data, that at least 80% of each segment
consisted of N-N interbeat intervals (n=121); or incomplete data on leisure-time activity,
exercise intensity, or walking habits (n=63). After these exclusions, 985 participants had 24-hour
recordings in both 1989-90 and 1994-95 for longitudinal analyses of time-domain HRV, and 909
for frequency-domain and nonlinear HRV. For cross-sectional analyses at baseline in 1989-90,
DOI: 10.1161/CIRCULATIONAHA.113.005361
1,219 participants had 24-hour recordings for time-domain HRV and 1,150 for frequencydomain and nonlinear HRV.
Statistical Analysis
HRV measures were tested for normality through numeric and graphical methods and natural log
transformed as needed to facilitate parametric comparisons. We used linear regression to assess
associations of PA measures with HRV indices. PA measures were assessed as categorical
(indicator) variables, with PA categories entered as continuous variables in tests for trend. We
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
assessed longitudinal associations of PA in years 1989-90 to 1993-94 with HRV indices in 199495. To assess long-term effects, we performed cumulative averaging of PA measures in years
also
assessed
1989-90 to 1993-94. To evaluate the potential relevance of changes in PA, we al
lso ass
sssesse
essse
sed
d hhow
ow
differences in PA between baseline and 1993-94 related to HRV in 1994-95, as well as to
changes
HRV
ch
han
ange
gess in H
ge
RV bbetween
etween baseline and 1994-95. In
et
I additional analyses,
anal
alyses
al
es,, we
we assessed crosssectional
baseline.
ecttio
i nal associations
assoociiatioons between
beetwe
etween
en PA
PA measures
measuuress and
meas
and 24-hour
24-hoour
ur HRV
HRV
RV indices
inddic
ices
ess aatt ba
ase
seli
line
nee.
basis
To minimize
minim
inim
mizze ppotential
ottent
ten iaal confounding,
co
onf
nfooun
oundin
ndin
ing,
g, covariates
cov
ovar
ov
arriaate
t s were
wer
eree included
incl
in
clud
udded oon
n th
thee ba
asi
siss of
of cclinical
linnica
nicall
relevance
elevance as factors
fac
acto
to
ors that
tha
h t may
may influence
in
nfl
flue
u nc
ue
n e exposures
exp
xpos
osur
os
ures
ur
ess and
and
n outcomes,
out
u co
come
mes,
me
s, previously
pre
revi
viou
vi
ousl
ou
slyy published
sl
puubl
blis
ishe
is
hedd
he
associations, or associations with exposures/outcomes in the current data set. The final
multivariable model was adjusted for age (years), sex (male, female), race (white, nonwhite),
enrollment center (4 sites), education (<high school, high school, college), income ($25,000,.
>$25,000), smoking (never, former, current), alcohol (<1 drink/week, 1-2 drinks/week, 3-7
drinks/week, 8-14 drinks/week, >14 drinks/week), and consumption of fish (<1 serving/month,
1-3 servings/month, 1-2 servings/week, 3-4 servings/week, 5+ servings/week), dietary fiber
(quintiles, g/day), and total calories (quintiles, kcal/day). Missing covariate values (all <7%)
were imputed with simple regression using age, race, gender, education, BMI, diabetes, and
DOI: 10.1161/CIRCULATIONAHA.113.005361
absence of coronary heart disease, congestive heart failure, hypertension, diabetes, or left
ventricular hypertrophy. Given the influence of PA on resting HR and the importance of resting
separate
HR for HRV, we also evaluated resting HR as a potential confounder or mediator
orr iin
n a se
sepa
para
pa
rate
ra
te
analysis. Potential effect modification by age, sex, resting HR, or presence or absence of
cardiovascular
ca
ard
dio
iovvasc
vasccul
ulaar ddisease
iseease
is
ea (coronary heart disease, congestive
con
o gestive heart failure)
failurre) was
fa
was assessed by adding
statistical
Wald
a multiplicative
mu
ve iinteraction
ntterractiionn tterm
erm
er
m an
andd aassessing
ssessiing sta
atiisticcal
cal significance
sign
si
gnif
gn
ific
iccan
nce of
of th
thee Wa
W
ld
d ttest.
est. We
est.
We also
a lso
evaluated
pair-wise
combinations
PA
measures
model
independent
ev
val
alua
uate
ua
t d pa
te
pair
ir-w
-w
wis
isee co
ombin
mbinat
atio
at
ions
ns of P
A me
meas
asur
u ess iin
ur
n the
the sa
same
mee m
oddel
e tto
o asse
aassess
sseesss ttheir
heeirr ind
nddep
pen
ndeent
n
influence
HRV.
Potential
dose-response
was
using
nfluence on
nH
RV.. Po
RV
Pote
tent
te
n ia
iall do
dos
see reesp
spon
onse
on
se w
as eevaluated
v lu
va
l atted ssemi-parametrically
emiem
i pa
ip ram
amet
etri
et
rica
ri
call
ca
l y us
usin
in
ng re
rrestricted
stricted
cubic splines45, assessing the cumulative average of leisure-time activity from baseline to 199394 in relation to HRV in 1994-95. Analyses were performed using Stata 10.0 (College Station,
TX), two-tailed =0.05.
Results
At baseline, average (meanSD) age was 715 years, 60% of participants were women and mean
nighttime HR was 659 bpm (Table 2). The median (P25, P75) level of leisure-time activity was
630 (158, 1485) kcal/wk. The various HRV indices had variable intercorrelations at baseline
DOI: 10.1161/CIRCULATIONAHA.113.005361
a trend toward higher SDNN (P=0.06), and with less erratic HRV as assessed by DFA1 (ptrend=0.003) and by Poincare ratio (p-trend=0.02). In contrast, self-reported intensity of exercise
(medium/high
medium/high compared with none/low) was not significantly associated with any
an
ny HRV
HRV indices
indi
in
dice
di
ces
(data
data not shown). PA variables were not significantly associated with other HRV indices,
including
rMSSD,
NLF,NHF,
ncllud
udin
ingg rM
in
rMS
SSD,
D N
D,
LF,NHF, LF/HF ratio, or VLF
VLF power.
nighttime
indices
were
results
When sspecific
pecifi
fiic ni
igh
ghtttim
ttim
imee HR
HRV
V in
ndicees wer
re evaluated,
re
eva
vaaluatted
d, re
esuult
lts we
were
re ggenerally
eneral
ener
a ly ssimilar,
al
imil
im
ilar
il
arr,
with
for
leisure-time
wi
ith a few
few exceptions.
excep
xcep
ptiion
onss. For
For example,
exxamp
mple
le,, inn contrast
le
con
onttras
trasst too findings
fin
i ding
ding
ngss fo
or ooverall
ver
eral
a l HR
al
HRV,
V, ggreater
r at
re
ater
er le
eisuure
re--tim
me
related
activity was llongitudinally
ongi
on
gitu
gi
tudi
tu
d na
nallyy re
rela
l te
tedd to
o hhigher
ighe
ig
herr nighttime
he
nigh
ni
ghtt
gh
t im
tt
imee HRV
HRV indices
in
ndi
dice
c s such
such as
as SDNNIDX
SD
DNN
NNID
IDX (pID
(
trend=0.04) and VLF (p-trend=0.04) and walking distance was longitudinally associated with a
trend toward less nighttime erratic HRV assessed by nighttime DFA1 (p-trend=0.06). Similar to
the findings for overall HRV faster walking pace was associated with higher nighttime LF/HF
ratio (p-trend=0.003), and with less erratic HRV as assessed by DFA1 (p-trend=0.003) and the
Poincare ratio (p-trend=0.06) (data not shown).
Results were also not appreciably altered in several sensitivity analyses, including further
adjustment for baseline or 1994-95 characteristics that could be either confounders or mediators
of these relationships (see Methods) or further adjustment for baseline HR; or exclusion of
DOI: 10.1161/CIRCULATIONAHA.113.005361
participants reporting fair or poor overall health status (data not shown). For nearly all
relationships, there was also little evidence for effect modification by age, sex, prevalent
cardiovascular disease, or resting HR measured at either baseline or 1994-95 (p-interaction>0.10
for each). For walking distance and SDNN, we found borderline evidence for effect modification
by prevalent cardiovascular disease in 1994-95 (p-interaction=0.04), but these findings should be
interpreted cautiously due to the multiple comparisons. Additionally, in post hoc subgroup
analyses, findings were generally similar among individuals taking or not taking beta-blockers or
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
10
DOI: 10.1161/CIRCULATIONAHA.113.005361
changes in walking habits and walking pace, but not leisure-time activity, were associated with
specific HRV indices. Across quintiles of changes in walking distance, those in the highest
quintile (increase of at least 25 block/wk, N=182) had significantly higher SDNN
(meanSD=116.38.2), compared with those in the lowest quintile (decrease of 23 blocks/wk or
more, N=193; 109.47.2; p-trend<0.0001). Similarly, those in the highest quintile had
significantly higher ULF power, compared with the lowest quintile (11.51.5 vs. 10.1 1.2; ptrend=0.001). Across categories of walking pace, those that increased walking pace had
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
11
DOI: 10.1161/CIRCULATIONAHA.113.005361
trend=0.001) and higher ULF (p-trend<0.0001). Similarly, walking distance and walking pace
were each associated with higher SDNN (p-trend = 0.005 and 0.02, respectively) and higher
ULF (p-trend = 0.01 and 0.11, respectively).
Discussion
In this large prospective study among older adults with average age 71 years at baseline, PA was
both cross-sectionally and longitudinally associated with specific, more favorable indices of
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
HRV. Furthermore, over 5 years, those who increased their walking pace or walking distance
had more favorable HRV indices when compared with those that decreased their walking pace or
population-based
walking distance. To our knowledge, this is the first large, prospective, populati
ion-b
on-b
bassed sstudy
tudy
tu
dy to
demonstrate independent associations of PA measures with 24-hour time- and frequency-domain
well
ass w
elll as nnonlinear
el
on
nline
neaar
ne
ar indices of HRV among older
er ppersons.
ersons.
In prospective
prosp
speecttive
ve analysis,
ana
nallysi
na
lysi
siss,
s, leisure-time
leiisuure
re-t
-tiime activity
acctiviity
y and
nd walking
walking
king
g distance
dis
istaance
nce were
were
wer
re ssignificantly
igni
ig
nifi
ni
fica
fi
c nt
ca
ntly
ly
y
related
ela
late
teed to SDNN
SDN
D N and
and ULF,
UL
LF,
F whereas
wheere
reas
as walking
wal
a ki
king
ng pace
pac
a e was
wass positively
posi
po
sittive
si
vely
ve
ly
y related
rel
elat
a ed
at
ed to
to DFA1
DFA11 and
and inversely
innver
nverssely
sely
related
linear,
suggesting
that
elated to thee Poincar
Poi
o nc
ncar
a ratio.
ar
r tiio.
ra
o Those
Tho
hosse associations
a soociiat
as
atio
ions
io
ns appeared
app
ppea
e reed li
ea
line
n ar
ne
ar,, sug
ugge
ug
gest
ge
stin
st
ingg th
in
hat aany
ny PA
PA is betterr
than none, and more is better. Our findings are consistent with previous evidence, largely from
middle-aged populations, that PA is associated with a more favorable HRV profile.9-13 The
specific associations with indices that might reflect circadian variation (SDNN and ULF),
combined activity of sympathetic and parasympathetic modulation (NLF), vagal control of HR
and also renin-angiotensin system neurohormonal modulation (VLF), and less abnormal (erratic)
HR patterns (Poincar ratio and DFA1) suggests relatively selective effects of PA on the biologic
pathways influencing these parameters. While the biologic interpretation of these indices is
complex e.g., SDNN and ULF may reflect multiple inputs beyond circadian variation our
12
DOI: 10.1161/CIRCULATIONAHA.113.005361
findings make clear that PA is not related to all HRV indices similarly in older adults. The
specific observed patterns are supported by other epidemiological and clinical evidence on the
effects of regular PA on biologic pathways. For instance, reduction in sympathetic activity and
increased vagal activity have been implicated as possible pathways by which regular PA
provides cardioprotective benefits46. Enhanced circadian variation in HRV, as possibly reflected
by SDNN and ULF, is consistent with PA induced neurohormonal modulation.47, 48 PA may also
directly affect cardiomyoctyes by leading to improved contractile capacity49 and by enhancing
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
cardiac electrical stability.50 Such potential electrophysiologic stabilizing effects could explain
the higher DFA1 and lower Poincar ratio, observed with greater PA, that are indicative of
diminished random (erratic) sinus firing. Overall, the present findings have impli
implications
liica
cati
t on
ti
ns fo
forr ho
hhow
ow
PA may reduce arrhythmic risk in humans.
Ourr results
Ou
reesuult
ltss also
a so provide further support for
al
for
o clinical
clinical benefits
beneffitts off PA,
PA, even later in life. For
example,
ex
xample,
am
m
greater
great
ater
err PA
PA was
was associated
asso
as
socciat
so
ciated
ed with
wit
ithh higher
hiigh
gher SDNN
SD N and
andd ULF,
ULF
F, indices
indi
in
dicces
di
ces which
whiich
wh
ich predict
pred
pred
edic
ict risk
ic
riskk of
of
myocardial
and
heart
failure.
my
yoc
ocar
ardi
ar
dial
all iinfarction
nfar
nf
arrct
ctio
ionn an
nd he
eart fa
fail
iluure
il
ure.
e.6 PA
PA was
was aalso
lsso
so rrelated
ellat
ateed
ed tto
o mo
more
re ffavorable
avor
av
orab
able
ab
le nnonlinear
onlline
on
neearr iindices
nd
dicces
e
DFA1, Poincar
Poinc
nccar
ar Ratio),
Rati
Ra
tio)
o),, wh
o)
whic
ichh pr
ic
pred
ed
dic
i t ca
card
rdio
rd
iova
io
vasc
va
s ul
sc
ular
a eevents
ar
vent
ve
ntss and
nt
and total
to
otaal mortality
mort
mo
rtal
rt
alit
i y6 aand
it
nd enhance
(DFA1,
which
predict
cardiovascular
the overall predictive value of HRV.51, 52
Past studies of PA and HRV among older adults have generally assessed only short-term
(ECG-derived), time-domain indices.10, 12, 14, 17, 20, 21, 23, 24 Results have been mixed, with most
studies being small in size and based on either cross-sectional evaluations or short-term
interventions. Studies evaluating PA and 24-hour HRV indices11, 13, 15, 18, 19, 21, 22 were mostly
performed in middle-aged populations and were also generally small and either cross-sectional
evaluations or short-term interventions. Our findings build upon and extend these prior
observations by longitudinally evaluating long-term, cumulatively updated PA and changes in
13
DOI: 10.1161/CIRCULATIONAHA.113.005361
pace with NLF and NHF were in opposite directions than might be expected. Yet, these indices
represent complex interactions of autonomic processes, and increased walking pace was also
associated with less erratic HRV, which could partly account for these results. T
he ddifferences
he
ifffe
fere
renc
re
nces
nc
es
The
we observed may be clinically relevant and could partly account for benefits of PA. For example
example,
baaseed on
on rresults
e ul
es
u tss ffrom
rom
ro
m the Framingham Heart Study,
Stu
udy
d ,53 the higher va
vvalues
lues
ess ooff SDNN that we
based
observed
ob
bserv
seer ed in th
the
he hhighest
ighes
ghesst ve
ver
versus
rsus llowest
rsus
ow
wes
estt qu
quartile
uartiile of lleisure-time
eiisurre--ti
tim
me aactivity
me
cttiv
vit
ityy w
would
oul
uldd co
corr
correspond
rres
e po
es
p nd
n tto
o ~1
~11%
lower
owe
werr risk
r sk of
ri
of cardiac
carrdiac events.
ca
eveents
ev
e nt .
Our an
analysis
nal
alys
ysis
ys
is had
had sev
several
ever
ev
e all strengths.
er
str
tren
e gt
en
gths
hs.. Data
hs
Dataa on P
PA,
A, H
HRV,
RV,, an
RV
andd ot
othe
other
herr ri
he
risk
sk ffactors
acto
ac
tors
to
rs w
were
ere
prospectively assessed using standardized methods. Participants were randomly selected and
enrolled from Medicare eligibility lists in several US communities, providing a population-based
sample of older adults. Long-term (24-hour) HRV including time-domain, frequency-domain,
and and nonlinear indices provided a comprehensive evaluation of HRV parameters. Serial
measures of PA allowed evaluation of long-term effects, reducing misclassification and
providing the best measure of habitual PA, and also allowing assessment of changes in PA.
Prospective analyses reduced the potential for reverse causation, and adjustment for a wide range
of covariates minimized the potential impact of confounding.
14
DOI: 10.1161/CIRCULATIONAHA.113.005361
were obtained from self-report, which likely appropriately reflect the relative ordering (ranking)
of subjects but not the precise quantitative levels of energy expenditure. Although a range of
were
similar
covariates were available and evaluated as potential confounders, and findings w
eree si
imi
mila
larr in
la
in
several
everal sensitivity analyses, residual confounding due to unknown or incompletely measured
factors
parameters
fa
acttor
orss cannot
cann
ca
nnot
nn
o bbee excluded.
ot
ex
The assessments of bot
bboth
oth PA and HRV
V pa
araameters were subject to
random
HRV)
which
would
anddom
o error aand
nd ((in
in
n tthe
hee ccase
ase of
ase
of H
RV) bbiological
RV)
ioloogical
og l vvariability,
ariab
abil
ab
ilit
itty, w
hicch
ch w
ould
ou
d aattenuate
ttten
enua
uate
ua
tee ffindings
i di
in
ding
nggs
toward
The
owa
ward
rd the
the null
nul
ulll and
andd result
resu
re
suult in
in underestimation
u de
un
dere
rest
re
stim
im
mat
atio
ionn of the
io
the
he magnitude
mag
agni
nitu
ni
tude
tu
de of
of th
thee associations.
asssociiat
a io
ions
nss. T
hee
associations
cumulatively
PA
with
HRV
partly
prospective as
sso
soci
ciiat
atio
ions
io
ns off cu
cumu
mula
mu
l ti
la
tive
vely
ly uupdated
pdat
pd
a ed P
at
A wi
ith H
RV ccould
ould
ou
ld
d aalso
lsoo pa
ls
part
rtly
rt
ly rreflect
efle
ef
lect the
effects of PA earlier in life; in contrast, the associations of changes in PA with HRV would not
be confounded by PA at younger ages.
Results were attained among older, predominantly white Americans and may not be
directly generalizable to other populations.
Our results suggest that leisure-time activity and walking are prospectively associated
with specific patterns of more favorable HRV, including certain time- and frequency-domain as
well as nonlinear indices, among older adults. In addition, older adults who increased their
walking pace or distance over 5 years of follow up had more favorable HRV when compared
15
DOI: 10.1161/CIRCULATIONAHA.113.005361
with those that decreased their walking pace or distance. This suggests not only that regular PA
later in life is beneficial, but that certain beneficial adaptations may be lost upon cessation of
PA54, supporting the need to maintain modest PA throughout the aging process. Our results
support cardiovascular benefits and provide insights into plausible biologic pathways of effects
of modest PA, including walking, among older adults.
Acknowledgments: The authors express their gratitude to the CHS participants. A full list of
participating CHS investigators and institutions is at http://www.chs-nhlbi.org.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
References:
1. Stein PK, Barzilay JI, Chaves PH, Domitrovich PP, Gottdiener JS. Heart rate variability and
its changes over 5 years in older adults. Age Ageing. 2009;38:212-218.
2. Tsuji H, Venditti FJ, Jr., Manders ES, Evans JC, Larson MG, Feldman CL, Levy D. Reduced
heart rate variability and mortality risk in an elderly cohort. The framingham heart study.
Circulation. 1994;90:878-883.
16
DOI: 10.1161/CIRCULATIONAHA.113.005361
3. Heart rate variability: Standards of measurement, physiological interpretation and clinical use.
Task force of the european society of cardiology and the north american society of pacing and
electrophysiology. Circulation. 1996;93:1043-1065.
4. Saffitz JE. Sympathetic neural activity and the pathogenesis of sudden cardiac death. Heart
Rhythm. 2008;5:140-141.
5. Vanoli E, De Ferrari GM, Stramba-Badiale M, Hull SS, Jr., Foreman RD, Schwartz PJ. Vagal
stimulation and prevention of sudden death in conscious dogs with a healed myocardial
infarction. Circ Res. 1991;68:1471-1481.
6. Kleiger RE, Stein PK, Bigger JT. Heart rate variability: Measurement and clinical utility.
A.N.E. 2005;10:1-14.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
7. Swain DP, Franklin BA. Comparison of cardioprotective benefits of vigorous versus moderate
intensity aerobic exercise. Am J Cardiol. 2006;97:141-147.
8. Hull SS, Jr., Vanoli E, Adamson PB, Verrier RL, Foreman RD, Schwartz PJ. Exercise
Exeerc
r is
isee training
trai
tr
aini
ning
ischemia.
Circulation.
confers anticipatory protection from sudden death during acute myocardial isch
hemia
emia
i . Ci
C
irc
rcul
rc
ulat
ul
ation
1994;89:548-552.
9. Borghi-Silva
Borgh
g i-Silva A,, Ross A, Viviane C, Rodrigo Polaquini Se, Luis Eduardo Barreto M,
Aparecida
Maria
autonomic
Apar
Ap
arec
ar
ecid
ec
idaa Ma
id
M
riaa C,
C, Dirceu C. Aerobic exercise ttraining
r ining improves
ra
ess aut
ton
onoomic nervous control in
patients
with
2009;103:1503-1510.
pa
ati
tieentts wit
i h copd.
co
opd
pd. Respir
Resp
Re
spir
sp
irr Med.
Med
d. 20
2009
09;1
; 03
;1
03:1
:150
5033-115110.
10.
Piquard
Brandenberger
Heart
rate
10
0. Buchheit
Buchheit
it MS,
MS, Simon
Sim
monn C,
C, Viola
Vi la AU,
Viol
A , Doutreleau
AU
Douttreeleau
u S, P
iqua
iq
uard
ua
d F,, B
randdennbe
nberg
erger
ger G.
G. H
eart rat
te
variability
with
Sports
Exerc.
vari
va
riab
ri
abil
ab
i it
il
ityy in sportive
spo
orttiv
ivee elderly:
elde
el
derl
r y: Relationship
Rel
elaati
tion
on
nshhip w
ithh dail
it
ddaily
ailyy pphysical
hyssica
hy
sicaal aactivity.
cti
tivi
ti
v ty
vi
y. Me
Medd Sc
Scii Sp
Spo
ortts
ts E
xerrc.
xe
rc.
2004
04 336:601-605.
6:60
601-60
605.
11. Di
Dietrich
DF,
Ackermann-Liebrich
U, Sc
Schindler
C. E
Effect
activity
heart
11
Diet
etri
rich
ch D
F A
cker
ck
erma
mann
nn-Lie
Liebr
briich U
Schi
hind
ndle
lerr C
ffec
ff
ectt of pphysical
hysi
hy
sica
call ac
acti
tivi
vity
ty oon
n he
hear
artt ra
rate
te
variability in normal weight, overweight and obese subjects: Results from the sapaldia study. Eur
J Appl Physiol. 2008;104:557-565.
12. Earnest CP, Lavie CJ, Blair SN, Church TS. Heart rate variability characteristics in sedentary
postmenopausal women following six months of exercise training: The drew study. PLoS ONE.
2008;3:e2288.
13. Garet M, Degache F, Pichot V, Duverney D, Costes F, DA Costa A, Isaaz K, Lacour JR,
Barthlmy JC, Roche F. Relationship between daily pa and ans activity in patients with chf.
Med Sci Sports Exerc. 2005;37:1257-1263.
14. Gulli G, Cevese A, Capelletto P, Gasparini G, Schena F. Moderate aerobic training improves
autonomic cardiovascular control in older women. Clin Auton Res. 2003;13:196-202.
15. Jensen-Urstad K, Bouvier F, Saltin B, Jensen-Urstad M. High prevalence of arrhythmias in
elderly male athletes with a lifelong history of regular strenuous exercise. Heart. 1998 79:161-
17
DOI: 10.1161/CIRCULATIONAHA.113.005361
164.
16. Karavirta L, Tulppo MP, Laaksonen DE, Nyman KAI, Laukkanen RT, Kinnunen H,
Hakkinen A, Hakkinen K. Heart rate dynamics after combined endurance and strength training in
older men. Med Sci Sports Exerc. 2009;41:1436-1443.
17. Melo RC, Santos MD, Silva E, Quitrio RJ, Moreno MA, Reis MS, Verzola IA, Oliveira L,
Martins LE, Gallo-Junior L, Catai AM. Effects of age and physical activity on the autonomic
control of heart rate in healthy men. Braz J Med Biol Res. 2005 38:1331-1338.
18. Pichot V, Roche F, Denis C, Garet M, Duverney D, Costes F, Barthlmy JC. Interval
training in elderly men increases both heart rate variability and baroreflex activity. Clin Auton
Res. 2005;15:107-115.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
19. Piotrowicz E, Rafa B, Ma0141Gorzata P, Tomasz ZS, Ryszard P. Variable effects of physical
training of heart rate variability, heart rate recovery, and heart rate turbulence in chronic heart
failure. Pacing Clin Electrophysiol. 2009;32:S113-S115.
chronic
20. Reland S, Ville NS, Wong S, Senhadj L, Carr Fo. Does the level of chr
hrron
nic pphysical
hysi
hy
sica
si
c l
ca
2004;107:29-35.
activity alter heart rate variability in healthy older women? Clin Sci. 2004;107:29
29-3
-355.
-3
21. Schuit
S huit AJ,
Sc
J van Amelsvoort LG, Verheij TC, Rijneke
Rijn
j eke RD, Maan AC,
C, Swenne CA, Schouten
EG.
Med
Sci
EG
G. Exercise
Exer
Ex
erci
cise
ci
s training
traain
inin
ing and heart rate variability in
in older
older people. Me
M
d Sc
ci SSports
ports Exerc.
1999;31:816-821.
19
999
9;3
31:81
166-82
21..
22.
exercise
22
2. Stein
Stein PK,
PK
K, Ehsani
Eh
hsaani AA,
AA, Domitrovich
Dom
mittro
r vi
v ch
h PP,
PP, Kleiger
Kleiger
er RE,
RE
E, Rottman
Rottm
Ro
mann JN.
JN. Effect
Effec
Eff
fect of
of exe
erccise
cisee ttraining
rain
nin
ng
variability
healthy
Heart
1999;138:567-576.
onn hheart
eart
ea
r rrate
rt
atte va
var
riab
bil
ilit
itty in hea
ealt
ea
ltthy oolder
ldeer
ld
er aadults.
dults.
du
s Am
s.
mH
eart J.
eart
J. 19
1999
99
9;1
138
3 :5667-5
67-5
576
6.
23. Ueno LM,
Moritani
Effects
long-term
exercise
M, M
orit
or
itan
it
an
ni T.. E
f eccts ooff lo
ff
ong
ng-t
-tter
erm
m ex
exer
erci
er
c see ttraining
ci
rain
ra
inin
in
ingg on
in
n ccardiac
ardi
ar
diac
di
ac aautonomic
u on
ut
onom
omic
om
i nervous
activities
sensitivity.
acti
ac
tivi
viti
ties
es and
and baroreflex
bar
aror
oref
efle
lexx se
sens
nsit
itiv
ivit
ityy Eur
Eur J Ap
Appl
pl Physiol.
Phy
hysi
siol
ol 2003;89:109-114.
2003
20
03;8
;89:
9:10
1099-11
1144
24. Verheyden B, Eijnde BO, Beckers F, Vanhees L, Aubert AE. Low-dose exercise training
does not influence cardiac autonomic control in healthy sedentary men aged 55-75 years. J
Sports Sci. 2006;24:1137-1147.
25. Sandercock GRH, Hardy-Shepherd D, Nunan D, Brodie D. The relationships between selfassessed habitual physical activity and non-invasive measures of cardiac autonomic modulation
in young healthy volunteers. J Sports Sci. 2008;26:1171-1177.
26. Karavirta L, Tulppo MP, Laaksonen DE, Nyman K, Laukkanen RT, Kinnunen H, Hakkinen
A, Hakkinen K. Heart rate dynamics after combined endurance and strength training in older
men. Med Sci Sports Exerc. 2009;41:1436-1443.
27. Stein PK, Reddy A. Non-linear heart rate variability and risk stratification in cardiovascular
disease. Indian Pacing Electrophysiol J. 2005;5:210-220.
18
DOI: 10.1161/CIRCULATIONAHA.113.005361
28. Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH,
Manolio TA, Mittelmark MB, Newman A. The cardiovascular health study: Design and rationale.
Ann Epidemiol. 1991;1:263-276.
29. Tell GS, Fried LP, Hermanson B, Manolio TA, Newman AB, Borhani NO. Recruitment of
adults 65 years and older as participants in the Cardiovascular Health Study. Ann Epidemiol.
1993;3:358-366.
30. Psaty BM, Kuller LH, Bild D, Burke GL, Kittner SJ, Mittelmark M, Price TR, Rautaharju
PM, Robbins J. Methods of assessing prevalent cardiovascular disease in the cardiovascular
health study. Ann Epidemiol. 1995 5:270-277.
31. Guthrie JR. Physical activity: Measurement in mid-life women. Acta Obstet Gynecol Scand.
2002;81:595-602.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
32. Taylor HL, Jacobs DR, Schucker B, Knudsen J, Leon AS, Debacker G. A questionnaire for
the assessment of leisure time physical activities. J Chronic Dis. 1978;31:741-755.
33. Richardson MT, Leon AS, Jacobs DR, Ainsworth BE, Serfass R. Comprehen
Comprehensive
nsi
sive
v eevaluation
valu
va
luat
lu
a io
at
ion
of the minnesota leisure time physical activity questionnaire. J Clin Epidemiol. 19
1994;47:271-281.
994
94;4
;47:
7:27
27127
1-28
2811
28
34. Geffken
M, Burke GL, Polak JF,
G ffken DF,, Cushman
Ge
Cu
J , Sakkinen PA, Tracy
JF
cy RP. Association
between
healthy
elderly
beetw
wee
eenn physical
phys
ph
ysical
ys
all activity
act
c ivity and markers of inflammation
inflam
mma
m tion in a healt
thy
h eld
ld
der
erly
l population. Am J
Epidemiol.
Epid
Ep
demiol.
l 2001;153:242-250.
2001;
001;;15
1 3:
3:24
24224
2-25
2 0.
25
35.
FM,
Visser
Body
composition
35
5. Mitchell
Mitchelll D,
D, Haan
Ha MN,
MN, Steinberg
Steein
i be
b rg
gF
M, V
issser M.
M B
odyy co
od
comp
possitionn in
in the
thee elderly:
eldeerly
erly
y: The
Th
influence
activity.
Aging.
nfl
flue
uenc
ue
n e of nutritional
nc
nut
utrriti
ri ion
onaal factors
fac
actor
orrs and
and physical
physic
phy
ysic
ical
al act
ctiv
ct
iviityy.
y. J Nutr
Nutr Health
Hea
e lt
l h Ag
Agi
ingg. 2003;7:130-139.
2003
20
03;7
03
;7:1
;7
:1
130
30--139
3 .
39
36. Mozaffarian
Mozaffarria
iann D, Furberg
Fur
u be
b rg CD,
CD, Psaty
Psaaty BM,
BM, Siscovick
Sis
isco
co
ovi
v ck
c D. Physical
Phy
hysi
sica
si
cal activity
ca
a ti
ac
tivi
v ty and
vi
andd incidence
inc
ncid
iden
id
e ce of atrial
en
atriaal
fibrillation
older
adults:
cardiovascular
study.
fibr
fi
bril
illa
lati
tion
on in
in ol
olde
derr ad
adul
ults
ts:: Th
Thee ca
card
rdio
iova
vasc
scul
ular
ar hhealth
ealt
ea
lthh st
stud
udyy Circulation.
Circ
Ci
rcul
ulat
atio
ionn 2008;118:800-807.
2008
20
08;1
;118
18:8
:800
00-807
807
37. Siscovick DS, Fried L, Mittelmark M, Rutan G, Bild D, O'Leary DH, Cardiovascular Health
Study Research G. Exercise intensity and subclinical cardiovascular disease in the elderly: The
cardiovascular health study. Am. J. Epidemiol. 1997;145:977-986.
38. van Stralen KJ, Doggen CJM, Lumley T, Cushman M, Folsom AR, Psaty BM, Siscovick D,
Rosendaal FR, Heckbert SR. The relationship between exercise and risk of venous thrombosis in
elderly people. J Am Geriatr Soc. 2008;56:517-522.
39. Fried LP, Borhani NO, Enright P, Furberg CD, Gardin JM, Kronmal RA, Kuller LH,
Manolio TA, Mittelmark MB, Newman A. The cardiovascular health study: Design and rationale.
Ann Epidemiol. 1991;1:263-276.
40. Task F. Heart rate variability: Standards of measurement, physiological interpretation and
clinical use. Circulation. 1996;93:1043-1065.
19
DOI: 10.1161/CIRCULATIONAHA.113.005361
41. Stein PK, Domitrovich PP, Hui N, Rautaharju P, Gottdiener J. Sometimes higher heart rate
variability is not better heart rate variability: Results of graphical and nonlinear analyses. J
Cardiovasc Electrophysiol. 2005;16:954-959.
42. Roach D, Wilson W, Ritchie D, Sheldon R. Dissection of long-range heart rate variability:
Controlled induction of prognostic measures by activity in the laboratory. J Am Coll Cardiol.
2004;43:2271-2277.
43. Raj SR, Roach DE, Koshman ML, Sheldon RS. Activity-responsive pacing produces longterm heart rate variability. J Cardiovasc Electrophysiol. 2004;15:179-183.
44. Soares-Miranda L, Stein PK, Imamura F, Sattelmair J, Lemaitre RN, Siscovick DS, Mota J,
Mozaffarian D. Trans-fatty acid consumption and heart rate variability in 2 separate cohorts of
older and younger adults. Circ Arrhythm Electrophysiol. 2012;5:728-738.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
45. Durrleman S, Simon R. Flexible regression models with cubic splines. Stat Med. 1989;8:551561.
autonomic
46. Carter JB, Banister EW, Blaber AP. Effect of endurance exercise on autonom
mic control
mic
cont
co
ntro
nt
roll of
ro
of
heart rate. Sports Med. 2003;33:33-46.
hypothalamic
norepinephrine.
47. Lambert GW,
GW, Jonsdottir
J nsdottir IH. Influence of voluntary exercise on hyp
Jo
ypothalamic norepinephrine
Physiol.
1998;85:962-966.
J Ap
Appl
pl P
hysi
hy
siool. 19
si
1998
9 ;85:962-966.
98
Neurobiol.
448.
8. Nybo
Nybo L, Secher
Sech
Se
c er NH.
ch
NH.. Cerebral
Cer
ereb
ebra
eb
rall perturbations
pert
pe
rtur
urrbatiionns provoked
provokked
ed by
by prolonged
prooloonge
pr
ongedd exercise.
exeerc
ex
ercise
cise. Pr
Prog
ogg N
eurrob
eu
robio
2004
20
0044 72:223-261.
72:2233-26
2 1.
Wisloff
Ellingsen
O,, Ke
OJ.
High-intensity
interval
maximize
49. Wi
slof
lofff U,, E
llin
ings
gsen O
Kemi
mi O
J. H
ighig
h-in
inte
tens
n ityy in
inte
terv
rval
al tr
ttraining
aiini
ning
ng tto
o ma
m
ximi
xi
mize ccardiac
ardiiac
ar
exercise
Exerc
Rev.
2009;37:139-146.
benefits of ex
xer
erci
c se ttraining?
ci
raain
nin
ng? Ex
xer
ercc Sp
SSport
o t Sc
or
Scii Re
R
v. 20
009
9;3
;37:
7 13
1399 14
9146.
6
6.
50. Billman GE. Cardiac autonomic neural remodeling and susceptibility to sudden cardiac death:
Effect of endurance exercise training. Am J Physiol Heart Circ Physiol. 2009;297:H1171-1193.
51. Mozaffarian D, Stein PK, Prineas RJ, Siscovick DS. Dietary fish and {omega}-3 fatty acid
consumption and heart rate variability in us adults. Circulation. 2008;117:1130-1137.
52. Stein PK, Kleiger RE. Insights from the study of heart rate variability. Annu Rev Med.
1999;50:249-261.
53. Tsuji H, Larson MG, Venditti FJ, Jr., Manders ES, Evans JC, Feldman CL, Levy D. Impact
of reduced heart rate variability on risk for cardiac events. The framingham heart study.
Circulation. 1996;94:2850-2855.
54. Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, Minson CT, Nigg CR, Salem GJ,
Skinner JS. American college of sports medicine position stand. Exercise and physical activity
for older adults. Med Sci Sports Exerc. 2009;41:1510-1530.
20
DOI: 10.1161/CIRCULATIONAHA.113.005361
Variable
Description and Potential Physiologic Correlation
Long-term recordings (24-hour) indices
Time-domain indices
Standard deviation of all N-N intervals (from the entire recording). Possibly reflects longer-term circadian differences and total
SDNN, ms
HRV, higher values are considered healthier. Moreover, lower values for this index are associated with risk of mortality in patient
populations.
rMSSD, ms Root mean square of successive differences between N-N intervals. Reflects the average of daytime and nighttime
parasympathetically-mediated beat-to-beat changes in N-N intervals. Higher values generally reflect higher parasympathetic (vagal)
influence but can reflect a greater degree of erratic rhythm, especially in older adults41.
ian
n rhyt
hyt
y hm
m. Us
Usua
uall
ua
lly,
ll
y, hhigher
igh
circadian
rhythm.
Usually,
N-index, Averaged 5-min SDNN. Reflects combined sympathetic and vagal activity but independent of circadia
SDNN-index,
values are considered healthier.
ms
uency-domain indices
Frequency-domain
NLF, %
Precise interpretation of this index is controversial. However there is evidence that normalized LF can be a measure of sympathetic
sympa
modulation of heart rate. LF band is between 0.04 and 0.15 Hz.
NHF, %
Relative
R la
Re
lati
tive
ti
ve vagal
vagal
gal modulation
mod
o ulation of heart rate in response to respiration.
resspi
pira
ration. Higher values reflectt hi
hhigher
gher
gh
h parasympathetic (vagal) influen
influence or
greater
degree
of
erractic
rhythm.
HF
band
is
between
0.15
and
0.4
Hz.
Normally,
higher
values
are
considered
healthier.
grea
eate
ter
er de
d grree f
a
15 n
y hi
y,
h ghher val
alues
Fluctuat
ations
at
ns iin
n RR-R
R in
inte
t rv
te
rval
alss wi
al
with
th un
unde
derl
de
rlyi
ying
yi
ng ccycle
yccle llength
e gt
en
gth of >
5-mi
min
mi
n an
and
d
24
4-hour
u r. P
redo
re
domi
do
m nant
ntly
ly ccircadian
irca
ir
cadi
ca
dian
an rrhythm
hyth
hy
t m bu
th
butt ot
oother
her
Fluctuations
intervals
underlying
>5-min
24-hour.
Predominantly
ULF, m
mss2
es iincluding
n lu
nc
udiing
g act
ctiv
ct
ivit
iv
ity
it
y an
aand
d ne
eur
u oend
endoc
o ri
rine rhy
h th
thms. UL
ULF
L ba
andd iiss be
bbelow
low 0.
lo
00.003Hz.
003H
3Hz.
z. A
dditio
dd
onall
naally,
y iitt is rrelated
ellat
ated
ted tto
o HR aand
nd
influences
activity
neuroendocrine
rhythms.
band
Additionally,
of vvariation,
ariation,
n, wh
ich
ch ma
m
y refl
flec
ect fu
ffunctional
n tional
nc
n ccapacity
apacity
c y42, 433.
coefficient of
which
may
reflect
VLF may
maay reflect
refl
f ect both
t vagal
vagall control
contr
ntr
t ol
o of heart
h ar
he
artt ra
rate
a aand
nd also
nd
also the
h ef
effe
fect
c ooff th
thee re
rrenin-angiotensin
enin--angi
ng otensin
n n sy
syst
s em.
st
m Hi
H
ghherr vvalues
alue
lu s are be
bbelieved
lieve to
effect
system.
Higher
VLF, m
mss2
refl
re
fllec
ectt be
bbetter
t err autnonomic
tt
aut
u no
ut
nono
n mi
no
m c fu
ffunction.
ncti
cti
t on.
n
reflect
May
y re
rreflect
flec
e t rela
ati
tive
ve sym
mpa
pathet
etic
ic-p
-par
aras
asympathet
hetic
ic aactivity.
ctiv
ct
vitty. How
wev
ever
er, thi
hiss is nnot
ot tot
otal
ally
y ccorrect
o rect
or
ct sin
nce llower
ower
ow
e ffrequency
requ
re
q en
ncy ffluctuations
luctua
lu
uatio
relative
sympathetic-parasympathetic
However,
this
totally
since
F ra
ati
tio
o Ma
LF/HF
ratio
may
ma
y be
be related
rellatted
d to
to both
both sympathetic
sym
ympa
path
thet
eti
tic
ic and
andd parasympathetic
para
pa
rasy
symp
mpat
ath
theti
heti
ticc activity.
acti
ac
tivi
vit
ity.
ty Also,
Also
Al
so, during
duri
du
ring
ing eexercise,
xerc
xe
rciise
se, ov
over
eral
all
ll hear
hheart
eartt ra
rat
te vvariability
te
ariiabi
ar
iabi
bili
litty
li
ty ddecreases
ecre
ec
reas
asees
overall
rate
(inclu
udi
d ng L
F)..
F)
(including
LF).
near iindices
ndi
dices
Nonlinear
Organization of heart rate patterns based on the ratio of the axes of an ellipse fitted to the scatter plot of N-N vs. N-N+1 intervals.
SD12,
Higher values can reflect a greater degree of erratic rhythm. Increases in this index can be considered to reflect more disorganized
Poincare
heart rate activity.
Ratio
Short-term fractal scaling exponent. Reflects randomness or correlatedness of the N-N intervals pattern. Totally random N-N
DFA1
intervals pattern has a value of 0.5, whereas a totally correlated pattern has a value of 1.5. Decreases in this index are considered to
reflect a more disorganized heart rate activity a marker of less healthy cardiac autonomic functioning.
DFA1= short-term fractal scaling exponent, NLF= normalized low-frequency power, NHF =normalized high-frequency power; SD12= Poincar plot ratio, rMSSD=square-root-of
the-mean-of-the-squares-of-successive-R-R-intervals differences, SDNN=standard-deviation-of-the-R-R-intervals, SDNN-index, ULF= ultra-low-frequency power, VLF= very
low-frequency power. In older adults, rMSSD, NLF and NHF can be exaggerated by erratic HR patterns, ie., unhealthy sinus arrhythmia of nonrespiratory origin. rMSSD, NLF,
NHF and LF/HF ratio were evaluated among individuals with lower erratic HRV.
21
DOI: 10.1161/CIRCULATIONAHA.113.005361
Table 2. Baseline characteristics of 985 older US adults in the Cardiovascular Health Study with
longitudinal assessment of PA and HRV.
Characteristic
Age, years
Gender, % male
Race, % white
Education
< High school, %
High school, %
> High school, %
Income $25,000, %
Smoking habits
Former smoker, %
Current smoker,%
Alcohol use 1/wk, %
Body mass index, kg/m
Resting heart rate, bpm
Systolic blood pressure, mm Hg
Diastolic blood pressure, mm Hg
Pr
rev
eval
alen
al
entt coronary
en
coro
co
r nar
nary
ry heart disease, %
Prevalent
Pr
rev
vallentt co
ong
n es
esti
t ve h
ti
ea
art
r ffailure,
ailu
ai
lure
lu
re,, %
Prevalent
congestive
heart
Prevalence
hypertension,
P
reevalence
ev
off hy
hyp
pertten
pert
nsion
siion
n, %
D
Diab
Diabetes
iab
betes m
mellitus,
elllitu
u s, %
Prevalence
hypertrophy,
Prev
Pr
eval
ev
alen
al
ence
en
ce ooff le
left
ft vventricular
entr
en
trric
i ullar h
yper
yp
ertr
trop
tr
ophy
op
hy,, %
hy
B-Blocker
use,
B
BBloc
Bl
ocke
oc
kerr us
ke
use
e, %
e,
Anti-depressants
Anti-depre
ess
s an
ants
ts use,
use
se, %
Anti-arrhythmic use, %
Daytime heart rate, bpm
Nighttime heart rate, bpm
Baseline leisure-time activity, kcal/wk*
Baseline walking blocks, blocks/wk*
Baseline walking pace, mph
< 2mph, %
> 2 mph, %
715
40
66
26
38
36
38
42
10
30
275
27
5
671
67
111
1
6711
13621
7412
23
3
4
45
116
6
5
13
2
1
7810 (71-85)
659 (59-71)
630 (158-1485)
12 (4-38)
24
76
22
DOI: 10.1161/CIRCULATIONAHA.113.005361
Table 3. Longitudinal associations of leisure-time activity with HRV at 1994-95, assessed using 24-hour Holter, among 985 older US
adults.
I: <135
(n=199)
Time-domain indices
SDNN, ms
DNNIDX, ms
SDNNIDX,
rMSSD**,
MSSD**, ms
Frequency-domain
uency-domain indices
NLF**,
LF**, %
NHF**,
HF**, %
VLF,
LF, ms
m
ms
ULF,
LF,, 1000
LF
1000 ms
ms
LF/HF**
F//HF
HF**
** ratioo
Nonlinear
nlinea
nl
l ea
e r indices
oincar
oi
ca e ratio (SD1
2))
Poincare
(SD12)
FA1
FA
A
DFA1
P for Trend
104.8 (6.4)
36.5 (2.3)
19.9 (2.1)
107.7 (6.6)
38.9 (2.7)
20.1 (2.1)
106.0 (5.9)
39.8 (2.6)
20.2 (2.2)
115.3 (5.7)
39.9 (2.3)
20.2 (2.1)
115.5 (7.0)
41.0 ((2.7)
2.7)
2.
7
7)
19
19.4
9.44 ((1.9)
1.9)
1.
9)
0.009
00.07
. 7
.0
00.75
.75
65.2 (1.9)
22.0 (1.8)
694.7 (111.5)
9.2 (1.1)
4.33 (0
4.
(0.5
(0.5)
. )
64.7 (1.7)
21.8 (1.8)
818.1 (137.5)
10.0 (1.2)
4.33 (0.4)
4.
(0.4
(0
.4))
66.7 (2.0)
20.5 (2.1)
855.5 (136.1)
99.3
.3 (1.0)
44.6
4.
6 (0
(0.5)
0.5
.5))
66.6 (2.0)
20.3 (1.9)
852.8 (126
(126.3)
6.3
3)
111.1
1.1 (1
1.
(1.0)
1.0
.0)
0)
44.8
4.
8 (0.5
(0.5)
0.5))
67.1 (2.0)
19.7 (1.9)
913.7 (146.7)
11.2 (1.3)
5.00 (0
5.
(0.5
(0.5)
.5))
0.49
0.42
0.10
0.02
0.34
0.
0.29
00.29
2 (0.
(0.03)
.03
0 )
1.08
0 (0
(0.07)
0.07
0 )
07
00.
0.27
27 ((0.02)
27
0.02))
1.
1.12
12 ((0.07)
12
0.07))
0.27
.2 ((0.03)
0.03
0.03
03))
1.13
.13 ((0.07)
0 07
0.
0 )
00.
0.26
26 (0.
26
(0.02)
02)
02
11.14
1.
14 (0.
(0.07)
07)
07
0.26 ((0.02)
0 02
0.
0 )
11.17
1.
1 ((0.07)
17
0 07
0.
0 )
00.
0.54
.54
00.25
0.
.25
Measures
res (except
(ex
e cept
ceept NLF,
NLF
LF,, NHF,
N F, and
NH
and
d DFA1)
DFA1
A ) were
w re llog-transformed
we
og-trans
og
g nssfo
f rm
rmed
ed for
for
o aanalysis
naly
na
lysiss an
ly
andd th
then
hen
n eexponentiated.
xpon
xp
ponen
enti
en
tiat
ti
a ed.
at
d Va
V
Values
lu
uess are
are m
mean
e n (S
ea
(SD),
D),
D)
), adju
adjusted
dju
just
sted
e ffor
ed
o age
or
g (ye
ge
(years),
y ar
ye
ars)
s),
) sex
ex (m
((male,
female),
), rac
race
acee (w
(whi
(white,
hite
hi
t , nonwhite),
te
nonwhi
no
hite
te)), eenrollment
nrol
nr
ollm
lmentt ce
centerr ((4
4 si
site
sites),
ite
t s)
s), educ
education
atio
tionn (<
(<hi
(<high
high
hi
gh sch
school,
choo
o l, hhigh
ighh sc
ig
scho
school,
h ol
ho
ol,, co
coll
college),
llege),
) in
inco
income
come
me (($25,000,.
$2
$25,
5 00
0 0,
0,. >$25,000),
>$25
>$
$25,000
00),
), smoking
smo
moki
king
ng (never,
(neeve
ver
former,
current),
alcohol
drink/week,
consumption
r, cu
curr
rren
rr
ent)
en
t),, al
t)
alco
coho
co
holl (<
ho
(<11 dr
drin
ink/
in
k/we
k/
week
we
ek,, 1ek
11-22 drinks/week,
drin
dr
inks
in
ks/w
ks
/wee
/w
eekk,
ee
k, 3-7
3-7
7 drinks/week,
dri
rink
nks/
nk
s/we
s/
week
we
ek,, 8-14
ek
8 14 ddrinks/week,
8rink
ri
nks/
nk
s/we
s/
week
we
ek,, >14
ek
>14 drinks/week),
drin
dr
inks
in
ks/w
ks
/wee
/w
eek)
ee
k),, and
k)
and co
cons
nsum
ns
umpt
um
ptio
pt
ionn of
io
of fi
fish
sh ((<1
<1 sserving/month,
ervi
er
ving
vi
ng/m
ng
/mon
/m
on 13 servings/month,
g/day),
ngs/month, 1-2 servings/week,
serv
erv
vin
ings
g /w
gs
wee
e k, 33-4
- sservings/week,
-4
ervvin
ings
gs/w
gs
/ eeek,
/w
k 55+
+ se
sservings/week),
rvvin
ings
gss/w
/wee
eek)
ee
k)), dietary
d ettar
di
a y fiber
fiibe
b r (quintiles,
(q
qui
u nt
ntil
i es
il
es,, g/
g/da
day)
da
y , and
and total
tota
to
tall calories
ta
calo
ca
lori
lo
r es
e (quintiles,
(qu
quin
inti
in
t le
ti
les,
s,, kcal/day).
kcal/day).
For values
transformed
upper
SD.
lues logg transform
ed
d we report
repo
p rtt tthe
he upp
pper S
D
D.
Numbers are shown for time-domain measures (n=985); slightly fewer individuals (n=909) had frequency-domain and nonlinear measures available.
*Cumulative average of leisure-time activity using all measures between baseline and year 4 (see text for details).
** rMSSD, NLF, NHF and LF/HF ratio were evaluated among individuals with lower erratic HRV (DFA>median(1.127); n=493).
23
DOI: 10.1161/CIRCULATIONAHA.113.005361
Table 4. Longitudinal associations of walking habits with HRV at 1994-95, assessed using 24-hour Holter, among 985 older US
adults.
I:0-6
(n=201)
Time-domain indices
103.8 (5.9)
SDNN, ms
37.9 (2.7)
SDNNIDX, ms
20.3 (2.0)
rMSSD**,
SD**, ms
Frequency-domain
uency-domain indices
64.2 (1.8)
NLF**,
F**, %
22.7 (1.8)
NHF**,
F**, %
735.5 (133.3)
VLF,
F, ms
8.9 (1.0)
ULF,
F, 1000 ms
4.11 (0
4.
(0.5)
LF/HF
**ratio
HF **r
rat
atio
Nonlinear
in
nea
earr indices
indi
in
dice
di
cess
0.29
0.
29 (0.03)
(0.
0.03))
Poincare
car
a e ratio
r ti
ra
tioo (SD12)
(SD1
D12)
2)
11.07
.07
0 (0.07)
07
(0.
0.07
07))
07
DFA1
A1
1
V:60.4
(n=190)
P For
Trend
109.2 (6.4)
39.3 (2.7)
20.1 (2.0)
111.5 (5.9)
39.9 (2.6)
19.4 (1.9)
106.6 (6.2)
38.2 (2.6)
19.2 (2.1)
118.8 (7.0)
40.9 (3.0)
20.9 (2.3)
0.02
0.71
0.40
105.0 (6.9)
37.5 (2.8)
19.6 (2.0)
111.4 (6.7)
39.8 (2.7)
(2.1)
220.3
0.33 (2
0.
(2.1
.1))
0.06
0.12
00.42
65.6 (1.9)
21.1 (1.9)
820.2 (139.0)
9.9 (1.1)
4.5 (0.5)
65.7 (1.9)
21.1 (1.9)
853.6 (129.4)
10.2 (1.0)
4.5 (0.5)
67.5 (1.8)
19.7 (1.9)
810.3 (126.3)
9.9 (1.1)
4.9 (0
4.
((0.5)
.5)
67.1 (1.9)
19.7 (2.0)
916.8 (153.0)
12.0 (1.3)
4.9 (0.5)
0.08
0.06
0.30
0.008
0.100
65.4 (2.2)
2)
21.2 (2.3))
743.3 (142.9)
9.4 (1.2)
4.3 (0.5)
66.5
(2.1)
66.
6 5 (2
6.
(2.1
.1
1)
(2.1)
220.7
0.77 (2
.1
1)
853.4 (141.1)
10.4 (1.2)
4.7 (0.6)
00.11
00.36
00.11
00.16
00.16
0.27
0. (0.02)
(0.
0.02
02))
1.12
1.12
1 (0.07)
(0.
0 07
07))
0.27
2 (0.02)
(0.
0.02)
1.14
1 14 (0.06)
1.
(0.
0 06
6)
0.26
0.26
6 (0.02)
(0.
0 02)
1.15
1 15
1.
5 (0.06)
(0.
0 06))
0.26
0 26
0.
2 (0.02)
(0.02
02))
1.16
1 16 (0.07)
1.
(0.
0.07
07))
07
0.24
0..24
0.13
0 13
0.
0.29 (0.03)
(0.0
(0
03)
1.07
1.07
07 (0.07)
(0.0
(0
.07)
07)
0.26
0.2
. 6 (0.02)
( .0
(0
02)
1.15
1.1
15 (0.07)
(0.0
(0
.0
07)
7)
0.02
0
0.003
0.0
Measures
res
e (ex
(except
e cept NLF, NHF, aand
ex
nd D
DFA1)
FA1) w
were
eree na
natu
natural
tura
rall lo
loglog-transformed
g tr
tran
ansf
s orrme
sf
m d forr analysis
analysis
s aand
nd then
then
e exponentiated.
e ponent
ex
ntia
iate
ted.
d. Values
Val
alue
ues ar
aaree me
m
mean
an
n ((SD),
S ),
SD
), adju
adjusted
djusted
u ed
e ffor
or aage
g (years),
ge
(ye
y ars)
s), sex
seex (m
(mal
(male,
a e, fema
female),
race (white,
hite, nnonwhite),
onwhite), enrollme
enrollment
ntt ce
center
ent
n er (44 si
sites),
tes), educ
te
education
u at
ation (<hi
(<high
high
gh scho
school,
h ol
ol, high sschool,
chool,
h
col
college),
o lege
eg ), inc
income
n ome
om
me (($25,000,.
$25,000,
2
0 . >$
>
>$25,000),
25,000
00), ssmoking
moki
k ngg (never, forme
former,
m r,, ccurrent),
urrent), aalcohol
ur
lcoh
co ol (<1
drink/week,
week
we
e , 111-22 drinks/week,
drinks/week,
ek, 3-7 drinks/week,
drink
n s/week
ee , 8-14 drinks/week,
drinks/
ks we
w ek, >14
>114 drinks/week),
drinks
dr
nk /week)
ek , and
a d consumption
an
con
o sump
mpti
tiion
o ooff fi
ffish
sh (<1
sh
< sserving/month,
ervi
r ng
ng/month,
h 111-3
-3 serv
servings/month,
e ings
ngs/m
/ onth
t , 1th
11-2
2 sservings/week,
ervin
vi gs/week,
w k 3-4
servings/week,
s/w
/wee
we k, 55+
+ servings/week),
s rv
se
vin
ings
gs/w
gs
/wee
/w
eek)
k), dietary
diet
e ar
aryy fiber
fibeer (q
fi
(qui
(quintiles,
u ntil
iles
il
e , g/
g/da
g/day),
day)
y),, an
y)
aand
d total
o l ca
calo
calories
lori
ries
ie ((quintiles,
quin
qu
inti
t le
l s,, kkcal/day).
cal/
al/day)
ay)
y.
For values
ues llog
og ttransformed
rans
ra
nsfform
formed
ed w
wee re
report
epo
p rt tthe
h uupper
he
p er S
pp
SD.
D
D.
Numbers
rs ar
are
re sh
show
shown
o n fo
for
or ti
time
time-domain
m -ddom
omain me
meas
measures
asur
u es (n=
(n=985);
=98
9 5);
); sl
slig
slightly
ight
h ly ffewer
ewer indi
individuals
divi
vidu
dual
alss (n
(n=9
(n=909)
=909
09) had fr
fre
frequency-domain
equ
quen
ency
y-d
-dom
omai
ainn andd no
nonlinear
nlin
linea
ear me
measures.
eas
asuurees.
*Cumulative
lat
ativ
ivee av
iv
aver
average
erag
er
agee of lleisure-time
ag
eisu
ei
sure
su
re-t
-tim
tim
imee ac
acti
activity
tivvit
ti
ityy us
usin
using
ingg al
in
alll me
meas
measures
asur
as
ures
ur
es bbetween
etwe
et
ween
we
en bbaseline
asel
as
elin
el
inee an
in
andd ye
year
ar 11993-94
9933-94
99
3-94 ((see
seee te
se
text
xt ffor
or ddetails).
etai
et
ails
ai
ls)).
ls
** rMSSD,
SD, NLF, NHF and
nd L
LF/HF
F/HF
F/
H rratio
HF
atio
at
io w
were
eree ev
er
eevaluated
alua
uaate
tedd am
aamong
ongg in
on
iindividuals
diivi
v dual
alss wi
al
w
with
t lower
th
low
ower
e erratic
er
err
rrat
a ic HR
at
HRV
V (D
(DFA>median(1.127);
DFA
FA>m
>med
>m
ediaan(
ed
n(1.
1 12
1.
1 7)); n=
n=49
n=493).
4 3)
49
3)..
24
DOI: 10.1161/CIRCULATIONAHA.113.005361
Table 5. Associations of changes in walking habits with HRV at 1994-95, assessed using 24-hour Holter, among older US adults.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
I:-300/-23
(n=193)
V:25/420
(n=182)
P For
Trend
Decreased
(n=253)
Time-domain indices
109.4 (7.2)
107.7 (7.6)
107.4 (7.9)
110.1 (7.3)
116.3 (8.2) <0.0001 109.9 (9.1)
SDNN, ms
38.7 (2.9)
38.9 (2.9)
37.9 (2.9)
41.0 (3.0)
39.8 (2.8)
0.11
38.7 (3.9)
SDNNIDX, ms
19.7 (2.3)
20.0 (2.7)
20.2 (2.1)
20.0 (2.1)
19.9 (2.4)
0.36
18.8 (2.4)
rMSSD, ms**
Frequency-domain
ency domain indices
66.5 (1.9)
65.9 (2.1)
65.5 (1.9)
65.4 (2.5)
66.9 (2.2)
0.60
66.1 (2.4)
NLF,
F, %**
20.7 (1.8)
21.2 (2.2)
20.7 (2.0)
21.6 (2.4)
20.1 (2.3)
0.52
20.2 (2.2)
NHF,
F, %**
825.3 (145.4) 759.8 (132.7) 773.7 (153.5) 909.9 (156.3) 880.9 (143.6) 0.02 808.4 (200.9)
VLF,
F, ms
10.1 (1.2)
0.001
9.6 (1.3)
9.9 (1.4)
10.0 (1.2)
11.5 (1.5)
10.0 (1.6)
ULF,
F, 1000 ms
4.7 (0.5)
4.5 (0.5)
4.5 (0.6)
4.4 (0.5)
4.9 (0.6)
0.57
4.6 (0.6)
LF/HF
HF ** ratio
Nonlinear
indices
near in
ndices
0.27 (0.03)
0.28 (0.03)
0.27 (0.02)
0.26 (0.02)
0.28
0.26
00.27
0.
.27 (0.02)
(0.
0 02)
(0.
0 02)
2)
0.266 (0.02)
Poincare
caree ratio
rati
ra
tioo (SD12)
ti
(SD1
(S
D12)
D1
2)
1.11 (0.08))
1.11 (0.09)
1.12 (0.07)
1.16 (0.07)
0.38
1.14
1.14
1.
14 (0.06)
(0.
0.06))
(0.
0.07)
7)
(0.07)
0 38
0.
1.1
.144 (0.07))
DFA1
A1
1
117.6 (9.3)
39.6 (3.9)
20.3 (2.1)
66.5
68.5
66.
6..5 (2
((2.5)
.5
.5)
5)
68
.55 ((2.2)
2 2)
2.
20.1
20.
0 1 (2
((2.1)
.1))
18.6
18
.66 ((2.0)
2.0)
2.
0)
878.2
883.0
(205.6)
878
78.2
2 ((203.2)
203.
20
3.2)) 88
883
3.0 (2
(205
05.6
05
.66)
11.7 (2.0)
10.9 (1.7)
4.9 (0.7)
5.1 (0.7)
0.26 (0.02)
1.15 (0.08)
P For
Trend
0.006
0.18
0.48
0.001
0.007
0.08
0.007
0.001
0.25 (0.03)
1.18 (0.07)
Measures
res
e ((except
es
exce
ex
xc pt NLF, NHF,
HF and
nd D
DFA1)
F 1) w
FA
were
eree lo
er
log-transformed
og-tr
tran
tr
ansf
sfor
orrme
medd fo
fforr an
analysis
nal
alys
ysis
ys
is aand
nd tthen
henn ex
he
expone
exponentiated.
n nt
ntiated.
d. V
Values
alu
al
lues aare
re m
mean
ean (SD)
ea
(S
(SD),
SD), ad
adju
adjusted
just
sted
te for
f r aage
ge ((years),
year
ye
ars)
ar
s),, se
sexx (m
(male,
ale,
al
e, ffemale),
emal
em
ale)
al
e , race
e)
(white, no
nnonwhite),
nwhi
wh te), enrollmentt ce
cent
center
nter ((4
nt
4 sit
si
sites),
ites)
tes , ed
education
duca
cati
t onn ((<high
ti
< ig
<h
i h sc
sschool,
hool, hi
ho
hhigh
gh sch
gh
school,
ool, ccollege),
oo
ollege), iincome
ncome
m (($25,000,.
$2
$25,
5 000,
00,. >$
>$25
>$25,000),
$25,0000
00)), ssmoking
m ki
mo
k ng ((never,
n ve
ne
v r, fformer,
orme
or
rme
m r, ccurrent),
urre
ur
r nt),
re
t), aalcohol
lcoh
lc
ohol
oh
ol (<1
drink/week,
week,
e , 1-2
1 2 drinks/week, 3-7 drinks/week,
1drink
n s/week
ee , 8-14 drinks/week,
dri
rinks/
ks week,
k >14
> 4 drinks/week),
>1
drinks
dr
nk /week)
ek , and
a d consumption
an
con
o sumption
ion of
of fish
fish
s (<1
< serving/month,
servi
r ng
ng/month,
h 111-3
-3 serv
servings/month,
er in
ings/m
n / onth
nt , 111-2
2 sservings/week,
ervin
vings/w
/week,
w k 3-4
s/w
/ eek,
ek 5+ servings/week),
servings/w
/week), dietary
diet
e ary fiber
fiberr (qui
fi
int
ntil
i es,
s, g/da
ay), an
andd to
ttotal
otal calo
ries
ie ((quintiles,
quintile
l s, kcal/
/da
day)
y)..
y)
servings/week,
(quintiles,
g/day),
calories
kcal/day).
For values
ues llog
ogg ttransformed
ransformed
dw
wee re
report
port the uupper
po
pperr SD.
pp
in wa
w
lkin
kin
ingg di
dist
stan
ancee and
and pace
pac
a e we
were
re restricted
restr
tric
ricte
tedd to iindividuals
n ivid
nd
vid
idua
uals
lss w
ithh in
it
iinformation
form
fo
orm
rmat
atio
at
i n on w
a king
al
ng hab
abitss at bbaseline
ab
aselin
lin
inee an
andd ye
ear
a 11994-95
9994 95
99495 ((n=960;
n=96
n=
960;; nn=638
=638
=6
3 rrespectively).
38
e pe
es
p ct
c ivel
e y)
el
y).. Nu
N
mber
mb
bers aare
Changee in
walking
distance
with
walking
habits
year
Numbers
shown for ti
time-domain
ime-d
-dom
dom
omai
ainn m
ai
measures
easuure
ress (n
((n=960;
n=9
=960
960
6 ; 63
6388 re
resp
respectively);
spective
vely
ly);
); sslightly
ligh
li
ghtl
t y fe
fewer
wer
er iindividuals
ndiv
nd
ivid
idua
duals
ls ((n=886;
n=88
n=
8866;
88
6; 5596
96 rrespectively)
espe
es
p ctiv
pe
ivel
elly) had ffrequency-domain
requ
re
quen
ency
cy-do
doma
main
n and
and
d nonlinear
non
onli
line
li
n arr measures.
mea
easu
sure
res.
s.
** rMSSD,
SD,, NL
NLF,, NHF
HF and LF
LF/HF ra
rati
ratio
tioo we
were eval
evaluated
alua
uated am
amongg in
individuals wi
with
th low
lower
ower erratic
errratic HR
HRV
V (D
(DFA
(DFA>median(1.127);
FA>m
>med
edia
ian((1.12
127)
7);; n=
n=49
n=490;
490;; n=3
n=349
=349 re
respectively).
sppec
ectively
y).
25
0.16
0.03
DOI: 10.1161/CIRCULATIONAHA.113.005361
Table 6. Associations of changes in leisure-time activity with HRV at 1994-95, assessed using 24-hour Holter, among older US adults.
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
I:-8768/-538.
(n=196)
Time-domain indices
SDNN, ms
SDNNIDX, ms
rMSSD,
MSSD, ms **
Frequency-domain
uency-domain indices
NLF,
LF, %**
HF, %**
NHF,
VLF,
LF, ms
ULF,
LF,, 1000
10
000
0 ms
ms
LF/HF**
F//HF
HF**
*** ratio
ratio
tio
Nonlinear
nlin
nl
i ea
in
ear
a indices
i dicess
in
oincar
oi
ca e ratio (SD1
D12)
2)
Poincare
(SD12)
DFA1
F 1
FA
116.3 (6.0)
40.7 (2.7)
19.6 (2.4)
109.5 (6.4)
38.7 (3.1)
20.2 (2.5)
104.1 (5.7)
38.4 (3.2)
20.4 (1.9)
67.2 (1.8)
19.3 (1.7)
898.5 (150.4)
11.2 (1.1)
5.0 (0.5)
65.5 (2.0)
21.0 (1.7)
789.7 (143.5)
10.0 (1.1)
4.5 (0.4)
64.8 (1.8)
22.5 (1.6)
790.6 (161.7)
9.2 (1.0)
4.2 (0.4)
00.26
.26 ((0.02)
0..02
0 )
1.
1.16
16
6 (0.
(0.07)
0 07)
0.
00.
0.27
27 ((0.02)
0.02
.0 )
11.13
1.
13 ((0.07)
0.07
.0 )
00.
0.29
29
9 ((0.03)
0 03
0.
03))
1.08
1.
08 (0.
(0.07)
0 07))
115.0 (7.8)
40.4 (3.3)
18.9 (2.2)
112.6 (6.3)
38.9 (3.0)
20.1
20
. ((2.4)
.1
2.4)
2.
4)
66.5 (1.8)
68.2
68.22 (1.8)
(1.
1 8)
19.7 (1.8)
19.3 (1.7)
890.7 (167.8) 823.4 (146.7)
11.2
11
. (1.5)
10.6 (1.2)
4.8 (0.5)
4.8
5.1 (0.5)
00.
0.26
26 (0.
(0.02)
0..02
02))
11.15
.15 (0.
(0.06)
0 06
06)
00.
0.25
.25
5 ((0.02)
0.02
02))
02
1.
1.18
18 ((0.07)
18
0.07
0 )
0.50
0.86
0.9
0.99
0.19
0.
0.40
0.4
0.9
0.96
0.3
0.35
0.2
0.27
0.3
0.37
0.4
0.44
Measures
res
e (except
(ex
e cept NLF, NHF,
ex
NH and
and
d DFA1)
DFA1) were
w re llog-transformed
we
o -t
og
-tra
trans
n form
rm
med for
forr ana
analysis
lysis
y s and th
then
hen
n exp
exponentiated.
xpon
ponen
enti
tiat
ated.
d Va
Values
luess are m
lu
mean
e n (S
ea
(SD),
SD)
D , ad
aadjusted
justed
ju
d ffor
o age
or
age
ge (ye
(years),
y ars),
) sex
e (m
(male,
female),
), rrace
a e (w
ac
(whi
(white,
hite
te,, nonwhite),
nonw
no
n hi
nw
hite
te),
) eenrollment
n ollm
nr
ment
nt ce
cent
center
err ((4
4 si
site
sites),
tes)
te
s),, ed
s)
educ
education
ucat
a ion
on (<
(<hi
(<high
h ghh sschool,
choo
ch
o l,, hhigh
igh
gh scho
school,
cho
hool
ol,, coll
ol
college),
oll
lleg
ege)
e , in
inco
income
comee (
($25,000,.
$2
$25,
5,00
0 0,
00
0,.. >$25,000),
>$25
>$
25,0
25
,000
,0
0 ),, smoking
smo
moki
king
ngg (never,
(ne
neve
v r
former,
current),
alcohol
consumption
r, curr
rren
ent)
t),, al
t)
alco
coho
co
h l (<
ho
(<11 dri
ddrink/week,
rink/
ink/
k/we
week
ek,
k 111-2
-22 drinks/week,
d in
dr
nks
ks/w
/wee
/w
eek,
ee
k 3-7
k,
3-7 drinks/week,
drink
ink
nks/
s/we
s/
week
ek,, 8-14
ek
8 14
84 ddrinks/week,
rink
ri
inkks/
s/we
week
ek, >14
ek
>1
14 drinks/week),
d inks
dr
ks//wee
/week)
k),
), and
and co
cons
nsumpt
mpt
ptio
ionn of
io
of fi
fish
shh ((<1
<1 sserving/month,
ervi
ving
ing/m
/ on 1/m
3 servings/month,
servings/week),
ngs
g /m
/mon
onth
th,, 1-2
1-2 servings/week,
serv
se
rvin
ings
g /w
gs
/wee
eek,
k,, 33-4
-44 sservings/week,
ervvin
er
ings
g /w
gs
/wee
eek,
k,, 55+
+ se
serv
rvin
ings
g /w
gs
/wee
eek)
k)), dietary
diet
di
etar
aryy fiber
fibe
fi
berr (quintiles,
(qui
(q
uint
ntil
iles
es,, g/
gg/day),
day)
da
y),, and
y)
and total
tota
to
tall calories
calo
ca
lori
ries
es (quintiles,
(qu
q in
inti
tile
les,
s,, kcal/day).
kca
cal/
l/da
day)
y)..
y)
For values
transformed
report
SD.
lues log transform
med
d we
we re
epo
port
r the
rt
h uupper
he
pperr S
D
D.
*Change
ge in leisure-timee activity
acti
ac
tivi
ti
viity
t analysis
ana
naly
l siis we
ly
were rrestricted
estr
es
t icted
ed to
t iindividuals
ndiv
nd
vid
idua
duals
ls with
wit
ithh information
info
info
f rm
rmat
atio
ionn onn leisure-time
lei
eisu
sure
re-tim
timee ac
aactivity
tivi
viity aatt ba
bbaseline
seli
se
l ne an
li
andd ye
year
ar 11994-95
9 499
4 95 (n=976).
Numbers are shown
for
measures.
h
f time-domain
i
d
i measures (n=976);
( 9 6) slightly
li h l ffewer iindividuals
di id l ((n=901)
901) hhad
d ffrequency-domain
d
i and
d nonlinear
li
** rMSSD, NLF, NHF and LF/HF ratio were evaluated among individuals with lower erratic HRV (DFA>median(1.127); n=490).
26
DOI: 10.1161/CIRCULATIONAHA.113.005361
Figure Legends:
Figure 2. Longitudinal associations of leisure-time activity and walking distance with specific
HRV indices at 1994-95, assessed using 24-hour Holter, among 985 older US adults. Dots and
vertical lines represent the adjusted mean differences and the respective SD 95%
% between
b tw
be
wee
eenn
participants in a given quintile; quintile of walking distance ( images in the left) or leirure-time
activity
ac
ctiivi
vity
ty (images
(im
imag
ag
gess in
in the
t e right) . Adjusted for age (years),
th
(yeaars), sex (male,
(y
(malee, female),
fema
malle),
ma
le race (white,
enrollment
(<high
income
nnonwhite),
onnwhite),
nw
enr
nrol
ollm
ol
lm
men
entt center
ceent
nter
err (4
(4 sites),
siite
tes)
s , eeducation
ducaatiion (<
(<h
high
gh sschool,
ch
hool,
l hi
high
gh sschool,
choo
ch
ool,
l, ccollege),
olle
ol
leegee),
) inc
ncom
nc
om
me
($25,000,.
$2
$25,
5,00
0 0,
00
0,.. >$25,000),
>$25
>$
255,0000
00)), smoking
smo
m king
kingg (never,
(ne
neve
veer, former,
for
orme
mer,
me
r current),
r,
currreent
nt),
), aalcohol
lcoh
lc
ohool ((<1
oh
< ddrink/week,
<1
riink
nk/w
/wee
/w
e k, 11-2
ee
-2
drinks/week),
consumption
drinks/week,
k,, 33-7
-7
7 ddrinks/week,
rink
ri
nk
ks/
s/we
weeek
ek, 888-14
1 ddrinks/week,
14
riink
nks/
s we
s/
week
ek,, >1
ek
>144 dr
rin
inks
ks/w
ks
/wee
/w
eek)
ee
k)), an
andd co
cons
n um
ns
umpt
ptio
pt
ionn of fish (<11
io
serving/month, 1-3 servings/month, 1-2 servings/week, 3-4 servings/week, 5+ servings/week),
dietary fiber (quintiles, g/day), and total calories (quintiles, kcal/day).
Figure 3. Longitudinal associations of leisure-time activity and walking distance with specific
HRV indices at 1994-95, assessed using 24-hour Holter, among 985 older US adults. Solid lines
represent restricted cubic splines (smoothed fits); dashed lines represent 95% confidence
intervals. Adjusted for age (years), sex (male, female), race (white, nonwhite), enrollment center
(4 sites), education (<high school, high school, college), income ($25,000,. >$25,000), smoking
27
DOI: 10.1161/CIRCULATIONAHA.113.005361
(never, former, current), alcohol (<1 drink/week, 1-2 drinks/week, 3-7 drinks/week, 8-14
drinks/week, >14 drinks/week), and consumption of fish (<1 serving/month, 1-3 servings/month,
1-2 servings/week, 3-4 servings/week, 5+ servings/week), dietary fiber (quintiles, g/day), and
total calories (quintiles, kcal/day). P nonlinearity for leisure-time activity: SDNN= 0.30; ULF=
0.67; DFA1=0.96) and walking distance P nonlinearity: SDNN= 0.66; ULF= 0.86; DFA1=0.97.
28
LeisureTimeActivity*
ExerciseIntensity
Walking**
198990
24hourH
24hourHRV
HRV
(n=1,219)
Figure 1
LeisureTimeActivity*
ExerciseIntensity
Walking**
Walking**
Walking**
Walking
Walking**
g**
199091
199192
199293
199394
1994
199495
24hourr
24hourHRV
(n=985)
Figure 2
Figure 3
Physical Activity and Heart Rate Variability in Older Adults: The Cardiovascular Health Study
Luisa Soares-Miranda, Jacob Sattelmair, Paulo Chaves, Glen Duncan, David S. Siscovick, Phyllis K.
Stein and Dariush Mozaffarian
Circulation. published online May 5, 2014;
Downloaded from http://circ.ahajournals.org/ by guest on October 23, 2016
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 2014 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/early/2014/04/15/CIRCULATIONAHA.113.005361
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/
Supplemental Material
Supplementary-Table 1. Measures of HRV and their inter-correlations at baseline in the Cardiovascular Health Study.
HRV Indices
Mean (SD)
SDNN
rMSSD
SDNNIDX
NLF
NHF
VLF
ULF
LF/HF
Poincare
Ratio
DFA1
121.6(34.7)
rMSSD, ms
26.5(17.0)
0.44
SDNNIDX, ms
44.1(15.1)
0.65
0.76
NLF, %
68.5(6.0)
0.01
-0.14
0.07
NHF, %
19.0(4.7)
0.008
0.33
0.06
-0.85
VLF, ms
2288(38801)
0.67
0.59
0.93
0.08
0.02
ULF, 1000 ms
14.5(27.1)
0.98
0.35
0.54
0.02
-0.01
0.58
LF/HF ratio
1.29 (0.56)
0.07
-0.56
-0.07
0.95
-0.97
0.14
0.11
Poincare Ratio
0.27(0.10)
-0.11
0.57
0.002
-0.43
0.61
-0.19
-0.13
-0.83
DFA1
1.04(0.18)
0.11
-0.49
0.05
0.77
-0.81
0.24
0.12
0.92
-0.91
*24-hour HRV indices capture resting, activity, sleep, and circadian influences.
Values are Spearman correlation coefficients (P<0.01 for all correlations >0.1). All measures were derived from 24-hour Holter recordings (n=1219).
Supplementary-Table 2. Associations of changes in walking habits with concurrent changes in HRV, assessed using24-hour Holter, among older US adults.
Walking distance, blocks/wk *
Q1
Q2
Q3
Q4
Q5
P for
(n=158)
(n=164)
(n=141)
(n=134)
(n=153)
Trend
SDNN, ms
-5.3 (12.1)
-5.0 (12.9)
-4.4 (12.7)
-2.1 (12.8)
-3.1 (13.0)
0.13
SDNNIDX, ms
-1.4 (3.6)
-0.6 (4.2)
-2.1 (3.5)
1.6 (3.5)
-0.8 (4.0)
0.59
rMSSD, ms**
0.3 (2.0)
0.8 (2.2)
-2.4 (2.3)
1.1 (2.6)
0.5 (2.4)
0.49
NLF, %**
-2.7 (1.6)
-2.5 (1.6)
-1.2 (1.8)
-2.9 (1.4)
-2.1 (1.9)
0.71
NHF, %**
1.4 (1.7)
0.7 (1.8)
-0.5 (1.8)
1.3 (1.6)
0.6 (1.8)
0.74
VLF, 1000ms
1.2 (11.5)
-2.2 (15.5)
0.5 (13.7)
-9.1 (13.4)
-1.4 (13.2)
0.65
ULF, 1000 ms
0.4 (9.4)
-10.2 (9.1)
0.9 (10.0)
-6.6 (105.8)
-0.3 (120.0)
0.86
LF/HF** ratio
-0.3 (0.6)
-0.3 (0.5)
0.0 (0.7)
-0.4 (0.6)
-0.2 (0.6)
0.42
0.04 (0.02)
0.02 (0.02)
0.01 (0.03)
0.03 (0.03)
0.03 (0.03)
0.83
DFA1
0.06 (0.05)
0.08 (0.05)
0.10 (0.05)
0.08 (0.04)
0.07 (0.05)
0.39
Time-domain indices
Frequency-domain indices
Nonlinear indices
Table shows changes in walking distance and concurrent changes in HRV indices. Values are mean (SD), adjusted for age (years), sex (male, female), race
(white, nonwhite), enrollment center (4 sites), education (<high school, high school, college), income ($25,000,. >$25,000), smoking (never, former,
current), alcohol (<1 drink/week, 1-2 drinks/week, 3-7 drinks/week, 8-14 drinks/week, >14 drinks/week), and consumption of fish (<1 serving/month, 1-3
servings/month, 1-2 servings/week, 3-4 servings/week, 5+ servings/week), dietary fiber (quintiles, g/day), and total calories (quintiles, kcal/day).
Numbers are shown for time-domain measures (n=620); slightly fewer individuals (n=555) had frequency-domain and nonlinear measures.
** rMSSD, NLF, NHF and LF/HF ratio were evaluated among individuals with lower erratic HRV (baseline DFA>median(1.044) and baseline
DFA>median(1.127) ; n=285).
Supplementary-Table 3. Associations of changes in leisure-time activity with concurrent changes in HRV, assessed using24-hour Holter, among older US
adults.
Quintiles of Leisure-Time Activity, kcal/wk*
Q1
Q2
Q3
Q4
Q5
P for
(n=170)
(n=135)
(n=98)
(n=195)
(n=147)
Trend
SDNN, ms
-3.9 (12.8)
-4.7 (11.8)
-0.8 (11.5)
-5.0 (12.1)
-4.1 (12.5)
0.44
SDNNIDX, ms
0.5 (3.9)
-0.8 (4.3)
-1.0 (3.1)
-0.9 (3.5)
-1.0 (3.4)
0.37
rMSSD, ms**
0.5 (2.3)
1.1 (2.5)
0.7 (2.9)
-1.1 (3.1)
0.8 (2.8)
0.13
NLF, %**
-2.8 (1.7)
-1.7 (1.6)
-2.7 (1.7)
-1.8 (1.7)
-2.3 (1.6)
0.97
NHF, %**
1.2 (1.7)
0.3 (1.6)
1.8 (1.8)
-0.2 (1.8)
0.5 (1.8)
0.48
VLF, 1000ms
-2.3 (13.8)
1.6 (13.7)
0.09 (15.0)
-8.8 (14.7)
3.0 (15.2)
0.10
ULF, 1000 ms
-0.4 (10.2)
0.7 (11.2)
0.8 (11.6)
-6.7 (10.1)
-0.3 (12.3)
0.20
LF/HF ** ratio
-0.5 (0.6)
-0.1 (0.5)
-0.5 (0.5)
0.0 (0.5)
-0.3 (0.5)
0.33
0.03 (0.02)
0.04 (0.03)
0.04 (0.03)
0.02 (0.03)
0.02 (0.03)
0.26
DFA1
0.07 (0.05)
0.07 (0.04)
0.08 (0.05)
0.08 (0.05)
0.09 (0.05)
0.64
Time-domain indices
Frequency-domain indices
Nonlinear indices
Table shows changes in leisure-time activity and concurrent changes in HRV indices. Values are mean (SD), adjusted for age (years), sex (male, female), race
(white, nonwhite), enrollment center (4 sites), education (<high school, high school, college), income ($25,000,. >$25,000), smoking (never, former,
current), alcohol (<1 drink/week, 1-2 drinks/week, 3-7 drinks/week, 8-14 drinks/week, >14 drinks/week), and consumption of fish (<1 serving/month, 1-3
servings/month, 1-2 servings/week, 3-4 servings/week, 5+ servings/week), dietary fiber (quintiles, g/day), and total calories (quintiles, kcal/day).
Numbers are shown for time-domain measures (n=624); slightly fewer individuals (n=559) had frequency-domain and nonlinear measures.
** rMSSD, NLF and NHF were evaluated among individuals with lower erratic HRV (baseline DFA>median(1.044) and baseline DFA>median(1.127) ; n=285).
Supplementary-Table 4. Cross-sectional associations at baseline of leisure-time activity, exercise intensity, and walking habits with HRV, assessed using 24hour Holter, among 1219 older US adults.
Quintiles of Leisure-Time Activity, kcal/wk
I: <150
II: 150-540
III:540-1100
IV:1100-2205
V: >2205
P for
None
Low
Medium
High
P for
(n=244)
(n=253)
(n=235)
(n=244)
(n=243)
Trend
(n=68)
(n=568)
(n=446)
(n=137)
Trend
SDNN, ms
114.0 (7.1)
111.7 (7.6)
115.6 (7.5)
119.3 (7.4)
125.2 (8.2)
0.001
114.5 (7.2)
117.8 (8.1)
114.3 (8.1)
123.6 (8.3)
0.83
SDNNIDX, ms
40.7 (3.9)
40.4 (3.7)
41.2 (3.8)
43.1 (3.9)
44.5 (3.9)
0.13
39.0 (3.9)
42.2 (4.1)
41.4 (4.1)
44.4 (4.2)
0.78
rMSSD**, ms
18.1 (1.6)
18.7 (1.7)
18.5 (1.8)
19.3 (1.7)
20.1 (1.9)
0.18
16.4 (1.5)
18.5 (1.6
19.1 (1.8)
17.8 (1.6)
0.97
Time-domain indices
Frequency-domain indices
NLF**, %
68.8 (1.7)
67.9 (1.7)
68.5 (1.8)
68.3 (1.7)
68.5 (1.7)
0.30
70.5 (1.7)
69.2 (1.7)
69.3 (1.7)
69.2 (1.7)
0.47
NHF**, %
19.0 (1.5)
19.5 (1.4)
18.8 (1.5)
19.0 (1.5)
18.9 (1.5)
0.70
17.7 (1.2)
18.2 (1.3)
18.4 (1.4)
18.0 (1.4)
0.60
VLF, ms
890 (205)
913 (201)
965 (218)
1065 (229)
1144 (234)
0.01
839 (200)
997 (239)
977 (237)
1141 (251)
0.46
ULF, ms
11.1 (1.5)
10.8 (1.5)
11.6 (1.5)
12.2 (1.6)
13.4 (1.7)
<0.0001
11.3 (1.4)
11.9 (1.7)
11.1 (1.5)
13.1 (1.7)
0.81
LF/HF** ratio
5.1 (0.5)
5.1 (0.5)
5.3 (0.5)
5.2 (0.5)
5.2 (0.5)
0.83
5.4 (0.5)
5.1 (0.5)
5.2 (0.5)
5.3 (0.5)
0.97
0.27 (0.03)
0.26 (0.02)
0.25 (0.02)
0.25 (0.03)
0.24 (0.02)
0.03
0.28 (0.02)
0.26 (0.02)
0.25 (0.02)
0.24 (0.03)
0.04
DFA1
1.01 (0.06)
1.03 (0.06)
1.07 (0.06)
1.06 (0.06)
1.08 (0.06)
0.02
0.99 (0.06)
1.03 (0.06)
1.05 (0.06)
1.08 (0.06)
0.02
Nonlinear indices
0-6
7-12
13-30
31-75
75
P For
< 2 mph
2-3mph
>3 mph
P For
(n=266)
(n=233)
(n=239)
(n=238)
(n=243)
Trend
(n=287)
(n=521)
(n=411)
Trend
Time-domain indices
SDNN, ms
112.6 (7.4)
113.9 (7.8)
119.1 (8.4)
117.2 (7.4)
123.0 (7.7)
0.005
113.8 (7.9)
115.9 (7.9)
120.6 (7.8)
0.02
SDNNIDX, ms
39.4 (3.7)
41.4 (3.9)
43.0 (4.0)
42.9 (3.8)
43.8 (3.7)
0.02
39.6 (3.8)
42.0 (3.9)
43.6 (4.1)
0.002
rMSSD**, ms
18.1 (1.7)
18.7 (1.8)
19.3 (1.9)
19.3 (1.8)
19.5 (1.8)
0.25
18.8 (1.7)
18.6 (1.8)
19.4 (1.9)
0.66
Frequency-domain indices
NLF**, %
67.4 (1.8)
68.6 (1.6)
68.3 (1.7)
69.0 (1.6)
68.7 (1.6)
0.18
68.0 (1.8)
68.3 (1.7)
68.7 (1.6)
0.29
NHF**, %
19.7 (1.5)
19.1 (1.4)
19.1 (1.5)
18.8 (1.6)
18.6 (1.5)
0.09
19.1 (1.5)
19.1 (1.5)
18.9 (1.5)
0.49
VLF, ms
ULF, 1000 ms
11.2 (1.5)
11.1 (1.5)
12.1 (1.7)
11.6 (1.5)
13.0 (1.7)
0.01
11.3 (1.6)
11.5 (1.5)
12.4 (1.6)
0.11
LF/HF** ratio
5.0 (0.5)
5.2 (0.4)
5.2 (0.5)
5.3 (0.5)
5.2 (0.5)
0.27
5.2(0.5)
5.1 (0.5)
5.3 (0.5)
0.42
0.26 (0.03)
0.26 (0.02)
0.25 (0.02)
0.25 (0.03)
0.24 (0.02)
0.04
0.26 (0.03)
0.25 (0.02)
0.24 (0.03)
0.11
1.01 (0.06)
1.03 (0.06)
1.06 (0.06)
1.06 (0.06)
1.08 (0.06)
0.01
1.02 (0.06)
1.04 (0.06)
1.07 (0.06)
0.03
Nonlinear indices
Poincare ratio
(SD12)
DFA1
Measures (except NLF, NHF, and DFA1) were log-transformed for analysis and then exponentiated. Values are mean (SD), adjusted for age (years), sex
(male, female), race (white, nonwhite), enrollment center (4 sites), education (<high school, high school, college), income ($25,000,. >$25,000), smoking
(never, former, current), alcohol (<1 drink/week, 1-2 drinks/week, 3-7 drinks/week, 8-14 drinks/week, >14 drinks/week), and consumption of fish (<1
serving/month, 1-3 servings/month, 1-2 servings/week, 3-4 servings/week, 5+ servings/week), dietary fiber (quintiles, g/day), and total calories (quintiles,
kcal/day).
For values log transformed we report the upper SD.
Numbers are shown for time-domain measures (n=1219); slightly fewer individuals (n=1150) had frequency-domain and nonlinear measures.
** rMSSD, NLF, NHF and LF/HF ratio were evaluated among individuals with lower erratic HRV (DFA>median(1.044); n=675).