Content uploaded by Michael Rosenbaum
Author content
All content in this area was uploaded by Michael Rosenbaum on Jun 09, 2015
Content may be subject to copyright.
Volume 332 MARCH 9, 1995 Number 10
Copyright, 1995, by the Massachusetts Medical Society
CHANGES IN ENERGY EXPENDITURE RESULTING FROM ALTERED BODY WEIGHT
R
UDOLPH
L. L
EIBEL
, M.D., M
ICHAEL
R
OSENBAUM
, M.D.,
AND
J
ULES
H
IRSCH
, M.D.
Abstract
Background.
No current treatment for obesi-
ty reliably sustains weight loss, perhaps because com-
pensatory metabolic processes resist the maintenance of
the altered body weight. We examined the effects of ex-
perimental perturbations of body weight on energy ex-
penditure to determine whether they lead to metabolic
changes and whether obese subjects and those who
have never been obese respond similarly.
Methods.
We repeatedly measured 24-hour total en-
ergy expenditure, resting and nonresting energy expend-
iture, and the thermic effect of feeding in 18 obese sub-
jects and 23 subjects who had never been obese. The
subjects were studied at their usual body weight and after
losing 10 to 20 percent of their body weight by underfeed-
ing or gaining 10 percent by overfeeding.
Results.
Maintenance of a body weight at a level 10
percent or more below the initial weight was associated
with a mean (
!
SD) reduction in total energy expenditure
of 6
!
3 kcal per kilogram of fat-free mass per day in
the subjects who had never been obese (P
"
0.001) and
8
!
5 kcal per kilogram per day in the obese subjects
(P
"
0.001). Resting energy expenditure and nonresting
energy expenditure each decreased 3 to 4 kcal per kilo-
gram of fat-free mass per day in both groups of subjects.
Maintenance of body weight at a level 10 percent above
the usual weight was associated with an increase in total
energy expenditure of 9
!
7 kcal per kilogram of fat-free
mass per day in the subjects who had never been obese
(P
"
0.001) and 8
!
4 kcal per kilogram per day in the
obese subjects (P
"
0.001). The thermic effect of feeding
and nonresting energy expenditure increased by approx-
imately 1 to 2 and 8 to 9 kcal per kilogram of fat-free mass
per day, respectively, after weight gain. These changes in
energy expenditure were not related to the degree of ad-
iposity or the sex of the subjects.
Conclusions.
Maintenance of a reduced or elevated
body weight is associated with compensatory changes in
energy expenditure, which oppose the maintenance of a
body weight that is different from the usual weight. These
compensatory changes may account for the poor long-
term efficacy of treatments for obesity. (N Engl J Med
1995;332:621-8.)
From the Laboratory of Human Behavior and Metabolism, Rockefeller Univer-
sity, 1230 York Ave., New York, NY 10021, where reprint requests should be ad-
dressed to Dr. Leibel.
Supported in part by grants from the National Institutes of Health (DK30583
and GCRCRR00102) and the Weight Watchers Foundation. During part of the
study period, Dr. Rosenbaum was an Amparo Rugarcia Clinical Scholar, and Dr.
Leibel was an Established Investigator of the American Heart Association.
O
BESITY is a common and intractable problem in
some modern societies. Body weight is normally
regulated by integrated, coordinate effects on food in-
take and energy expenditure.
1
The high rate of recidi-
vism among obese people who lose weight may reflect
the operation of such regulatory processes.
2-4
In humans, total energy expenditure is accounted for
by resting energy expenditure (approximately 60 per-
cent of total energy expenditure), which is the metabol-
ic cost of processes such as the maintenance of trans-
membrane ion gradients and resting cardiopulmonary
activity; the thermic effect of feeding (approximately
10 percent of total energy expenditure), which is the
energy expended in the digestion, transport, and depo-
sition of nutrients; and nonresting energy expenditure
(approximately 30 percent of total energy expendi-
ture), which is all the remaining expenditure of energy,
mainly in the form of physical activity.
5
In an earlier
study, we found a persistent 28 percent decrease in the
energy expended per unit of body-surface area in for-
merly obese patients with a stable reduced weight,
6
suggesting a metabolic resistance to the maintenance
of a reduced body weight. The present study was de-
signed to examine the components of energy expendi-
ture during the maintenance of usual and altered body
weight in obese subjects and subjects who had never
been obese.
M
ETHODS
Subjects
Eighteen obese subjects (11 women and 7 men; mean [
!
SD] age,
29
!
10 years [range, 21 to 45]) and 23 subjects who had never been
obese (7 women and 16 men; mean age, 26
!
10 years [range, 19 to
41]) were recruited through physicians’ referrals or advertisements
(Table 1). All subjects were at their maximal lifetime weight and had
maintained this weight within a range of 2 kg for at least six months.
None were taking medications or on special diets. Subjects whose
body-mass index (expressed as the weight in kilograms divided by
the square of the height in meters) was higher than 28.0 were clas-
sified as obese.
7
All subjects had normal findings on physical exami-
nation and laboratory evaluations, including thyroid-function tests,
complete blood count, tests for hepatitis A and B and human im-
munodeficiency virus infection, biochemical tests, and urinalysis. Six
of the obese subjects and two of those who had never been obese
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
622 THE NEW ENGLAND JOURNAL OF MEDICINE March 9, 1995
smoked 2 to 10 cigarettes daily throughout the study. The protocol
was approved by the Rockefeller University Hospital Institutional Re-
view Board, and written informed consent was obtained from all the
subjects.
Study Design
The subjects were admitted to the Clinical Research Center at
Rockefeller University and fed a liquid formula (40 percent fat [corn
oil], 45 percent carbohydrate [glucose polymer], and 15 percent pro-
tein [casein hydrolysate]) supplemented with 5.0 g of iodized sodium
chloride, 1.9 g of potassium ions as a potassium salt, and 2.5 g of cal-
cium carbonate per day, 1 mg of folic acid twice weekly, and 36 mg
of ferrous iron every other day. The mean caloric content of this for-
mula, measured with a bomb calorimeter, was 1.36 kcal per gram.
With the use of standard digestibility quotients, the content of metab-
olizable calories was 1.25 kcal per gram.
8
Fecal calorie and urinary
nitrogen losses were measured at all weight plateaus to confirm that
they did not change (see below). The caloric intake was adjusted until
the body weight was constant (slope of the regression line of body
weight [grams] vs. time [days],
"
10 g per day) for at least 14 days.
All subjects then underwent studies of energy expenditure and body
composition during approximately a 10-day period while continuing
to ingest the same quantity of dietary formula. Body composition was
analyzed by hydrodensitometry
9
; stool and urine samples, collected
for eight days, were analyzed to determine fecal calorie loss (by bomb
calorimetry) and urinary nitrogen excretion
10
; and resting energy
expenditure and the thermic effect of feeding were determined by
indirect calorimetry with the use of a Beckman MMC Horizon Met-
abolic Cart (Beckman Instruments, Fullerton, Calif.) with a ventilat-
ed hood fitted snugly around the subject’s neck
11
(and unpublished
data).
Resting energy expenditure at 8 a.m. in the postabsorptive state
was calculated from oxygen consumption corrected for the respirato-
ry quotient and the daily rate of nitrogen excretion based on the ni-
trogen content of the dietary formula and the rate of urinary nitrogen
excretion. To measure the thermic effect of feeding, at 9 a.m. the sub-
jects were given dietary formula with a caloric content equal to 60
percent of the 24-hour resting energy expenditure measured that
morning. Oxygen consumption and carbon dioxide production were
measured in the hood calorimeter for 30 minutes 2 and 4 hours after
the feeding. The area of the polygon whose base is the prefeeding val-
ue of resting energy expenditure and whose other vertexes are resting
energy expenditure at 9 a.m., 11 a.m., and 1 p.m. represents the in-
crease in energy expenditure during the four hours after ingestion of
food; this area was used to calculate the percentage of calories oxi-
dized after ingestion of the formula.
Since no technique is available for the direct measurement of non-
resting energy expenditure, we calculated this component of energy
expenditure as the difference between total energy expenditure and
the sum of the resting components of total energy expenditure: non-
resting energy expenditure
#
total energy expenditure
$
(resting en-
ergy expenditure
%
thermic effect of feeding). In a subgroup of sub-
jects studied at multiple weight plateaus, total energy expenditure
was also determined by the differential excretion rates
12
of two sta-
ble isotopes of water (
2
H
2
O and H
218
O) and by in direct calorime-
try performed at the Clinical Diabetes and
Nutrition Section, National Institute of Dia-
betes and Digestive and Kidney Diseases
(Phoenix, Ariz.), in a respiration chamber
equipped with wall-mounted radar detectors
to monitor physical activity.
13
These subjects
also underwent measurement of body compo-
sition by isotope dilution
12
and by dual-pho-
ton absorptiometry
14
to validate the caloric ti-
tration and hydrodensitometric methods used
in the study (unpublished data).
After the completion of studies at their
initial weight, 11 of the obese subjects and
13 of the subjects who had never been obese
(hereafter referred to as nonobese) were giv-
en the maximally tolerated amount of self-
selected foods (generally 5000 to 8000 kcal
per day) until they had gained 10 percent of
their initial body weight (Fig. 1). No formula
was ingested during the period of weight
gain, which ranged from 4 to 6 weeks in the
nonobese subjects and from 6 to 10 weeks in
the obese subjects. At the new weight pla-
teau, the dietary formula was reinstated,
and the quantity of formula was titrated to
maintain the weight. When the weight had
been stable for at least 14 days, the studies
of energy expenditure and body composi-
tion, described above, were repeated. Eight
obese women who had undergone and main-
tained a 10 percent weight gain were then
fed 800 kcal per day of the formula until
they had returned to their initial weight.
They were given the same number of kilo-
calories of formula that had been required to
maintain their initial weight, and the studies
*Values are means
!
SD.
†P
"
0.001 for the comparison with the initial weight.
‡P
#
0.022 for the comparison with the initial weight.
§P
#
0.003 for the comparison with the initial weight.
¶P
"
0.001 for the comparison with nonobese subjects at the same weight plateau.
!
P
#
0.004 for the comparison with the initial weight.
**P
#
0.002 for the comparison with the initial weight.
††P
#
0.047 for the comparison with the initial weight.
‡‡P
#
0.021 for the comparison with the initial weight.
Table 1. Characteristics and Body Composition of Subjects at Initial and Altered
Weights.
*
S
UBJECTS
A
GE
W
EIGHT
F
AT
-
FREE
M
ASS
F
AT
M
ASS
% W
EIGHT
C
HANGE
AS
F
AT
M
ASS
yr kilograms
Nonobese (n
#
13)
Initial weight
10% gain
Obese (n
#
11)
Initial weight
10% gain
Obese (n
#
8)
Initial weight
Return to initial
weight
Nonobese (n
#
11)
Initial weight
10% loss
Obese (n
#
9)
Initial weight
10% loss
27
!
7
28
!
8
29
!
9
25
!
7
32
!
8
66.5
!
11.8
73.2
!
13.3†
131.2
!
25.3¶
143.1
!
25.6†¶
129.7
!
36.8
129.1
!
36.4
70.5
!
11.7
63.7
!
10.1
132.1
!
26.9¶
114.3
!
21.5†¶
54.5
!
12.1
56
!
12.6‡
62.8
!
10.8¶
65.6
!
13.0¶
!
60.8
!
11.9
61.6
!
15.5
53.0
!
10.4
50.6
!
9.5
64.1
!
11.3¶
59.7
!
9.1¶ ††
12.0
!
4.5
17.1
!
4.7§
68.4
!
18.6¶
74.9
!
18.6¶**
68.9
!
38.8
67.5
!
30.0
17.5
!
12.6
13.1
!
3.5
68.0
!
19.8¶
54.6
!
14.7¶‡‡
—
80.1
!
25.2
—
57.9
!
34.8
—
—
—
63.7
!
27.5
—
83.6
!
23.8
Obese (n
#
10)
Initial weight
20% loss
31
!
8 124.8
!
29.6
95.6
!
22.5†
60.8
!
11.2
57.5
!
10.5†
64.4
!
24.8
39.0
!
17.2†
—
82.1
!
25.5
Figure 1. Study Design.
All subjects were studied at their initial weight and after at least
one change in weight.
% o
f
I
n
i
t
i
a
l
W
e
i
g
h
t
110
100
90
80
70
10% gain
Initial
weight
Return to
initial weight
Time
10% loss 10% loss
20% loss
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
Vol. 332 No. 10 CHANGES IN ENERGY EXPENDITURE RESULTING FROM ALTERED BODY WEIGHT 623
of energy metabolism were repeated once their weight had been
stabilized.
After the initial-weight studies described above, 9 obese subjects
and 11 nonobese subjects were fed 800 kcal of the dietary formula
per day until their body weight had been reduced to a level that was
10 percent below their initial weight (Fig. 1). (Eight of the obese and
one of the nonobese subjects had previously undergone studies after
a 10 percent gain in weight.) The period of weight loss ranged from
4 to 7 weeks for the nonobese subjects and from 6 to 14 weeks for the
obese subjects. When the subjects had lost 10 percent of their initial
weight, the formula was reinstated and adjusted to maintain the
weight for at least 14 days; the studies described above were then re-
peated. Ten obese subjects who had undergone initial-weight studies
(seven of whom were also studied after a 10 percent weight loss) were
fed 800 kcal of dietary formula per day until they had lost 20 percent
of their initial weight. The studies described above were performed
after their weight had stabilized at the lower level.
Resting energy expenditure was determined in some subjects at
the end of each period of weight gain or loss, when the intended level
of weight had been achieved (10 percent higher or lower than the ini-
tial weight or after a return to the initial weight) but the subjects
were still gaining or losing weight. These studies were performed to
assess the degree to which the metabolic status was carried over from
a period of changing weight to a period of stable weight.
Statistical Analysis
The results are presented as means
!
SD. Energy expenditure at
each weight plateau is expressed as the absolute number of kilocalo-
ries per day, as well as the number per kilogram of fat-free mass, to
allow comparisons among different groups of subjects at the same
weight plateau and among different weight plateaus for the same
subject. The thermic effect of feeding and fecal calorie loss are ex-
pressed as percentages of ingested calories that were oxidized and
lost in stool, respectively. Comparisons of energy expenditure in the
same subjects at different weight plateaus were made by a one-way
analysis of variance with repeated measures.
15
The effects of sex and
adiposity on measures of energy expenditure at different weight pla-
teaus were determined by a multivariate analysis of variance with re-
peated measures
15
in which sex and adiposity were treated as dichot-
omous variables (male vs. female and nonobese vs. obese).
At the usual body weight, resting energy expenditure is closely cor-
related with measures of metabolic mass (e.g., fat-free mass).
13
Re-
gression lines relating energy expenditure to metabolic mass do not
have zero intercepts.
5
Thus, subjects with values on the same regres-
sion line can have different values for the ratio of energ y expenditure
to metabolic mass. To correct for this possibility, regression equations
relating measures of energy expenditure to fat-free mass and fat
mass at the initial weight were calculated at that weight and used to
determine the predicted value of energy expenditure in the same
subject at each new weight plateau. The observed-minus-predicted
values were then tested against the null hypothesis that the observed-
minus-predicted value was 0, to determine whether the observed val-
ues differed significantly from the predicted values for each subject.
All statistical tests were two-tailed.
R
ESULTS
Energy Expenditure
The rates of energy expenditure changed in both the
obese and the nonobese subjects after changes in body
weight (Table 2 and Fig. 2 and 3). A 10 percent in-
crease or decrease in the usual weight was accompa-
nied by a 16 percent increase or 15 percent decrease,
*Values are means
!
SD. The thermic effect of feeding is expressed as the percentage of metabolizable calories ingested (metabolizable caloric density of formula, 1.25 kcal per gram); fecal
calorie loss is expressed as the percentage of total calories ingested (caloric density of formula by bomb calorimetry, 1.36 kcal per gram).
†P
"
0.001 for the comparison with the same subjects at their initial weight. ‡P
"
0.001 for the comparison with 0.
§P
#
0.003 for the comparison with 0. ¶P
#
0.002 for the comparison with the same subjects at their initial weight.
!
P
"
0.001 for the comparison with nonobese subjects. **P
#
0.016 for the comparison with nonobese subjects.
††P
#
0.067 for the comparison with nonobese subjects. §§P
#
0.004 for the comparison with the same subjects at their initial weight.
¶¶P
#
0.042 for the comparison with the same subjects at their initial weight.
!!
P
#
0.003 for the comparison with the same subjects at their initial weight.
***P
#
0.025 for the comparison with 0. †††P
#
0.004 for the comparison with 0.
‡‡‡P
#
0.037 for the comparison with the same subjects at their initial weight. §§§P
#
0.004 for the comparison with nonobese subjects.
¶¶¶P
#
0.058 for the comparison with 0.
!!!
P
#
0.023 for the comparison with 0.
****P
#
0.023 for the comparison with the same subjects at their initial weight.
Table 2. Measures of 24-Hour Energy Expenditure.*
SUBJECTS TOTAL ENERGY EXPENDITURE RESTING ENERGY EXPENDITURE NONRESTING ENERGY EXPENDITURE
FECAL
CALORIE
LOSS
THERMIC
EFFECT OF
FEEDING
kcal
kcal/kg
fat-free
mass
observed minus
predicted kcal kcal
kcal/kg
fat-free mass
observed minus
predicted kcal kcal
kcal/kg
fat-free
mass
observed minus
predicted kcal % kcal ingested
Nonobese
(n#13)
Initial weight
10% gain
Obese (n #11)
Initial weight
10% gain
Obese (n #8)
Initial weight
Return to ini-
tial weight
Nonobese
(n#11)
Initial weight
10% loss
Obese (n #9)
Initial weight
10% loss
2481!412
3110!527†
3162!712!
4034!746†!
3079!627
3079!627
2380!528
1952!402†
3100!648!
2549!554†!
47!7
54!18†
51!7**
59!6† ††
51!8
50!9
45!6
39!3§§
50!8!
42!5†
—
368!246‡
—
534!278‡
—
$35!347
—
$218!123§
—
$244!198†††
1463!270
1610!267
2127!427!
2261!446!
2015!402
2021!386
1511!304
1290!228¶
2068!359!
1778!416¶!
28!5
30!5
35!7!
33!5!
33!8
30!3
29!3
26!3¶¶
34!7!
30!4‡‡‡§§§
—
27!163
—
6!348
—
$79!307
—
$54!98
—
$137!305¶¶¶
976!239
1496!381†
1075!481
1764!468†**
1030!545
1109!452
864!278
658!240¶
1030!509
768!246†
18!7
28!8†
16!6
26!4†
16!7
17!7
16!4
13!3!!
16!7
13!3§§
—
360!288§
—
524!262‡
—
37!579
—
$158!183***
—
$165!194!!!
2!2
2!2
2!2
2!3
2!1
2!1
2!2
3!2
2!1
2!2
3!1
5!1¶
2!3
4!2§§
3!1
4!1
6!2
4!3
3!3
3!2
Obese (n #10)
Initial weight
20% loss
3129!735
2243!504
51!7
39!4†
—
$301!252†††
1984!342
1581!348†
32!5
28!3****
—
$79!294
1089!456
589!357†
17!6
10!5†
—
$273!336***
2!2
3!2
3!2
3!2
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
624 THE NEW ENGLAND JOURNAL OF MEDICINE March 9, 1995
respectively, in 24-hour total energy expenditure cor-
rected for body composition. Fat-free mass was signifi-
cantly related to total energy expenditure, resting and
nonresting energy expenditure, and the thermic effect
of feeding. Fat mass was significantly related to total
and resting energy expenditure.
In agreement with the results of other studies,16 total
energy expenditure and resting energy expenditure, ex-
pressed as kilocalories per kilogram of fat-free mass
per day, were significantly higher in the obese subjects
than in the nonobese subjects, whereas the thermic ef-
fect of feeding was lower in the obese subjects than in
the nonobese subjects.17,18 The higher resting energy
expenditure in the obese subjects probably reflects in-
creased cardiorespiratory work related to chest-wall
weight and a larger mass of adipose tissue. Smokers
did not differ significantly from nonsmokers for any of
the measures. Fecal calorie and urinary nitrogen losses,
expressed as percentages of ingested calories and pro-
tein, respectively, were not significantly affected by
changes in body weight and did not differ significantly
according to sex or prior adiposity.
Effects of Weight Gain
Total energy expenditure, nonresting energy expend-
iture, and the thermic effect of feeding were signifi-
cantly higher after a 10 percent gain in weight than at
the initial weight. Stabilization of body weight after a
10 percent gain resulted in significant increases in
observed-minus-predicted values for total energy ex-
penditure, nonresting energy expenditure, and the
thermic effect of feeding (Table 2). The magnitude of
these changes was not affected by sex or initial adipos-
ity. In 14 subjects (7 obese and 7 nonobese), the per-
centage of time spent in motion during a 23-hour peri-
od, measured in a respiration chamber, did not differ
significantly between the initial weight (9.1!2.0 per-
cent) and the 10 percent higher weight (8.6!2.1 per-
cent, P#0.47).
Return to Initial Weight
Eight obese women were studied at their initial
weight, at a weight 10 percent higher than their initial
weight, and after a return to their initial weight. No
significant differences in body composition or in any as-
pect of energy expenditure were noted between the
time of the initial-weight study and the return to the
initial weight (Tables 1 and 2).
Effects of Weight Loss
Total energy expenditure and nonresting and resting
energy expenditure were significantly lower at weights
10 and 20 percent below the initial weight than at the
initial weight (Table 2). Stabilization of body weight at
a level 10 percent below the initial weight was associ-
ated with negative observed-minus-predicted values for
total energy expenditure and nonresting and resting
energy expenditure. Stabilization of body weight at a
level 20 percent below the initial weight was associated
with negative observed-minus-predicted values for total
energy expenditure and nonresting energy expendi-
ture. The magnitude of these changes was not signifi-
cantly related to sex or initial adiposity. There were no
significant differences in energy expenditure at weights
10 and 20 percent below the initial weight, suggesting
that the maximal adaptation to the maintenance of a
reduced body weight was already attained at the 10
percent level. Among eight subjects (six nonobese and
two obese) the mean percentage of time spent in mo-
tion in the respiratory chamber was 9.2!2.0 percent at
the initial weight and 9.4!1.8 percent after a 10 per-
cent loss in weight (P#0.52).
Static Weight Maintenance versus Dynamic Weight Change
The high degree of weight stability among the sub-
jects (mean rate of weight change during the 10-day
testing periods, $1.2 g per day) suggests that body
composition was constant during weight maintenance.
When weight and body composition are stable, the res-
piratory quotient reflects mainly the composition of the
diet. As expected, the processes of weight gain and loss
resulted in increases and decreases, respectively, in the
respiratory quotient. However, the respiratory quotient
did not differ significantly from the quotient for the di-
etary formula (0.85) at any of the weight plateaus, in-
dicating that the subjects were in caloric balance at
each plateau, without a carryover effect of weight loss
or gain on caloric requirements or substrate use.
The effects of weight gain and loss on energy metab-
olism were also assessed by comparing resting energy
expenditure at the end of the dynamic phase of weight
change with that at the end of the period of mainte-
nance of the same weight (Table 3). The process of in-
creasing weight by overfeeding was accompanied by
approximately 12 percent more resting energy expend-
iture than a 10 percent weight gain maintained for 14
Figure 2. Mean (!SD) Observed-minus-Predicted Total Energy
Expenditure (Shaded Bars) Based on the Regression of Total
Energy Expenditure in a Model with a Variable Combining Fat-
free Mass and Fat Mass in the Same Subjects at Their Initial
Weight.
The components of total energy expenditure are given in Table 2.
Initial
weight
Observed-minus-Predicted
Total Energy Expenditure (kcal/day)
600
500
400
300
200
100
0
$100
$200
$300
$400
10%
gain
Return
to initial
weight
10%
loss
20%
loss
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
Vol. 332 No. 10 CHANGES IN ENERGY EXPENDITURE RESULTING FROM ALTERED BODY WEIGHT 625
Figure 3. Total and Resting Energy Expenditure According to Fat-free Mass at the Initial Weight and after a Gain or Loss in Weight.
The results are presented in terms of fat-free mass to facilitate comparisons among subjects studied at different weights. The diagonal
lines represent regression equations for energy expenditure as compared with fat-free mass at the initial weight and at an altered
weight in the same subjects. For subjects studied at their initial weight and after a 10 percent gain in weight (left-hand graphs), total
energy expenditure equaled 39.7 kg of fat-free mass plus 348.8 (r2#0.72, P"0.001), and resting energy expenditure equaled 13.1
kg of fat-free mass plus 670.1 (r2#0.42, P#0.004). For subjects studied at their initial weight and after a weight loss (right-hand
graphs), total energy expenditure equaled 56.8 kg of fat-free mass minus 496.1 (r 2#0.74, P"0.001), and resting energy expenditure
equaled 27.5 kg of fat-free mass plus 220.2 (r 2#0.44, P #0.004).
Plotted numbers denote individual subjects. Subjects 1 through 23 were not obese; subjects 24 through 38 were obese. Six women
and two men studied after a 10 percent weight loss had previously been studied after a 10 percent weight gain; six women studied
after a 20 percent weight loss had previously been studied after a 10 percent gain and a 10 percent loss. The number and direction
of previous weight changes did not significantly affect any measures of energy expenditure at a given weight plateau.
Total Energy Expenditure (kcal/day)
Fat-free Mass (kg) Fat-free Mass (kg)
Fat-free Mass (kg) Fat-free Mass (kg)
Total Energy Expenditure (kcal/day)
Resting Energy Expenditure (kcal/day)
6000 6000
5000 5000
4000 4000
3000 3000
2000 2000
1000
25 35 45 55 65 75 85 95 25 35 45 55 65 75 85 95
25 35 45 55 65 75 85 95 25 35 45 55 65 75 85 95
1000
Resting Energy Expenditure (kcal/day)
3000
2000
1000
3000
2000
1000
Initial weight
20% loss
10% loss
Initial weight
10% gain
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
626 THE NEW ENGLAND JOURNAL OF MEDICINE March 9, 1995
days. Conversely, the process of losing weight on a diet
of 800 kcal per day resulted in 10 to 15 percent less
resting energy expenditure than a stabilized weight
loss of 10 percent.
Body Composition
A 10 percent gain in weight resulted in increases in
adipose tissue and fat-free body mass, and a weight
loss of 10 percent resulted in significant decreases in
both these measures. These changes in body composi-
tion, expressed as the change in the percentage of body
fat, were statistically significant within all subgroups of
subjects. There was no evidence that sex or prior adi-
posity affected the distribution of weight gained or lost
between fat mass and fat-free mass (Table 1). There
was a trend for the obese subjects to gain a lower per-
centage of weight as fat than the nonobese subjects
(P#0.11) and to lose a higher percentage of weight as
fat (P#0.13).
DISCUSSION
We found that energy expenditure adjusted for met-
abolic mass increased with a weight gain and decreased
with a weight loss. These changes in energy expendi-
ture were evident during periods of stable altered body
weight and were in a direction tending to return the
subject to his or her initial weight; their magnitude was
similar in nonobese and obese subjects. After a 10 per-
cent gain in weight, the increase in total energy ex-
penditure reflected a large increase in the absolute
number of kilocalories of nonresting energy expendi-
ture per day and a small increase in the absolute num-
ber of kilocalories per day attributed to the thermic ef-
fect of feeding. After a 10 or 20 percent loss in weight,
the decline in total energy expenditure reflected similar
decreases in both nonresting and resting energy ex-
penditure.
In some studies, energy expenditure has been higher
than that predicted for metabolic mass during weight
gain or maintenance of a higher body weight in non-
obese subjects.19 In other studies,20,21 such an increase
has not occurred. Resting energy expenditure declines
during a period of weight loss,22 but whether a similar
decline occurs when weight loss has stabilized is the
subject of considerable debate.23-26 Likewise, there is a
lack of agreement concerning the effect of weight re-
duction on total energy expenditure; both decreases (of
8 to 22 percent) and no change or increases (of approx-
imately 9 percent, as compared with the values for
weight-matched nonobese controls) have been report-
ed.27-29
Several considerations are important in drawing con-
clusions from our findings. The alterations in energy
expenditure do not reflect carryovers from the dynamic
periods of weight change, since the measures of energy
expenditure at the initial weight and after a return to
that weight were similar, the anticipated changes in
resting energy expenditure occurred during periods
of weight change, and respiratory quotients at the var-
ious weight plateaus equaled that predicted for a stable
weight while subjects were ingesting the dietary formu-
la. Subtle shifts in body composition during periods of
stable weight could have masked differences in energy
expenditure among the weight plateaus. Such changes
are unlikely, since they would have led to discrepancies
between total energy expenditure as measured by for-
mula titration and total energy expenditure as meas-
ured by elimination rates for 2H2O and H218O. Yet
these two measures were highly correlated and did not
differ significantly at any weight plateau (unpublished
data). In addition, respiratory quotients during all peri-
ods of stable weight equaled that predicted on the basis
of the nutrient composition of the formula diet, suggest-
ing that there was no net storage or catabolism of fat.
Although we did not examine the permanence of
these changes in rates of energy expenditure, a reduced
level of energy expenditure has been reported to per-
sist in subjects who have maintained a reduced body
weight for periods ranging from six months to more
than four years.6 The aspect of body composition that
mediates or signals these changes in energy expendi-
ture is not known. In our study, the largest changes
in body composition with weight alteration occurred in
fat mass. A substantial body of literature suggests that
the mass of adipose tissue or the size of adipocytes
is the aspect of body composition that is regulated,30
but the feedback mechanism for the effect of fat mass
on energy metabolism is not known. A candidate gene
for such a signal from fat has recently been cloned.31
The metabolic variable most affected by weight
change was nonresting energy expenditure. Since it
was not measured directly, the nature of the changes in
this variable cannot be identified. Differences in the en-
ergy needed to move a larger or smaller body mass ac-
count for only some of the differences in energy ex-
penditure, as suggested by Weigle and Brunzell, in
whose study lost weight was replaced with backpack
loads.32 This cannot be the entire explanation, however,
since obese and lean subjects at their usual body
*Results are means !SD. All P values are for the comparison between weight gain or loss
and weight maintenance.
†P"0.001. ‡P#0.03.
§P#0.009. ¶P #0.029.
!P#0.043. **P#0.014.
Table 3. Comparison of Resting Energy Expenditure and Respi-
ratory Quotient at the End of a Period of Weight Gain or Loss
and during Maintenance of the Altered Weight.*
PERIOD
NO. OF
SUBJECTS WEIGHT
RESTING ENERGY
EXPENDITURE
RESPIRATORY
QUOTIENT
kg kcal/day
Weight gain 14 136.2!41.1 2391!750† 0.92!0.07‡
Weight maintenance
(10% gain)
136.5!41.5 2102!573 0.86!0.07
Weight loss 8 127.9!25.4 1704!450§ 0.78!0.08¶
Weight maintenance
(return to initial
weight)
129.1!26.5 2059!384 0.85!0.08
Weight loss 9 112.8!20.6 1598!385!0.73!0.06**
Weight maintenance
(10% loss)
114.2!21.5 1747!416 0.83!0.08
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
Vol. 332 No. 10 CHANGES IN ENERGY EXPENDITURE RESULTING FROM ALTERED BODY WEIGHT 627
weight have nearly equal nonresting energy expendi-
ture, corrected for both fat-free mass and fat mass.
One possibility is that the efficiency with which skeletal
muscle performs mechanical work is different at differ-
ent weight plateaus (decreased with a 10 percent gain
in weight and increased with a 10 percent loss). The ef-
ficiency of muscular work before and after weight gain
does not change, nor does the energy expenditure as-
sociated with moderate exercise.33,34 The effect of a
change in weight on the energy expended during mild
physical exertion, however, has not been systematically
studied, and this minimal level of exertion may be most
representative of the physical activity of sedentary
adults. The possibility that changes in skeletal muscle
have a role in mediating the alterations in energy ex-
penditure that occur with weight loss may be the rea-
son exercise is helpful in maintaining a reduced body
weight.
Body weight in adults is remarkably stable for long
periods of time. In the Framingham Study the body
weight of the average adult increased by only 10 per-
cent over a 20-year period.35 Such a fine balance is ev-
idence of the presence of regulatory systems for body
weight.4,36 Whatever the mechanism (or mechanisms),
the weight at which regulation occurs differs from one
person to another, and these differences are almost cer-
tainly due in part to genetic37,38 and developmental39 in-
fluences.
Our results have immediate implications for the clin-
ical management of obesity. Many obese people who
lose weight have metabolic alterations similar to those
observed in our subjects. The reduction in energy ex-
penditure to a level 15 percent below that predicted for
body composition, as a result of a 10 percent (or larger)
decrease in body weight, is large when one considers
that an average daily intake of 2500 kcal would be as-
sociated with a positive energy balance of approximate-
ly 375 kcal per day. In addition, the sense of hunger or
dysphoria that may accompany this state of reduced
energy expenditure will promote increased food intake,
further widening the gap between energy output and
intake.3 Physicians should be aware that for some
obese patients the achievement of what is considered to
be a more healthful body weight may be accompanied
by metabolic alterations that make it difficult to main-
tain the lower weight. Nevertheless, the beneficial
effect of even a modest weight loss on lipid and carbo-
hydrate metabolism in obese patients40,41 justifies per-
sistent efforts at weight reduction and maintenance of
a reduced body weight for the treatment of obesity.
We are indebted to Drs. Elio Presta, Streamson C. Chua, and Lisa
C. Hudgins, Mr. David Markel, Ms. Rachael Kolb, Ms. Eileen Mul-
len, Ms. Jennifer Ziedonis, Ms. Alice Murphy, and the members of
the nursing staff of the Rockefeller University Hospital Clinical
Research Center for their help with the care of the subjects; to Ms.
Cynthia Seidman and her staff of research dietitians for supervising
the preparation and testing of the dietary formula; to Drs. Steven
Heymsfield and Steven Lichtman at St. Luke’s–Roosevelt Hospital
Medical Center for supervising the body-composition studies; to Dr.
Eric Ravussin at the National Institute of Diabetes and Digestive and
Kidney Diseases in Phoenix, Arizona, for supervising the chamber
respirometry studies and for his helpful suggestions on the man-
uscript; and to Dr. Dwight Matthews and Mr. Chuck Gilker for
performing the mass spectrometric analysis of urine for 2H2O and
H218O.
REFERENCES
1. Keesey RE, Corbett SW. Metabolic defense of the body weight set-point. In:
Stunkard AJ, Stellar E, eds. Eating and its disorders. Vol. 62 of Research
publications: association for research in nervous and mental disease. New
York: Raven Press, 1984:87-96.
2. Leibel RL. A biologic radar system for the assessment of body mass: the
model of a geometry sensitive endocrine system is presented. J Theor Biol
1977;66:297-306.
3. Weigle DS. Appetite and the regulation of body composition. FASEB J
1994;8:302-10.
4. Leibel RL. Is obesity due to a heritable difference in ‘set point’ for adipos-
ity? West J Med 1990;153:429-31.
5. Ravussin E, Bogardus C. Relationship of genetics, age and physical fitness
to daily energy expenditure and fuel utilization. Am J Clin Nutr 1989;49:
968-75.
6. Leibel RL, Hirsch J. Diminished energy requirements in reduced-obese pa-
tients. Metabolism 1984;33:164-70.
7. Epstein FH, Higgins M. Epidemiology of obesity. In: Bjorntorp P, Brodoff
BN, eds. Obesity. Philadelphia: J.B. Lippincott, 1992:330-42.
8. Department of Agriculture. Composition of foods: raw, processed, prepared.
Agricultural handbook no. 8 series. Washington, D.C.: Government Printing
Office, 1963:159-69.
9. Heymsfield SB, Wang J, Kehayias JJ, Heshka S, Lichtman S, Pierson RN Jr.
Chemical determination of human body density in vivo: relevance to hy-
drodensitometry. Am J Clin Nutr 1989;50:1282-9.
10. Benedict FG, Fox EL. A method for determination of the energy values of
foods and excreta. J Biol Chem 1925;66:783-99.
11. Ferrannini E. The theoretical bases of indirect calorimetry: a review. Metab-
olism 1988;37:287-301.
12. Schoeller DA, van Santen E. Measurement of energy expenditure in humans
by doubly labeled water method. J Appl Physiol 1982;53:955-9.
13. Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus C. Determinants
of 24-hour energy expenditure in man: methods and results using a respira-
tory chamber. J Clin Invest 1986;78:1568-78.
14. Mazess RB, Peppler WW, Gibbons M. Total body composition by dual-pho-
ton (153Gd) absorptiometry. Am J Clin Nutr 1984;40:834-9.
15. SAS/STAT user’s guide: version 6. 4th ed. Cary, N.C.: SAS Institute, 1990.
16. Ravussin E, Burnand B, Schutz Y, Jequier E. Twenty-four-hour energy ex-
penditure and resting metabolic rate in obese, moderately obese, and control
subjects. Am J Clin Nutr 1982;35:566-73.
17. Blondheim SH, Mendelson B. Dietary thermogenesis: why the conflicting
results? In: Berry EM, Blondheim H, Eliahou HE, Shafrir E, eds. Recent ad-
vances in obesity research V. London: John Libbey, 1987:151-4.
18. Ravussin E, Swinburn BA. Energy metabolism. In: Stunkard AJ, Wadden
TA, eds. Obesity: theory and therapy. 2nd ed. New York: Raven Press, 1993:
97-123.
19. Sims EAH, Danforth E Jr, Horton ES, Bray GA, Glennon JA, Salans LB.
Endocrine and metabolic effects of experimental obesity in man. Recent
Prog Horm Res 1973;29:457-96.
20. Norgan NG, Durnin JVGA. The effect of 6 weeks of overfeeding on the
body weight, body composition, and energy metabolism of young men. Am
J Clin Nutr 1980;33:978-88.
21. Tremblay A, Despres JP, Theriault G, Fournier G, Bouchard C. Overfeeding
and energy expenditure in humans. Am J Clin Nutr 1992;56:857-62.
22. Forbes GB, Welle SL. Lean body mass in obesity. Int J Obes 1983;7:99-107.
23. Elliot DL, Goldberg L, Kuehl KS, Bennett WM. Sustained depression of the
resting metabolic rate after massive weight loss. Am J Clin Nutr 1989;49:93-6.
24. Doré C, Hesp R, Wilkins D, Garrow JS. Prediction of energy requirements of
obese patients after massive weight loss. Hum Nutr Clin Nutr 1982;36C:41-8.
25. Froidevaux F, Schutz Y, Christin L, Jequier E. Energy expenditure in obese
women before and during weight loss, after refeeding, and in the weight-
relapse period. Am J Clin Nutr 1993;57:35-42.
26. Amatruda JM, Statt MC, Welle SL. Total and resting energy expenditure in
obese women reduced to ideal body weight. J Clin Invest 1993;92:1236-42.
27. deBoer JO, van Es AJH, Roovers LCA, van Raaij JM, Hautvast JG. Adap-
tation of energy metabolism of overweight women to low-energy intake,
studied with whole-body calorimeters. Am J Clin Nutr 1986;44:585-95.
28. Weigle DS, Sande KJ, Iverius PH, Monsen ER, Brunzell JD. Weight loss
leads to a marked decrease in nonresting energy expenditure in ambulatory
human subjects. Metabolism 1988;37:930-6.
29. Astrup A, Buemann B, Christensen NJ, Madsen J. 24-Hour energy expend-
iture and sympathetic activity in postobese women consuming a high-carbo-
hydrate diet. Am J Physiol 1992;262:E282-E288.
30. Faust IM. Role of the fat cell in energy balance physiology. In: Stunkard AJ,
Stellar E, eds. Eating and its disorders. Vol. 62 of Research publications: as-
sociation for research in nervous and mental disease. New York: Raven
Press, 1984:97-107.
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
628 THE NEW ENGLAND JOURNAL OF MEDICINE March 9, 1995
31. Zhang Y, Proenca R, Maffei M, et al. Positional cloning of the mouse obese
gene and its human homologue. Nature 1994;372:425-32.
32. Weigle DS, Brunzell JD. Assessment of energy expenditure in ambulatory
reduced-obese subjects by the techniques of weight stabilization and exog-
enous weight replacement. Int J Obes 1990;14:Suppl 1:69-77.
33. Weigle DS. Contribution of decreased body mass to diminished thermic ef-
fect of exercise in reduced-obese men. Int J Obes 1988;12:567-78.
34. Geissler CA, Miller DS, Shah M. The daily metabolic rate of the post-obese
and the lean. Am J Clin Nutr 1987;45:914-20.
35. Belanger AJ, Cupples LA, D’Agostino RB. Means at each examination and
inter-examination consistency of specified characteristics: Framingham
Heart Study, 30-year followup. In: Kannel WB, Wolf PA, Garrison RJ, eds.
The Framingham Study: an epidemiological investigation of cardiovascular
disease. Sect. 36. Washington, D.C.: Government Printing Office, 1988.
(NIH publication no. 88-2970.)
36. Harris RB. Role of set-point theory in regulation of body weight. FASEB J
1990;4:3310-8.
37. Bogardus C, Lillioja S, Ravussin E, et al. Familial dependence of the resting
metabolic rate. N Engl J Med 1986;315:96-100.
38. Bouchard C, Perusse L. Genetics of obesity. Annu Rev Nutr 1993;13:337-54.
39. Knittle JL, Timmers K, Ginsberg-Fellner F, Brown RE, Katz DP. The growth
of adipose tissue in children and adolescents: cross-sectional and longitudi-
nal studies of adipose cell number and size. J Clin Invest 1979;63:239-46.
40. Olefsky J, Reaven GM, Farquhar JW. Effects of weight reduction on obesity:
studies of lipid and carbohydrate metabolism in normal and hyperlipopro-
teinemic subjects. J Clin Invest 1974;53:64-76.
41. Jimenez J, Zuniga-Guajardo S, Zinman B, Angel A. Effects of weight loss
in massive obesity on insulin and C-peptide dynamics: sequential changes
in insulin production, clearance, and sensitivity. J Clin Endocrinol Metab
1987;64:661-8.
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.
New England Journal of Medicine
CORRECTION
Changes in Energy Expenditure Resulting from
Altered Body Weight
Changes in Energy Expenditure Resulting from Altered Body Weight .
On page 623, in Table 2, the initial weight results and the results after
10 percent weight loss in the nonobese and the obese subjects were
incorrectly aligned, and should have been transposed down one line.
We regret the error.
N Engl J Med 1995;333:399
Copyright © 1995 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org on January 22, 2009 . For personal use only. No other uses without permission.