Aerobic or Resistance Exercise, or Both, in Dieting Obese Older Adults

The n e w e ng l a n d j o u r na l of m e dic i n e
Original Article
Aerobic or Resistance Exercise, or Both,

in Dieting Obese Older Adults
DennisT. Villareal, M.D., Lina Aguirre, M.D., A.Burke Gurney, Ph.D., P.T.,
DebraL. Waters, Ph.D., DavidR. Sinacore, Ph.D., P.T., Elizabeth Colombo, M.D., Ph.D.,
Reina ArmamentoVillareal, M.D., and Clifford Qualls, Ph.D.
A BS T R AC T
BACKGROUND
Obesity causes frailty in older adults; however, weight loss might accelerate age- From the Division of Endocrinology, Dia-
related loss of muscle and bone mass and resultant sarcopenia and osteopenia. betes, and Metabolism, Baylor College of
Medicine, and the Center for Translational
METHODS Research on Inflammatory Diseases, Mi-
chael E. DeBakey Veterans Affairs (VA)
In this clinical trial involving 160 obese older adults, we evaluated the effectiveness Medical Center both in Houston (D.T.V.,
of several exercise modes in reversing frailty and preventing reduction in muscle R.A.-V.); Medicine Care Line, New Mexico
and bone mass induced by weight loss. Participants were randomly assigned to a VA Health Care System (L.A., D.L.W.,
E.C.), and the Department of Internal
weight-management program plus one of three exercise programs aerobic train- Medicine (L.A., E.C.), the Division of
ing, resistance training, or combined aerobic and resistance training or to a Physical Therapy (A.B.G.), and the De-
control group (no weight-management or exercise program). The primary outcome partment of Mathematics and Statistics
(C.Q.), University of New Mexico School
was the change in Physical Performance Test score from baseline to 6 months of Medicine both in Albuquerque; the
(scores range from 0 to 36 points; higher scores indicate better performance). Department of Medicine, School of Phys-
Secondary outcomes included changes in other frailty measures, body composi- iotherapy, University of Otago, Dunedin,
New Zealand (D.L.W.); and the Program
tion, bone mineral density, and physical functions. in Physical Therapy, Washington Univer-
RESULTS sity School of Medicine, St. Louis (D.R.S.).
Address reprint requests to Dr. Villareal
A total of 141 participants completed the study. The Physical Performance Test score at Baylor College of Medicine, Michael E.
increased more in the combination group than in the aerobic and resistance groups DeBakey VA Medical Center, 2002 Hol-
(27.9 to 33.4 points [21% increase] vs. 29.3 to 33.2 points [14% increase] and 28.8 to combe Ave., Houston, TX 77030, or at
[email protected].
32.7 points [14% increase], respectively; P=0.01 and P=0.02 after Bonferroni correc-
tion); the scores increased more in all exercise groups than in the control group N Engl J Med 2017;376:1943-55.
DOI: 10.1056/NEJMoa1616338
(P<0.001 for between-group comparisons). Peak oxygen consumption (milliliters per Copyright 2017 Massachusetts Medical Society.
kilogram of body weight per minute) increased more in the combination and aerobic
groups (17.2 to 20.3 [17% increase] and 17.6 to 20.9 [18% increase], respectively) than
in the resistance group (17.0 to 18.3 [8% increase]) (P<0.001 for both comparisons).
Strength increased more in the combination and resistance groups (272 to 320 kg
[18% increase] and 288 to 337 kg [19% increase], respectively) than in the aerobic
group (265 to 270 kg [4% increase]) (P<0.001 for both comparisons). Body weight
decreased by 9% in all exercise groups but did not change significantly in the control
group. Lean mass decreased less in the combination and resistance groups than in
the aerobic group (56.5 to 54.8 kg [3% decrease] and 58.1 to 57.1 kg [2% decrease],
respectively, vs. 55.0 to 52.3 kg [5% decrease]), as did bone mineral density at the
total hip (grams per square centimeter; 1.010 to 0.996 [1% decrease] and 1.047 to
1.041 [0.5% decrease], respectively, vs. 1.018 to 0.991 [3% decrease]) (P<0.05 for all
comparisons). Exercise-related adverse events included musculoskeletal injuries.
CONCLUSIONS
Of the methods tested, weight loss plus combined aerobic and resistance exercise
was the most effective in improving functional status of obese older adults. (Funded
by the National Institutes of Health; LITOE ClinicalTrials.gov number, NCT01065636.)
n engl j med 376;20nejm.org May 18, 2017 1943
The New England Journal of Medicine
Downloaded from nejm.org by R LOPEZ MATA on May 19, 2017. For personal use only. No other uses without permission.
Copyright 2017 Massachusetts Medical Society. All rights reserved.
M
ore than a third of persons 65 Affairs Health Care System. The study was ap-
years of age or older in the United States proved by the institutional review board of the
are obese,1 and this group constitutes University of New Mexico School of Medicine
a population vulnerable to adverse outcomes, be- and was monitored by an independent data and
cause obesity exacerbates the age-related decline safety monitoring board; all participants pro-
A Quick Take is
available at in physical function and causes frailty.2-5 How- vided written informed consent. All the authors
NEJM.org ever, appropriate management of obesity in older had access to the data and vouch for the integ-
adults remains controversial, given the reported rity, accuracy, and completeness of the data and
reduction in relative health risks associated with analyses and for the fidelity of the study to the
increasing body-mass index in this group.6 More- protocol. The first author wrote the first draft of
over, an important concern is that weight loss the manuscript; all the authors participated in
could worsen frailty by accelerating the age- subsequent drafts and made the decision to sub-
related decline in muscle and bone mass and mit the manuscript for publication. No commer-
resultant sarcopenia and osteopenia.7,8 cial support was received.
Given the positive effects of exercise on physi-
cal function, healthy aging in obese older adults Participants
might require an intervention that involves regu- Participants were recruited through advertise-
lar exercise.9 We reported previously that exer- ments and underwent comprehensive medical
cise (combined aerobic and resistance training) screening. Persons were eligible for inclusion if
in combination with weight loss was associated they were 65 years of age or older, were obese
with greater improvement in physical function (body-mass index [the weight in kilograms divid-
than weight loss alone or exercise alone.10 How- ed by the square of the height in meters] 30),
ever, exercise attenuated but did not prevent loss were sedentary (regular exercise <1 hour per
in muscle and bone mass induced by weight loss week), and had had a stable body weight (loss or
and ameliorated but did not reverse frailty. The gain of no greater than 2 kg) and stable medica-
physiologic adaptations to aerobic and resistance tion use for 6 months before enrollment. All
exercise are distinctly different: aerobic exercise participants had mild-to-moderate frailty, as de-
improves cardiovascular adaptations that increase fined by a score of 18 to 31 on the modified
peak oxygen consumption without significantly Physical Performance Test (scores range from 0 to
changing strength, whereas resistance exercise 36 points, with higher scores indicating better
improves neuromuscular adaptations that increase performance).15 Persons who had severe cardio-
strength without significantly changing peak pulmonary disease (e.g., recent myocardial infarc-
oxygen consumption.11 These physiologic adap- tion or unstable angina), musculoskeletal or neu-
tations may interfere with each other when the romuscular impairments that precluded exercise
two types of training are performed together.12-14 training, or cognitive impairments or who used
In the current clinical trial, we compared the drugs that affect bone metabolism were excluded.
effectiveness of aerobic exercise, resistance exer-
cise, and combined exercise in reversing frailty Study Outcomes
and preserving muscle and bone mass during The primary outcome was the change in score
weight loss in obese older adults. We hypothe- on the Physical Performance Test from baseline
sized that weight loss plus resistance exercise to 6 months. Secondary outcomes were changes
would improve physical function more than in other frailty measures, body composition, bone
weight loss plus aerobic exercise or weight loss mineral density, specific physical functions, and
plus combined aerobic and resistance exercise. quality of life.
Intervention
Me thods
In this 26-week study, participants were ran-
Study Oversight domly assigned, with stratification according to
We conducted the Lifestyle Intervention Trial in sex, to one of four groups a control protocol
Obese Elderly (LITOE) from April 2010 through that included neither a weight-management nor
June 2015 at the University of New Mexico an exercise intervention, an aerobic group that
School of Medicine and New Mexico Veterans participated in a weight-management program
1944 n engl j med 376;20nejm.org May 18, 2017

Exercise Tr aining in Dieting Obese Older Adults
and aerobic exercise training, a resistance group weekly. The sessions were 75 to 90 minutes long
that participated in a weight-management pro- and included 10 minutes of flexibility exercises,
gram and resistance exercise training, and a followed by 30 to 40 minutes of aerobic exer-
combination group that participated in a weight- cises, 30 to 40 minutes of resistance exercises,
management program and combined aerobic and 10 minutes of balance exercises. To test the
and resistance exercise training. interference effect,12-14 aerobic and resistance
The control group was asked not to partici- training were balanced between groups: the lon-
pate in external weight-loss or exercise programs. ger duration of exercise in the combination group
However, this group attended group educational allowed the participants to perform an amount
sessions about a healthful diet during monthly of aerobic exercise that was equivalent to that of
visits. the aerobic group and an amount of resistance
The aerobic group participated in a weight- exercise that was equivalent to that of the resis-
management program, in which the participants tance group.
were prescribed a balanced diet that provided an Exercise sessions were supervised by exercise
energy deficit of 500 to 750 kcal per day and trainers. Participants were advised to maintain
contained approximately 1 g of high-quality pro- their usual physical activity outside of exercise
tein per kilogram of body weight per day.2 Par- sessions. All participants received supplements
ticipants met weekly with a dietitian for dietary to ensure an intake of approximately 1500 mg of
adjustments and behavioral therapy (diet therapy). calcium per day and approximately 1000 IU of
They were instructed to set weekly behavioral vitamin D per day.2
goals and attend weigh-in sessions. Food diaries
were reviewed, and goals were set on the basis Baseline Assessments
of diary reports. The goal was to achieve a Physical Function
weight loss of approximately 10% at 6 months. Frailty was assessed with the Physical Perfor-
They also participated in aerobic exercise train- mance Test. The Physical Performance Test in-
ing sessions three times weekly. The sessions cludes seven standardized tasks (walking 15.2 m
were approximately 60 minutes long and included [50 ft], putting on and removing a coat, picking
10 minutes of flexibility exercises, followed by up a penny, standing up from a chair, lifting a
40 minutes of aerobic exercises and 10 minutes book, climbing one flight of stairs, and perform-
of balance exercises. The aerobic exercises con- ing a progressive Romberg test) plus two addi-
sisted of treadmill walking, stationary cycling, tional tasks (going up and down four flights of
and stair climbing. Participants exercised at ap- stairs and making a 360-degree turn). The score
proximately 65% of their peak heart rate, which for each task ranges from 0 to 4, with higher
was gradually increased to 70 to 85%. scores indicating better physical performance; a
The resistance group participated in the same perfect score would be 36.15 Frailty was also as-
weight-management program as the aerobic sessed by measurement of peak oxygen consump-
group, as well as resistance exercise training tion and by administration of the Functional
sessions three times weekly; the sessions were Status Questionnaire. Peak oxygen consumption
approximately 60 minutes long and included 10 was assessed during graded treadmill walking,
minutes of flexibility exercises, followed by 40 min- as described previously.3 Ability to perform ac-
utes of resistance exercises and 10 minutes of tivities of daily living was assessed with the use
balance exercises. The resistance training con- of the Functional Status Questionnaire (scores
sisted of nine upper-body and lower-body exer- range from 0 to 36, with higher scores indicat-
cises using weight-lifting machines. The initial ing better functional status).16 We also assessed
sessions were 1 to 2 sets of 8 to 12 repetitions strength, balance, gait speed, and one-repetition
at 65% of the one-repetition maximum. This was maximum (the maximum weight a participant
increased progressively to 2 to 3 sets at approxi- can lift, in one attempt, in the biceps curl, bench
mately 85% of the one-repetition maximum. press, seated row, knee extension, knee flexion,
The combination group participated in the and leg press). We assessed static balance by
same weight-management program as the other measuring the time a participant could stand on
exercise groups, as well as combined aerobic and a single leg3 and dynamic balance by measuring
resistance exercise training sessions three times the time needed to complete an obstacle course.15

Fast gait speed was assessed by measurement of changes in the aerobic group would differ from
the time needed to walk 7.6 m (25 ft). those in the control group, that changes in the
resistance group would differ from those in the
Body Composition and Bone Mineral Density control group, that changes in the aerobic group
Fat mass, lean mass, and bone mineral density would differ from those in the resistance group,
of the whole body and at the lumbar spine and and that changes in the combination group would
total hip were measured with the use of dual- differ from those in the aerobic group and from
energy x-ray absorptiometry (Lunar DPX [Gen- those in the resistance group. For the Physical
eral Electric] or Discovery A [Hologic] scanner), Performance Test score, Bonferroni correction
as described previously.3,17 For each participant, was used to adjust for these comparisons. Within-
baseline and follow-up scans were obtained with group changes were analyzed with the use of
the use of the same instrument. Thigh muscle repeated-measures analysis of variance. Second-
and fat volumes were measured by magnetic ary outcomes were grouped into five domains. In
resonance imaging (Magnetom Avanto [Siemens]), accordance with a gatekeeping strategy,21 a sig-
as described previously.18 nificant group-by-time interaction (P<0.01) and
at least one significant difference between an
Quality of Life exercise group and the control group and at least
We used the physical and mental component sub- one significant difference among the exercise
scales of the Medical Outcomes Study 36-Item groups in the change in score on the Physical
Short-Form Health Survey (SF-36), version 2, to Performance Test were required to continue to
evaluate quality of life.19 Scores on the physical testing of secondary outcomes; comparisons of
and mental component subscales of the SF-36 the exercise groups with the control group were
range from 0 to 100, with higher scores indicat- performed with Dunnetts test and comparisons
ing better health status; the minimal clinically among the intervention groups were performed
important difference is 2 points.20 with the FisherHayter test22 (Fig. S1 in the Sup-
plementary Appendix, available with the full text
Follow-up Assessments of this article at NEJM.org). Sensitivity analyses
All baseline assessments were repeated at 6 months. that validated the statistical results included mul-
The Physical Performance Test was also repeated tiple imputation for missing fitness data. Data
at 3 months. Assessors were unaware of the for change scores and percentage changes are
study-group assignments. presented as least-squaresadjusted means (SE).
Statistical Analysis
R e sult s
We estimated that a sample size of 40 partici-
pants per group would provide 80% power to Study Population
detect a mean (SD) clinically important differ- A total of 160 participants underwent random-
ence between groups of 1.82.5 in the change in ization, and 141 participants (88%) completed
score on the Physical Performance Test, at an the study (Fig.1). Nineteen participants discon-
alpha level of 0.05. Intention-to-treat analyses tinued the intervention and were included in the
were performed with SAS software, version 9.4 intention-to-treat analyses (follow-up data were
(SAS Institute). Baseline characteristics were obtained for all 19 participants at 3 months but
compared with the use of analysis of variance or not at 6 months). There were no significant dif-
Fishers exact test. Longitudinal changes between ferences among the groups in baseline charac-
groups were tested with the use of mixed-model teristics (Table1).
repeated-measures analysis of variance, with ad- Median attendance at the diet-therapy sessions
justment for baseline values and sex. The pri- was 96% (interquartile range, 87 to 100) in the
mary focus of the analyses was the 6-month aerobic group, 100% (interquartile range, 90 to
change in outcome in the four groups. When the 100) in the resistance group, and 97% (inter-
overall P value for the interaction between group quartile range, 89 to 100) in the combination
and time was less than 0.05, prespecified com- group. Median attendance at exercise sessions
parisons were used to test five hypotheses: that was 96% (interquartile range, 84 to 100) in the

258 Persons were assessed for eligibility
98 Were excluded
73 Did not meet inclusion criteria
25 Declined to participate
160 Underwent randomization
40 Were assigned to the 40 Were assigned to the 40 Were assigned to the 40 Were assigned to the
control group aerobic group resistance group combination group
4 Discontinued intervention 5 Discontinued intervention 5 Discontinued intervention 5 Discontinued intervention

3 Lacked interest 2 Had difficulty adhering to 1 Had difficulty adhering to 3 Had personal reasons
1 Had medical reasons intervention intervention 1 Had a job-related reason
2 Had personal reasons 1 Had personal reasons 1 Had medical reasons
1 Had medical reasons 3 Had medical reasons
40 Were included in the analyses 40 Were included in the analyses 40 Were included in the analyses 40 Were included in the analyses
Figure 1. Screening, Randomization, and Follow-up.

The study groups included a control group that received neither a weight-management nor an exercise intervention and three exercise
groups: a group that received aerobic exercise training (aerobic group), a group that received resistance exercise training (resistance
group), and a group that received combined aerobic and resistance exercise training (combination group); all three exercise groups also
participated in a weight-management program.
aerobic group, 98% (interquartile range, 81 to aerobic or resistance groups: 27.9 to 33.4 points
100) in the resistance group, and 93% (inter- (a change of 5.50.4 points [21% increase from
quartile range, 83 to 100) in the combination the least-squares adjusted mean at baseline]) ver-
group. sus 29.3 to 33.2 points (a change of 3.90.4 points
[14% increase]) and 28.8 to 32.7 points (a change
Adverse Events of 3.90.4 points [14% increase]), respectively;
Exercise-related adverse events included falling, scores in all three exercise groups increased more
shoulder pain, and back pain in the aerobic than scores in the control group (4% increase)
group; atrial fibrillation, shoulder pain, and knee (Table2 and Fig.2A). Peak oxygen consumption
pain in the resistance group; and shoulder injury, (measured as milliliters per kilogram of body
knee pain, back pain, and hip pain in the com- weight per minute) increased more in the com-
bination group. There were no other differences bination and aerobic groups than in the resis-
in adverse events relative to the control group tance group: 17.2 to 20.3 (a change of 3.10.3
(Table S1 and S2 in the Supplementary Appendix). [17% increase]) and 17.6 to 20.9 (a change of
3.30.3 [18% increase]), respectively, versus 17.0
Physical Performance Test and Other Frailty to 18.3 (a change of 1.30.3 [8% increase]) (Table2
Measures and Fig.2B). Functional Status Questionnaire
The scores on the Physical Performance Test in- scores increased more in the combination group
creased more in the combination group than in than in the aerobic and resistance groups: 29.8

Table 1. Baseline Characteristics of the Participants.*
Control Aerobic Resistance Combination

Characteristic (N=40) (N=40) (N=40) (N=40)
Age yr 705 704 705 705
Sex no. (%)
Male 12 (30) 14 (35) 15 (38) 16 (40)
Female 28 (70) 26 (65) 25 (62) 24 (60)
Race no. (%)
White 36 (90) 36 (90) 33 (82) 35 (88)
Black 1 (2) 0 3 (8) 2 (5)
Other 3 (8) 4 (10) 4 (10) 3 (8)
Ethnic group no. (%)
Hispanic or Latino 13 (32) 13 (32) 12 (30) 12 (30)
Not Hispanic or Latino 27 (68) 27 (68) 27 (68) 28 (70)
Unknown 0 0 1 (2) 0
Education no. (%)
Less than a college degree 14 (35) 13 (32) 15 (38) 10 (25)
College degree 20 (50) 20 (50) 12 (30) 19 (48)
Graduate school 6 (15) 7 (18) 13 (32) 11 (28)
Marital status no. (%)
Single 8 (20) 4 (10) 9 (22) 5 (12)
Married 15 (38) 25 (62) 22 (55) 23 (58)
Divorced 9 (22) 7 (18) 6 (15) 8 (20)
Widowed 8 (20) 4 (10) 3 (8) 4 (10)
Body-mass index 36.75.0 35.94.4 36.75.8 35.84.5
Chronic diseases no. 3.23.4 2.62.7 3.63.9 2.72.0
Routine medications no. 5.54.3 5.73.4 6.25.7 5.74.0
* Plusminus values are means SD. There were no significant between-group differences in baseline characteristics.
The study groups included a control group that received neither a weight-management nor an exercise intervention and
three exercise groups: a group that received aerobic exercise training (aerobic group), a group that received resistance
exercise training (resistance group), and a group that received combined aerobic and resistance exercise training (combi-
nation group); all three exercise groups also participated in a weight-management program. Percentages may not total
100 because of rounding.
Race and ethnic group were reported by the participants.
The body-mass index is the weight in kilograms divided by the square of the height in meters.
Chronic diseases included hypertension, diabetes, coronary artery disease, congestive heart failure, arthritis, and chronic
lung disease.
Routine medications included antihypertensive, antidiabetic, antidyspeptic, antianginal, diuretic, antiarthritic, antilipidemic,
and antidepressant medications.
to 33.4 points (a change of 3.60.3 points [14% in the resistance group (101.8 to 93.3 kg; a
increase]) versus 30.1 to 32.1 points (a change of change of 8.50.5 kg [9% decrease]), and in the
2.00.3 points [7% increase]) and 29.3 to 31.6 combination group (99.0 to 90.5 kg; a change of
points (a change of 2.30.3 points [8% increase]), 8.50.5 kg [9% decrease]), but there was no
respectively (Table2 and Fig.2C). significant change in body weight in the control
group (97.9 to 97.0 kg; a change of 0.90.5 kg
Body Composition [<1% decrease]) (Table2). The time course of
Body weight decreased in the aerobic group (96.9 weight loss is shown in Figure3. Lean mass de-
to 87.9 kg; a change of 9.00.6 kg, [9% decrease]), creased less in the combination group (56.5 kg

to 54.8 kg; a change of 1.70.3 kg [3% decrease]) increase]) than in the aerobic group (48.6 to 55.1
and in the resistance group (58.1 to 57.1 kg; a points; a change of 6.50.7 points [14% increase])
change of 1.00.3 kg [2% decrease]) than in and the resistance group (51.0 to 58.4 points; a
the aerobic group (55.0 to 52.3 kg; a change of change of 7.40.8 points [17% increase]) (Table
2.70.3 kg [5% decrease]) (Table2 and Fig.2D). S3 in the Supplementary Appendix).
Fat mass decreased by 6.30.5 kg in the aerobic
group (41.9 to 35.6 kg [16% decrease]), 7.30.4 kg Discussion
in the resistance group (44.3 to 37.0 kg [17%
decrease]), and 7.00.5 kg in the combination Our randomized, controlled trial involving obese
group (42.5 to 35.5 kg [17% decrease]). The adults 65 years of age or older indicated that
changes in thigh muscle and thigh fat among weight loss plus a combination of aerobic and
the exercise groups were similar to those ob- resistance exercise improved physical function
served for lean mass and fat mass, respectively. and reduced frailty more than weight loss plus
aerobic exercise or weight loss plus resistance
Bone Mineral Density exercise. Evidence-based data to guide treatment
Bone mineral density at the total hip did not of older adults with obesity are limited5,23,24 and
change significantly in the resistance group tend to rely on studies involving younger adults.2
(1.047 to 1.041 g per square centimeter; a Our study directly compared aerobic, resistance,
change of 0.0060.004 g per square centimeter and combined (aerobic and resistance) training
[<1% decrease]), whereas it decreased in the during weight loss in obese older adults. The
aerobic group (1.018 to 0.991 g per square centi- matched weight loss across groups facilitated the
meter; a change of 0.0270.004 g per square assessment of the independent and combined
centimeter [2.6% decrease]) and in the combina- effects of aerobic and resistance training. Despite
tion group (1.010 to 0.996 g per square centi- a negative energy balance, aerobic training im-
meter; a change of 0.0140.004 g per square proved cardiovascular fitness and resistance train-
centimeter [1.1% decrease]) (Table2 and Fig.2E). ing improved strength. Contrary to our hypoth-
Bone mineral density of the whole body and at esis, combined aerobic and resistance training
the lumbar spine did not change significantly in improved cardiovascular fitness to the same ex-
any of the study groups (Table S3 in the Supple- tent as aerobic training alone and strength to
mentary Appendix). the same extent as resistance training alone.
Therefore, combined aerobic and resistance
Strength, Balance, Gait, and Quality of Life training resulted in additive effects that trans-
Total one-repetition maximum strength increased lated into the greatest improvement among the
in the resistance group (288 to 337 kg; a change interventions in physical function and reduction
of 495 kg [19% increase]) and in the combina- of frailty. Both resistance training and com-
tion group (272 to 320 kg; a change of 485 kg bined resistance and aerobic training attenuat-
[18% increase]), whereas it was maintained in ed the loss of lean mass during aerobic train-
the aerobic group (265 to 270 kg; a change of ing. Moreover, although only resistance training
55 kg [4% increase]) (Table2 and Fig.2F). prevented the weight-lossinduced reduction in
Time needed to complete the obstacle course bone mineral density at the total hip, combined
decreased more in the combination group (17.0 aerobic and resistance training nonetheless at-
to 14.1 seconds; a change of 2.90.3 seconds tenuated the loss of bone mineral density at the
[13% decrease]) than in the aerobic group (15.5 total hip during aerobic training. Our data sug-
to 14.0 seconds; a change of 1.50.4 seconds gest that obese older adults can adapt and re-
[7% decrease]). Gait speed increased more in the spond to exercise training during an energy defi-
combination group (68.8 to 80.9 m per minute; cit and that combined aerobic and resistance
a change of 12.11.3 m per minute [14% increase]) training provides the greatest benefits with re-
than in the aerobic group (74.6 to 82.7 m per spect to physical function, with relative preserva-
minute; a change of 8.11.3 m per minute [9% tion of lean mass.
increase]). The SF-36 physical-component score Our findings in obese older adults expand
increased more in the combination group (45.9 observations of the positive effects of exercise
to 55.4 points; a change of 9.50.7 points [24% training without weight loss in nonobese older

1950
Table 2. Effect of Specific Exercise Modes, Added to Diet-Induced Weight Loss, on Outcomes.*
Control Aerobic Resistance Combination

Outcome (N=40) (N=40) (N=40) (N=40) P Value
Aerobic Resistance Aerobic Combination Combination

GroupTime vs. vs. vs. vs. vs.
Interaction Control Control Resistance Aerobic Resistance
Primary outcome
PPT score
Baseline 28.60.5 29.30.3 28.80.4 27.90.4
Change at 3 mo 0.40.3 2.90.4 3.20.4 4.00.4
Change at 6 mo 1.00.4 3.90.4 3.90.4 5.50.4 <0.001 <0.001 <0.001 0.87 0.002 0.004
The
Secondary outcomes
Other frailty measures
VO2peak (ml/kg/min)
Baseline 17.00.5 17.60.5 17.00.6 17.20.6
Change at 6 mo 0.10.3 3.30.3 1.30.3 3.10.3 <0.001 <0.001 0.007 <0.001 0.63 0.001
FSQ score
Baseline 29.80.5 30.10.5 29.30.6 29.80.6
Change at 6 mo 0.40.3 2.00.3 2.30.3 3.60.3 <0.001 0.002 <0.001 0.46 0.005 0.03
Body composition
n e w e ng l a n d j o u r na l
Body weight (kg)

of
Baseline 97.92.9 96.92.3 101.82.9 99.02.9

Change at 6 mo 0.90.5 9.00.6 8.50.5 8.50.5 <0.001 <0.001 <0.001 0.76 0.72 0.96
n engl j med 376;20nejm.org May 18, 2017

Lean mass (kg)
Baseline 54.92.3 55.01.9 58.12.3 56.51.8

m e dic i n e
Change at 6 mo 0.00.2 2.70.3 1.00.3 1.70.3 <0.001 <0.001 0.03 0.001 0.047 0.20
Fat mass (kg)
Baseline 43.01.5 41.91.3 44.31.5 42.51.6
Change at 6 mo 0.90.4 6.30.5 7.30.4 7.00.5 <0.001 <0.001 <0.001 0.18 0.36 0.67
Thigh muscle (cm3)
Baseline 130263 123462 119048 118666
Change at 6 mo 107 777 237 407 <0.001 <0.001 0.008 <0.001 0.005 0.21
Thigh fat (cm3)
Baseline 1774132 170098 1848108 1784125
Change at 6 mo 236 26035 28035 28835 <0.001 <0.001 <0.001 0.64 0.61 0.97
BMD at total hip (g/cm 2)
Baseline 1.0310.025 1.0180.019 1.0470.022 1.0100.025

Change at 6 mo 0.0040.004 0.0270.004 0.0060.004 0.0140.004 0.001 <.001 0.37 0.005 0.04 0.43
Strength, balance, and gait
Total 1RM (kg)
Baseline 26921 26520 28830 27216
Change at 6 mo 25 5.05 495 485 <0.001 0.99 <0.001 <0.001 <0.001 0.82
Obstacle course (sec)
Baseline 15.90.7 15.50.8 16.40.6 17.01.0
Change at 6 mo 0.00.3 1.50.4 2.20.3 2.90.3 <0.001 0.02 <0.001 0.20 0.01 0.20
Gait speed (m/min)
Baseline 75.02.6 74.62.0 74.32.3 68.82.2
Change at 6 mo 0.51.3 8.11.3 9.31.3 12.11.3 <0.001 <0.001 <0.001 0.56 0.03 0.09
* Plusminus values for the change scores are the least-squares adjusted means SE from the repeated-measures analyses of variance; plusminus values for the baseline values are the
observed means SE. Scores on the Physical Performance Test (PPT) (primary outcome) range from 0 to 36, with higher scores indicating better physical function; the minimal clinically
important difference is 1.8. Peak oxygen consumption (VO2peak) was assessed during graded treadmill walking. Scores on the Functional Status Questionnaire (FSQ) range from 0 to
36, with higher scores indicating better function. BMD denotes bone mineral density.
P values for the changes from baseline to 6 months in between-group comparisons were calculated with the use of mixed-model repeated-measures analyses of variance (with baseline
n engl j med 376;20nejm.org May 18, 2017

values and sex as covariates) and are reported when the overall P value was lower than 0.05 for the interaction among the four groups over time. In a Bonferroni correction to adjust for
the multiple comparisons in the PPT score (in which the P values were multiplied by 5 for the comparison with an alpha level of 0.05), the corrected P values were 0.01 for the combina-
tion group versus the aerobic group and 0.02 for the combination group versus the resistance group. In accordance with a gatekeeping strategy, a significant group-by-time interaction
(P<0.01) and at least one significant difference between an exercise group and the control group and at least one significant difference among the exercise groups in the change in PPT
score were required to continue to testing of the secondary outcomes; comparisons of the exercise groups with the control group were performed with Dunnetts test and comparisons
among the intervention groups were performed with the FisherHayter test. Secondary analyses included a comparison between the combination group and the control group; all P val-

ues were less than 0.05.
P<0.001 for the comparison of the value at the follow-up time with the baseline value within the group, as calculated with the use of mixed-model repeated-measures analysis of variance.
The between-group differences were not significant in the last three columns because the group-by-time interaction for the three intervention groups (step 1 of the FisherHayter test)
was not significant; the family-wise error rate of 0.05 or less is maintained for the FisherHayter multiple comparison procedure.
P<0.01 for the comparison of the value at the follow-up time with the baseline value within the group, as calculated with the use of mixed-model repeated-measures analysis of variance.
Total one-repetition maximum (1RM) is the total of the maximum weight a participant can lift, in one attempt, in the biceps curl, bench press, seated row, knee extension, knee flexion,
and leg press.
1951
A PPT Score B Peak Oxygen Consumption

25 25

20 *
Change from Baseline (%)

20
* 15
15 *
10 *
10
5
5
0
0 5
n
c
c
l
l
nc
nc
bi
bi
io
io
ro
ro
ro
ro
t
t
nt
nt
ta
ta
na
na
Ae
Ae
sis
sis
Co
Co
bi
bi
Re
Re
m
m
Co
Co
C FSQ Score D Lean Mass
16 1
14 0

12
1
10 *
* 2
8 *
3
6
4
4
2 5
*
0 6
e
n
c
c
l
l
nc
nc
bi
bi
tio
tio
ro
ro
ro
ro
nt
nt
ta
ta
na
na
Ae
Ae
sis
sis
Co
Co
bi
bi
Re
Re
m
m
Co
Co
E BMD at Total Hip F Strength
1 25
*
0 20
1 15
2 10
3 * 5
4 0
ce
ce
n
ic
ic
l
l
tio
tio
ro
ro
b
b
an
an
ro
ro
nt
nt
na
na
ist
ist
Ae
Ae
Co
Co
bi
bi
s
s
Re
Re
m
m
Co
Co
adults25,26 and support the results of previous the aerobic and resistance training to be moder-
studies that showed that exercise training was ate to vigorous in intensity to induce exercise
most beneficial in frail older adults in the ear- adaptations29 while keeping exercise volumes
lier stages of frailty.27 Given the exercise goals moderate.27 Using these exercise strategies, we
for our frail and obese participants,28 we designed found additive effects of aerobic and resistance

Figure 2 (facing page). Mean Percentage Changes in 2

Physical Function, Lean Mass, and BMD at the Total
Hip during the Interventions. 0
Measures of physical function included the Physical
Performance Test (PPT; scores range from 0 to 36, with 2
Weight Change (%)

higher scores indicating better functional status), peak
oxygen consumption, Functional Status Questionnaire 4
(FSQ; scores range from 0 to 36, with higher scores in-
dicating better functional status), and strength (measured 6
as total one-repetition maximum [i.e., the total of the
maximum weight a participant can lift, in one attempt, 8 Control
in the biceps curl, bench press, seated row, knee exten- Aerobic
sion, knee flexion, and leg press]). Scores on the PPT Resistance
10
were used as an objective measure of frailty (primary
Combination
outcome), and scores on the FSQ were used as a sub-
jective measure of frailty. The asterisk indicates P<0.05 12
0 2 4 6 8 10 12 14 16 18 20 22 24 26
for the comparison with the control group, the dagger
P<0.05 for the comparison with the aerobic group, and Weeks of Intervention
the double dagger P<0.05 for the comparison with the
resistance group. Percentage changes are presented as Figure 3. Mean Percent Changes in Body Weight during the Interventions.
least-squaresadjusted means; T bars indicate standard Percent changes are presented as least-squaresadjusted means; I bars in-
errors. BMD denotes bone mineral density. dicate standard errors.
training without interference effect from con- effect in that these reductions further diminish
current training.12-14 Adherence to exercise was tissue reserves superimposed on age-related
high despite frailty, and adverse events were losses. However, resistance training improved
relatively few and consistent with coexisting med- strength despite muscle loss induced by weight
ical conditions. Our findings suggest that the loss. Conversely, whether improved physical
recommendation by the American Heart Asso- function lowers the risks of falls and fractures
ciation and American College of Sports Medicine despite the decline in bone mineral density is
to combine aerobic exercise with resistance exer- currently unclear. In future studies, additional
cise for overall health30 extends to obese older strategies to preserve lean mass might include
adults undertaking weight loss. improving the efficiency of vitamin D and pro-
The improvements in objective measures of tein intake, increasing weight-bearing exercis-
frailty in our participants may have important es, and perhaps administering anabolic hor-
implications for preserving independent living. mone therapy.35,36 Another adverse effect was
The Physical Performance Test assesses multiple exercise-related musculoskeletal injuries, which
domains of physical function15 and predicts dis- could be minimized through individualized ex-
ability, loss of independence, and death.31,32 The ercises.
peak oxygen consumption relative to body weight Strengths of our study include the random-
is the best indicator of cardiovascular endur- ized, controlled trial design, the comprehensive
ance33 and is important for performing daily lifestyle programs, the high rate of adherence to
tasks with increased body weight.3,34 Improve- the trial interventions, the similar weight-loss
ments in the objective Physical Performance Test management that allowed for unbiased group
score and peak oxygen consumption were con- comparisons, and the use of objective and sub-
sistent with improvements in Functional Status jective measures of physical function. Because
Questionnaire and SF-36 scores, which indicate this was an efficacy study, the 6-month duration
subjective improvements in functional ability. was appropriate to determine which exercise was
Although combined aerobic and resistance most efficacious in improving physical function
training improved physical function the most during weight loss. Data from long-term studies
among the interventions, the reductions in lean that show whether weight loss plus combined
mass and bone mineral density that were attenu- aerobic and resistance training prolongs physi-
ated but not prevented might represent an adverse cal independence in obese older adults are cur-

rently lacking. The findings from our study may loss plus resistance training or aerobic training
have pragmatic implications, because Medicare improved physical function and ameliorated frail-
currently covers behavioral therapy for weight ty; however, weight loss plus combined aerobic
loss,37 and a growing number of Medicare plans and resistance training provided greater improve-
now offer gym memberships.38 Data from a pre- ment in physical function and reduction of frailty
vious trial show that older adults may be more than either intervention alone and was associated
successful in achieving long-term weight loss than with relative preservation of lean mass.
younger adults.39
The content is solely the responsibility of the authors and
Our study has limitations. First, in accordance does not necessarily represent the official views of the National
with the exclusion criteria, the participants in Institutes of Health and the Department of Veterans Affairs (VA).
our study were physically able to participate in a Supported by grants from the National Institutes of Health
(RO1-AG031176, UL1-TR000041, and P30-DK020579).
lifestyle program and thus may not be fully rep- Disclosure forms provided by the authors are available with
resentative of the general obese older adult popu- the full text of this article at NEJM.org.
lation. Second, our sample was not large enough We thank the participants for their cooperation, Kenneth Fowl-
er for study coordination, Brandy Martinez and Erik Faria for ex-
to analyze differences according to sex. Finally, ercise training, and Ronni Farris and Reed Vawter for weight-loss
most of the participants were women, white, and training. We also thank the members of the Alkek Foundation for
well educated, which limits broader generali their support. The findings reported in this article are the result
of work supported with resources and the use of facilities at the
zation. New Mexico VA Health Care System and Michael E. DeBakey VA
In conclusion, our study showed that weight Medical Center.
References
1. Flegal KM, Kruszon-Moran D, Carroll 11. Lambert CP, Evans WJ. Adaptations derly population. Age Ageing 1994; 23:
MD, Fryar CD, Ogden CL. Trends in obe- to aerobic and resistance exercise in the 182-4.
sity among adults in the United States, elderly. Rev Endocr Metab Disord 2005;6: 20. Maruish ME, ed. Users manual for
2005 to 2014. JAMA 2016;315:2284-91. 137-43. the SF-36v2 health survey. 3rd ed. Lin-
2. Villareal DT, Apovian CM, Kushner 12. Hickson RC. Interference of strength coln, RI:QualityMetric, 2011.
RF, Klein S. Obesity in older adults: tech- development by simultaneously training 21. Dmitrienko A, Tamhane AC. Mixtures
nical review and position statement of for strength and endurance. Eur J Appl of multiple testing procedures for gate-
the American Society for Nutrition and Physiol Occup Physiol 1980;45:255-63. keeping applications in clinical trials. Stat
NAASO, The Obesity Society. Am J Clin 13. Wilson JM, Marin PJ, Rhea MR, Wilson Med 2011;30:1473-88.
Nutr 2005;82:923-34. SM, Loenneke JP, Anderson JC. Concur- 22. Hayter AJ. The maximum familywise
3. Villareal DT, Banks M, Siener C, Sina- rent training: a meta-analysis examining error rate of Fishers least significant dif-
core DR, Klein S. Physical frailty and body interference of aerobic and resistance ex- ference test. J Am Stat Assoc 1986;12:
composition in obese elderly men and ercises. J Strength Cond Res 2012; 26: 1000-4.
women. Obes Res 2004;12:913-20. 2293-307. 23. Locher JL, Goldsby TU, Goss AM,
4. Blaum CS, Xue QL, Michelon E, Sem- 14. Cadore EL, Pinto RS, Lhullier FL, et al. Kilgore ML, Gower B, Ard JD. Calorie re-
ba RD, Fried LP. The association between Physiological effects of concurrent train- striction in overweight older adults: do
obesity and the frailty syndrome in older ing in elderly men. Int J Sports Med 2010; benefits exceed potential risks? Exp Geron-
women: the Womens Health and Aging 31:689-97. tol 2016;86:4-13.
Studies. J Am Geriatr Soc 2005;53:927-34. 15. Brown M, Sinacore DR, Binder EF, 24. Batsis JA, Gill LE, Masutani RK, et al.
5. Porter Starr KN, McDonald SR, Bales Kohrt WM. Physical and performance Weight loss interventions in older adults
CW. Obesity and physical frailty in older measures for the identification of mild to with obesity: a systematic review of ran-
adults: a scoping review of lifestyle inter- moderate frailty. J Gerontol A Biol Sci domized controlled trials since 2005. J Am
vention trials. J Am Med Dir Assoc 2014; Med Sci 2000;55:M350-M355. Geriatr Soc 2017;65:257-68.
15:240-50. 16. Jette AM, Cleary PD. Functional dis- 25. Liu CJ, Latham NK. Progressive resis-
6. Waters DL, Ward AL, Villareal DT. ability assessment. Phys Ther 1987; 67: tance strength training for improving phys-
Weight loss in obese adults 65 years and 1854-9. ical function in older adults. Cochrane
older: a review of the controversy. Exp 17. Villareal DT, Fontana L, Weiss EP, et al. Database Syst Rev 2009;3:CD002759.
Gerontol 2013;48:1054-61. Bone mineral density response to caloric 26. Binder EF, Schechtman KB, Ehsani
7. Roubenoff R. Sarcopenic obesity: the restriction-induced weight loss or exercise- AA, et al. Effects of exercise training on
confluence of two epidemics. Obes Res induced weight loss: a randomized con- frailty in community-dwelling older adults:
2004;12:887-8. trolled trial. Arch Intern Med 2006;166: results of a randomized, controlled trial.
8. Miller SL, Wolfe RR. The danger of 2502-10. J Am Geriatr Soc 2002;50:1921-8.
weight loss in the elderly. J Nutr Health 18. Weiss EP, Racette SB, Villareal DT, et al. 27. Theou O, Stathokostas L, Roland KP,
Aging 2008;12:487-91. Lower extremity muscle size and strength et al. The effectiveness of exercise inter-
9. Evans WJ. Exercise as the standard of and aerobic capacity decrease with caloric ventions for the management of frailty:
care for elderly people. J Gerontol A Biol restriction but not with exercise-induced a systematic review. J Aging Res 2011;
Sci Med Sci 2002;57:M260-M261. weight loss. J Appl Physiol (1985) 2007; 2011:569194.
10. Villareal DT, Chode S, Parimi N, et al. 102:634-40. 28. Aguirre LE, Villareal DT. Physical ex-
Weight loss, exercise, or both and physi- 19. Lyons RA, Perry HM, Littlepage BN. ercise as therapy for frailty. Nestle Nutr
cal function in obese older adults. N Engl Evidence for the validity of the Short- Inst Workshop Ser 2015;83:83-92.
J Med 2011;364:1218-29. form 36 Questionnaire (SF-36) in an el- 29. American College of Sports Medicine

Position Stand: the recommended quan- tery assessing lower extremity function: olism in obesity and weight loss. Annu
tity and quality of exercise for developing association with self-reported disability Rev Nutr 2012;32:287-309.
and maintaining cardiorespiratory and and prediction of mortality and nursing 37. Centers for Medicare and Medicaid
muscular fitness, and flexibility in healthy home admission. J Gerontol 1994;49:M85- Services. Decision memo for intensive be-
adults. Med Sci Sports Exerc 1998;30:975- M94. havioral therapy for obesity (CAG-00423N).
91. 33. Binder EF, Birge SJ, Spina R, et al. Baltimore:Centers for Medicare & Medic-
30. Nelson ME, Rejeski WJ, Blair SN, et al. Peak aerobic power is an important com- aid Services, November 21, 2011.
Physical activity and public health in older ponent of physical performance in older 38. Nguyen HQ, Maciejewski ML, Gao S,
adults: recommendation from the Ameri- women. J Gerontol A Biol Sci Med Sci Lin E, Williams B, Logerfo JP. Health care
can College of Sports Medicine and the 1999;54:M353-M356. use and costs associated with use of a
American Heart Association. Circulation 34. Shah K, Wingkun NJ, Lambert CP, health club membership benefit in older
2007;116:1094-105. Villareal DT. Weight-loss therapy improves adults with diabetes. Diabetes Care 2008;
31. Reuben DB, Siu AL. An objective mea- endurance capacity in obese older adults. 31:1562-7.
sure of physical function of elderly outpa- J Am Geriatr Soc 2008;56:1157-9. 39. Svetkey LP, Clark JM, Funk K, et al.
tients: the Physical Performance Test. J Am 35. Bouchonville MF, Villareal DT. Sarcope- Greater weight loss with increasing age in
Geriatr Soc 1990;38:1105-12. nic obesity: how do we treat it? Curr Opin the weight loss maintenance trial. Obesity
32. Guralnik JM, Simonsick EM, Ferrucci Endocrinol Diabetes Obes 2013;20:412-9. (Silver Spring) 2014;22:39-44.
L, et al. A short physical performance bat- 36. Shapses SA, Sukumar D. Bone metab- Copyright 2017 Massachusetts Medical Society.


Aerobic or Resistance Exercise, or Both, in Dieting Obese Older Adults

Uploaded by

Copyright:

Available Formats

Aerobic or Resistance Exercise, or Both, in Dieting Obese Older Adults

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aerobic or Resistance Exercise, or Both, in Dieting Obese Older Adults

Uploaded by

Copyright:

Available Formats

What was the primary outcome measured in the study?

What was the primary outcome measured in the study?

What exercise modalities were compared in the study?

What exercise modalities were compared in the study?

The n e w e ng l a n d j o u r na l of m e dic i n e

Aerobic or Resistance Exercise, or Both,

1944 n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

n engl j med 376;20nejm.org May 18, 2017 1945

1946 n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

258 Persons were assessed for eligibility

160 Underwent randomization

4 Discontinued intervention 5 Discontinued intervention 5 Discontinued intervention 5 Discontinued intervention

Figure 1. Screening, Randomization, and Follow-up.

n engl j med 376;20nejm.org May 18, 2017 1947

Table 1. Baseline Characteristics of the Participants.*

Control Aerobic Resistance Combination

1948 n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

n engl j med 376;20nejm.org May 18, 2017 1949

Control Aerobic Resistance Combination

Aerobic Resistance Aerobic Combination Combination

Body weight (kg)

The New England Journal of Medicine

Baseline 97.92.9 96.92.3 101.82.9 99.02.9

n engl j med 376;20nejm.org May 18, 2017

Copyright 2017 Massachusetts Medical Society. All rights reserved.

Baseline 1.0310.025 1.0180.019 1.0470.022 1.0100.025

n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

Copyright 2017 Massachusetts Medical Society. All rights reserved.

A PPT Score B Peak Oxygen Consumption

Change from Baseline (%)

Change from Baseline (%)

Change from Baseline (%)

Change from Baseline (%)

1952 n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

Figure 2 (facing page). Mean Percentage Changes in 2

Weight Change (%)

n engl j med 376;20nejm.org May 18, 2017 1953

1954 n engl j med 376;20nejm.org May 18, 2017

The New England Journal of Medicine

n engl j med 376;20nejm.org May 18, 2017 1955

You might also like