Original Article
Physical Activity Measured With Implanted Devices Predicts
Patient Outcome in Chronic Heart Failure
Viviane M. Conraads, MD*; Martijn A. Spruit, PhD; Frieder Braunschweig, MD;
Martin R. Cowie, MD; Luigi Tavazzi, MD; Martin Borggrefe, MD; Michael R.S. Hill, PhD;
Sandra Jacobs, PhD; Bart Gerritse, PhD; Dirk J. van Veldhuisen, MD
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Background—Physical activity (PA) predicts cardiovascular mortality in the population at large. Less is known about its
prognostic value in patients with chronic heart failure (HF).
Methods and Results—Data from 836 patients with implantable cardioverter defibrillator without or with cardiac
resynchronization therapy enrolled in the Sensitivity of the InSync Sentry OptiVol feature for the prediction of Heart
Failure (SENSE-HF)1 study and the Diagnostic Outcome Trial in Heart Failure (DOT-HF) were pooled. The devices
continuously measured and stored total daily active time (single-axis accelerometer). Early PA (average daily activity over
the earliest 30-day study period) was studied as a predictor of time to death or HF-related hospital admission (primary
end point). Data from 781 patients were analyzed (65±10 years; 85% men; left ventricular ejection fraction, 26±7%).
Older age, shorter height, ischemic cause, peripheral artery disease, atrial fibrillation, diabetes mellitus, rales, peripheral
edema, higher New York Heart Association class, lower diastolic blood pressure, and no angiotensin II receptor ­blocker/
angiotensin-converting enzyme inhibitor use were associated with reduced early PA. The primary end point occurred in
135 patients (15±7 months of follow-up). In multivariable analysis including baseline variables, early PA predicted death
or HF hospitalization, with a 4% reduction in risk for each 10 minutes per day additional activity (hazard ratio [HR], 0.96;
confidence interval [CI], 0.94–0.98; P=0.0002 compared with a model with the same baseline variables but without PA).
PA also predicted death (HR, 0.93; CI, 0.90–0.96; P<0.0001) and HF hospitalization (HR, 0.97; CI, 0.95–0.99; P=0.011).
Conclusions—Early PA, averaged over a 30-day window early after defibrillator implantation or cardiac resynchronization
therapy in patients with chronic HF, predicted death or HF hospitalization, as well as mortality and HF hospitalization
separately, accounting for baseline HF severity.
Clinical Trial Registration Information—URL: http://www.clinicaltrials.gov. Unique identifiers: NCT00400985, NCT00480077. (Circ Heart Fail. 2014;7:279-287.)
Key Words: heart failure ◼ mortality ◼ physical activity
D
derived during cardiorespiratory exercise testing, in particular the minute ventilation/CO2 production relationship (the
VE/VCO2 slope) in conjunction with peak O2 uptake, allows
prognostic stratification in patients with CHF.8 However,
repetitive cardiopulmonary exercise testing, which may be
required for a dynamic condition such as CHF, is demanding
in clinical practice.9
Daily physical activity (PA) has been shown to predict outcome in various chronic diseases.10–12 However, data for the
CHF population are scarce and are often based on qualitative
subjective patient and physician reports or questionnaires or
espite significant improvements in the treatment of chronic
heart failure (CHF), mortality remains as high as 20% to 30%
after 3 years.1–3 The costs for heart failure (HF) care approach 2%
of the total health care expenditure in Western countries, primarily
due to hospitalization for decompensation.4 As a result, the past
decade has witnessed a surge in finding reliable markers to identify patients at risk of early death and impending decompensation.
Clinical Perspective on p 287
Physical fitness predicts cardiovascular mortality in the
population at large.5–7 The collection of multiple parameters
Received May 6, 2013; accepted January 30, 2014.
From the Department of Cardiology, Antwerp University Hospital, Edegem, Belgium (V.V.C.); University of Antwerp, Antwerp, Belgium (V.V.C.);
Department of Research and Education, CIRO+, Center of Expertise for Chronic Organ Failure, Horn, the Netherlands (M.A.S.); REVAL–Rehabilitation
Research Center, BIOMED–Biomedical Research Institute, Faculty of Medicine and Life Sciences, Hasselt University, Diepenbeek, Belgium (M.A.S.);
Karolinska Institutet, Department of Cardiology, Karolinska University Hospital, Stockholm, Sweden (F.B.); National Heart and Lung Institute, Imperial
College, London, United Kingdom (M.R.C.); Maria Cecilia Hospital, GVM Care and Research, E.S. Health Science Foundation, Cotignola, Italy (L.T.);
Department of Cardiology, University Hospital Mannheim, Mannheim, Germany (M.B.); Medtronic Inc, Minneapolis, MN (M.R.S.H.); Medtronic Bakken
Research Center, Maastricht, the Netherlands (S.J., B.G.); and Department of Cardiology, University Medical Center Groningen, University of Groningen,
Groningen, the Netherlands (D.J.v.V).
*Dr Conraads died on December 12, 2013. She is greatly missed.
The Data Supplement is available at http://circheartfailure.ahajournals.org/lookup/suppl/doi:10.1161/CIRCHEARTFAILURE.113.000883/-/DC1.
Correspondence to Dirk J. van Veldhuisen, MD, Department of Cardiology, University Medical Center Groningen, University of Groningen, PO
Box 30.001, 9700RB Groningen, the Netherlands. E-mail [email protected]
© 2014 American Heart Association, Inc.
Circ Heart Fail is available at http://circheartfailure.ahajournals.org
279
DOI: 10.1161/CIRCHEARTFAILURE.113.000883
280 Circ Heart Fail March 2014
short-term measurement with pedometers and accelerometers.13–15
Continuous recording of PA, through accelerometers incorporated in implanted devices, may more accurately address the
question whether PA merely reflects disease severity16 or whether
it is an independent prognostic marker.
The Sensitivity of the InSync Sentry OptiVol feature for
the prediction of Heart Failure (SENSE-HF)1 study and the
Diagnostic Outcome Trial in Heart Failure (DOT-HF)17 are
recent prospective trials conducted to evaluate the utility of
intrathoracic impedance measurement, incorporated as a diagnostic feature in implantable cardioverter defibrillator without
(ICD) or with cardiac resynchronization therapy (CRT-D).
These devices also measured and stored total daily active time.
We analyzed the pooled patient data from SENSE-HF and
DOT-HF to determine whether and to what extent daily PA as
measured by implanted devices is related to patient outcome.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Methods
Patient Population
Detailed descriptions of the design and results of SENSE-HF and
DOT-HF have been published previously.18,19 Briefly, SENSE-HF18
enrolled 501 patients <34 days after implantation of an ICD or CRT-D
with the OptiVol feature (Medtronic Inc, Minneapolis, MN), which
measures intrathoracic impedance. After a waiting period of 34 days
to ensure correct measurement of intrathoracic impedance, patients
were followed up for 6 months to determine the sensitivity and the
positive predictive value of OptiVol intrathoracic fluid monitoring for
the detection of HF-related hospitalizations (phase I, blinded). Phase
I was followed by maximally 18 months with full access to the device
data. PA was recorded and stored during the entire study period, starting at day 35 postimplant.
DOT-HF17,19 tested whether physician access to device-based diagnostic information, including audible patient alerts for decreasing intrathoracic impedance, would improve outcome in patients with HF.
In total, 335 patients were randomized to an access versus a control
arm between 35 days and 6 months after implantation with an ICD or
CRT-D. Patients were followed up for 15 months, during which PA
was continuously measured and stored.
The Table in the Data Supplement describes end points and the
major inclusion and exclusion criteria for both studies. Both studies
were approved by the local ethics committees, and written informed
consent was obtained from all patients. Given the similarities in patient population and treatment, and because all devices in SENSE-HF
and DOT-HF recorded and stored PA during daily life on a continuous
basis, it was decided to pool and analyze patient data from both trials.
The authors had full access to the study data.
Study End Points
The primary end point of the present analysis was time to death or HFrelated hospital admission. Time to death and time to first ­HF-related
hospital admission were analyzed as secondary end points. HF hospitalizations were adjudicated by study-specific, independent, blinded
Adverse Event Advisory Committees.18,19
Device Measurement of Daily PA
The activity measurement in the devices is designed to capture normal daily activities in patients with HF, including walking at a slow
pace. A single-axis accelerometer is used to capture patient motion as
an electric signal. The number of minutes a patient is active per day is
counted, where a minute is considered active if a threshold is reached
that incorporates both the number and magnitude of deflections in the
accelerometer signal. The devices store the number of active minutes
for the most recent 425 days (Figure 1), and this information was
retrieved from device memory for this analysis. The sensor provides
a quantification of patient activity that, due to the specifics of the
algorithm, may numerically differ from measurements obtained from
other devices such as pedometers or external accelerometers.
Analyses Methods
We defined early PA as the average daily activity over the earliest
­30-day period in the study (the activity window). For patients in
SENSE-HF, this window started 35 days after implant and continued
until day 64. For patients in DOT-HF, the activity window started
after randomization, which was between day 35 and day 183 postimplant. We averaged 30 daily values to account for within-patient
variability. Patients were excluded from analysis if they died or had a
HF hospitalization during or before the 30-day activity window, or if
there were <7 days with valid activity data in the window.
Because lower activity was expected to be associated with a generally worse condition of patients, we investigated the relation of early
activity with baseline characteristics. The relation of early PA and outcome as well as the incremental value of activity as a risk marker when
added to known risk factors was assessed in line with the American
Heart Association scientific statement on evaluation of novel markers
of cardiovascular risk.20 As measures of known risk, we determined
the best predictive model based on baseline patient characteristics, and
we used The Candesartan in Heart failure Assessment of Reduction in
Mortality and morbidity (CHARM) programme risk score, a validated
score for risk of all-cause death in patients with HF.21 For both approaches, time-to-event models including and excluding early activity
are compared to determine incremental predictive value.
Statistical Analyses
Results are reported with mean and standard deviation for continuous, and counts and percentages for categorical variables. The relation between baseline characteristics and early PA was analyzed with
t test for dichotomous and linear regression for ordinal and continuous characteristics. Reported effect sizes (Table 1) are observed differences for dichotomous characteristics, and regression coefficients
otherwise. Multivariable linear regression with backward variable
elimination was used to assess joint association with early activity.
For calculation of the CHARM risk score, coefficients were taken
from the study by Pocock et al.21 Multiple imputation was used to account for missing values, using the multivariate normal specification.
For categorical variables, the imputed value was rounded to the nearest coded category value. Ten completed data sets were generated.
The CHARM risk score was calculated for each completion separately. Finally, Rubin rule was used to determine a single risk score
value per patient.
Cox proportional hazards regression was used to analyze the relation
between events and early PA, which was included as a continuous variable. Patients were considered at risk starting at the end of the activity
window (time zero). Hazard ratio (HR) with associated 95% confidence
interval (CI) is reported for 10 minutes per day incremental activity. The
incidence of end points is illustrated with Kaplan–Meier graphs with
patients divided into 2 or 3 equal-sized groups using medians or tertiles
as cut points. Multivariable proportional hazards models using model
reduction with backward selection are used to define predictive models
based on baseline characteristics. To assess the incremental predictive
value of early activity, multivariable models including and excluding
early activity are compared using likelihood ratio tests. The predictive
ability of models is quantified using the Harrell c-index, which extends
the area under the receiver operating characteristic curve to the context
of time-to-event data with censoring.22
All analyses were done using SAS (version 9.3; SAS Institute Inc,
Cary, NC) or R (version 3.0.1; the R project for statistical computing;
www.r-project.org), and P values <0.05 were considered significant.
No correction for multiple testing was applied.
Results
Patients
SENSE-HF and DOT-HF together enrolled 836 patients
between 2005 and 2009 in 100 centers across Europe (91%
Conraads et al Physical Activity and Outcome in Heart Failure 281
A
B
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Figure 1. Illustration of device-based activity
measurement. A, A stable and relatively active
patient. B, A patient with 2 heart failure hospitalizations. C, A patient who died.
C
of patients), the Middle East, and Asia. Fifty-five patients
were removed from analysis because they experienced
events before the end of the activity window (35 patients;
11 deaths and 28 HF hospital admissions), early study exit
(5 patients), or because device data were not available for
analysis (15 patients).
Demographic and clinical characteristics are presented in
Table 1. At enrollment, the majority of patients was in New York
Heart Association (NYHA) class II/III, with severely impaired
left ventricular ejection fraction (26±7%) and implanted with a
CRT-D (80%).
Correlates of Early PA
Early PA was 199±106 minutes per day and was comparable
between studies (SENSE-HF: 199 minutes per day; DOT-HF:
198 minutes per day; P=0.90; see Figure in the Data Supplement). As a result of the design of both studies, the activity
window started 35 days after implantation in all SENSE-HF
patients, and on average 86 days after implantation in DOT-HF
patients. Time between implant and start of the activity window did not affect early activity (linear regression; P=0.83) or
its prognostic value for DOT-HF patients.
Table 1 illustrates the relation between early PA and baseline characteristics. Multivariable analysis identified older
age, shorter height, ischemic cause, peripheral artery disease,
atrial fibrillation, diabetes mellitus, rales, peripheral edema,
higher NYHA class, lower diastolic blood pressure, and
absence of angiotensin receptor blocker (­ARB)/angiotensinconverting enzyme (ACE) inhibitor therapy as jointly associated with reduced early PA.
Patient Outcome and Early PA
During follow-up (15±7 months), 65 deaths and 141 HF hospitalizations were recorded in 100 patients. The primary end
282 Circ Heart Fail March 2014
Table 1. Baseline Characteristics and Relation to Early Physical Activity
Baseline Characteristics
Age, y
Analysis Cohort
(n=781)
Effect Size
(min/d)
Univariable
P Value
P Value From
Multivariable Model
P Value From Reduced
Multivariable Model
<0.0001
<0.0001
0.012
0.003
0.002
65±10
−3.19
<0.0001
Men
661 (85%)
16.01
0.13
Body height, cm
171±8
1.40
0.002
0.29
0.22
−0.32
0.70
6.99
0.56
0.61
Weight, kg
80±16
BMI, kg/m2
27±5
Underweight (BMI ≤ 21 kg/m2)
47 (6%)
HF hospitalization in last 6 mo*
406 (87%)
7.39
HF hospitalization in last 12 mo
774 (99%)
40.36
0.32
Ischemic cause
445 (57%)
−37.61
<0.0001
0.032
Previous myocardial infarction
364 (47%)
−30.20
<0.0001
0.79
Mitral insufficiency
360 (46%)
3.60
0.64
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Peripheral artery disease
45 (6%)
−41.23
0.011
Left bundle branch block
418 (54%)
−1.95
0.80
History of atrial fibrillation
292 (37%)
−31.26
Diabetes mellitus
232 (30%)
27 (3%)
Diabetes mellitus (type I)
Dyspnea
Dyspnea at rest
Rales
0.055
0.020
<0.0001
0.025
0.009
−35.86
<0.0001
0.09
0.015
−63.78
<0.001
0.18
552 (71%)
−33.08
<0.0001
0.08
8 (1%)
−45.11
0.24
48 (6%)
−54.21
<0.001
0.033
0.043
Peripheral edema
119 (15%)
−43.19
<0.0001
0.008
0.007
Pulmonary edema
64 (8%)
−31.87
0.021
0.12
−32.96
<0.0001
0.013
<0.001
0.53
0.033
0.003
NYHA classification
Class I
38 (5%)
Class II
381 (49%)
Class III
346 (44%)
Class IV
15 (2%)
Heart rate, beats/min
72±12
−0.21
Systolic blood pressure, mm Hg
117±18
0.39
0.064
Diastolic blood pressure, mm Hg
71±11
1.42
<0.0001
LVEF, %
26±7
0.23
0.66
Intrinsic QRS duration, ms
146±36
−0.12
0.27
MLHF score, points
36±23
−0.52
0.002
Current smoker
48 (6%)
3.05
0.71
CRT-D
621 (80%)
10.20
0.32
β-Blocker
689 (88%)
16.99
0.15
ACE inhibitor or ARB
682 (87%)
24.90
0.029
0.005
Total diuretics dose, mg/kg
0.88±1.02
−10.98
0.003
0.09
0.26
0.027
Characteristics were taken at baseline, which could be up to 6 mo after device implantation. Effect size is the average difference in early physical
activity for patients with characteristic versus without (dichotomous) or the average change in activity per unit increase (ordinal, continuous). ACE
indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BMI, body mass index; CRT-D, defibrillator with cardiac resynchronization
therapy; HF, heart failure; LVEF, left ventricular ejection fraction; MLHF, Minnesota Living with Heart Failure questionnaire; and NYHA, New York Heart
Association.
*Data not collected in Diagnostic Outcome Trial in Heart Failure (DOT-HF) so that denominator is 469, the number of Sensitivity of the InSync Sentry
OptiVol feature for the prediction of Heart Failure (SENSE-HF) study patients in the analysis cohort.
point of death or HF hospitalization occurred in 135 patients.
Early PA was significantly associated with the primary end
point, with 5% relative reduction of risk for each 10 minutes per
day additional activity (HR, 0.95; CI, 0.94–0.97; P<0.0001).
There were also significant associations with death (HR, 0.92;
CI, 0.89–0.95; P<0.0001) and HF hospitalization (HR, 0.97;
CI, 0.95–0.99; P=0.0011). Figure 2 shows event incidence
with patients stratified into 3 equal-sized groups according
to early PA. The cut points are 235 minutes per day between
high and medium activity and 146 minutes per day between
medium and low activity. The incidence of primary end point
at 18 months is 12.5% for high activity, 17.5% for medium
Conraads et al Physical Activity and Outcome in Heart Failure 283
A
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
B
C
Figure 2. Kaplan–Meier estimates for the incidence of clinical
events. A, All-cause death or heart failure (HF) hospitalization
(primary end point). B, All-cause death. C, HF hospitalization.
Patients are stratified in 3 equal-sized groups based on their early
activity with cut points 145.6 and 235.0 min/d (tertiles). Events are
counted after the activity window used to determine early activity.
activity, and 30.0% for low activity. Mortality is 2.5%, 9.9%,
and 18.2%, respectively. The incidence of HF hospitalization
is 11.7%, 10.2%, and 22.0%.
Outcome and Early PA Accounting for Disease
Severity
To assess the incremental value of early PA as a risk marker for
clinical events when added to other risk factors, 2 approaches
were used.
First, baseline characteristics were tested for their relation
with patient outcome. Fourteen characteristics were significantly related to the primary end point (age, height, weight,
underweight, left bundle branch block, diabetes mellitus, dyspnea, systolic blood pressure, diastolic blood pressure, QRS
duration, QRS ≥120 ms, QRS ≥150 ms, ARB/ACE inhibitor,
daily diuretic dosage). Additionally, peripheral edema was
predictive of mortality. These 15 variables were included in
multivariable models, from which nonsignificant parameters
were iteratively removed. In the final multivariable model for
death or HF hospitalization, age (P=0.0011), systolic blood
pressure (P=0.0001), daily diuretic dosage (P=0.0004), and
absence of ARB/ACE inhibitor (P=0.04) were retained as
joint predictors of outcome. The c-index for this model was
0.66 (CI, 0.61–0.71). The addition of early PA to the model
significantly improved the predictive ability (likelihood ratio
P=0.0002). The model with activity has a c-index of 0.68 (CI,
0.63–0.73). Accounting for the effect of other factors, there
was a 4% reduction in risk for each 10 minutes per day additional activity (HR, 0.96; CI, 0.94–0.98). These results are
summarized in Table 2, which also includes results for mortality and HF hospitalization separately.
Next, we determined the predictive value of the validated
CHARM risk score on the current patient cohort. On its own,
the score was significantly associated with the incidence of
death or HF hospitalization (P<0.0001; c-index 0.61; CI,
0.56–0.66; Table 2). Figure 3 shows event incidence with
patients stratified into low-, medium-, and high-risk groups.
The CHARM score and early PA were significantly correlated
(r=−0.33; P<0.0001).
The addition of early PA to the model significantly
improved the predictive ability (likelihood ratio P<0.0001),
establishing lower activity as a predictor of increased event
rates when accounting for baseline risk. The c-index increased
to 0.65 (CI, 0.60–0.70). With CHARM score in the model,
there was a 4% reduction in risk of death or HF hospitalization for each 10 minutes per day additional activity (HR, 0.96;
CI, 0.94–0.98). Figure 4 illustrates these findings by showing
the incidence of death or HF hospitalization for patients stratified by the CHARM score and early PA. For both parameters,
patients are classified as above or below the median value, creating 4 groups. Table 2 also shows that for death and HF hospitalization separately, early activity significantly improved
the predictive value when added to the CHARM score in
­time-to-event models.
Discussion
The main finding of the present study is that early PA, measured by the accelerometer sensor of implanted ICD and CRT-D
devices, strongly predicts outcome in patients with CHF. Early
PA was defined as the average activity over a 30-day window,
measured either starting 35 days after implant for S
­ ENSE-HF
patients or after randomization, and between 35 days and 6
months after implant for DOT-HF patients. Despite the expected
284 Circ Heart Fail March 2014
Table 2. Incremental Value of Early Physical Activity as Risk Marker
Risk Model Based on Study Data
Risk Model With CHARM Risk Score
Without Activity
With Activity
Without Activity
With Activity
Age, SBP, diuretic dose,
ARB/ACE absence
Age, SBP, diuretic dose,
ARB/ACE absence, early activity
CHARM
CHARM early activity
Death and HF hospitalization
Variables in model
Hazard ratio, activity increase
by 10 min/d
Harrell c-index
0.96 (0.94–0.98)
0.6585 (0.6064–0.7106)
P value comparing models
0.6803 (0.6282–0.7324)
0.96 (0.94–0.98)
0.6114 (0.5601–0.6627)
0.0002
0.6490 (0.5977–0.7003)
<0.0001
Death
Variables in model
Age, SBP, diuretic dose
Age, SBP, diuretic dose, early
activity
Hazard ratio, activity increase
by 10 min/d
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Harrell c-index
CHARM
CHARM early activity
0.93 (0.90–0.96)
0.7123 (0.6350–0.7897)
P value comparing models
0.7535 (0.6762–0.8309)
0.93 (0.90–0.96)
0.6108 (0.5348–0.6869)
<0.0001
0.7172 (0.6411–0.7932)
<0.0001
HF hospitalization
Variables in model
Weight, SBP, diabetes mellitus
Weight, SBP, diabetes mellitus,
early activity
Hazard ratio, activity increase
by 10 min/d
Harrell c-index
P value comparing models
CHARM
CHARM early activity
0.97 (0.95–0.99)
0.6246 (0.5643–0.6849)
0.6450 (0.5847–0.7053)
0.011
0.97 (0.95–1.00)
0.6031 (0.5441–0.6621)
0.6226 (0.5636–0.6816)
0.013
Cox proportional hazards models were used to assess the relation of baseline risk markers and early physical activity with outcome. The first 2 data columns report
on predictive models derived from the study data, comparing the model with only baseline risk factors to the model with these risk factors and early activity added.
Reported P values are from likelihood ratio tests. Similarly, the last 2 columns compare a model with only the CHARM risk score to a model with CHARM score and early
physical activity. ACE indicates angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CHARM, The Candesartan in Heart failure Assessment of Reduction
in Mortality and morbidity (CHARM) programme; HF, heart failure; and SBP, systolic blood pressure.
significant relation between early PA and HF severity, low early
PA was a significant predictor of the combined primary end
point of death and HF hospitalization, as well as of both components separately, when correcting for HF severity.
Figure 3. Kaplan–Meier estimates for the incidence of death
or heart failure (HF) hospitalization. Patients are stratified in 3
­equal-sized groups based on their baseline CHARM (The Candesartan in Heart failure Assessment of Reduction in Mortality and
morbidity programme) risk score value with cut points 2.0 and
8.0 (tertiles).
Population-based studies show that PA and a high level of
cardiorespiratory fitness predict longevity.5–7,23,24 Literature
on the prognostic impact of daily PA in patients with HF is
scarce and almost exclusively based on subjective evaluation
of patients’ reports23 or data gathered by intermittent application of pedometers12 and external accelerometers.13,14
Continuous measurement of daily PA with accelerometers
incorporated in implanted devices can exclude nonadherence
and avoid bias of deliberate changes in patients’ habits at the
time of planned assessments. Surprisingly, few studies have
investigated the clinical utility of continuous measurements
and have mainly concentrated on short-term prediction of acute
decompensated HF.25,26 The Program to Access and Review
Trending Information and Evaluate Correlation to Symptoms
in Patients with Heart Failure (PARTNERS-HF) demonstrated
that a combination of device-measured parameters, including
activity, predicted a 5.5-fold higher HF hospitalization risk
in CRT-D patients, but results for activity are not reported
separately.27 Only Shoemaker15 retrospectively related PA
recordings in 102 ICD or CRT-D patients to long-term outcome.
They found a moderate correlation between mean daily activity, calculated for a 2-week period, and estimated short- and
­long-term outcome, based on the Seattle Heart Failure Score.28
The results of the present study support the notion that daily
PA reflects disease severity, measured by means of a validated
HF risk score. However, the fact that early PA predicted outcome, taking well-recognized traditional risk factors into
Conraads et al Physical Activity and Outcome in Heart Failure 285
Figure 4. Kaplan–Meier estimates for the incidence of death or heart failure (HF) hospitalization.
Patients are stratified in 4 groups based on the
combination of early activity and CHARM (The
Candesartan in Heart failure Assessment of Reduction in Mortality and morbidity programme) risk
score. For both parameters, the median value
(188 min/d and 4.8 points, respectively) is used to
group patients in the low or high category.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
account, is novel and may be of use in the management of
patients with CHF.
As shown by a subanalysis of the Heart Failure: A
Controlled Trial Investigating Outcomes of Exercise Training
(HF-ACTION), improved patient outcome, as well as increased
exercise capacity, depends on exercise volume.29 Our data support a similar dose–response effect with regard to daily PA.
Patients in the lowest tertile for early activity (<146 minutes
per day) had 2.3 times higher risk for the combined end point
(HR, 2.3), compared with those in the highest tertile (>235 minutes per day), and even a 5-fold risk for mortality (HR, 5.0).
The advice provided to patients by SENSE-HF16 and DOT-HF17
physicians based on stored PA data was not part of the study
protocols. Therefore, a prospective study is necessary to confirm
the current findings and to establish clinical recommendations.
In this study, several patient and disease characteristics, as
well as comorbities, were associated with early PA. A significant effect of age and NYHA functional class on 6-minute
walking distances30 and cardiorespiratory fitness,31 as well as
a detrimental effect of raised jugular venous pressure on PA,
is described.30 The decline of cardiorespiratory fitness with
age32 likely increases exercise-induced symptoms of fatigue
and dyspnea in patients with HF. These effects may hamper
the motivation and ability of patients with HF to engage in
daily PA. A similar argument may explain lower levels of PA
in patients presenting with higher NYHA functional class,
rales, and peripheral edema.
After multivariable analysis, height was retained as a predictor of PA. Without providing a physiological justification,
Witham et al30 described a significant effect of height on
accelerometry-based quantification of PA in patients with HF
in univariate analysis. One plausible explanation may involve
the larger excursion of the body during movements in taller
people, which may be sensed by the accelerometer.
The fact that atrial fibrillation significantly determined low
levels of PA may be related to lower cardiac output during
exertion due to poorly controlled ventricular rate and atrioventricular dyssynchrony. The functional substudy of the Atrial
Fibrillation Follow-Up Investigation of Rhythm Management
(AFFIRM) trial33 demonstrated a significantly larger increase
in functional status, assessed with the ­6-minute walk test, in
HF patients with atrial fibrillation allocated to the rhythm versus the rate control group. Similar findings were reported in
the Pharmacological Intervention in Atrial Fibrillation (PIAF)
trial through 2 years of follow-up.34 Recently, Izawa et al35
described a significantly reduced step count and estimated
daily PA energy expenditure obtained with uniaxial accelerometers in diabetic versus nondiabetic HF patients. Skeletal
muscle wasting as well as bioenergetic changes at the level
of peripheral skeletal musculature are plausible mechanisms.36
Exercise-induced pain and significant lower limb ischemia
change gait pattern and lead to lower walking performance
as well as poor daily PA.37 HF patients with peripheral artery
disease have lower exercise capacity and lower response to
exercise training intervention.38 Patients with ischemic cause,
compared with those with nonischemic cause, demonstrate
lower peak oxygen uptake and a steeper VE/VCO2 slope.31
Impaired PA may result from chronotropic incompetence,
exercise-induced myocardial ischemia, and mitral regurgitation.39 Treatment with an ACE inhibitor improves exercise
duration40 as well as peak aerobic capacity.41 There is a direct
benefit on cardiac performance through the inhibition of the
renin–angiotensin–aldosterone system. Additionally, facilitated alveolar–capillary gas transfer,42 as well as restoration of
peripheral endothelial function, have been described.43
Limitations
For this post hoc analysis, we pooled data from 2 studies with
different design. As a result, the 30-day activity window was
collected for a period between 35 and 183 days. Nevertheless, prognostic information derived from this pooled data set
remains strong and seems not affected by this approach.
The devices used the same type of accelerometer, and it
is not known whether the present results can be extrapolated
to other activity sensors. Literature on the validation of the
device’s activity sensor is limited. Gulati et al44 found that the
6-minute walk test and daily activity measured within the preceding or subsequent 24 hours were correlated (r=0.42 and
286 Circ Heart Fail March 2014
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
0.49, respectively). A similar relation was found in a subset
of 83 patients enrolled in SENSE-HF in whom the 6-minute
walk test was conducted (r=0.43; P<0.0001). Braunschweig
et al45 calculated weekly means of PA in a group of 56 patients
after the novo device implant using the same accelerometer.
There was a steady increase in weekly PA, which was clearly
related to NYHA functional class with absolute values after 4
weeks that were well within the range of current observations
(283±97, 206±132, and 186±53 minutes per day for NYHA II,
III, and IV, respectively).
The CHARM risk score was developed from a large number of nondevice patients across a broad spectrum of HF.21
Nevertheless, the risk score strongly predicted outcome for
the currently analyzed cohort of patients, characterized by
broad QRS complexes, impaired systolic left ventricular function, and implanted with either an ICD or a CRT-D.
The studies analyzed here did not enroll patients in the
United States. Therefore, clinical recommendations for daily
PA derived from the present analyses may not be applicable to
the American HF population. Finally, >30 tests were performed
to assess the relation between activity and baseline characteristics so that some of the significant findings may be spurious.
Conclusions
Early PA, averaged over a 30-day window early after ICD or
CRT-D implant in patients with CHF, predicted death or HF
hospitalization, as well as mortality and HF hospitalization
separately, also corrected for disease severity. These findings
underscore the validity of spontaneous PA as a clinical status
marker. Whether regular follow-up of PA, provided by accelerometers in implanted devices, will aid physicians in modifying patients’ lifestyle and whether this will improve outcome
may need to be prospectively studied.
Sources of Funding
This study was funded by the Medtronic Bakken Research Center.
M.R.C. receives salary support from the National Institute for Health
Research Cardiovascular Biomedical Research Unit at the Royal
Brompton Hospital. D.J.v.V. is an established investigator of the
Netherlands Heart Foundation (grant No. D97.017). V.M.C. was supported by Fonds Wetenschappelijk Onderzoek-Flanders as a clinical
postdoctoral fellow.
Disclosures
Drs Borggrefe, Braunschweig, Conraads, Cowie, van Veldhuisen, and
Tavazzi received consultancy fees or research grants from Medtronic
and from other device companies. Drs Conraads, Cowie, and Tavazzi
served on the Steering Committee of the SENSE-HF study. Drs
Borggrefe, Braunschweig, Conraads, Cowie, and van Veldhuisen
served on the Steering Committee for DOT-HF. Drs Hill, Jacobs, and
Gerritse are employees of Medtronic. Dr Gerritse reports having equity in Medtronic. Dr Spruit has no conflicts to report.
References
1. Böhm M, Swedberg K, Komajda M, Borer JS, Ford I, Dubost-Brama
A, Lerebours G, Tavazzi L; SHIFT Investigators. Heart rate as a risk
factor in chronic heart failure (SHIFT): the association between heart
rate and outcomes in a randomised placebo-controlled trial. Lancet.
2010;376:886–894.
2. Meta-analysis Global Group in Chronic Heart F. The survival of patients
with heart failure with preserved or reduced left ventricular ejection fraction:
an individual patient data meta-analysis. Eur Heart J. 2012;33:1750–1757.
3. Braunschweig F, Cowie MR, Auricchio A. What are the costs of heart
failure? Europace. 2011;13(suppl 2):ii13–ii17.
4. Paffenbarger RS Jr, Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB. The
association of changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J Med. 1993;328:538–545.
5. Blair SN, Kohl HW III, Barlow CE, Paffenbarger RS Jr, Gibbons LW,
Macera CA. Changes in physical fitness and all-cause mortality. A prospective study of healthy and unhealthy men. JAMA. 1995;273:1093–1098.
6. Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE.
Exercise capacity and mortality among men referred for exercise testing.
N Engl J Med. 2002;346:793–801.
7. Guazzi M, Adams V, Conraads V, Halle M, Mezzani A, Vanhees L, Arena
R, Fletcher GF, Forman DE, Kitzman DW, Lavie CJ, Myers J; European
Association for Cardiovascular Prevention and Rehabilitation; American
Heart Association. EACPR/AHA Scientific Statement. Clinical recommendations for cardiopulmonary exercise testing data assessment in specific patient populations. Circulation. 2012;126:2261–2274.
8. Arena R, Myers J, Abella J, Pinkstaff S, Brubaker P, Kitzman D, Peberdy
MA, Bensimhon D, Chase P, Forman DE, Guazzi M. Defining the optimal prognostic window for cardiopulmonary exercise testing in patients
with heart failure. Circ Heart Fail. 2010;3:405–411.
9. Garcia-Aymerich J, Lange P, Benet M, Schnohr P, Antó JM. Regular
physical activity reduces hospital admission and mortality in chronic
obstructive pulmonary disease: a population based cohort study. Thorax.
2006;61:772–778.
10. Johansen KL, Kaysen GA, Dalrymple LS, Grimes BA, Glidden DV,
Anand S, Chertow GM. Association of physical activity with survival
among ambulatory patients on dialysis: the Comprehensive Dialysis
Study. Clin J Am Soc Nephrol. 2013;8:248–253.
11. Waschki B, Kirsten A, Holz O, Müller KC, Meyer T, Watz H, Magnussen
H. Physical activity is the strongest predictor of all-cause mortality in patients with COPD: a prospective cohort study. Chest. 2011;140:331–342.
12. Walsh JT, Charlesworth A, Andrews R, Hawkins M, Cowley AJ. Relation
of daily activity levels in patients with chronic heart failure to long-term
prognosis. Am J Cardiol. 1997;79:1364–1369.
13. Jehn M, Schmidt-Trucksäss A, Schuster T, Weis M, Hanssen H, Halle
M, Koehler F. Daily walking performance as an independent predictor
of advanced heart failure: prediction of exercise capacity in chronic heart
failure. Am Heart J. 2009;157:292–298.
14. Howell J, Strong BM, Weisenberg J, Kakade A, Gao Q, Cuddihy P,
Delisle S, Kachnowski S, Maurer MS. Maximum daily 6 minutes of
activity: an index of functional capacity derived from actigraphy and
its application to older adults with heart failure. J Am Geriatr Soc.
2010;58:931–936.
15. Shoemaker MJ, Curtis A, Vangsnes E, Dickinson M, Paul R. Analysis of
daily activity data from implanted cardiac defibrillators: the minimum
clinically important difference and relationship to mortality/life expectancy. World J Cardiovasc Dis. 2012;2:129–135.
16. Conraads VM, Tavazzi L, Santini M, Oliva F, Gerritse B, Yu CM, Cowie
MR. Sensitivity and positive predictive value of implantable intrathoracic impedance monitoring as a predictor of heart failure hospitalizations:
the SENSE-HF trial. Eur Heart J. 2011;32:2266–2273.
17. van Veldhuisen DJ, Braunschweig F, Conraads V, Ford I, Cowie MR,
Jondeau G, Kautzner J, Aguilera RM, Lunati M, Yu CM, Gerritse B,
Borggrefe M; DOT-HF Investigators. Intrathoracic impedance monitoring, audible patient alerts, and outcome in patients with heart failure.
Circulation. 2011;124:1719–1726.
18. Cowie MR, Conraads V, Tavazzi L, Yu CM; SENSE-HF Investigators.
Rationale and design of a prospective trial to assess the sensitivity and
positive predictive value of implantable intrathoracic impedance monitoring in the prediction of heart failure hospitalizations: the SENSE-HF
study. J Card Fail. 2009;15:394–400.
19. Braunschweig F, Ford I, Conraads V, Cowie MR, Jondeau G, Kautzner
J, Lunati M, Munoz Aguilera R, Man Yu C, Marijianowski M, Borggrefe
M, van Veldhuisen DJ; DOT-HF Steering Committee and Investigators.
Can monitoring of intrathoracic impedance reduce morbidity and mortality in patients with chronic heart failure? Rationale and design of the
Diagnostic Outcome Trial in Heart Failure (DOT-HF). Eur J Heart Fail.
2008;10:907–916.
20. Hlatky MA, Greenland P, Arnett DK, Ballantyne CM, Criqui MH, Elkind
MS, Go AS, Harrell FE Jr, Hong Y, Howard BV, Howard VJ, Hsue PY,
Kramer CM, McConnell JP, Normand SL, O’Donnell CJ, Smith SC Jr,
Wilson PW; American Heart Association Expert Panel on Subclinical
Atherosclerotic Diseases and Emerging Risk Factors and the Stroke
Council. Criteria for evaluation of novel markers of cardiovascular risk:
Conraads et al Physical Activity and Outcome in Heart Failure 287
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
a scientific statement from the American Heart Association. Circulation.
2009;119:2408–2416.
21. Pocock SJ, Wang D, Pfeffer MA, Yusuf S, McMurray JJ, Swedberg KB,
Ostergren J, Michelson EL, Pieper KS, Granger CB. Predictors of mortality and morbidity in patients with chronic heart failure. Eur Heart J.
2006;27:65–75.
22.Harrell FE Jr. Regression Modeling Strategies. Springer: New York;
2001.
23. Aspenes ST, Nilsen TI, Skaug EA, Bertheussen GF, Ellingsen Ø, Vatten
L, Wisløff U. Peak oxygen uptake and cardiovascular risk factors in 4631
healthy women and men. Med Sci Sports Exerc. 2011;43:1465–1473.
24. Myers J, Kaykha A, George S, Abella J, Zaheer N, Lear S, Yamazaki
T, Froelicher V. Fitness versus physical activity patterns in predicting
mortality in men. Am J Med. 2004;117:912–918.
25. Adamson PB, Smith AL, Abraham WT, Kleckner KJ, Stadler RW, Shih
A, Rhodes MM; InSync III Model 8042 and Attain OTW Lead Model
4193 Clinical Trial Investigators. Continuous autonomic assessment in
patients with symptomatic heart failure: prognostic value of heart rate
variability measured by an implanted cardiac resynchronization device.
Circulation. 2004;110:2389–2394.
26. Small RS, Wickemeyer W, Germany R, Hoppe B, Andrulli J, Brady PA,
Labeau M, Koehler J, Sarkar S, Hettrick DA, Tang WH. Changes in intrathoracic impedance are associated with subsequent risk of hospitalizations for acute decompensated heart failure: clinical utility of implanted
device monitoring without a patient alert. J Card Fail. 2009;15:475–481.
27. Whellan DJ, Ousdigian KT, Al-Khatib SM, Pu W, Sarkar S, Porter CB,
Pavri BB, O’Connor CM. Combined heart failure device diagnostics
identify patients at higher risk of subsequent heart failure hospitalizations: results from partners hf (program to access and review trending
information and evaluate correlation to symptoms in patients with heart
failure) study. J Am Coll Cardiol. 2010;55:1803–1810.
28. Levy WC, Mozaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp
AB, Anand I, Maggioni A, Burton P, Sullivan MD, Pitt B, Poole-Wilson
PA, Mann DL, Packer M. The Seattle Heart Failure Model: prediction of
survival in heart failure. Circulation. 2006;113:1424–1433.
29. Keteyian SJ, Leifer ES, Houston-Miller N, Kraus WE, Brawner CA,
O’Connor CM, Whellan DJ, Cooper LS, Fleg JL, Kitzman DW,
Cohen-Solal A, Blumenthal JA, Rendall DS, Piña IL; HF-ACTION
­
Investigators. Relation between volume of exercise and clinical outcomes
in patients with heart failure. J Am Coll Cardiol. 2012;60:1899–1905.
30.Witham MD, Argo IS, Johnston DW, Struthers AD, McMurdo ME.
Predictors of exercise capacity and everyday activity in older heart failure patients. Eur J Heart Fail. 2006;8:203–207.
31. De Feo S, Franceschini L, Brighetti G, Cicoira M, Zanolla L, Rossi A, Golia
G, Zardini P. Ischemic etiology of heart failure identifies patients with more
severely impaired exercise capacity. Int J Cardiol. 2005;104:292–297.
32. Forman DE, Clare R, Kitzman DW, Ellis SJ, Fleg JL, Chiara T, Fletcher
G, Kraus WE; HF-ACTION Investigators. Relationship of age and exercise performance in patients with heart failure: the HF-ACTION study.
Am Heart J. 2009;158(4 suppl):S6–S15.
33. Chung MK, Shemanski L, Sherman DG, Greene HL, Hogan DB, Kellen
JC, Kim SG, Martin LW, Rosenberg Y, Wyse DG; AFFIRM Investigators.
Functional status in rate-versus rhythm-control strategies for atrial fibrillation: results of the Atrial Fibrillation Follow-Up Investigation of
Rhythm Management (AFFIRM) Functional Status Substudy. J Am Coll
Cardiol. 2005;46:1891–1899.
34. Hohnloser SH, Kuck KH, Lilienthal J. Rhythm or rate control in atrial
fibrillation–Pharmacological Intervention in Atrial Fibrillation (PIAF): a
randomised trial. Lancet. 2000;356:1789–1794.
35.Izawa KP, Watanabe S, Oka K, Osada N, Omiya K, Brubaker PH,
Shimizu H. Diabetes mellitus may lower daily physical activity in heart
failure patients. Int J Cardiol. 2013;168:4882–4883.
36.Scheuermann-Freestone M, Madsen PL, Manners D, Blamire AM,
Buckingham RE, Styles P, Radda GK, Neubauer S, Clarke K. Abnormal
cardiac and skeletal muscle energy metabolism in patients with type 2
diabetes. Circulation. 2003;107:3040–3046.
37. Crowther RG, Spinks WL, Leicht AS, Quigley F, Golledge J. Relationship
between temporal-spatial gait parameters, gait kinematics, walking performance, exercise capacity, and physical activity level in peripheral arterial disease. J Vasc Surg. 2007;45:1172–1178.
38. Jones WS, Clare R, Ellis SJ, Mills JS, Fischman DL, Kraus WE, Whellan
DJ, O’Connor CM, Patel MR. Effect of peripheral arterial disease on
functional and clinical outcomes in patients with heart failure (from
­HF-ACTION). Am J Cardiol. 2011;108:380–384.
39. Lapu-Bula R, Robert A, Van Craeynest D, D’Hondt AM, Gerber BL,
Pasquet A, Melin JA, De Kock M, Vanoverschelde JL. Contribution
of exercise-induced mitral regurgitation to exercise stroke volume and
exercise capacity in patients with left ventricular systolic dysfunction.
Circulation. 2002;106:1342–1348.
40. Abdulla J, Abildstrom SZ, Christensen E, Kober L, Torp-Pedersen C.
A meta-analysis of the effect of angiotensin-converting enzyme inhibitors on functional capacity in patients with symptomatic left ventricular
systolic dysfunction. Eur J Heart Fail. 2004;6:927–935.
41. Guazzi M, Arena R. The impact of pharmacotherapy on the cardiopulmonary exercise test response in patients with heart failure: a mini review. Curr Vasc Pharmacol. 2009;7:557–569.
42. Guazzi M, Melzi G, Marenzi GC, Agostoni P. Angiotensin-converting
enzyme inhibition facilitates alveolar-capillary gas transfer and improves
ventilation-perfusion coupling in patients with left ventricular dysfunction. Clin Pharmacol Ther. 1999;65:319–327.
43. Poelzl G, Frick M, Lackner B, Huegel H, Alber HF, Mair J, Herold
M, Schwarzacher SP, Pachinger O, Weidinger F. Short-term improvement in submaximal exercise capacity by optimized therapy with
ACE inhibitors and beta blockers in heart failure patients is associated with restoration of peripheral endothelial function. Int J Cardiol.
2006;108:48–54.
44. Gulati SK, McKenzie J, Crossley G, Papp MA, Sims J, Andriulli J.
Device measured physical activity: is it the new 6-minute hall walk?
J Cardiac Failure. 2009;15:S119–S120.
45. Braunschweig F, Mortensen PT, Gras D, Reiser W, Lawo T, Mansour
H, Sogaard P, Stegemann B, Bruns HJ, Linde C; InSync III Study
Investigators. Monitoring of physical activity and heart rate variability
in patients with chronic heart failure using cardiac resynchronization devices. Am J Cardiol. 2005;95:1104–1107.
Clinical Perspective
This study shows that patient activity measured by implanted devices predicts death and hospitalization for worsening
heart failure. It included 781 heart failure patients with implanted defibrillators (most with resynchronization therapy). The
patient activity measured by device accelerometers was averaged over a 30-day period shortly after device implantation. The
resulting early physical activity (PA) level was strongly predictive of later clinical events. Each 10 minutes per day of incremental activity was associated with 8% reduction of the risk of death (P<0.0001), 3% reduction of the risk of heart failure
hospitalization (P=0.001), and 5% reduction of the risk of the composite event (P<0.0001). Similar results were observed
after correction for known heart failure risk factors. Risk stratification based on device-measured early PA may contribute
to efficient patient management. However, further research is needed to assess if device-measured activity can be used to
monitor patient status, or whether and to what extent increasing activity levels by exercise programs or patient counseling
will result in better outcomes.
Downloaded from http://circheartfailure.ahajournals.org/ by guest on June 15, 2017
Physical Activity Measured With Implanted Devices Predicts Patient Outcome in Chronic
Heart Failure
Viviane M. Conraads, Martijn A. Spruit, Frieder Braunschweig, Martin R. Cowie, Luigi
Tavazzi, Martin Borggrefe, Michael R.S. Hill, Sandra Jacobs, Bart Gerritse and Dirk J. van
Veldhuisen
Circ Heart Fail. 2014;7:279-287; originally published online February 11, 2014;
doi: 10.1161/CIRCHEARTFAILURE.113.000883
Circulation: Heart Failure is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX
75231
Copyright © 2014 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-3289. Online ISSN: 1941-3297
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circheartfailure.ahajournals.org/content/7/2/279
Data Supplement (unedited) at:
http://circheartfailure.ahajournals.org/content/suppl/2014/02/11/CIRCHEARTFAILURE.113.000883.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Heart Failure 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: Heart Failure is online at:
http://circheartfailure.ahajournals.org//subscriptions/
SUPPLEMENTAL MATERIAL
Physical activity, measured with implanted devices,
predicts patient outcome in Chronic Heart Failure.
Viviane M Conraads, MD1, Martijn A Spruit, PhD2,3, Frieder Braunschweig, MD4, Martin R
Cowie, MD5, Luigi Tavazzi, MD6, Martin Borggrefe, MD7, Michael RS Hill, PhD8, Sandra Jacobs,
PhD9, Bart Gerritse, PhD9, Dirk J van Veldhuisen, MD10
1
Department of Cardiology, Antwerp University Hospital, Edegem, Belgium
University of Antwerp, Antwerp, Belgium
2
Department of Research & Education, CIRO+, center of expertise for chronic organ failure,
Horn, The Netherlands
3
REVAL - Rehabilitation Research Center, BIOMED - Biomedical Research Institute, Faculty of
Medicine and Life Sciences, Hasselt University, Diepenbeek, Belgium
4
Karolinska Institutet, Dept of Cardiology, Karolinska University Hospital, Stockholm, Sweden
5
National Heart & Lung Institute, Imperial College, London, UK
6
Maria Cecilia Hospital, GVM Care and Research, E.S. Health Science Foundation, Cotignola,
Italy
7
Department of Cardiology, University Hospital Mannheim, Mannheim, Germany
8
Medtronic Inc., Minneapolis, MN, USA
9
Medtronic Bakken Research Center, Maastricht, The Netherlands
10
Department of Cardiology, University Medical Center Groningen, University of Groningen,
Groningen, The Netherlands
Methods
The table below (eTable 1) provides additional details of the designs of SENSE-HF and DOTHF. Investigators may have influenced patient behavior based on physical activity records,
which were accessible after Phase I for SENSE-HF (unblinding after Phase I) and for patients in
the access arm of DOT-HF. We therefore investigated the interaction effect between access to
device diagnostic data and early physical activity. Both studies were approved by the local
ethics committees, and written informed consent was obtained from all patients. The authors
had full access to study data
To evaluate the possible confounding effect of a positive response to CRT, in terms of exercise
tolerance, as well as outcome, we compared patients implanted with an ICD only to those with
an implanted CRT-D device.
Influences of access to diagnostics and CRT were assessed from interaction terms.
Results:
The average early physical activity is 199 minutes per day, with a standard deviation of 106
min/day. Further descriptive statistics are: range: 3-579 min/day; and quartiles 120 and 259
min/day. eFigure 1 provides the frequency distribution.
Access to device diagnostic data for patients enrolled in DOT-HF did not influence the relation
between early activity and the primary endpoint (HR=0.96 in patients with, and HR=0.95 in
patients without access, interaction p=0.58). Patients with ICD and CRT-D were comparable
with respect to early activity (p=0.32) and incidence of primary endpoints (p=0.17). The
association between early activity and primary endpoints did also not differ significantly
(HR=0.92 in ICD patients and HR=0.96 in CRT-D patients, interaction p=0.15).
eTable 1. Design of SENSE-HF and DOT-HF
SENSE-HF
DOT-HF
Trial registration
NCT00400985
NCT00480077
Design
single arm
randomized parallel two arm
Treatment
ICD/CRT-D device
ICD/CRT-D device
Main inclusion criteria
chronic heart failure with
impaired LVEF (≤ 35%)
despite OMT, HF
hospitalization in last 12
months, recent implant (≤ 34
days) of ICD or CRT-D with
OptiVol
chronic heart failure with
impaired LVEF (≤ 35%)
despite OMT, HF
hospitalization in last 12
months, recent implant (≤ 6
months) of ICD or CRT-D with
OptiVol
Main exclusion
criteria
post heart transplant or on
waiting list, COPD, pulmonary
hypertension, dialysis
post heart transplant or on
waiting list, COPD, pulmonary
hypertension, dialysis, cardiac
surgery in last 90 days
Access to device data
phase I: 6 months, no access
phase II/III: 18 months, full
access
control arm: no access
access arm: full access
Primary endpoint
phase I: sensitivity and PPV of
OptiVol threshold crossing for
detection of HF hospitalization
phase II/III: PPV of first
detected OptiVol alert for
worsening HF
all-cause death and HF
hospitalization
Enrolment
4 Mar 2005 - 9 Sep 2008
27 Mar 2007- 2 Feb 2009
Abbreviations: COPD=Chronic Obstructive Pulmonary Disease; CRT=Cardiac
Resynchronization Therapy; CRT-D=Defibrillator with CRT; HF=Heart Failure; ICD=Implantable
Cardioverter Defibillator; LVEF=Left Ventricular Ejection Fraction; OMT=Optimal Medical
Therapy; PPV=Positive Predictive Value
eFigure 1. Histogram for early physical activity.
Histogram of early physical activity
12
10
Percent
8
6
4
2
0
0
40
80
120 160
200 240
280 320 360
400 440
Early Physical Activity (min/day)
480 520 560
600