0021-972X/00/$03.00/0 Vol. 85, No.

3
The Journal of Clinical Endocrinology & Metabolism Printed in U.S.A.
Copyright © 2000 by The Endocrine Society

Energy Expenditure, Fat Oxidation, and Body Weight

Regulation: A Study of Metabolic Adaptation to Long-
Term Weight Change
CHRISTIAN WEYER, RICHARD E. PRATLEY, ARLINE D. SALBE,
CLIFTON BOGARDUS, ERIC RAVUSSIN, AND P. ANTONIO TATARANNI
Clinical Diabetes and Nutrition Section, National Institute of Diabetes and Digestive and Kidney
Diseases, National Institutes of Health, Phoenix, Arizona 85016

ABSTRACT
Relatively low rates of energy expenditure and fat oxidation predict
body weight gain. Weight gain, in turn, is associated with increases
in energy expenditure and fat oxidation that may oppose further
weight change. In response to experimental weight gain induced by
overfeeding, increases in energy expenditure and fat oxidation are
overcompensatory, i.e. greater than predicted for the change in body
composition. To determine whether such metabolic adaptation occurs
in response to spontaneous long term weight change, we conducted a
longitudinal study in which 24-h energy expenditure (24-EE) and 24-h
respiratory quotient (24-RQ; i.e. fat to carbohydrate oxidation) were
repeatedly measured in 102 Pima Indians at baseline and after a
mean follow-up of 3.6 ⫾ 2.7 yr, during which changes in body weight
varied widely (⫺21 to ⫹28 kg). We found that changes in 24-EE and
24-RQ in response to weight change were related to the amount of
weight change, even after adjustment for body composition (partial
r ⫽ 0.23 and ⫺0.30, respectively; both P ⬍ 0.05). For a 15-kg weight
gain, the increases in 24-EE (⫹244 Cal/day) and 24-h fat oxidation
(⫹152 Cal/day) were 33 and 53 Cal/day greater than predicted from
the cross-sectional relationship between both measures and body
weight. Changes in 24-EE and 24-RQ varied substantially among
individuals. Thus, on the average, spontaneous long term weight
changes are accompanied by small metabolic adaptations in both
energy expenditure and fat oxidation. The metabolic responses to
weight changes are highly variable among individuals, however.
(J Clin Endocrinol Metab 85: 1087–1094, 2000)

T HE PREVALENCE of overweight and obesity has in-

creased considerably in the United States in recent
years, and it is estimated that some 22% of Americans are
in energy expenditure that is greater than predicted for the
changes in body size and composition (14 –22). Similarly,
most underfeeding studies reveal that in the short term,
currently obese (1). To develop successful prevention and intentional weight loss leads to a decrease in energy expen-
treatment strategies, it is important to understand the phys- diture beyond predicted values (23–36). Such overcompen-
iological mechanisms underlying the long term regulation of satory metabolic changes act to oppose further weight
body weight. change and have thus been referred to as metabolic adap-
Differences in energy metabolism may play a role in tation (13, 24, 37).
long term body weight regulation and the pathogenesis of Although metabolic adaptation thus seems to occur in
human obesity (2–13). Several (2–7), but not all (8 –10), response to large perturbations in body weight over rel-
prospective studies have shown that a relatively low en- atively short periods of time, it is unknown whether sim-
ergy expenditure (2–5) and a relatively high respiratory ilar adaptive mechanisms also occur in response to spon-
quotient, i.e. a low fat to carbohydrate oxidation rate (6, 7) taneous long term weight changes in free living conditions.
predict body weight gain. Longitudinal studies, however, To examine this question, we analyzed data from an on-
in which energy metabolism was assessed not only at going longitudinal study of the pathogenesis of obesity ini-
baseline but also at follow-up, indicate that upon gaining tiated in 1985 among the Pima Indians of Arizona, a popu-
weight, energy expenditure and fat oxidation increase (2, lation with a very high prevalence of obesity, in whom low
6, 13). Metabolic propensity to obesity might thus depend rates of energy expenditure and fat oxidation predict body
not only on initial rates of energy expenditure and fat weight gain (2, 6). We present results from over 100 subjects
oxidation, but also on how these measures change in re- in whom 24-h energy expenditure and 24-h substrate oxi-
sponse to weight change (13). dation were repeatedly measured in a whole body respira-
Results from most overfeeding studies indicate that short tory chamber before and after an average follow-up of 3.6 yr
term experimental weight gain is accompanied by an over- during which changes in body weight varied widely. The
compensatory increase in energy expenditure, i.e. an increase aims of this study were 1) to test whether metabolic adap-
tation in 24-h energy expenditure and 24-h substrate oxida-
Received October 14, 1999. Revision received November 16, 1999. tion occur in response to spontaneous long term weight
Accepted November 23, 1999. change, 2) to quantify and explain the variability in these
Address all correspondence and requests for reprints to: Christian
changes among individuals, and 3) to determine the rela-
Weyer, M.D., Clinical Diabetes and Nutrition Section, National Institutes
of Health, 4212 N. 16th Street, Room 5-41, Phoenix, Arizona 85016. tionship between changes in energy expenditure and sub-
E-mail: cweyer@phx.niddk.nih.gov. strate oxidation in response to weight gain and weight loss.

1087

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
1088 WEYER ET AL. JCE & M • 2000
Vol 85 • No 3

Subjects and Methods
Subjects
Since 1985, Pima Indians have been admitted to the metabolic ward
of the Clinical Diabetes and Nutrition Section of the NIH in Phoenix,
Arizona, for an ongoing longitudinal study of the pathogenesis of obe-
sity that includes the repeated assessment of 24-h energy expenditure
and 24-h substrate oxidation in a whole body respiratory chamber.
Subjects with diabetes mellitus and subjects who were not in energy
balance (energy intake ⫾20% of expenditure during the stay in the
chamber) or who had participated in studies involving diet, exercise, or
any other interventions affecting body weight or energy metabolism,
were excluded from the analysis. Among the 491 subjects meeting these
criteria, 102 subjects had been studied on at least 2 occasions. In subjects
studied more than twice, the visit with the greatest weight change was
selected for follow-up. Among the 102 subjects, 31 subjects had lost
weight and 71 had gained weight at follow-up. All subjects were be-
tween 18 –50 yr of age at baseline and follow-up, healthy according to
a physical examination and routine laboratory tests, and did not smoke
or take medications at baseline or follow-up (Table 1).
All subjects stayed on the metabolic ward for at least 3 days, where
they were fed a weight-maintaining diet (50%, 30%, and 20% of daily
calories provided as carbohydrate, fat, and protein, respectively) and
abstained from strenuous physical activity before evaluation in the
chamber. Glucose tolerance was assessed by a 75-g oral glucose tolerance
test (38). The study protocol was approved by the Institutional Review
Board of the NIDDK and by the Tribal Council of the Gila River Indian
Community, and all subjects provided written informed consent before
participation.
Body composition
Body composition was estimated by underwater weighing, with de-
termination of residual lung volume by helium dilution (39), or by total
body dual energy x-ray absorptiometry (DPX-L, Lunar Corp., Madison,
WI) (40). Percent body fat, fat mass (FM), and fat-free mass (FFM) were
calculated as previously described (41), and a conversion equation (42)
was used to make measurements comparable between the two methods.
Waist and thigh circumferences were measured at the umbilicus and the
gluteal fold in the supine and standing positions, respectively, and the
waist to thigh ratio (WTR) was calculated as an index of body fat
distribution (43).
Respiratory chamber
The measurement of energy expenditure and substrate oxidation in
the respiratory chamber has previously been described (44) and did not
differ at baseline and follow-up. In brief, volunteers entered the chamber
at 0745 h after an overnight fast and remained there until 0700 h the
following morning. Subjects were fed a standardized diet with the
amount of calories calculated according to previously determined equa-
tions to achieve energy balance (45). Meals were provided at 0800, 1130,
and 1700 h, and an evening snack was given at 2000 h. The rate of energy
expenditure was measured continuously, calculated for each 15-min
interval of the 23 h in the chamber, and then extrapolated to 24 h (24-h
energy expenditure, 24-EE). Spontaneous physical activity (SPA) was
detected by radar sensors and expressed as the percentage of time over
the 24-h period in which activity was detected (44). Carbon dioxide
production (VCO2) and oxygen consumption (VO2) were calculated at
15-min intervals, summed for the 23 h in the chamber, and then ex-
trapolated to 24 h. The 24-h respiratory quotient (24-RQ) was calculated
as the ratio of 24-h VCO2 and 24-h VO2 and adjusted for the 24-h energy
balance (24-h energy intake ⫺ 24-EE during the stay in the chamber) in
a multiple regression analysis. Based upon 24-RQ, 24-EE, and 24-h
urinary nitrogen excretion, the rates of 24-h fat, carbohydrate, and pro-
tein oxidation were determined as previously described (46).
Statistical analyses
Statistical analyses were performed using the procedures of the
SAS Institute, Inc. (Cary, NC) (47). Results are given as the mean ⫾
sd. Data from the entire group of 491 subjects were used to assess the
cross-sectional relationships between 24-EE and 24-RQ vs. body
weight (prediction line and its 95% confidence interval) and to cal-
culate the adjusted values of 24-EE (FFM, FM, WTR, age, and sex) and
24-RQ (percentage of body fat, age, and sex; multiple linear regres-
sion). Changes in anthropometric and metabolic parameters were
assessed in the subset of 102 subjects with follow-up measurements.
Changes (⌬) in 24-EE and 24-RQ were calculated as the difference
between follow-up and baseline measurements for both the unad-
justed and the adjusted values. Paired t tests were used to test
whether measurements at follow-up were significantly different from
those at baseline. Pearson correlation coefficients were calculated to
assess the relation of the changes in unadjusted and adjusted 24-EE
and 24-RQ to the change in body weight. Stepwise and general linear
regression models were used to assess determinants of ⌬ 24-EE and
⌬ 24-RQ, the percentage of variance explained by these determinants
(r2), and the residual variance that remains after adjustment
(公MSE ⫽ root of the mean square error). The changes in 24-EE and
24-RQ predicted for a 15-kg weight loss or 15-kg weight gain were
determined from the regression equation of the relationships between
⌬ 24-EE and ⌬ 24-RQ vs. ⌬ weight. These changes were superimposed
onto the 95% confidence intervals of the prediction lines for the
cross-sectional relationships between 24-EE and 24-RQ vs. body
weight, as assessed in the entire study population of 491 subjects. The
residuals of the relationships between ⌬ 24-EE and ⌬ 24-RQ vs. ⌬
weight were calculated using general linear regression models.

TABLE 1. Physical and metabolic characteristics of the entire study population and of the subset of 102 subjects with follow-up

Entire population (n ⫽ 491) Subjects with follow-up (n ⫽ 102)

baseline Baseline Follow-up ⌬ (min/max) P value
Females/males 183/308 32/70
NGT/IGT 363/128 91/11
Age (yr) 28.6 ⫾ 8.7 26.9 ⫾ 5.7 30.5 ⫾ 6.2 3.6 ⫾ 2.7 (0.6/11.9) ⬍0.0001
Body weight (kg) 93.1 ⫾ 22.6 93.2 ⫾ 23.9 97.7 ⫾ 25.3 4.5 ⫾ 9.0 (⫺20.9/27.5) ⬍0.0001
Body fat (%) 32 ⫾ 9 32 ⫾ 9 32 ⫾ 8 0 ⫾ 5 (⫺12/11) n.s.
Fat mass (kg) 35.5 ⫾ 15.8 35.5 ⫾ 16.5 37.4 ⫾ 16.6 1.9 ⫾ 6.9 (⫺15.3/21.2) ⬍0.01
Fat-free mass (kg) 57.6 ⫾ 11.6 57.7 ⫾ 11.9 60.3 ⫾ 12.0 2.6 ⫾ 5.4 (⫺15.0/17.6) ⬍0.0001
Waist to thigh ratio 1.64 ⫾ 0.16 1.66 ⫾ 0.16 1.67 ⫾ 0.16 0.01 ⫾ 0.14 (⫺0.61/0.28) n.s.
24-h energy expenditure (Cal/day) 2316 ⫾ 394 2327 ⫾ 389 2364 ⫾ 395 36 ⫾ 192 (⫺386/561) ⬍0.06
24-h energy intake (Cal/day) 2251 ⫾ 333 2243 ⫾ 363 2317 ⫾ 383 74 ⫾ 168 (⫺355/621) ⬍0.0001
24-h energy balance (Cal/day) ⫺65 ⫾ 141 ⫺84 ⫾ 170 ⫺47 ⫾ 165 38 ⫾ 197 (⫺612/713) n.s.
24-h respiratory quotient 0.852 ⫾ .022 0.854 ⫾ .023 0.853 ⫾ .022 ⫺0.001 ⫾ 0.003 (⫺0.062/0.062) n.s.
24-h fat oxidation (Cal/day) 920 ⫾ 248 916 ⫾ 287 951 ⫾ 284 35 ⫾ 299 (⫺562/662) n.s.
24-h carbohydrate oxidation (Cal/day) 1065 ⫾ 215 1067 ⫾ 256 1114 ⫾ 212 47 ⫾ 221 (⫺496/541) ⬍0.05
Values are the ⌬ (min/max). At baseline, all measurements in the follow-up group were comparable to those in the entire population. The
P values refer to the changes over time (determined by paired t-test). NGT/IGT, Normal/impaired glucose tolerance (33).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
 METABOLIC ADAPTATION TO WEIGHT CHANGE 1089

Results
The anthropometric and metabolic characteristics of the
entire study population and of the subset of individuals with
follow-up studies are summarized in Table 1. The baseline
anthropometric and metabolic characteristics of the 102 sub-
jects with repeated measurements of energy metabolism
were similar to those of the entire study population. The
follow-up duration ranged from 0.6 –11.2 yr, averaging 3.6 ⫾
2.7 yr. During this time, body weight changes varied widely
(ranging from ⫺21 to ⫹28 kg; Table 1).

Cross-sectional analysis (n ⫽ 491)

In the entire population of 491 subjects, 24-EE was posi-
tively related to body weight (r ⫽ 0.81; P ⬍ 0.0001; Fig. 1C),
whereas 24-RQ was negatively related to body weight (r ⫽
⫺0.22; P ⬍ 0.0001; Fig. 2C).

Longitudinal analysis (n ⫽ 102)

Changes in 24-h energy expenditure. There was a positive linear
correlation between the change in 24-h energy expenditure
(⌬ 24-EE) and the change in body weight (⌬ weight), but at
any given ⌬ weight there was considerable interindividual
variability in ⌬ 24-EE (sd, 192 Cal/day; Fig. 1A). The corre-
lation between ⌬ 24-EE and ⌬ weight remained significant
when baseline and follow-up 24-EE were adjusted for FFM,
FM, WTR, age, and sex (Fig. 1B). The ⌬ SPA measured in the
chamber was not related to ⌬ weight (r ⫽ 0.01; P ⫽ NS), and
the correlation between ⌬ 24-EE and ⌬ weight remained
significant when baseline and follow-up 24-EE was addi-
tionally adjusted for SPA (partial r ⫽ 0.23; P ⬍ 0.05). When
the changes in 24-EE predicted for a 15-kg weight loss and
a 15-kg weight gain were superimposed onto the cross-
sectional relationship between 24-EE and body weight as
assessed in the entire population of 491 subjects, 24-EE fell
below the 95% confidence interval after weight loss and
above the 95% confidence interval after weight gain (Fig. 1C).
For these weight changes, the calculated change in 24-EE
(⫾244 Cal/day) was 16% (33 Cal/day) greater than that
predicted from the cross-sectional relationship between
24-EE and body weight (⫾211 Cal/day; Fig. 1C).
The ⌬ weight alone explained 58% of the variability in ⌬
24-EE (r2 ⫽ 0.58). In a multiple regression analysis, the ⌬
FFM, ⌬ FM, ⌬ WTR, and ⌬ SPA were significant independent
determinants of ⌬ 24-EE: ⌬ 24-EE (Cal/day) ⫽ ⫺25 ⫹ 14.9 ⌬ FIG. 1. A, Relationship between ⌬ 24-EE and ⌬ weight over 3.6 ⫾ 2.7
FFM (kg) ⫹16.5 ⌬ FM (kg) ⫹ 22.2 ⌬ WTR (decimal) ⫹ 9.7 ⌬ yr in 102 nondiabetic Pima Indians. B, Relationship between ⌬ 24-EE
SPA (%). Together, these factors explained 63% of the vari- and ⌬ weight after adjustment of 24-EE for FFM, FM, WTR, and age.
C, Predicted changes in 24-EE attributable to a 15-kg weight loss and
ability in ⌬ 24-EE, reducing the variance from 192 to 118 a 15-kg weight gain (arrows) superimposed onto the cross-sectional
Cal/day (公MSE). Sex, age, glucose tolerance status (normal relationship between 24-EE and body weight as assessed in 491 non-
or impaired), initial body weight, and follow-up duration diabetic Pima Indians (prediction line with its upper and lower 95%
were not significant determinants of ⌬ 24-EE. confidence intervals is shown).

Changes in 24-h respiratory quotient and substrate oxidation.

There was a negative linear correlation between ⌬ 24-RQ and dicted for a 15-kg weight loss and a 15-kg weight gain were
⌬ weight, but for any given ⌬ weight there was considerable superimposed onto the cross-sectional relationship between
interindividual variability in ⌬ 24-RQ (sd, 0.030; Fig. 2A). The 24-RQ and body weight, 24-RQ fell above the 95% confidence
correlation between ⌬ 24-RQ and ⌬ weight remained signif- interval after weight loss and below the 95% confidence
icant when baseline and follow-up 24-RQ were adjusted for interval after weight gain (Fig. 2C). For this weight change,
percent body fat, age, and sex in addition to energy balance the calculated change in fat oxidation (⫾152 Cal/day) was
in the chamber (Fig. 2B). When the changes in 24-RQ pre- 54% (53 Cal/day) greater than predicted from the cross-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
1090 WEYER ET AL. JCE & M • 2000
Vol 85 • No 3

Responses in energy expenditure vs. responses in substrate oxi-

dation. Based on the relationship between ⌬ 24-EE and ⌬
weight (Fig. 1A), the changes in 24-EE can be divided into
those that were relatively high (positive residuals) and those
that were relatively low (negative residuals). The changes in
24-RQ (Fig. 2A) can be divided in the same manner, but
unlike 24-EE, positive residuals indicate relatively low
changes in fat oxidation, whereas negative residuals indicate
relatively high changes. In response to weight gain there was
a negative correlation between the residuals of ⌬ 24-EE and
the residuals of ⌬ 24-RQ (Fig. 3), i.e. subjects with metabolic
adaptation in 24-EE (positive residuals) also tended to have
metabolic adaptation in substrate oxidation (negative resid-
uals in 24-RQ) and vice versa. In response to weight loss, the
residuals in ⌬ 24-EE and ⌬ 24-RQ were unrelated, i.e. met-
abolic adaptation in 24-EE did not tend to be accompanied
by metabolic adaptation in substrate oxidation (Fig. 3).

FIG. 2. A, Relationship between changes in 24-RQ, adjusted for en-

ergy balance, and ⌬ weight over 3.6 ⫾ 2.7 yr in 102 nondiabetic Pima
Indians. B, Relationship between ⌬ 24-RQ and ⌬ weight after ad-
justment of 24-RQ for percent body fat, age, and sex in addition to
energy balance. C, Predicted changes in 24-RQ attributable to a 15-kg
weight loss and a 15-kg weight gain (arrows) superimposed onto the
cross-sectional relationship between 24-RQ and body weight as as-
sessed in 491 nondiabetic Pima Indians (prediction line with its upper
and lower 95% confidence intervals is shown).

sectional relationship between fat oxidation and body weight

(⫾99 Cal/day).
The ⌬ weight alone explained only 9% of the variance in
⌬ 24-RQ (r2 ⫽ 0.09). In a multiple regression analysis, the only
additional independent determinant of ⌬ 24-RQ was age at
baseline, explaining another 3% of the variance: ⌬ 24-RQ ⫽
⫺0.023 ⫺0.001 ⌬ weight (kg) ⫹ 0.001 age (yr) [residual vari-
ance (公MSE), 0.029]. Changes in percent body fat (P ⫽ 0.09)
or WTR (P ⫽ 0.13) were not significant independent deter- FIG. 3. Relationship between residual ⌬ 24-RQ and residual ⌬ 24-EE
in response to either weight gain (upper panel; n ⫽ 71) or weight loss
minants of ⌬ 24-RQ, nor was sex, age, glucose tolerance (lower panel; n ⫽ 31) in 102 nondiabetic Pima Indians in whom body
status, initial body weight, or follow-up duration. weight changed widely over 3.6 ⫾ 2.7 yr.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
 METABOLIC ADAPTATION TO WEIGHT CHANGE 1091

Discussion expenditure of only 70 Cal/day was associated with marked

In the present longitudinal study we examined the differences in subsequent weight gain.
changes in 24-h energy expenditure and 24-h substrate ox- Second, the present study was observational in design,
idation associated with spontaneous long term weight which has both advantages and disadvantages. On the one
changes in more than 100 Pima Indians who spent 24-h in a hand, we have no information on the exact causes of the
respiratory chamber at baseline and after a mean follow-up weight changes. In some individuals, weight loss might have
of 3.6 yr, during which changes in body weight varied been secondary to illness, although this is unlikely because
subjects in our studies typically remain in close contact with
widely.
the research unit and receive a comprehensive medical ex-
The results indicate that metabolic adaptation, i.e. changes
amination before each admission. An advantage of the ob-
in energy expenditure and substrate oxidation greater than
servational design, on the other hand, is that it allows us to
predicted for the change in body size and composition, can
examine the metabolic responses to spontaneous long term
occur in response to spontaneous long term weight changes.
weight changes that probably more closely resemble the
On the average, the metabolic changes were only slightly
typical pattern of weight change under free living conditions
greater than predicted, but varied substantially among in-
than imposed by over- and underfeeding regimens. The fact
dividuals. Finally, we found that in response to weight gain,
that the magnitude of metabolic adaptation in response to
adaptations in energy expenditure and substrate oxidation
such gradual weight change was small, on the average,
were related to one another, such that subjects with the most
agrees with cross-sectional findings indicating that energy
pronounced metabolic adaptation in energy expenditure also
expenditure is only marginally reduced in formerly obese
had the most pronounced metabolic adaptation in fat oxi-
individuals who had returned to a normal body weight and
dation and vice versa. This was not the case for weight loss. had successfully maintained the weight loss over months or
Most previous intervention studies have demonstrated years (postobese individuals) (50). Some previous interven-
metabolic adaptation in response to experimental short term tion studies suggest that the suppression in energy expen-
weight change induced by controlled over- and underfeed- diture in response to weight loss might be larger shortly after
ing regimens (14 –36). Whether similar overcompensatory a more rapid decrease in body weight (26, 27, 29, 31, 36).
changes in energy expenditure and fat oxidation occur in the It is also important to point out that energy expenditure
natural history of weight changes has been a matter of con- in the present study was measured in the restricted en-
tention (9 –13, 16, 26). vironment of a respiratory chamber, which significantly
The present study demonstrates, for the first time, that reduces physical activity. Although nonexercise activity
metabolic adaptation can occur in response to spontaneous thermogenesis, of which spontaneous physical activity is
long term weight changes, but also reveals that, on the av- a component, has recently been suggested to play an im-
erage, these overcompensatory changes are small. We esti- portant role in the adaptation to overfeeding (21), our
mate that a 15-kg weight change is accompanied by a change findings do not suggest a major role of spontaneous phys-
in 24-h energy expenditure of 244 Cal/day, which is only 33 ical activity (i.e. fidgeting) in the metabolic response to
Cal/day greater than predicted from the cross-sectional re- long term weight change. To what extent changes in vo-
lationship between 24-h energy expenditure and body litional physical activities such as exercise habits contrib-
weight (211 Cal/day). The change in 24-h fat oxidation after ute to the overall metabolic responses to long term weight
a 15-kg weight change was 53 Cal/day greater than pre- change remains unknown. Our study also provides no
dicted from the cross-sectional data. In practical terms, these information on the role of spontaneous adaptations in
adaptations translate into the caloric content of approxi- energy intake. Because an increase in body weight of 15 kg
mately one half of an apple, one fifth of a bagel, or one tenth resulted in an increase in energy expenditure of about
of a cheeseburger (for the adaptation in 24-h energy expen- 200 –250 Cal/day, a similar increase in energy intake must
diture) or the fat content of two teaspoons of peanut butter have occurred to maintain the higher body weight. Thus,
or seven potato chips (for the metabolic adaptation in 24-h fat as with the metabolic adaptation in energy expenditure,
oxidation), respectively. small differences in the adaptation in energy intake may
These data indicate that in the long term, the defense mech- play an important role in determining whether body
anisms of the body to resist weight gain by an overcompen- weight remains stable or continues to increase.
satory increase in energy expenditure and/or fat oxidation are Although the adaptations in energy metabolism were
relatively weak and easy to offset by small changes in food small, on the average, the present study reveals that the
intake. The results also indicate that even a large decrease in responses of both 24-h energy expenditure (sd, ⫾192 Cal/
body weight over several years is, on the average, not accom- day) and 24-h fat oxidation (sd, ⫾286 Cal/day) to weight
panied by a profound slowing of energy metabolism, as occa- change vary substantially among individuals. Thus, weight
sionally implied to explain the high rate of weight recidivism gain/loss is not universally accompanied by small adaptive
in the medical treatment of obesity. However, several aspects increases/decreases in energy expenditure and/or fat oxi-
need to be considered in this respect. dation. Rather, some individuals will experience relatively
First, it has been demonstrated that even small differences large overcompensatory responses, whereas others will have
in energy expenditure and/or substrate oxidation that may subnormal responses. Such interindividual variability in
appear trivial on a daily basis can have an important impact metabolic responses has also been found in response to ex-
on body weight over the long term (2, 6). In a previous perimental over- and underfeeding and has been used to
prospective study (2), we found that a difference in energy explain why the amount of weight gained or lost under

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
1092 WEYER ET AL. JCE & M • 2000
Vol 85 • No 3

standardized dietary regimens can differ substantially The above findings may have important implications for
among individuals (14, 19, 22, 26). The large number of our understanding of the role of energy metabolism in the
subjects in the present study allowed us to quantify the long term regulation of body weight and the pathogenesis of
interindividual variability in metabolic responses to weight human obesity. To illustrate this, we have developed a sche-
changes and to search for possible underlying determinants. matic model that integrates previous cross-sectional and pro-
We found that the change in 24-h energy expenditure was spective findings with those from the present longitudinal
explained not only by the changes in FFM and FM, but also study (Fig. 4).
independently by the changes in body fat distribution and Cross-sectionally, energy expenditure and the rate of fat to
spontaneous physical activity. The only additional determi- carbohydrate oxidation increase with increasing body size
nant of the change in 24-h respiratory quotient was age at (prediction line), but at any given body size, both measures
baseline. The effect of these additional factors was small, vary considerably among individuals (44, 51, 57). Several
however, and as much as 37% of the variability in ⌬ 24-h prospective studies revealed the relative importance of this
energy expenditure and 89% of the variability in ⌬ 24-h variation, in that relatively low rates of energy expenditure
respiratory quotient remained unexplained. As only a small (2–5) and/or fat oxidation (6, 7) predispose individuals to
part of this variability can be attributed to the variability of weight gain (arrows 1A and 1B in Fig. 4). Our present lon-
the method (2, 6, 44), other factors must be involved. These gitudinal data indicate that upon gaining weight, the initially
may include, for example, changes in plasma thyroid or sex low rates of energy expenditure and fat oxidation tend to
hormone concentrations (49), autonomic nervous system ac- normalize, on the average, but as with the cross-sectional
tivity (51, 52), mitochondrial uncoupling activity (53), plasma relationship, there is substantial interindividual variability in
insulin and free fatty acid concentrations, and glucose tol- these responses. Accordingly, the metabolic drive to weight
erance (54), each of which was found to be related to energy gain may soon diminish in some individuals, thereby lim-
expenditure and/or substrate oxidation in cross-sectional iting the amount of weight gain (metabolic adaptation; arrow
studies. Moreover, there is strong evidence from overfeeding 1A), whereas it may be sustained in others, who will thus be
studies in identical twins that the metabolic responses to predisposed to gain further weight (arrow 1B). Conversely,
weight changes are in part genetically determined (55, 56). individuals with relatively high rates of energy expenditure
Another interesting observation in the present study was and/or fat oxidation are predisposed to spontaneously lose
that in response to weight gain, the changes in 24-h energy weight, but the amount of weight loss will depend upon the
expenditure and 24-h substrate oxidation were related to one occurrence (arrow 2A) or not (arrow 2B) of metabolic adap-
another, in that individuals with the most pronounced ad- tation. Of note, these considerations apply to spontaneous
aptation in energy expenditure also tended to have the most long term weight changes. If, for instance, an obese individ-
pronounced adaptation in fat oxidation and vice versa. In- ual decides to intentionally lose weight, his or her initial
terestingly, this was not the case in response to weight loss, energy expenditure and/or fat oxidation may fall below
where adaptations in energy expenditure and substrate ox- rather than above the normal range.
idation were unrelated. Further studies are needed to confirm the role of low

FIG. 4. Schematic model integrating cross-sectional, prospective, and longitudinal findings to illustrate the potential role of energy expenditure
and substrate oxidation in the long term regulation of body weight. Arrows 1A/2A, Pronounced metabolic adaptation3 quick decrease in
the metabolic drive3 small weight change; arrows 1B/2B, no metabolic adaptation3 sustained metabolic drive3 large weight change (detailed
explanation provided in Discussion).

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
 METABOLIC ADAPTATION TO WEIGHT CHANGE
energy expenditure and fat oxidation as predictors of weight
gain and to formally test the effect of metabolic adaptation
on further weight change. It will also be important to ex-
amine the role of adaptation in energy and substrate intake
to weight change. These are probably complex and could
include changes in the perception of hunger and satiation as
well as in caloric intake and food preferences.
In summary, the results of this longitudinal study indicate
that the changes in 24-h energy expenditure and 24-h respi-
ratory quotient (i.e. in substrate oxidation) associated with
long term weight changes 1) are greater than those predicted
for the change in body size and composition, 2) vary sub-
stantially among individuals, and 3) are related to one an-
other in response to weight gain.
We conclude that metabolic adaptation can occur not only
in response to experimental short term perturbations in body
weight, but also in response to spontaneous long term weight
changes. These responses, albeit small on the average, vary
substantially among individuals and may thus play a role in
the long term regulation of body weight and the pathogen-
esis of human obesity.
Acknowledgments
We gratefully acknowledge Mr. Tom Anderson, Mrs. Carol Massen-
gill, and the nurses of the Clinical Research Unit as well as the staff of
the metabolic kitchen for their care of the patients in the studies, and the
Clinical Diabetes and Nutrition Section technical staff for assisting with
the chamber measurements and laboratory analyses. We thank the mem-
bers and leaders of the Gila River Indian Community, without whose
continuing cooperation this study would not have been possible.
References
1. Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL. 1998 Overweight and
obesity in the United States: prevalence and trends, 1960 –1994. Int J Obes.
22:39 – 47.
2. Ravussin E, Lillioja S, Knowler WC, et al. 1988 Reduced rate of energy
expenditure as a risk factor for body weight gain. N Engl J Med. 318:467– 482.
3. Buscemi S, Di Maggio O, Blunda G, Maneri R, Verga S, Bompiani GD. 1998
A low resting metabolic rate is associated to body weight gain in adult Cau-
casian subjects: preliminary results of a 8 –10 year longitudinal study. Int J
Obes. 22(Suppl 1):S75.
4. Griffiths M, Payne PR, Stunkard AJ, Rivers JPW, Cox M. 1990 Metabolic rate
and physical development in children at risk for obesity. Lancet. 336:76 –77.
5. Roberts SB, Savage J, Coward WA, Chew B, Lucas A. 1990 Energy expen-
diture and intake in infants born to lean and overweight mothers. Lancet.
318:461– 466.
6. Zurlo F, Lillioja S, Puente A, et al. 1991 Low ratio of fat to carbohydrate
oxidation as a predictor of weight gain: study of 24-RQ. Am J Physiol.
259:E650 –E657.
7. Seidell JA, Muller DC, Sorkin JD, Andres R. 1992 Fasting respiratory ex-
change ratio and resting metabolic rate as predictors of weight gain: the
Baltimore Longitudinal Study on Aging. Int J Obes. 16:667– 674.
8. Davies PSW, Day JME, Lucas A. 1991 Energy expenditure in early infancy and
later body fatness. Int J Obes. 15:727–731.
9. Weinsier RL, Nelson KM, Hensrud DD, Darnell BE, Hunter GR, Schutz Y.
1995 Metabolic predictors of obesity. Contributions of resting energy expen-
diture, thermic effect of food, and fuel utilization to four-year weight gain in
post-obese and never-obese women. J Clin Invest. 95:980 –985.
10. Amatruda JM, Statt MC, Welle SL. 1993 Total and resting energy expenditure
in obese women reduced to ideal body weight. J Clin Invest. 92:1236 –1242.
11. Flatt JP. 1988 Importance of nutrient balance in body weight regulation. Di-
abetes Metab Rev. 6:571–581.
12. Ravussin E, Swinburn BA. 1996 Energy expenditure and obesity. Diabetes
Rev. 4:403– 422.
13. Ravussin E, Swinburn BA. 1993 Energy metabolism. In: Stunkard AJ, Wadden
TA, eds. Obesity: theory and therapy, 2nd Ed. New York: Raven Press; 97–123.
14. Miller DS, Mumford P, Stock MJ. 1967 Gluttony. II. Thermogenesis in over-
eating man. Am J Clin Nutr. 20:1223–1229.
15. Sims EAH, Danforth Jr E, Horton ES, Bray GA, Glennon JA, Salans LB. 1973
Endocrine and metabolic effects of experimental obesity in man. Recent Prog
Horm Res. 29:457– 496.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
1093
16. Garrow JS. 1978 The regulation of energy expenditure in man. In: Bray GA,
ed. Recent advances in obesity research. London: Newman; vol 2:200 –210.
17. Norgan NG, Durnin JVGA. 1980 The effect of 6 weeks of overfeeding on the
body weight, body composition, and energy metabolism of young men. Am J
Clin Nutr. 33:978 –988.
18. Ravussin E, Schutz Y, Acheson KJ, Bourquin L, Jequier E. 1985 Short-term,
mixed-diet overfeeding in man: no evidence for “luxuskonsumption.” Am J
Physiol. 249:E470 –E474.
19. Tremblay A, Despres JP, Theriault G, Fournier G, Bouchard C. 1992 Over-
feeding and energy expenditure in humans. Am J Clin Nutr. 56:857– 862.
20. Deriaz O, Fournier G, Tremblay A, Despres JP, Boucahrd C. 1992 Lean-
body-mass composition and resting energy expenditure before and after long-
term overfeeding. Am J Clin Nutr. 56:840 – 847.
21. Levine JA, Eberhardt NL, Jensen M. 1999 Role of nonexercise activity ther-
mogenesis in resistance to fat gain in humans. Science. 283:212–214.
22. Bouchard C, Tremblay A, Despres JP, et al. 1996 Overfeeding in identical
twins: 5-year postoverfeeding results. Metabolism. 45:1042–1050.
23. Benedict FG, Miles WR, Roth P, Smith HM. 1919 Human vitality and effi-
ciency under a prolonged restricted diet. Washington DC: Carnegie Institute;
Publication 280.
24. Keys A, Brozek J, Henschel A, Mickelsen O, Taylor HL. 1950 The biology of
human starvation, vol 1. Minneapolis, University of Minneapolis Press.
25. Grande F, Anderson JT, Keys A. 1958 Changes of basal metabolic rate in man
in semi-starvation and refeeding. J Appl Physiol. 12:230 –238.
26. Leibel RL, Rosenbaum M, Hirsch J. 1995 Changes in energy expenditure
resulting from altered body weight. N Engl J Med. 332:621– 628.
27. Welle SL, Campbell RG. 1986 Decrease in resting metabolic rate during rapid
weight loss is reversed by low dose thyroid hormone therapy. Metabolism.
35:289 –291.
28. DeGrott L, van Es AJH, van Raaij JMA, Vogt JE, Hautvast JGAJ. 1990 Energy
metabolism of overweight women 1 mo and 1 y after an 8-week slimming
period. Am J Clin Nutr. 51:578 –583.
29. Mansell PI, Macdonald IA. 1988 The effect of underfeeding on the physio-
logical response to food ingestion in normal weight women. Br J Nutr.
60:39 – 48.
30. Bessard T, Schutz Y, Jequier E. 1983 Energy expenditure and postprandial
thermogenesis in obese women before and after weight loss. Am J Clin Nutr.
38:680 – 693.
31. DeBoer JO, van Es AJH, Roovers A, van Raaij JMA, Hautvast JGAJ. 1986
Adaptations of energy metabolism of overweight women to low-energy intake,
studied with whole body calorimeters. Am J Clin Nutr. 44:585–595.
32. Ravussin E, Burnand B, Schutz Y, Jéquier E. 1985 Energy expenditure
before and during energy restriction in obese patients. Am J Clin Nutr.
41:753–759.
33. Froidevaux F Schutz Y, Christin L, Jéquier E. 1993 Energy expenditure in
obese women before and during weight loss, after refeeding, and in the weight-
relapse period. Am J Clin Nutr. 57:35– 42.
34. Elliot DL, Goldberg L, Kuehl KS, Bennett WM. 1989 Sustained suppression
of the resting metabolic rate after massive weight loss. Am J Clin Nutr.
49:93–96.
35. Buscemi S, Caimi G, Verga S. 1996 Resting metabolic rate and postabsorptive
substrate oxidation in morbidly obese subjects before and after massive weight
loss. Int J Obes. 20:41– 46.
36. Heshka S, Yang MU, Wand J, Burt P, Pi-Sunyer FX. 1990 Weight loss and
change in resting metabolic rate. Am J Clin Nutr. 52:981–986.
37. Food and Agriculture Organization/WHO/United Nations University. 1985
Energy and protein requirements. Geneva: WHO; WHO Tech Rep Ser 724.
38. WHO. 1985 Diabetes mellitus: report of a WHO study group. Geneva; WHO;
WHO Tech Rep Ser 17:45–51.
39. Goldman RF, Buskirk ER. 1961 A method for underwater weighing and the
determination of body density. In: Brozek J, Herschel A, eds. Techniques for
measuring body composition. Washington DC: National Academy of Sciences;
78 –106.
40. Mazess RB, Barden HS, Bisek JP, Hanson J. 1990 Dual-energy x-ray absorp-
tiometry for total-body and regional bone-mineral and soft-tissue composition.
Am J Clin Nutr. 51:1106 –1112.
41. Siri WE. 1961 Body composition from fluid spaces and density: analysis of
methods. In: Brozek J, Herschel A, eds. Techniques for measuring body com-
position. Washington DC: National Academy of Sciences; 223–244.
42. Tataranni PA, Ravussin E. 1995 Use of dual-energy x-ray absorptiometry in
obese individuals. Am J Clin Nutr. 62:730 –734.
43. Tataranni PA, Larson DE, Ravussin E. 1994 Body fat distribution and energy
metabolism in obese men and women. J Am Coll Nutr. 13:569 –574.
44. Ravussin E, Lillioja S, Anderson TE, Christin L, Bogardus C. 1986 Deter-
minants of 24-hour energy expenditure in man: methods and results using a
respiratory chamber. J Clin Invest. 78:1568 –1578.
45. Abbott WGH, Howard BV, Ruoloto G, Ravussin E. 1990 Energy expenditure
in humans: effects of dietary fat and carbohydrate. Am J Physiol
258:E347–E351.
46. Jequier E, Acheson K, Schutz Y. 1987 Assessment of energy expenditure and
fuel utilization in man. Annu Rev Nutr. 7:187–208.
1094
47. SAS Institute. 1990 SAS/STAT user’s guide, version 6, 4th Ed. Cary: SAS
Institute.
48. Deleted in proof.
49. Deleted in proof.
50. Astrup A, Gotzsche PC, van de Werken K, Ranneries C, Toubro S, Raben A,
Bueman B. 1999 Meta-analysis of resting metabolic rate in formerly obese
subjects. Am J Clin Nutr. 69:1117–1122.
51. Toubro S, Sorensen TIA, Ronn B, Christensen NJ, Astrup A. 1996 Twenty-
four-hour energy expenditure: the role of body composition, thyroid status,
sympathetic activity, and family membership. J Clin Endocrinol Metab.
81:2670 –2674.
52. Spraul M, Ravussin E, Fontvieille AM, Rising R, Larson DE, Anderson EA.
1993 Reduced sympathetic nervous activity. A potential mechanism predis-
posing to body weight gain. J Clin Invest. 92:1730 –1735.
53. Schrauwen P, Xia J, Bogardus C, Pratley RE, Ravussin E. 1998 Skeletal muscle
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 October 2015. at 22:27 For personal use only. No other uses without permission. . All rights reserved.
WEYER ET AL. JCE & M • 2000
Vol 85 • No 3
uncoupling protein 3 expression is a determinant of energy expenditure in
Pima Indians. Diabetes. 48:146 –149.
54. Weyer C, Bogardus C, Pratley RE. 1999 Metabolic factors contributing to
increased resting metabolic rate and decreased insulin-induced thermogenesis
during the development of type 2 diabetes. Diabetes. 48:1607–1614.
55. Bouchard C, Tremblay A, Despres JP, et al. 1990 The response to long-term
overfeeding in identical twins. N Engl J Med. 322:1477–1482.
56. Bouchard C, Perusse L, Deriaz O, Despres JP, Tremblay A. 1993 Genetic
influence on energy expenditure in humans. Crit Rev Food Sci Nutr.
33:345–350.
57. Weyer C, Snitker S, Rising R, Bogardus C, Ravussin E. 1999 Determinants
of energy expenditure and fuel utilization in man: effects of body compo-
sition, age, sex, ethnicity, and glucose tolerance in 916 subjects. Int J Obes.
23:715–722.