This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the author's
institution and sharing with colleagues.

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party
websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier's archiving and manuscript policies are
encouraged to visit:

http://www.elsevier.com/authorsrights

Listen to this manuscript’s
audio summary by
JACC Editor-in-Chief
Dr. Valentin Fuster.
Author's Personal Copy
JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 71, NO. 10, 2018
ª 2018 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION
PUBLISHED BY ELSEVIER
Sustained Physical Activity,
Not Weight Loss, Associated With
Improved Survival in Coronary Heart Disease
Trine Moholdt, PHD,a,b Carl J. Lavie, MD,c Javaid Nauman, PHDa,d
ABSTRACT
BACKGROUND Individuals with coronary heart disease (CHD) are recommended to be physically active and to maintain
a healthy weight. There is a lack of data on how long-term changes in body mass index (BMI) and physical activity (PA)
relate to mortality in this population.
OBJECTIVES This study sought to determine the associations among changes in BMI, PA, and mortality in individuals
with CHD.
METHODS The authors studied 3,307 individuals (1,038 women) with CHD from the HUNT (Nord-Trøndelag Health Study)
with examinations in 1985, 1996, and 2007, followed until the end of 2014. They calculated the hazard ratio (HR) for all-
cause and cardiovascular disease (CVD) mortality according to changes in BMI and PA, and estimated using Cox proportional
hazards regression models adjusted for age, smoking, blood pressure, diabetes, alcohol, and self-reported health.
RESULTS There were 1,493 deaths during 30 years of follow-up (55% from CVD, median 15.7 years). Weight loss,
classified as change in BMI <–0.10 kg/m2/year, associated with increased all-cause mortality (adjusted HR: 1.30; 95%
confidence interval [CI]: 1.12 to 1.50). Weight gain, classified as change in BMI $0.10 kg/m2/year, was not associated with
increased mortality (adjusted HR: 0.97; 95% CI: 0.87 to 1.09). Weight loss only associated with increased risk in those
who were normal weight at baseline (adjusted HR: 1.38; 95% CI: 1.11 to 1.72). There was a lower risk for all-cause
mortality in participants who maintained low PA (adjusted HR: 0.81; 95% CI: 0.67 to 0.97) or high PA (adjusted HR: 0.64;
95% CI: 0.50 to 0.83), compared with participants who were inactive over time. CVD mortality associations were similar
as for all-cause mortality.
CONCLUSIONS The study observed no mortality risk reductions associated with weight loss in individuals
with CHD, and reduced mortality risk associated with weight gain in individuals who were normal weight at baseline.
Sustained PA, however, was associated with substantial risk reduction. (J Am Coll Cardiol 2018;71:1094–101)
© 2018 by the American College of Cardiology Foundation.
S ubstantial evidence suggests causality between
obesity and development of coronary heart dis-
ease (CHD) (1–4). In line with this, guidelines
for secondary prevention of CHD from the American
Cardiology Foundation recommend that patients
should maintain or achieve a body mass index (BMI)
within the normal range (18.5 to 24.9 kg/m2) (5).
However, despite the strong association between
Heart Association and the American College of obesity and development of CHD, results from large
From the aDepartment of Circulation and Medical Imaging, Faculty of Medicine and Health Sciences, Norwegian University of
Science and Technology, Trondheim, Norway; bWomen’s Clinic, St. Olav’s University Hospital, Trondheim, Norway; cDepartment
of Cardiovascular Diseases, John Ochsner Heart and Vascular Institute, Ochsner Clinical School, University of Queensland School
of Medicine, New Orleans, Louisiana; and the dInstitute of Public Health, College of Medicine and Health Sciences, United Arab
Emirates University, Al-Ain, United Arab Emirates. This study was supported by a grant from the Norwegian Health Association
(Dr. Moholdt). The authors were also supported by grants from the K. G. Jebsen Foundation, Norway (Dr. Nauman); and from the
Liaison Committee between the Central Norway Regional Health Authority and the Norwegian University of Science and Tech-
nology, Trondheim, Norway (Drs. Nauman and Moholdt). The HUNT (Nord-Trøndelag Health) study is a collaboration between the
HUNT Research Centre (Faculty of Medicine, Norwegian University of Science and Technology), Nord-Trøndelag county Council,
Central Norway Health Authority, and the Norwegian Institute of Public Health. Dr. Lavie is author of the book The Obesity
Paradox. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Manuscript received October 18, 2017; revised manuscript received November 30, 2017, accepted January 4, 2018.
ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2018.01.011
 Author's Personal Copy
JACC VOL. 71, NO. 10, 2018
MARCH 13, 2018:1094–101
meta-analyses indicate that subjects with established
CHD who have BMI above the normal range have
better prognosis, often termed the “obesity paradox”
(6,7). Additionally, a recent study of CHD patients
indicated worse prognosis associated with weight
fluctuations over time (8). Furthermore, among CHD
patients with a high cardiorespiratory fitness (9–11),
or who have a high level of physical activity (PA)
(12), there is no obesity paradox. Previous studies
addressing the obesity paradox have, for the most
part, been limited to 1 single measure
body weight or PA at a baseline visit, without fully
determining how changes in weight and PA levels
associate with survival.
SEE PAGE 1102
The most studied nonpharmacological therapy in
CHD is cardiac rehabilitation. Exercise-based cardiac
rehabilitation is associated with a 26% reduction in
mortality post–myocardial infarction (MI) (13). How-
ever, the patients who participate in such programs
are predominately men, are younger, and have less
comorbidity than do the patients not referred to car-
diac rehabilitation. Furthermore, participation rates
are often low, varying from 10% to 60% (14–16).
Therefore, there is a paucity of data on the associa-
tions between survival and changes in cardiorespira-
tory fitness or PA, as well as in body composition, in a
nonselected population of subjects with established
CHD.
In this study, we investigated how long-term
changes in BMI and PA associated with all-cause and
cardiovascular disease (CVD) mortality in subjects
with CHD. We hypothesized that maintaining or
achieving a high level of PA would associate with
improved survival, and that weight loss would asso-
ciate with improved survival in overweight and obese
subjects with CHD.
METHODS
STUDY DESIGN AND PARTICIPANTS. To date, the
HUNT (Nord-Trøndelag Health Study) has been con-
ducted in 3 waves, with data collection in 1984 to
1986 (HUNT1), in 1995 to 1997 (HUNT2), and in 2006 to
2008 (HUNT3) (17). All inhabitants 20 years of age or
older in Nord-Trøndelag County in Norway were
invited to participate in the HUNT study. Participants
attended a clinical examination and filled out
detailed questionnaires about their health and life-
style. The overall participant rates in the HUNT1,
HUNT2, and HUNT3 waves were 88%, 70%, and 56%,
respectively. Our analysis included men and women
reporting to have CHD, as either angina pectoris (AP)
Moholdt et al. 1095
Exercise, Weight Change, and Mortality in CHD
or MI. We included only those with data on ABBREVIATIONS
PA, BMI, diabetes mellitus, self-reported AND ACRONYMS
health, blood pressure, smoking, and alcohol
AP = angina pectoris
consumption in 2 or 3 HUNT study waves
BMI = body mass index
(Figure 1). We excluded participants with a
2 CHD = coronary heart disease
BMI of <18.5 kg/m due to limited subject
numbers and, hence, statistical power in this CI = confidence interval
group. The study protocol was approved by CVD = cardiovascular disease
the Regional Committee for Medical and HR = hazard ratio
Health Research Ethics in Central Norway MI = myocardial infarction
of (2014/1493). PA = physical activity
ASSESSMENT OF BMI. Height and weight were
measured and BMI calculated as weight in kilograms
divided by the square of height in meters. We used
the World Health Organization categorization of
BMI—normal weight (18.5 to 24.9 kg/m2), overweight
(25.0 to 29.9 kg/m2 ), and obese ($30.0 kg/m 2)—and
categorized changes in BMI over time as loss
(<–0.10 kg/m 2/year), stable (–0.10 to 0.09 kg/m 2/year),
and gain ($0.10 kg/m 2/year) (18).
ASSESSMENT OF PA. At each HUNT study wave, the
participants answered questions about frequency,
duration, and intensity of leisure time PA. We
grouped the participants into 3 levels of PA—inactive,
low PA, and high PA—based on a previously published
index (19). We made these categories to group par-
ticipants according to the current recommendations
on PA to promote health in adults (20). The inactive
category comprises those participants who reported
no PA, the low category comprises those who re-
ported PA below the recommended level, and the
high category is those who reported to fulfill or
exceed the recommendations. We then categorized
changes in PA into 9 categories (inactive-inactive,
inactive-low, inactive-high, low-inactive, low-low,
low-high, high-inactive, high-low, and high-high).
ASCERTAINMENT OF OUTCOMES. The primary
outcome was all-cause mortality, with CVD mortality
(International Classification of Diseases-Ninth Revi-
sion: 390 to 459; International Statistical Classifica-
tion of Diseases-10th Revision: I00 to I99) as a
secondary outcome. Follow-up ended on December
31, 2014. Norwegian physicians and public health
officers are directed to report all deaths to the
National Cause of Death Registry in Norway, and thus
our study had a complete follow-up.
ASSESSMENT OF COVARIABLES. The participants
answered detailed questions about various health
variables and lifestyle. We adjusted for smoking
(current, former, or never), alcohol consumption over
a 14-day period (abstainer, 0 drinks, 1 to 4 drinks,
or $5 drinks), self-reported health status (bad,
not so good, good, or very good), diabetes mellitus
 Author's Personal Copy
1096 Moholdt et al. JACC VOL. 71, NO. 10, 2018

Exercise, Weight Change, and Mortality in CHD MARCH 13, 2018:1094–101

categories). Results are reported as model 2 adjusted

F I G U R E 1 Selection of Participants for the Study
HR: with 95% CI if not otherwise stated. We also
performed stratified analyses for changes in BMI ac-
Participants with self-reported coronary heart disease in cording to baseline BMI category, and for changes in
HUNT 1 & HUNT 2, n = 5690
HUNT 2 & HUNT 3, n = 2690 PA according to baseline PA. All variables were
HUNT 1 & HUNT 3, n = 2754 updated over time in the analyses; therefore, for
HUNT 1 & HUNT 2 & HUNT 3, n = 2545
participants who attended all 3 HUNT study waves,
changes in BMI and PA were updated over time.
In separate analyses, we excluded deaths occurring
Participants with self-reported coronary
heart disease in at least one HUNT wave during the first 3 years from the last HUNT study
n = 9463 wave the participants attended to minimize the
chance of bias due to reversed causality. We also
additionally adjusted for MI and AP in our sensitivity
analyses, as well as including only those who re-
ported to have had an MI. To minimize the competing
Exclusions:
90 BMI <18.5 kg/m2 risk bias, we repeated our analyses using competing
3819 missing on changes in BMI risk survival regression models (21). We used Stata
1740 missing on changes in
physical activity
software version 13.1 (StataCorp, College Station,
507 missing on smoking, alcohol, Texas) for all analyses, with all tests 2 sided and
diabetes, self-reported p values of <0.05 considered significant.
health status, and hypertension
RESULTS

Participants with self-reported
coronary heart disease in at least
2 waves, n = 3307
(1038 women, 2269 men)
The number of participants from the HUNT (Nord-Trøndelag Health) study with
self-reported coronary heart disease. To be included in this study, the participants had to
have self-reported coronary heart disease at participation in at least 2 HUNT study waves
and to have valid data on physical activity, body mass index (BMI), diabetes mellitus,
self-reported health, blood pressure, smoking, and alcohol consumption. Participants
with a BMI of <18.5 kg/m2 were excluded due to limited number of individuals (n ¼ 90).
(yes or no), and hypertension (yes or no), which was
defined as systolic blood pressure $140 mm Hg or
diastolic blood pressure $90 mm Hg or taking blood
pressure medications.
STATISTICAL ANALYSIS. Baseline characteristics of
participants according to BMI categories were
compared using linear regression for continuous
variables and chi-square tests for categorical
variables.
We used Cox proportional hazards models to esti-
mate hazard ratios (HRs) with 95% confidence in-
tervals (CIs), conditioning on sex. In the crude
analysis (model 1), we adjusted for attained age as
time scale and examination year. In model 2, we
additionally adjusted for smoking status, alcohol
consumption, hypertension, and self-reported health
status. In model 3, we additionally adjusted for
change in BMI (in categories), and change in PA (in
Figure 1 shows the participant flow in the study. Of
the 3 307 who participated in 2 or more HUNT study
waves, 1493 died during 30 (median 15.7) years of
follow-up. Of these, 199 died during the first 3 years
of follow-up. Table 1 shows the proportion of partic-
ipants in each BMI category and selected baseline
characteristics at HUNT1 study wave. Characteristics
of participants according to changes in BMI are pre-
sented in Online Table 1. There were 1,507 (45.6%)
participants with AP only, 929 (28.1%) with MI only,
and 871 (26.3%) with both AP and MI. Almost one-half
of the participants were inactive at baseline, with
a higher percentage of inactive and women in the
obese category.
BMI CHANGE AND ALL-CAUSE MORTALITY.
Compared with having a stable BMI, those who lost
weight (BMI <–0.10 kg/m2 /year) had a 30% increased
all-cause mortality risk (HR: 1.30; 95% CI: 1.12 to 1.50),
whereas weight gain (BMI $0.10 kg/m 2/year) was not
significantly associated with mortality risk (Table 2).
The association between weight loss and increased
risk was still significant when excluding deaths
occurring the first 3 years of follow-up, with an HR of
1.26 (95% CI: 1.08 to 1.47) (Online Table 2).
In stratified analyses, we observed different asso-
ciations between changes in BMI and mortality risk
for the different BMI categories (Figure 2A). In
normal-weight subjects, weight loss associated with
38% increased risk (HR: 1.38; 95% CI: 1.11 to 1.72),
whereas weight gain was associated with 25%
 Author's Personal Copy
JACC VOL. 71, NO. 10, 2018 Moholdt et al. 1097
MARCH 13, 2018:1094–101 Exercise, Weight Change, and Mortality in CHD

decreased risk (HR: 0.75; 95% CI: 0.56 to 0.99).

T A B L E 1 Baseline Characteristics of the Participants According to Body Mass Index
Adjusting also for changes in PA did not affect esti-
mates (Online Table 3). When excluding deaths Body Mass Index

occurring the first 3 years of follow-up, weight loss Total 18.5–24.9 kg/m2 25.0–29.9 kg/m2 $30.0 kg/m2 p
(N ¼ 2,821) (n ¼ 1,000) (n ¼ 1,352) (n ¼ 469) Value*
was no longer significantly associated with mortality
Female 1,035 (36.7) 345 (34.5) 440 (32.5) 250 (53.3) <0.01
in normal-weight subjects (HR: 1.25; 95% CI: 0.98 to
Age, yrs 68.7  9.6 69.9  9.7 68.2  9.6 67.5  9.1 <0.01
1.60) (Online Table 4). In overweight and obese Weight, kg 74.9  12.8 64.6  8.4 77.3  8.6 90.0  12.0 <0.01
subjects, neither weight loss nor weight gain was Height, cm 167.6  9.0 167.9  8.9 168.3  8.8 165.0  9.5 <0.01
associated with mortality risk (Figure 2A), neither Physical activity
when adjusting for changes in PA nor when excluding Inactive 1,381 (48.9) 486 (48.6) 628 (46.5) 267 (56.9)

the first 3 years of follow-up (Online Tables 3 and 4). Low 917 (32.5) 308 (30.8) 473 (35.0) 136 (29.0)
Recommended 341 (12.1) 116 (11.6) 177 (13.1) 48 (10.2)
We observed no material changes to the estimates
High 182 (6.5) 90 (9.0) 74 (5.4) 18 (3.8) <0.01
when we repeated our analyses with adjustments for
Diabetes status
AP and MI (Online Table 5) or when we included only Yes 295 (10.5) 102 (10.2) 136 (10.1) 57 (12.2)
individuals with an MI (Online Table 6). No 2,526 (89.5) 898 (89.8) 1,216 (89.9) 412 (87.8) 0.42
Smoking status
BMI CHANGE AND CVD MORTALITY. A total of 819
Never 1,162 (41.2) 395 (39.5) 544 (40.2) 223 (47.5)
(55%) deaths were due to CVD, and of these 198
Current 636 (22.5) 283 (28.3) 273 (20.2) 80 (17.1)
occurred during the first 3 years of follow-up. Weight Former 1,023 (36.3) 322 (32.2) 535 (39.6) 166 (35.4) <0.01
loss associated with 36% increased CVD mortality Alcohol consumption†
(HR: 1.36; 95% CI: 1.12 to 1.65), whereas weight gain Abstainer 597 (21.2) 218 (21.8) 276 (20.4) 103 (21.9)
did not associate with CVD mortality (Table 2, 0 1,527 (54.1) 550 (55.0) 720 (53.3) 257 (54.8)

Online Figure 1). Weight loss was still significantly 1–4 531 (18.8) 175 (17.5) 267 (19.7) 89 (19.0)
$5 166 (5.9) 57 (5.7) 89 (6.6) 20 (4.3) 0.44
associated with CVD mortality when adjusting also
Hypertension status‡
for changes in PA (Table 2) and when excluding
Yes 2,242 (79.5) 735 (73.5) 1,093 (80.8) 414 (88.3)
deaths occurring during the first 3 years of follow-up
No 579 (20.5) 265 (26.5) 259 (19.2) 55 (11.7) <0.01
(Online Table 2). Health status§
Figure 2B shows the associations between changes Bad 264 (9.4) 98 (9.8) 119 (8.8) 47 (10.0)
in BMI and CVD mortality according to baseline BMI Not so good 1,814 (64.3) 628 (62.8) 864 (63.9) 322 (68.7)
category. We observed increased CVD mortality risk Good 711 (25.2) 260 (26.0) 352 (26.0) 99 (21.1)

associated with weight loss in normal-weight subjects Very good 32 (1.1) 14 (1.4) 17 (1.3) 1 (0.2) 0.10

(HR: 1.47; 95% CI: 1.09 to 1.98) (Online Table 3), but
Values are n (%) or mean  SD. *For linear trend, regression analyses were used for continuous variables; chi-
this was no longer significant when we excluded square tests were used for proportions of categorical variables. †Based on consumption over a 2-week period.
‡Hypertension was defined as systolic blood pressure $140 mm Hg or diastolic blood pressure $90 mm Hg or
deaths occurring during the first 3 years of follow-up taking blood pressure medications. §Self-reported health status.
(HR: 1.31; 95% CI: 0.94 to 1.83) (Online Table 4). In
participants who were overweight and obese at
baseline, there were no significant associations be-
tween changes in BMI and CVD mortality risk (Online
Table 3). Furthermore, the results of competing risk T A B L E 2 HRs of Mortality According to Change in Body Mass Index
survival regression were not different from our main
Model 1 HR Model 2 HR Model 3 HR
results (Online Table 7, Online Figure 1). Deaths* (95% CI) (95% CI) (95% CI)

All cause
CHANGE IN PA AND ALL-CAUSE MORTALITY. Those
Loss 305 1.39 (1.21–1.60) 1.30 (1.12–1.50) 1.27 (1.10–1.47)
who reported high PA over time had a 36% lower all-
Stable 586 Reference Reference Reference
cause mortality risk (HR: 0.64; 95% CI: 0.50 to 0.83) Gain 602 0.96 (0.85–1.08) 0.97 (0.87–1.09) 0.96 (0.85–1.07)
compared with those who were inactive over time CVD
(Central Illustration). We also observed significantly Loss 169 1.44 (1.19–1.75) 1.36 (1.12–1.65) 1.34 (1.11–1.63)
reduced all-cause mortality risk in those who changed Stable 317 Reference Reference Reference

from low PA to inactive (HR: 0.82; 95% CI: 0.70 to Gain 333 0.97 (0.83–1.14) 0.98 (0.83–1.14) 0.96 (0.82–1.13)

0.96), who maintained low PA over time (HR: 0.81;

95% CI: 0.67 to 0.97), and in those who changed from
high PA to low PA (HR: 0.74; 95% CI: 0.60 to 0.92).
The estimates were similar when also adjusting
for changes in BMI (Online Table 8) and when
excluding deaths occurring during the first 3 years
Model 1 was adjusted for age and examination year, and stratified by sex. Model 2 was adjusted
for age, examination year, smoking status, diabetes mellitus, alcohol consumption, hypertension,
and health status, and stratified by sex. Model 3 was model 2 plus change in physical activity. Loss
was defined as <–0.10 kg/m2/year, stable as –0.10 to 0.09 kg/m2/year, and gain as $0.10 kg/m2/
year. *Refers to the count at the last HUNT (Nord-Trøndelag Health) study wave attendance
before meeting the endpoint or before censoring.
CI ¼ confidence interval; CVD ¼ cardiovascular disease; HR ¼ hazard ratio.
 Author's Personal Copy
1098 Moholdt et al. JACC VOL. 71, NO. 10, 2018

Exercise, Weight Change, and Mortality in CHD MARCH 13, 2018:1094–101

F I G U R E 2 Change in BMI and Mortality Risk

A Normal weight Overweight Obese

2.2 2.2 2.2
2.0 2.0 2.0
Mortality Hazard Ratio

Mortality Hazard Ratio

Mortality Hazard Ratio

1.8 1.8 1.8
1.6 1.6 1.6
1.4 1.4 1.4
1.2 1.2 1.2
1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
Loss Stable Gain Loss Stable Gain Loss Stable Gain

B Normal weight Overweight Obese

2.2 2.2 2.2
2.0 2.0 2.0
Mortality Hazard Ratio

Mortality Hazard Ratio

Mortality Hazard Ratio

1.8 1.8 1.8
1.6 1.6 1.6
1.4 1.4 1.4
1.2 1.2 1.2
1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
Loss Stable Gain Loss Stable Gain Loss Stable Gain

(A) All-cause mortality hazard ratio according to change in body mass index (BMI), stratified for baseline BMI. (B) Cardiovascular disease mortality hazard ratio
according to change in BMI, stratified for baseline BMI. Boxes represent hazard ratios and vertical lines represent 95% confidence intervals, after adjusting for age,
examination year, smoking status, diabetes status, alcohol consumption, hypertension, and health status, and stratifying for sex. Dotted lines represent a hazard ratio
of 1.0. Normal weight was defined as a BMI of 18.5 to 24.9 kg/m2, overweight was a BMI of 25.0–29.9 kg/m2, and obese was a BMI of $30.0 kg/m2. Loss was defined
as <–0.10 kg/m2/year, stable as –0.10 to 0.09 kg/m2/year, and gain as $0.10 kg/m2/year.

of follow-up (Online Table 9). Furthermore, we
observed no substantial change in estimates when we
included only individuals with an MI (Online Table 6).
CHANGE IN PA AND CVD MORTALITY. A significantly
reduced CVD mortality risk was observed only in
those who maintained a high level of PA over time
(HR: 0.62; 95% CI: 0.43 to 0.89) and in those who
changed from inactive to high PA (HR: 0.68; 95% CI:
0.47 to 0.97) (Online Figure 2). Additionally, adjust-
ing for changes in BMI did not affect the estimates
(Online Table 8). When we excluded the first 3 years
of follow-up, only maintaining a high level of PA over
time was significantly associated with reduced CVD
mortality (HR: 0.59; 95% CI: 0.39 to 0.89) (Online
Table 9). Again, the results of competing risk sur-
vival regression were not different from our main
results (Online Table 10).
DISCUSSION
Contrary to our original hypothesis, we found an as-
sociation between weight loss and increased risk for
all-cause mortality, as well as for CVD mortality, in
individuals with CHD (Figure 2). When stratifying for
BMI, this association was only observed in those who
were normal weight at baseline, whereas weight gain
was associated with reduced all-cause mortality.
Maintaining a high level of PA over 2 or 3 decades was
associated with substantial reductions in mortality
risk, compared with being inactive over time (Central
Illustration).
Although the obesity paradox in CHD has been
described in numerous cohorts over the last decades,
there are fewer studies assessing weight change (22–27).
In clinical decision making, the fundamental question
is not whether being normal weight is beneficial in
CHD, but rather whether weight loss is associated
with improved prognosis (6). Patients in the clinical
setting would like to know whether trying to lose
weight is beneficial and worth the effort. Common of
previous studies on the prognostic importance of
weight change is a relatively short follow-up time,
with the longest follow-up period being 7 years (22).
Pack et al. (28) performed a systematic review and
meta-analysis of the prognostic effects of weight loss
in patients with CHD on a composite outcome of
 Author's Personal Copy
JACC VOL. 71, NO. 10, 2018 Moholdt et al. 1099
MARCH 13, 2018:1094–101 Exercise, Weight Change, and Mortality in CHD

C ENTR AL I LL U STRA T I O N Change in Physical Activity Level and Mortality Risk

1.4
1.2
Mortality Hazard Ratio

1.0
0.8
0.6
0.4
0.2
0.0
Inactive Inactive Inactive Low Low Low High High High
⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓ ⇓
Inactive Low High Inactive Low High Inactive Low High
Moholdt, T. et al. J Am Coll Cardiol. 2018;71(10):1094–101.

All-cause mortality hazard ratio (bars) with 95% confidence intervals (vertical lines) according to change in physical activity (PA) level, after adjusting for age,
examination year, smoking status, diabetes status, alcohol consumption, hypertension, and health status, and stratifying for sex. Blue bars represent participants who
were inactive at baseline, orange bars participants with low physical activity level at baseline, and gray bars participants with high PA level at baseline. The inactive
category comprises those who reported no PA, the low category those who reported PA below the recommended level, and the high category those who fulfilled or
exceeded the recommendations. The dotted line represents a hazard ratio of 1.0.

all-cause mortality, cardiovascular mortality, and
major adverse cardiovascular events, including 35,335
patients. They found a 5% body weight loss over a
mean of 3.2 years to be associated with a 30% greater
risk of the composite outcome, compared with keep-
ing a stable weight. However, a presumed intentional
weight loss (i.e., in the presence of programmed
therapeutic lifestyle changes) was associated with a
33% risk reduction. In contrast, observational weight
loss, analyzed from 10 cohorts, associated with a 62%
increased risk (28). We observed a lower (26%)
increased risk of mortality associated with weight loss
compared with that meta-analysis (28), likely due to
our longer follow-up time. It is, however, difficult
to compare data from different cohorts due to large
variations in weight loss definitions, weight loss time
intervals, adjustments for covariates, population
characteristics, and follow-up time. One possible
explanation for higher short-term mortality risk
associated with weight loss
However, when we excluded deaths occurring during
the first 3 years of follow-up, the estimated HR
did not change substantially, indicating that the
association is not merely a result of reversed
causality. In our opinion, purposeful weight loss
could be beneficial in individuals who are overweight
or obese, although there is little data to support this
in CHD populations.
Our findings suggest that PA needs to be regular
and sustained to confer the largest cardiovascular
benefits. However, compared with being inactive
over time, all patterns of PA change had lower all-
cause mortality risk estimates. Furthermore, we
observed greater risk reductions associated with
taking up or maintaining high PA, compared with low
PA. This was particularly evident for CVD mortality
risk, where maintaining a low PA level over time was
not associated with reduced risk. Prospective studies
of change in PA and mortality in CHD are sparse (29).
Wannamethee et al. (29) included older men with and
without diagnosed CVD and studied the relations
between changes in PA and all-cause mortality and
found that maintaining or taking up light or moderate
PA reduced mortality over 4 years of follow up. In line
with this, men without known CVD who maintain or
improve cardiorespiratory fitness (30,31) and healthy
men and women who increase their PA level over
is occult disease. time (32,33) reduce their risk of all-cause and CVD
mortality. Indeed, cardiorespiratory fitness is a strong
predictor of mortality, independent from traditional
CVD risk factors, in healthy individuals and those
with CVD (34). Cardiorespiratory fitness and PA
markedly impact the obesity paradox, with no obesity
paradox seen in those who are relatively fit or who
report high levels of PA (12,35,36). Physical inactivity
and low aerobic capacity has largely been overlooked
 Author's Personal Copy
1100 Moholdt et al. JACC VOL. 71, NO. 10, 2018

Exercise, Weight Change, and Mortality in CHD MARCH 13, 2018:1094–101

as a risk factor in primary and secondary prevention
of CVD (37), and is currently the only major risk factor
not routinely assessed in clinical practice (38). The
time has come for health care providers to promote
PA in their patients with CHD.
This study presents novel data about the impor-
tance of changes in body weight and PA on mortality
risk in individuals with CHD. We included a relatively
large number of participants, both men and women
across a wide age range, had a complete mortality
follow-up for 30 years, and controlled extensively for
potential confounders. Furthermore, given that both
changes in PA and in BMI associate with mortality,
mutually adjusting for each other strengthen our
results.
STUDY LIMITATIONS. The CHD diagnosis used as
basis for inclusion in our analyses was based on self-
report and a validation with hospital records has not
been undertaken. The validity of self-reported MI in a
Norwegian population has been found to be accept-
able, with a sensitivity of 91.1% and a specificity of
99.5% (39). We believe that we can be more certain
about the diagnosis in those who reported to have
had an MI compared with AP (40). In the sensitivity
analysis adjusting for MI and AP, or including in-
dividuals with an MI only, we observed no substantial
difference in the estimates compared with the main
analysis. We used The National Cause of Death Reg-
istry in Norway to ascertain cause of death. We did
not have the opportunity to quality control death
certificates to ensure correct causes of death in this
study, and this must be regarded as a limitation in our
secondary outcome analysis (CVD mortality). Another
limitation to our study is that BMI was used as the
only measure of body composition; however, the
same obesity paradox has been observed when
comparing patients with high and low body fat
percent as when comparing high and low BMI
(9,23,41). Furthermore, we could not distinguish be-
tween intentional and unintentional weight loss.
Reversed causality can be a problem in observational
studies, especially when studying diseased cohorts.
We limited the chance of reversed causality by
adjusting for known confounders, such as self-
reported health status, smoking, hypertension, and
diabetes, and performed secondary analyses where
we excluded the deaths occurring during the first
3 years of follow-up, with only a minimal effect on the
outcomes. Our study included mostly elderly partic-
ipants and we are unsure if the results extrapolate to
younger populations.
CONCLUSIONS
In this large cohort of subjects with CHD, we observed
an increased all-cause and cardiovascular mortality in
individuals who lost weight, compared with those
being weight stable, especially in those who had a
normal weight at baseline. Maintaining or taking up
PA was associated with substantial reductions in all-
cause and CVD mortality risk, with larger reductions
seen with high PA levels compared with low levels.
Increased attention should be placed on strategies to
increase PA in secondary prevention of CHD.
ADDRESS FOR CORRESPONDENCE: Dr. Trine
Moholdt, Department of Circulation and Medical
Imaging, Norwegian University of Science and
Technology, Medisinsk Teknisk Forskningssenter,
Post box 8905, 7491, Trondheim, Norway. E-mail:
trine.moholdt@ntnu.no.
PERSPECTIVES
COMPETENCY IN PATIENT CARE AND PROCEDURAL
SKILLS: In observational studies, there is no reduc-
tion in mortality risk associated with overweight or
obesity among individuals with high cardiorespiratory
fitness. Sustained PA is associated with reduced mor-
tality risk, in contrast to weight loss, which is associ-
ated with increased mortality.
TRANSLATIONAL OUTLOOK: Further studies are
needed to determine the impact of deliberate weight
loss on patients with various forms of CVD.

REFERENCES

1. Li TY, Rana JS, Manson JE, et al. Obesity as
compared with physical activity in predicting risk
of coronary heart disease in women. Circulation
2006;113:499–506.
2. Rexrode KM, Buring JE, Manson JE. Abdominal
and total adiposity and risk of coronary heart
disease in men. Int J Obes Relat Metab Disord
2001;25:1047–56.
3. Lavie CJ, Arena R, Alpert MA, Milani RV,
Ventura HO. Management of cardiovascular dis-
eases in patients with obesity. Nat Rev Cardiol
2018;15:45–56.
4. Ortega FB, Lavie CJ, Blair SN. Obesity and car-
diovascular disease. Circ Res 2016;118:1752–70.
5. Smith SC Jr., Benjamin EJ, Bonow RO, et al.
AHA/ACCF Secondary Prevention and Risk
Reduction Therapy for Patients with Coronary and
other Atherosclerotic Vascular Disease: 2011 up-
date: a guideline from the American Heart Asso-
ciation and American College of Cardiology
Foundation. J Am Coll Cardiol 2011;58:2432–46.
6. Romero-Corral A, Montori VM, Somers VK, et al.
Association of bodyweight with total mortality
and with cardiovascular events in coronary artery

JACC VOL. 71, NO. 10, 2018
MARCH 13, 2018:1094–101
disease: a systematic review of cohort studies.
Lancet 2006;368:666–78.
7. Wang ZJ, Zhou YJ, Galper BZ, Gao F, Yeh RW,
Mauri L. Association of body mass index with
mortality and cardiovascular events for patients
with coronary artery disease: a systematic review
and meta-analysis. Heart 2015;101:1631–8.
8. Bangalore S, Fayyad R, Laskey R, DeMicco DA,
Messerli FH, Waters DD. Body-weight fluctuations
and outcomes in coronary disease. N Engl J Med
2017;376:1332–40.
9. McAuley PA, Artero EG, Sui X, et al. The obesity
paradox, cardiorespiratory fitness, and coronary
heart disease. Mayo Clin Proc 2012;87:443–51.
10. Goel K, Thomas RJ, Squires RW, et al. Com-
bined effect of cardiorespiratory fitness and
adiposity on mortality in patients with coronary
artery disease. Am Heart J 2011;161:590–7.
11. Oktay AA, Lavie CJ, Kokkinos PF, Parto P,
Pandey A, Ventura HO. The interaction of cardio-
respiratory fitness with obesity and the obesity
paradox in cardiovascular disease. Prog Cardiovasc
Dis 2017;60:30–44.
12. Moholdt T, Lavie CJ, Nauman J. Interaction of
physical activity and body mass index on mortality
in coronary heart disease: data from the Nord-
Trondelag Health study. Am J Med 2017;130:
949–57.
13. Lawler PR, Filion KB, Eisenberg MJ. Efficacy of
exercise-based cardiac rehabilitation post-
myocardial infarction: a systematic review and
meta-analysis of randomized controlled trials. Am
Heart J 2011;162:571–84.e2.
14. Aragam KG, Moscucci M, Smith DE, et al.
Trends and disparities in referral to cardiac reha-
bilitation after percutaneous coronary interven-
tion. Am Heart J 2011;161:544–51.e2.
15. Golwala H, Pandey A, Ju C, et al. Temporal
trends and factors associated with cardiac reha-
bilitation referral among patients hospitalized
with heart failure: findings from Get With The
Guidelines-Heart Failure Registry. J Am Coll Car-
diol 2015;66:917–26.
16. Brown TM, Hernandez AF, Bittner V, et al.
Predictors of cardiac rehabilitation referral in cor-
onary artery disease patients: findings from the
American Heart Association’s Get With The
Guidelines Program. J Am Coll Cardiol 2009;54:
515–21.
17. Krokstad S, Langhammer A, Hveem K, et al.
Cohort profile: the HUNT study, Norway. Int J
Epidemiol 2013;42:968–77.
18. Klenk J, Rapp K, Ulmer H, Concin H, Nagel G.
Changes of body mass index in relation to mor-
tality: results of a cohort of 42,099 adults. PLoS
One 2014;9:e84817.
Author's Personal Copy
19. Moholdt T, Wisloff U, Lydersen S, Nauman J.
Current physical activity guidelines for health are
insufficient to mitigate long-term weight gain:
more data in the fitness versus fatness debate
(The HUNT study, Norway). Br J Sports Med 2014;
48:1489–96.
20. Haskell WL, Lee IM, Pate RR, et al. Physical
activity and public health: updated recommenda-
tion for adults from the American College of
Sports Medicine and the American Heart Associa-
tion. Circulation 2007;116:1081–93.
21. Gutierrez R. Competing-risk regression - Stata.
Available at: https://www.stata.com/meeting/
boston10/boston10_gutierrez.pdf. Accessed
November 16, 2017.
22. Myers J, Lata K, Chowdhury S, McAuley P,
Jain N, Froelicher V. The obesity paradox and
weight loss. Am J Med 2011;124:924–30.
23. Lavie CJ, Milani RV, Artham SM, Patel DA,
Ventura HO. The obesity paradox, weight loss, and
coronary disease. Am J Med 2009;122:1106–14.
24. Kocz R, Hassan MA, Perala PR, Negargar S,
Javadzadegan H, Nader ND. The effect of weight
loss on the outcome after coronary artery bypass
grafting in obese patients. Ann Card Anaesth 2012;
15:190–8.
25. Sierra-Johnson J, Romero-Corral A,
Somers VK, et al. Prognostic importance of weight
loss in patients with coronary heart disease
regardless of initial body mass index. Eur J Car-
diovasc Prev Rehabil 2008;15:336–40.
26. Kennedy LM, Dickstein K, Anker SD, et al.
Weight-change as a prognostic marker in 12 550
patients following acute myocardial infarction or
with stable coronary artery disease. Eur Heart J
2006;27:2755–62.
27. Lopez-Jimenez F, Wu CO, Tian X, et al. Weight
change after myocardial infarction–the Enhancing
Recovery in Coronary Heart Disease patients
(ENRICHD) experience. Am Heart J 2008;155:
478–84.
28. Pack QR, Rodriguez-Escudero JP, Thomas RJ,
et al. The prognostic importance of weight loss in
coronary artery disease: a systematic review and
meta-analysis. Mayo Clin Proc 2014;89:1368–77.
29. Wannamethee SG, Shaper AG, Walker M.
Changes in physical activity, mortality, and inci-
dence of coronary heart disease in older men.
Lancet 1998;351:1603–8.
30. Lee DC, Sui X, Artero EG, et al. Long-term
effects of changes in cardiorespiratory fitness and
body mass index on all-cause and cardiovascular
disease mortality in men: the Aerobics Center
Longitudinal Study. Circulation 2011;124:
2483–90.
31. Blair SN, Kohl HW 3rd, Barlow CE,
Paffenbarger RS Jr., Gibbons LW, Macera CA.
Moholdt et al. 1101
Exercise, Weight Change, and Mortality in CHD
Changes in physical fitness and all-cause mortal-
ity. A prospective study of healthy and unhealthy
men. JAMA 1995;273:1093–8.
32. Petersen CB, Gronbaek M, Helge JW,
Thygesen LC, Schnohr P, Tolstrup JS. Changes in
physical activity in leisure time and the risk of
myocardial infarction, ischemic heart disease, and
all-cause mortality. Eur J Epidemiol 2012;27:91–9.
33. Byberg L, Melhus H, Gedeborg R, et al. Total
mortality after changes in leisure time physical
activity in 50 year old men: 35 year follow-up of
population based cohort. BMJ 2009;338:b688.
34. 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.
35. Lavie CJ, De Schutter A, Parto P, et al. Obesity
and prevalence of cardiovascular diseases and
prognosis-the obesity paradox updated. Prog
Cardiovasc Dis 2016;58:537–47.
36. Lavie CJ, Cahalin LP, Chase P, et al. Impact of
cardiorespiratory fitness on the obesity paradox in
patients with heart failure. Mayo Clin Proc 2013;
88:251–8.
37. Myers J, McAuley P, Lavie CJ, Despres JP,
Arena R, Kokkinos P. Physical activity and cardio-
respiratory fitness as major markers of cardiovas-
cular risk: their independent and interwoven
importance to health status. Prog Cardiovasc Dis
2015;57:306–14.
38. Ross R, Blair SN, Arena R, et al. Importance of
assessing cardiorespiratory fitness in clinical
practice: a case for fitness as a clinical vital sign: a
scientific statement from the American Heart As-
sociation. Circulation 2016;134:e653–99.
39. Eliassen BM, Melhus M, Tell GS, et al. Validity
of self-reported myocardial infarction and stroke
in regions with Sami and Norwegian populations:
the SAMINOR 1 Survey and the CVDNOR project.
BMJ Open 2016;6:e012717.
40. Misirli G, Bamia C, Dilis V, Benetou V, Zilis D,
Trichopoulou A. Validation of self-reported inci-
dent cardiovascular disease events in the Greek
EPIC cohort study. Italian J Pub Health 2012;9:
e75381–9.
41. Lavie CJ, De Schutter A, Patel D, Artham SM,
Milani RV. Body composition and coronary heart
disease mortality–an obesity or a lean paradox?
Mayo Clin Proc 2011;86:857–64.
KEY WORDS body mass index, exercise,
HUNT, mortality, myocardial infarction,
obesity paradox
A PP END IX For supplemental figures and
tables, please see the online version of
this article.