Combined assessment of heart rate recovery and T-wave alternans during routine exercise testing improves prediction of total and cardiovascular mortality: the Finnish Cardiovascular Study.

Abstract

Identification of individuals who are at risk for cardiovascular death remains a pressing public health challenge. Derangements in autonomic function acting upon an electrically unstable substrate are thought to be critical elements in triggering cardiovascular events.
The purpose of this study was to analyze heart rate recovery (HRR) in combination with T-wave alternans (TWA) to improve risk assessment.
The Finnish Cardiovascular Study (FINCAVAS) enrolled consecutive patients (N = 1,972 [1,254 men and 718 women], age 57 +/- 13 years [mean +/- SD]) with a clinically indicated exercise test using bicycle ergometer. TWA was analyzed continuously with the time-domain modified moving average method. Maximum TWA at heart rates <125 bpm was derived.
During 48 +/- 13 months of follow-up (mean +/- SD), 116 patients died; 55 deaths were cardiovascular. In multivariable Cox analysis after adjustment for common coronary risk factors, high exercise-based TWA (> or =60 microV) and low HRR (< or =18 bpm) yielded relative risks for all-cause mortality of 5.0 (95% confidence 2.1-12.1, P <.01) and for cardiovascular mortality of 12.3 (95% confidence interval 4.3-35.3, P <.01). High recovery-based TWA (> or =60 microV) and low HRR (< or =18 bpm) yielded relative risks for all-cause death of 6.1 (95% confidence interval 2.8-13.2, P <.01) and for cardiovascular mortality of 8.0 (95% confidence interval 2.9-22.0, P <.01). Prediction by HRR and TWA, both singly and in combination, exceeded that of standard cardiovascular risk factors.
Reduced HRR and heightened TWA powerfully predict risk for cardiovascular and all-cause death in a low-risk population. This novel approach could aid in screening of general populations during routine exercise protocols as well as improve insights into pathophysiology.

Full text

Page 1

Combined assessment of heart rate recovery and T-wave alternans
during routine exercise testing improves prediction of total and
cardiovascular mortality: The Finnish Cardiovascular Study
Johanna Leino, BMS,* Mikko Minkkinen, BMS,* Tuomo Nieminen, MD, PhD,†Terho Lehtimäki, MD, PhD,*‡
Jari Viik, PhD,¶Rami Lehtinen, PhD,*§?Kjell Nikus, MD,** Tiit Kööbi, MD, PhD,*§Väinö Turjanmaa, MD, PhD,*§
Richard L. Verrier, PhD,††Mika Kähönen, MD, PhD*§
From the *Medical School, University of Tampere, Tampere, Finland,†Department of Pharmacological Sciences, Medical
School, University of Tampere, and Department of Internal Medicine, Päijät-Häme Central Hospital, Lahti, Finland,
‡Laboratory of Atherosclerosis Genetics, Department of Clinical Chemistry, Tampere University Hospital, Tampere, Finland,
¶Ragnar Granit Institute, Department of Biomedical Engineering, Tampere University of Technology, Tampere, Finland,
§Department of Clinical Physiology, Tampere University Hospital, Tampere, Finland,?Tampere Polytechnic University of
Applied Sciences, Tampere, Finland, **Heart Centre, Department of Cardiology, Tampere University Hospital, Tampere,
Finland, and††Harvard Medical School, Beth Israel Deaconess Medical Center, Boston, Massachusetts.
BACKGROUND Identification of individuals who are at risk for
cardiovascular death remains a pressing public health challenge.
Derangements in autonomic function acting upon an electrically
unstable substrate are thought to be critical elements in trigger-
ing cardiovascular events.
OBJECTIVE The purpose of this study was to analyze heart rate
recovery (HRR) in combination with T-wave alternans (TWA) to
improve risk assessment.
METHODS The Finnish Cardiovascular Study (FINCAVAS) enrolled
consecutive patients (N ? 1,972 [1,254 men and 718 women], age
57 ? 13 years [mean ? SD]) with a clinically indicated exercise
test using bicycle ergometer. TWA was analyzed continuously with
the time-domain modified moving average method. Maximum TWA
at heart rates ?125 bpm was derived.
RESULTS During 48 ? 13 months of follow-up (mean ? SD), 116
patients died; 55 deaths were cardiovascular. In multivariable Cox
analysis after adjustment for common coronary risk factors, high
exercise-based TWA (?60 ?V) and low HRR (?18 bpm) yielded
relative risks for all-cause mortality of 5.0 (95% confidence 2.1–
12.1, P ?.01) and for cardiovascular mortality of 12.3 (95%
confidence interval 4.3–35.3, P ?.01). High recovery-based TWA
(?60 ?V) and low HRR (?18 bpm) yielded relative risks for
all-cause death of 6.1 (95% confidence interval 2.8–13.2,
P ?.01) and for cardiovascular mortality of 8.0 (95% confi-
dence interval 2.9–22.0, P ?.01). Prediction by HRR and TWA,
both singly and in combination, exceeded that of standard
cardiovascular risk factors.
CONCLUSION Reduced HRR and heightened TWA powerfully pre-
dict risk for cardiovascular and all-cause death in a low-risk pop-
ulation. This novel approach could aid in screening of general
populations during routine exercise protocols as well as improve
insights into pathophysiology.
KEYWORDS Exercise test; Heart rate recovery; Mortality; Progno-
sis; T-wave alternans
ABBREVIATIONS EF ? ejection fraction; FINCAVAS ? Finnish
Cardiovascular Study; HRR ? heart rate recovery; SCD ? sudden
cardiac death; TWA ? T-wave alternans
(Heart Rhythm 2009;6:1765–1771) © 2009 Heart Rhythm Society.
All rights reserved.
Introduction
An abnormal autonomic nervous system response in terms
of heart rate recovery (HRR) during or after clinical exer-
cise testing predicts all-cause and cardiovascular mortality
in a variety of relatively low-risk cohorts,1–7including
ours.8The reduction in heart rate during the first 30 to 60
seconds after exercise appears to be caused principally by
reactivation of the parasympathetic nervous system but sub-
sequently by withdrawal of sympathetic tone.9
T-wave alternans (TWA) is an ECG phenomenon indi-
cating an electrically unstable myocardial substrate.10This
beat-to-beat alternation in the shape, amplitude, or timing of
the ST segment and the T wave has been found to predict
sudden cardiac death (SCD) and cardiovascular and total
mortality independent of standard factors in relatively low-
Dr. Verrier is co-inventor of the modified moving average method for
T-wave alternans analysis, with patent assigned to Beth Israel Deaconess
Medical Center and licensed by GE Healthcare. Financial support was
received from the Medical Research Fund of Tampere University Hospital,
Tampere Tuberculosis Foundation, and Finnish Cultural Foundation. Ad-
dress reprint requests and correspondence: Dr. Mika Kähönen, Depart-
ment of Clinical Physiology, Tampere University Hospital, FI-33520,
Tampere, Finland. E-mail address: mika.kahonen@uta.fi.
March 15, 2009; accepted August 12, 2009.)
(Received
1547-5271/$ -see front matter © 2009 Heart Rhythm Society. All rights reserved.doi:10.1016/j.hrthm.2009.08.015

Page 2

risk populations,11,12including ours13,14as well as in higher-
risk groups.15–19We applied the time-domain modified
moving average method,20which permits TWA measure-
ment during routine symptom-limited exercise.13,14
HRR and TWA reflect different pathophysiologic mech-
anisms. The aims of this study were to determine whether
the combined analysis of HRR and TWA during routine
exercise testing enhances their predictive power for cardio-
vascular and all-cause mortality over independent assess-
ment of either variable and to compare their predictive
strength to that of other standard risk factors.
Methods
Study cohort
All consecutive patients who were referred for an exercise
stress test at Tampere University Hospital between October
2001andtheendof2004andwerewillingtoparticipateinThe
Finnish Cardiovascular Study (FINCAVAS)21were recruited.
A total of 1,972 patients (1,254 men and 718 women) with
technically successful exercise tests were enrolled in the study.
A test was considered technically adequate if storing the he-
modynamic data and continuous digital ECG signal was suc-
cessful. Patients with atrial fibrillation (N ? 31) were excluded
because atrial fibrillation is an exclusion criterion in HRR
studies.2,3The main indications for the exercise test were
suspicion of coronary heart disease (frequency 45%); testing
vulnerability to arrhythmia during exercise (22%); evaluation
of work capacity (18%) and the adequacy of treatment of
coronary heart disease (16%); and obtaining an exercise test
profile prior to an invasive procedure (13%) or after a myo-
cardial infarction (8%). Some patients had more than one
indication. The study protocol was approved by the Ethics
Committee of the Tampere University Hospital District, Fin-
land, and all patients gave informed consent prior to the inter-
view and measurements as stipulated in the Declaration of
Helsinki.
Study flow
After written informed consent was obtained, the medical
history of each patient was collected via a computer-based
questionnaire form. The exercise test then was performed.
Exercise test protocol
The subject lay down in the supine position for 10 minutes,
and the resting ECG was digitally recorded. The upright
routine exercise test then was performed using a bicycle
ergometer with electrical brakes. The lead system consisted
of the Mason-Likar modification of the standard 12-lead
system. The initial workload varied from 20 to 30 W, and
the load was increased stepwise by 10 to 30 W every
minute. Continuous ECGs were digitally recorded at 500 Hz
using the CardioSoft exercise ECG system (version 4.14,
GE Healthcare, Freiburg, Germany). During the test, heart
rate and ST segment deviation were continuously registered
on the ECG, while systolic arterial pressure and diastolic
arterial pressure were measured with a brachial cuff every 2
minutes.
Measurement of HRR
HRR was determined as the difference between maximum
heart rate during exercise minus heart rate during the first
minute following cessation of exercise. We used the HRR
cutpoint of ?18 bpm, which has been suggested for exercise
tests with an abrupt end.22Differences in recovery protocols
have not negated the predictive strength of HRR.22
Measurement of TWA
Assessing the relationship between TWA and mortality is
one of the original goals of FINCAVAS.21We used the
time-domain, Food and Drug Administration–cleared mod-
ified moving average method because of its intrinsic flexi-
bility and demonstrated capacity to measure TWA accu-
rately under dynamic conditions, including changing heart
rates, myocardial ischemia, exercise, activity, and behav-
ioral stress.11,13,14,16,19,23In brief, the modified moving av-
erage algorithm reports TWA as the maximum difference in
T-wave morphology between successive beats. It separates
odd from even beats, calculates average morphologies of
both the odd and even beat streams separately, and contin-
uously updates the result by a weighting factor of 1/8 of the
difference between the ongoing average and the new incom-
ing beat. The method performs at a resolution of 1 ?V and
has undergone extensive validation.20
TWA values were calculated automatically and contin-
uously by the released version of GE Healthcare’s modified
moving average algorithm during rest, exercise, and recov-
ery using all standard precordial leads (V1–V6). Maximum
TWA values at heart rates ?125 bpm were derived. TWA
values at higher heart rates were excluded because inaccu-
racies in TWA measurement can result at heart rates ex-
ceeding this range. Precordial leads have been shown to be
optimum for TWA measurement.24,25The exercise-based
TWA cutpoint of 60 ?V, which yielded excellent Cox
regression results in our previous study,14was used. Recov-
ery-based TWA values were analyzed according to cut-
points 20 ?V and 60 ?V.14,26TWA cutpoint of 20 ?V was
chosen because it has shown the highest sensitivities com-
pared with other cutpoints.26
Left ventricular ejection fraction
Measurement of left ventricular ejection fraction (EF) is not
routine for patients referred for a clinical exercise test.
However, EF was determined for 1,200 (55%) of the study
patients using echocardiography or isotope techniques
within 6 months of the exercise test. More than one fifth
(N ? 408 [21%]) of the patients were examined with cor-
onary angiography.
Follow-up
Death certificates were received from the Causes of Death
Register, maintained by Statistics Finland, in May 2007, a
source that has been shown to be reliable.27The certificates
included causes of death using the tenth revision of the
International Classification of Diseases (ICD-10). The diag-
nosis numbers and certificate texts were used to classify the
1766 Heart Rhythm, Vol 6, No 12, December 2009

Page 3

deaths as all cause or cardiovascular. The investigators who
analyzed TWA test results were blinded to events.
Statistical analysis
The t-test for independent samples was used to compare
continuous parameters of patient characteristics (Table 1)
and exercise test variables (Table 2) for survivors and non-
survivors. The Chi-square test was applied for dichotomous
variables. P values were derived with the t-test and the
Chi-square test for independent samples. Relative risks for
total and cardiovascular mortality were analyzed for HRR,
TWA, and their combinations as well as for ST-segment
deviation by Cox regression analysis after adjustment by
standard covariates (Table 3). The proportionality assump-
tion for all covariates was checked by using correlations of
the survival rankings with the Schoenfeld residuals. All of
the covariatesfulfilledtheproportionalityassumption.Harrell’sC
indices also were calculated (Table 4). The calculations for
combination variables were based on three categories: no pa-
rameter positive, either parameter positive, and both parame-
ters positive. Harrell’s C index is a generalization of the area
under the receiver operator characteristic (ROC) curve for
survival data with censored cases. Values above 0.5 show
better than random prediction, and a value of one represents
perfect concordance between predicted and observed numbers.
Statistics were analyzed using SPSS release 14.0 for Win-
dows (SPSS, Inc., Chicago, IL, USA) and Stata 10.1 for Win-
dows (StataCorp LP, College Station, TX, USA). All statistical
tests were two-tailed and used an alpha level ?.05.
Results
Baseline characteristics
During the follow-up period of 48 ? 13 months (mean ?
SD) in our study population of 1,972 consecutive patients
referred for clinical exercise testing, there were 116 deaths
(5.9% of the population), including 55 (2.8% of the popu-
Table 1
Patient characteristics of the study population
Parameter
Survivors (N ? 1,856) Deaths (N ? 116)
P valueMeanSD Mean SD
Age (years)
Weight (kg)
Height (cm)
Body mass index
56.5
80.5
171.0
27.5
13.1
15.2
9.3
4.5
65.1
79.0
171.6
26.7
11.3
15.8
9.3
4.1
?.01
.27
.49
.07
N%N%
P value
Sex: female
Smoking: yes
Nitrates
Beta-blockers
Hypercholesterolemia
Diabetes
Coronary heart disease
Left ventricular hypertrophy
History of myocardial infarction
776
534
695
37.8
26.0
33.9
57.3
49.9
11.6
38.2
4.4
20.7
30
41
66
99
64
23
64
7
44
23.6
32.3
52.0
78.0
50.4
18.1
50.4
5.5
34.6
?.01
.12
?.01
?.01
.91
.03
.01
.57
?.01
1,174
1,024
238
784
91
425
Table 2
Exercise test variables of the study population
Parameter
Survivors (N ? 1,856) Deaths (N ? 116)
P value Mean SD MeanSD
Duration of test (minutes)
Age-adjusted expected maximum HR (bpm)
Reached maximum HR (bpm)
Maximum SAP during the exercise (mmHg)
Maximum DAP during the exercise (mmHg)
HR at rest (bpm)
SAP at rest (mmHg)
DAP at rest (mmHg)
Maximum TWA at rest before exercise (?V)
Maximum TWA during exercise (?V)
Maximum TWA during recovery (?V)
Maximum left ventricular ejection fraction
HRR at 1 minute postexercise (bpm)
ST-segment deviation during exercise (mV)
7.52.1
6.7
26.4
28.5
11.9
11.6
18.6
9.6
11.5
21.8
23.4
13.8
11.5
0.10
6.2 2.0
5.6
29.6
32.0
13.4
14.5
25.8
11.4
17.2
23.3
19.5
15.6
13.8
0.14
?.01
?.01
?.01
?.01
?.01
.19
.62
?.01
?.01
.04
.03
?.01
?.01
.01
176.8
146.6
193.6
92.3
63.1
136.1
79.7
19.4
35.8
26.7
65.9
24.7
0.08
172.5
127.0
175.6
85.1
64.5
135.2
75.7
25.5
39.9
31.3
60.2
18.2
0.11
DAP ? diastolic arterial pressure; HR ? heart rate; HRR ? heart rate recovery; SAP ? systolic arterial pressure; TWA ? T-wave alternans.
1767Leino et alCombined Heart Rate Recovery and T-Wave Alternans for Risk Stratification

Page 4

lation; 47.4% of all deaths) that were classified as cardio-
vascular deaths. Thus, the cardiovascular mortality of the
present patients was 0.7% per year. Patient characteristics
and exercise test variables for survivors (N ? 1,856) and
nonsurvivors (N ? 116) are given in Tables 1 and 2,
respectively.
Mortality, HRR, and TWA
HRR was abnormal in 29.5% (N ? 590) of the population.
Exercise-based TWA ?60 ?V was found in 5.2% (N ?
107). During recovery, 51.3% (N ? 1,063 patients) had
TWA ?20 ?V, including 3.9% (N ? 81 patients) with
TWA ?60 ?V. Thus, the present approach classified the
majority of the patients as low risk. Combined Cox propor-
tional hazard analysis of depressed HRR and heightened
exercise- or recovery-based TWA more than doubled the
prognostic capacity for total and cardiovascular mortality
after adjustment for standard risk factors and exceeded
exercise-induced ST-segment deviation (Table 3). In addi-
tion to standard covariates, maximum left ventricular EF,
blood pressures at rest, maximum blood pressures during ex-
ercise, and resting heart rate were added to the multivariate
analysis with the combination of HRR and TWA. None of
these factors exceeded the predictive power of the combination
of HRR and TWA. Incidence rates of all-cause and cardiovas-
cular deaths in subgroups are shown in Figure 1.
Harrell’s C indices were calculated for all single and
combination parameters as well as for ST-segment devia-
tion (Table 4). For the single parameters, HRR provided the
highest C index for both total and cardiovascular mortality.
Adding exercise-based TWA ?60 ?V to reduced HRR
yielded highest C index for all-cause and cardiovascular
mortality, although confidence intervals overlapped with
HRR alone.
Survival curves depict events across 4 years of follow-up
for the combined analysis of reduced HRR and elevated
TWA during exercise (Figure 2) and recovery (Figure 3).
Discussion
Our study is the first to demonstrate that the presence of
high levels of TWA during exercise or recovery adds sig-
nificantly to the prognostic power of poor HRR for all-cause
and cardiovascular mortality. Because both markers are
automated and widely used parameters that can be moni-
Table 3
mortality
Results of Cox multivariable regression analysis (N ? 1,972) of relative risks for all-cause mortality and cardiovascular
All-cause mortality Cardiovascular mortality
RR
95% CI
P valueRR
95% CI
P valueLowerUpper LowerUpper
HRR ?18 bpm
Exercise-based TWA ?60 ?V
Recovery-based TWA ?20 ?V
Recovery-based TWA ?60 ?V
HRR ?18 bpm and exercise-based TWA ?60 ?V
HRR ?18 bpm or exercise-based TWA ?60 ?V
HRR and recovery-based TWA ?20 ?V
HRR or recovery-based TWA ?20 ?V
HRR and recovery-based TWA ?60 ?V
HRR or recovery-based TWA ?60 ?V
ST-segment deviation (0.1 mV) during exercise
2.5
2.5
1.1
2.4
5.0
2.8
3.0
2.0
6.1
2.3
1.3
1.6
1.4
0.7
1.3
2.1
1.8
1.6
1.2
2.8
1.5
0.9
3.7
4.5
1.6
4.4
12.1
4.3
5.5
3.5
13.2
3.5
1.9
?.01
?.01
.73
?.01
?.01
?.01
?.01
.01
?.01
?.01
.18
2.3
5.8
1.5
3.5
12.3
3.4
5.2
3.5
8.0
2.2
1.8
1.3
3.1
0.8
1.6
4.3
1.8
1.8
1.3
2.9
1.2
1.0
4.2
11.1
2.5
7.9
35.3
6.6
14.4
9.1
22.0
4.2
3.0
.01
?.01
.18
?.01
?.01
?.01
?.01
.01
?.01
.01
.04
Results after adjustment for sex, age, body mass index, smoking (yes/no), use of beta-blockers (yes/no), reached maximum heart rate, and prior
diagnoses of coronary heart disease (yes/no), history of myocardial infarction (yes/no), diabetes (yes/no), and hypercholesterolemia (yes/no).
CI ? confidence interval; HRR ? heart rate recovery; RR ? relative risk; TWA ? T-wave alternans.
Table 4
Harrell’s C indices for cardiovascular and all-cause mortality
All-cause
death
95% CI
Cardiovascular death
95% CI
LowerUpperLower Upper
HRR ?18 bpm
Exercise-based TWA ?60 ?V
Recovery-based TWA ?20 ?V
Recovery-based TWA ?60 ?V
HRR ?18 bpm and/or exercise-based TWA ?60 ?V
HRR ?18 bpm and/or recovery-based TWA ?20 ?V
HRR ?18 bpm and/or recovery-based TWA ?60 ?V
ST-segment deviation (0.1 mV) in exercise test
0.655
0.539
0.555
0.526
0.677
0.655
0.661
0.558
0.609
0.507
0.508
0.499
0.631
0.608
0.614
0.510
0.702
0.570
0.601
0.553
0.723
0.702
0.709
0.606
0.650
0.594
0.606
0.550
0.713
0.691
0.671
0.580
0.583
0.535
0.544
0.535
0.648
0.633
0.602
0.511
0.718
0.653
0.668
0.653
0.777
0.749
0.740
0.649
Calculations for combination variables were based on three categories (0, 1, or 2 parameters positive).
CI ? confidence interval; HRR ? heart rate recovery; TWA ? T-wave alternans.
1768 Heart Rhythm, Vol 6, No 12, December 2009

Page 5

tored in conjunction with routine exercise testing, their
combination may serve as a new risk stratification tool for
screening low-risk patient populations.
Previous studies
The significant influence of autonomic nervous system ac-
tivity on cardiovascular and total mortality has been amply
demonstrated, most recently by baroreceptor sensitivity
(BRS)28and noninvasive assessment with heart rate vari-
ability,28,29heart rate turbulence,30and HRR.1–8The latter
is a strong predictor of cardiovascular mortality in asymp-
tomatic patients5,7,8and in broad populations4as well as of
SCD.1Importantly, impaired HRR is not attributable to
ischemic burden3or lipid abnormalities.6Treadmill exercise
scores strongly predict mortality among intermediate- to
high-risk patients if HRR is abnormal.6
The accuracy and utility of exercise-based TWA in predict-
ing arrhythmic events and death have been investigated.12–15
Most TWA studies have been performed in high-risk popula-
tions, such as patients with heart failure,15,17–19cardiomyopa-
thies,15,18or history of myocardial infarction.11,12,15–19We
previously reported in approximately 2,000 FINCAVAS pa-
tients that TWA analyzed with the modified moving average
method is a strong predictor of all-cause and cardiovascular
mortality as well as of SCD in this low-risk population.14
Especially high specificity when compared with other car-
diovascular parameters has characterized the prognostic
value of elevated TWA,13suggesting suitability to confirm
suspected risk.
Figure 2
exercise-based T-wave alternans (TWA) ?60 ?V and heart rate recovery
(HRR) ?18 bpm (top curve in both panels), TWA ?60 ?V or HRR ?18
bpm (middle curve in both panels), and TWA ?60 ?V and HRR ?18 bpm
(bottom curve in both panels) for all-cause mortality (A) and cardiovascu-
lar mortality (B). Note that the scale for the y-axis is from 0.75 to 1.00.
Adjusted survival curves by Cox regression for subjects with
Figure 3
recovery-based T-wave alternans (TWA) ?60 ?V and heart rate recovery
(HRR) ?18 bpm (top curve in both panels), TWA ?60 ?V or HRR ?18
bpm during recovery (middle curve in both panels), and TWA ?60 ?V and
HRR ?18 bpm during recovery (bottom curve in both panels) for all-cause
mortality (A) and cardiovascular mortality (B). Note that the scale for the
y-axis is from 0.75 to 1.00.
Adjusted survival curves by Cox regression for subjects with
Figure 1
1000 person-years among patients according to exercise-based T-wave
alternans (TWA) and heart rate recovery (HRR).
Incidence rate of all-cause and cardiovascular mortality per
1769 Leino et al Combined Heart Rate Recovery and T-Wave Alternans for Risk Stratification

Page 6

The potential to improve prediction of cardiovascular
and total mortality by combining TWA with the ambulatory
ECG-based autonomic marker of heart rate turbulence was
recently confirmed in a high-risk population of postmyocar-
dial infarction patients with left ventricular dysfunction.16
The present study, which enrolled a 6.9-fold larger, lower-
risk population of almost 2,000 patients, demonstrated fur-
ther improvements in odds ratio.
Current investigation
The present study confirms and extends the findings of our
previous investigations of TWA13,14and HRR8in the low-
risk FINCAVAS patient population. When analyzed to-
gether, TWA and HRR provide high relative risk ratios for
all-cause death and for cardiovascular mortality after adjust-
ment for standard risk factors (Table 3), indicating a marked
independent prognostic capacity and exceeding the predic-
tive value of either parameter alone or ST-segment devia-
tion. The combinations of reduced HRR with heightened
TWA were superior to exercise-induced ST-segment devi-
ation in our low-risk population using Cox proportional
hazards models (Table 3) and Harrell’s C indices (Table 4).
The incidence rate of all-cause as well as cardiovascular
deaths was clearly higher among patients with reduced HRR
and heightened TWA compared to patients with normal
values (Figure 1).
The mechanistic basis for the improvement in prediction
resulting from combined analysis of HRR and TWA is
unclear. A plausible explanation is that a more complete
picture of underlying pathophysiologic factors is rendered
by information regarding both autonomic function and car-
diac electrical instability. As HRR is thought to reflect the
dynamic interplay between sympathetic and parasympa-
thetic nerve activity as influenced by changes in barorecep-
tor gain,1a reduced HRR may indicate autonomic imbal-
ance as a basis for cardiovascular events. Moreover, HRR
may reflect aerobic capacity and physical fitness, which
have been linked to prognosis.31The independent associa-
tion between increased risk for all-cause and cardiovascular
mortality and TWA13,14is consistent with the finding that
TWA indicates increased heterogeneity of repolariza-
tion.10,32Although the incidence of SCD was not evaluated
in the current investigation, because both TWA13,14and
reduced HRR1have been independently associated with
SCD in low-risk populations, it is possible that a number of
the cardiovascular deaths were arrhythmic in origin. Ath-
erosclerotic heart disease, typical of 29 (48%) of patients
who died of cardiovascular causes, predisposes to ventric-
ular fibrillation and SCD.33Accordingly, reduced HRR
could indicate impaired vagus nerve activation and lessened
capacity to withdraw sympathetic nerve tone, both influ-
ences known to be arrhythmogenic.27,34Thus, the presence
of both abnormal HRR and elevated TWA, reflecting de-
rangements in autonomic function as well as in cardiac
electrical instability, would be expected to be associated
with the highest risk for cardiovascular events, as demon-
strated in the present study.
Study limitations
We do not have information on changes in parameters
affecting mortality risk (e.g., smoking, lifestyles, medica-
tion) during follow-up. In addition, data on EF were not
available for 45% of patients. It is likely that patients in
whom no need was found for EF determinations had even
better cardiovascular health than did those with EF mea-
surement. EF is an arrhythmia risk stratifier only when EF
levels are below normal.35
Conclusion
A broad implication of the study finding is that routine
exercise testing discloses increased risk for cardiovascular
as well as all-cause death among patients with both de-
pressed HRR and abnormal TWA who are not identified by
standard risk factors. In addition to improving predictivity,
the combined assessment of HRR and TWA may be helpful
in gaining insight into the pathophysiologic mechanisms on
an individual patient basis that could help to guide therapy.
In particular, patients with markedly depressed HRR could
be directed toward an exercise training regimen that im-
proves vagus nerve tone, BRS, and long-term prognosis.31
TWA results reflective of an unstable cardiac substrate
could signal the need for antiarrhythmic therapy. Finally,
particularly as the measurements can be performed nonin-
vasively during routine exercise testing in the typical flow
of clinical care, these parameters can readily be incorpo-
rated, either singly or in combination, into routine risk
assessment paradigms.
Acknowledgements
We thank the staff of the Department of Clinical Physiol-
ogy, Tampere University Hospital, for collecting the exer-
cise test data.
References
1. Jouven X, Empana JP, Schwartz PJ, Desnos M, Courbon D, Ducimetière P.
Heart-rate profile during exercise as a predictor of sudden death. N Engl J Med
2005;352:1951–1958.
2. Elhendy A, Mahoney DW, Khandheria BK, Burger K, Pellikka PA. Prognostic
significance of impairment of heart rate response to exercise: impact of left
ventricular function and myocardial ischemia. J Am Coll Cardiol 2003;42:823–
830.
3. Vivekananthan DP, Blackstone EH, Pothier CE, Lauer MS. Heart rate recovery
after exercise is a predictor of mortality, independent of the angiographic
severity of coronary disease. J Am Coll Cardiol 2003;42:831–838.
4. Cole CR, Blackstone EH, Pashkow FJ, Snader CE, Lauer MS. Heart-rate
recovery immediately after exercise as a predictor of mortality. N Engl J Med
1999;341:1351–1357.
5. Aktas MK, Ozduran V, Pothier CE, Lang R, Lauer MS. Global risk scores and
exercise testing for predicting all-cause mortality in a preventive medicine
program. JAMA 2004;292:1462–1468.
6. Nishime EO, Cole CR, Blackstone EH, Pashkow FJ, Lauer MS. Heart rate
recovery and treadmill exercise score as predictors of mortality in patients
referred for exercise ECG. JAMA 2000;284:1392–1398.
7. Mora S, Redberg RF, Cui Y, et al. Ability of exercise testing to predict
cardiovascular and all-cause death in asymptomatic women: a 20-year follow-up
of the lipid research clinics prevalence study. JAMA 2003;290:1600–1607.
8. Nieminen T, Leino J, Maanoja J, et al. The prognostic value of hemodynamic
parameters in the recovery phase of an exercise test. The Finnish Cardiovascular
Study. J Human Hypertens 2008;22:537–543.
9. Imai K, Sato H, Hori M, et al. Vagally mediated heart rate recovery after
exercise is accelerated in athletes but blunted in patients with chronic heart
failure. J Am Coll Cardiol 1994;24:1529–1535.
1770 Heart Rhythm, Vol 6, No 12, December 2009

Page 7

10. Verrier RL, Kumar K, Nearing BD. Basis for sudden cardiac death prediction by
T-wave alternans from an integrative physiology perspective. Heart Rhythm
2009;6:416–422.
11. Verrier RL, Nearing BD, La Rovere MT, et al. Ambulatory electrocardiogram-
based tracking of T wave alternans in postmyocardial infarction patients to
assess risk of cardiac arrest or arrhythmic death. J Cardiovasc Electrophysiol
2003;14:705–711.
12. Ikeda T, Yoshino H, Sugi K, et al. Predictive value of microvolt T-wave
alternans for sudden cardiac death in patients with preserved cardiac function
after acute myocardial infarction: Results of a collaborative cohort study. J Am
Coll Cardiol 2006;48:2268–2274.
13. Nieminen T, Lehtimäki T, Viik J, et al. T-wave alternans predicts mortality in
a population undergoing a clinically indicated exercise test. Eur Heart J 2007;
28:2332–2337.
14. Minkkinen M, Kähönen M, Viik J, et al. Enhanced predictive power of quan-
titative TWA during routine exercise testing in the Finnish Cardiovascular
Study. J Cardiovasc Electrophysiol 2009;20:408–415.
15. Gehi AK, Stein RH, Metz LD, Gomes JA. Microvolt T-wave alternans for the
risk stratification of ventricular tachyarrhythmic events: a meta-analysis. J Am
Coll Cardiol 2005;46:75–82.
16. Exner DV, Kavanagh KM, Slawnych MP, et al, for the REFINE Investigators.
Noninvasive risk assessment early after a myocardial infarction: the REFINE
study. J Am Coll Cardiol 2007;50:2275–2284.
17. Stein PK, Sanghavi D, Domitrovich PP, Mackey RA, Deedwania P. Ambulatory
ECG-based T-wave alternans predicts sudden cardiac death in high-risk post-MI
patients with left ventricular dysfunction in the EPHESUS Study. J Cardiovasc
Electrophysiol 2008;19:1037–1042.
18. Salerno-Uriarte JA, De Ferrari GM, Klersy C, et al. Prognostic value of T-wave
alternans in patients with heart failure due to nonischemic cardiomyopathy:
results of the ALPHA Study. J Am Coll Cardiol 2007;50:1896–1904.
19. Sakaki K, Ikeda T, Miwa Y, et al. Time-domain T-wave alternans measured
from Holter electrocardiograms predicts cardiac mortality in patients with left
ventricular dysfunction: a prospective study. Heart Rhythm 2009;6:332–337.
20. Nearing BD, Verrier RL. Modified moving average analysis of T-wave alternans
to predict ventricular fibrillation with high accuracy. J Appl Physiol 2002;92:
541–549.
21. Nieminen T, Lehtinen R, Viik J, et al. The Finnish Cardiovascular Study
(FINCAVAS): characterising patients with high risk of cardiovascular morbidity
and mortality. BMC Cardiovasc Disord 2006;6:9.
22. Shetler K, Marcus R, Froelicher VF, et al. Heart rate recovery: validation and
methodologic issues. J Am Coll Cardiol 2001;38:1980–1987.
23. Kop WJ, Krantz DS, Nearing BD, et al. Effects of acute mental stress and
exercise on T-wave alternans in patients with implantable cardioverter defibril-
lators and controls. Circulation 2004;109:1864–1869.
24. Nearing BD, Oesterle SN, Verrier RL. Quantification of ischaemia-induced
vulnerability by precordial T wave alternans analysis in dog and human. Car-
diovasc Res 1994;28:1440–1449.
25. Martínez JP, Olmos S, Wagner G, Laguna P. Characterization of repolarization
alternans during ischemia: time-course and spatial analysis. IEEE Trans Biomed
Eng 2006;53:701–711.
26. Slawnych MP, Nieminen T, Kähönen M, et al. Post-exercise assessment of
cardiac repolarization alternans in patients with coronary artery disease using the
modified moving average method. J Am Coll Cardiol 2009;53:1130–1137.
27. Pajunen P, Koukkunen H, Ketonen M, et al. The validity of the Finnish Hospital
Discharge Register and Causes of Death Register data on coronary heart disease.
Eur J Cardiovasc Prev Rehabil 2005;12:132–137.
28. La Rovere MT, Pinna GD, Hohnloser SH, et al. Baroreflex sensitivity and heart
rate variability in the identification of patients at risk for life-threatening ar-
rhythmias: implications for clinical trials. Circulation 2001;103:2072–2077.
29. Lombardi F. Clinical implications of present physiological understanding of
HRV components. Card Electrophysiol Rev 2002;6:245–249.
30. Schmidt G, Malik M, Barthel P, et al. Heart-rate turbulence after ventricular
premature beats as a predictor of mortality after acute myocardial infarction.
Lancet 1999;353:1390–1396.
31. La Rovere MT, Bersano C, Gnemmi M, Specchia G, Schwartz PJ. Exercise-
induced increase in baroreflex sensitivity predicts improved prognosis after
myocardial infarction. Circulation 2002;106:945–949.
32. Nearing BD, Verrier RL. Tracking heightened cardiac electrical instability by
computing interlead heterogeneity of T-wave morphology. J Appl Physiol 2003;
95:2265–2272.
33. Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to vulnerable patient:
a call for new definitions and risk assessment strategies: part II. Circulation
2003;108:1772–1778.
34. Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and sudden
cardiac death: experimental basis and clinical observations for post-myocardial
infarction risk stratification. Circulation 1992;85(Suppl I):I77–I91.
35. Huikuri HV, Castellanos A, Myerburg RJ. Sudden death due to cardiac arrhyth-
mias. N Engl J Med 2001;345:1473–1482.
1771 Leino et alCombined Heart Rate Recovery and T-Wave Alternans for Risk Stratification