Energy expenditure is very high in extremely obese women.
ABSTRACT To test the hypothesis that total energy expenditure (TEE) and resting energy expenditure (REE) are low in extremely obese individuals, factors that could contribute to maintenance of excess weight, a cross-sectional study was conducted in 30 weight stable, extremely obese women [BMI (mean +/- SEM) 48.9 +/- 1.7 kg/m(2)]. TEE was measured over 14 d using the doubly labeled water method, REE and the thermic effect of feeding (TEF) were measured using indirect calorimetry, and activity energy expenditure (AEE) was calculated as TEE - (REE + TEF). Body composition was determined using a 3-compartment model. Subjects were divided into tertiles of BMI (37.5-45.0; 45.1-52.0; and 52.1-77.0 kg/m(2)) for data analysis. TEE and REE increased with increasing BMI tertile: TEE, 12.80 +/- 0.5, 14.67 +/- 0.5, and 16.10 +/- 0.9 MJ/d (P < 0.01); REE, 7.87 +/- 0.2, 8.78 +/- 0.3, and 9.94 +/- 0.6 MJ/d (P < 0.001), and these values were 29-38% higher than published means of measured TEE in nonobese individuals. No significant differences were observed among BMI tertiles for AEE, TEF, or physical activity level (PAL = TEE/REE, overall mean 1.64 +/- 0.16). The Harris-Benedict and WHO equations provided the closest estimates of REE (within 3%), whereas the obese-specific equations of Ireton-Jones overpredicted (40%) and Bernstein underpredicted (21%) REE. Extremely obese individuals have high absolute values for TEE and REE, indicating that excess energy intake contributes to the maintenance of excess weight. Standard equations developed for nonobese populations provided the most accurate estimates of REE for the obese individuals studied here. REE was not accurately predicted by equations developed in obese populations.
- SourceAvailable from: Fernando Carrasco[Show abstract] [Hide abstract]
ABSTRACT: La obesidad constituye en la actualidad un problema de suma importancia, debido a su asociación con comorbilidades (hiper-tensión arterial, diabetes mellitus tipo II, dislipidemia, algunos ti-pos de cáncer, cardiopatía isquémica, problemas osteoarticula-res, apnea obstructiva del sueño) (1) ; su prevalencia en aumento a nivel mundial, alcanzando cifras de 23,3% en la población mayor de 17 años en Chile (2) ; y por su papel de factor de riesgo indepen-diente de mortalidad (3) . Las estrategias de intervención en la obesidad incluyen cambios en la alimentación, tratamiento farmacológico, terapia conduc-tual, ejercicio físico y técnicas de cirugía bariátrica. Parte del éxi-to de estas intervenciones radica en realizar una estimación ade-cuada de los requerimientos de energía (4,5) ; e independiente de la técnica que se utilice, el éxito en la perdida de peso va a depen-der del balance energético negativo que se logre mantener du-rante un período prolongado de tiempo. La calorimetría directa o indirecta y el método del agua doblemente marcada, constituyen los instrumentos más precisos para determinar el gasto energético; sin embargo, son de un costo relativamente elevado, requieren personal entrenado y un tiempo determinado para realizarlos (4) .Revista Hospital Clínico Universidad de Chile. 01/2005; 16(4):267-272.
- [Show abstract] [Hide abstract]
ABSTRACT: Stress places a metabolic burden on homeostasis and is linked to heightened sympathetic activity, increased energy expenditure and pathology. The yogic state is a hypometabolic state that corresponds with mind-body coherence and reduced stress. This study aimed to investigate metabolic responses to stress and different yoga practices in regular yoga practitioners (YP), non-yoga practitioners (NY) and metabolic syndrome patients (MS).BMC Complementary and Alternative Medicine 11/2014; 14(1):445. · 1.88 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: To assess the effect of a low calorie diet on the resting metabolic rate (RMR), substrate oxidation, body composition, and to compare measured and calculated RMR of obese Brazilian women, we selected 19 patients aged 31 ± 9 years, with a body mass index of 51 ± 8 kg/m2, for admission to the Metabolic Unit of the University Hospital for 8 weeks, who were then submitted to a 3.3 to 4.2 MJ/d (800-1000 kcal/d) diet. Weight, height, and circumferences were measured on the first and last days of the study. Body composition was assessed by bioelectrical impedance, and RMR and substrate oxidation rate by indirect calorimetry. A decrease in body weight (134 ± 23 kg vs 121 ± 21 kg, P < .05), waist (136 ± 17 cm vs 123 ± 17 cm, P < .05), and hip circumference (149 ± 14 cm vs 137 ± 16 cm, P < .05) occurred during the study. Mean RMR measured by indirect calorimetry (10.6 ± 1.7 MJ/d; 2540 ± 420 kcal/d) was 16% higher (P < .05) than that calculated by Harris-Benedict and World Health Organization equations (8.7 ± 0.9 MJ/d; 2070 ± 210 kcal/d and 9.0 ± 1.4 MJ/d; 2161 ± 344 kcal/d, respectively) at the beginning, but not at the end of the study. Lipid oxidation rate was 45% of RMR at the beginning of the study, reaching 59% at the end (P > .05). Present data suggest that equations to estimate RMR of obese females are reliable after a low-calorie diet and weight loss. Resting metabolic rate was correlated with fat-free mass and body fat. A low-calorie diet with balanced macronutrients is effective for weight loss, leading to a maintenance of lipid oxidation rate and to a reduction of carbohydrate and protein oxidation rates. The low-calorie diet reduced body fat and maintained lean mass.Nutrition Research - NUTR RES. 01/2006; 26(9):437-442.
Human Nutrition and Metabolism
Energy Expenditure Is Very High in Extremely Obese Women1
Sai Krupa Das,*2Edward Saltzman,*†Megan A. McCrory,* L. K. George Hsu,†
Scott A. Shikora,†Gregory Dolnikowski,* Joseph J. Kehayias,* and Susan B. Roberts*
*Jean Mayer U.S. Department of Agriculture Human Nutrition Research Center on Aging at Tufts University,
Boston, MA 02111 and†Tufts-New England Medical Center Hospital, Boston, MA 02111
low in extremely obese individuals, factors that could contribute to maintenance of excess weight, a cross-
sectional study was conducted in 30 weight stable, extremely obese women [BMI (mean ? SEM) 48.9 ? 1.7 kg/m2].
TEE was measured over 14 d using the doubly labeled water method, REE and the thermic effect of feeding (TEF)
were measured using indirect calorimetry, and activity energy expenditure (AEE) was calculated as TEE ? (REE ?
TEF). Body composition was determined using a 3-compartment model. Subjects were divided into tertiles of BMI
(37.5–45.0; 45.1–52.0; and 52.1–77.0 kg/m2) for data analysis. TEE and REE increased with increasing BMI tertile:
TEE, 12.80 ? 0.5, 14.67 ? 0.5, and 16.10 ? 0.9 MJ/d (P ? 0.01); REE, 7.87 ? 0.2, 8.78 ? 0.3, and 9.94 ? 0.6 MJ/d
(P ? 0.001), and these values were 29–38% higher than published means of measured TEE in nonobese
individuals. No significant differences were observed among BMI tertiles for AEE, TEF, or physical activity level
(PAL ? TEE/REE, overall mean 1.64 ? 0.16). The Harris-Benedict and WHO equations provided the closest
estimates of REE (within 3%), whereas the obese-specific equations of Ireton-Jones overpredicted (40%) and
Bernstein underpredicted (21%) REE. Extremely obese individuals have high absolute values for TEE and REE,
indicating that excess energy intake contributes to the maintenance of excess weight. Standard equations
developed for nonobese populations provided the most accurate estimates of REE for the obese individuals
studied here. REE was not accurately predicted by equations developed in obese populations.
To test the hypothesis that total energy expenditure (TEE) and resting energy expenditure (REE) are
J. Nutr. 134:
● extreme obesity ● energy expenditure ● fat-free mass ● prediction equations
The prevalence of extreme obesity, as defined by BMI ? 40
kg/m2, is now 5% in U.S. adults (1), representing a 3-fold
increase over the last 40 years. This previously rare condition
is associated with profound adverse health consequences; how-
ever, very little is known about factors that promote and
maintain extreme obesity. Studies using doubly labeled water
indicated that individuals with higher than recommended
body weight (BMI ? 25 kg/m2) have high absolute total
energy expenditure (TEE)3compared with individuals of lower
BMI (2–4). This implies that a high energy intake is required
to maintain the excess weight. However, most of those studies
were conducted in overweight or marginally obese individuals,
and there is very little information available on the extremely
obese individuals who may be more likely to have metabolic
abnormalities that could predispose to low energy require-
The aim of the present study was to test the hypothesis that
the weight maintenance energy expenditure of extremely
obese individuals is low relative to both published data for
nonobese individuals (3) and TEE predicted by the equations
of the U.S. Dietary Reference Intakes (DRIs) (5). In addition,
the accuracy of equations for predicting resting energy expen-
diture (REE) was assessed using prediction equations devel-
oped for use with nonobese (5–7) individuals and equations
developed for use in obese individuals (8,9).
SUBJECTS AND METHODS
Subjects. The subjects were 30 women who were awaiting gastric
bypass (GBP) surgery for weight reduction at the Tufts-New England
Medical Center Hospital (Table 1). Individuals were ineligible for
the study if they had diabetes, cancer, coronary heart disease, endo-
crine disorders, other acute or chronic diseases, or medication use
known to influence energy expenditure. Measurements were con-
ducted at the Clinical Study Unit (CSU) of the Tufts-New England
Medical Center Hospital and at the Jean Mayer USDA Human
Nutrition Research Center on Aging at Tufts University. The study
was approved by the Human Investigation Review Committee of
Tufts-New England Medical Center Hospital. All subjects gave writ-
ten, informed consent before participating.
Study design. All subjects were monitored for weight stability for
1 mo before the study and only those who were weight stable (defined
as body weight maintained within ? 2.3 kg of starting weight) were
allowed to participate. Two subjects whose BMI was slightly below 40
kg/m2and who were eligible for GBP surgery were included in the
1Supported by National Institutes of Health grants MH/DK54092–01A3,
M01-RR00054, and P30 DK 46200 and by the U.S. Department of Agriculture,
Agricultural Research Service under contract 53–3K06–5-10.
2To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
3Abbreviations used: AEE, activity energy expenditure; CSU, clinical studies
unit; DRI, dietary reference intakes; FFM, fat-free mass; FQ, food quotient; GBP,
gastric bypass; LTPA, leisure time physical activity; PAL, physical activity level;
REE, resting energy expenditure; RQ, respiratory quotient; TEE, total energy
expenditure; TEF, thermic effect of feeding.
0022-3166/04 $8.00 © 2004 American Society for Nutritional Sciences.
Manuscript received 8 December 2003. Initial review completed 4 January 2004. Revision accepted 29 March 2004.
at TUFTS UNIVERSITY on July 24, 2012
study population. The study was conducted over a 15-d period.
Subjects were free-living except for 2 overnight inpatient stays for
measurements of REE and the thermic effect of feeding (TEF), and an
additional 36-h period for other testing. Subjects were admitted to
the CSU the day before the start of the study. After an overnight fast,
measurement of TEE by doubly labeled water was started and REE
was determined followed by a 4-h TEF measurement. Body compo-
sition was also measured as described below. Subjects then returned
home with instructions on how to collect daily urine samples for the
measurement of TEE. They were readmitted on d 7–8 for additional
body composition and other measurements and then returned home
for the rest of the week and continued urine collections at home.
Subjects returned to the CSU on the evening of d 14, and a second
REE measurement was conducted on the morning of d 15. Other
measures during the study period included collection of blood samples
from fasting subjects and the administration of food frequency and
physical activity questionnaires.
A 15-d doubly labeled water study was con-
ducted to measure TEE. REE and TEF were measured using indirect
calorimetry (Deltatrac portable metabolic cart, Sensor Medics). The
Siri 3-compartment model was used to estimate the percentage of
body fat and body composition, and standard anthropometric mea-
sures were also obtained. Biochemical measures such as thyroid func-
tion tests and leptin were determined in blood from fasting subjects.
Insulin and glucose were also measured in fasting subjects, and the
homeostasis model assessment for insulin sensitivity (HOMA) was
calculated using these values (10,11). Self-reported leisure time ac-
tivity during the previous 12-mo period was determined by a struc-
tured interview using the Minnesota Leisure Time Physical Activity
(LTPA) questionnaire; occupational activity over the previous 12 mo
was also assessed using the self-administered Tecumseh Occupational
Activity Questionnaire. All of the above methods were detailed and
referenced previously (12,13).
Statistics. Statistical analyses were performed using SPSS 10.0.7
and SYSTAT 9.0.1 (SPPS) and SAS (version 8, SAS Institute).
Values are expressed as means ? SEM unless otherwise specified.
Given the wide BMI range (37.5–77 kg/m2) in our study subjects,
BMI was divided into tertiles to examine trends across different BMI
levels. Baseline characteristics were compared among tertiles using
1-way ANOVA, and energy expenditure variables were compared
using both ANOVA and analysis of covariance (ANCOVA), adjust-
ing for fat-free mass (FFM). Tukey’s Honestly Significant Difference
(HSD) test for multiple comparisons was used to determine signifi-
cance among groups. Linear multiple regression analysis was per-
formed to determine the best predictors of energy expenditure and
suitable models to predict TEE, REE, and activity energy expenditure
To test the hypothesis that TEE was low, measured TEE was
compared with published TEE data for nonobese individuals (3) and
data for nonobese from our own laboratory (14) using a 1-sample t
test. Using ANOVA and Tukey’s HSD test, measured TEE was also
compared with TEE predicted using the equations of the U.S. DRIs
(5). ANOVA with Dunnett’s post-hoc test was used to determine
whether predicted REE using the WHO equations (6), the Harris-
Benedict equation (7), and with 2 equations developed or recom-
mended for use in the obese (8,9) were significantly different from
measured REE. In addition, the sum of squared differences between
measured and predicted REE was used to assess the accuracy of the
predictions. For all tests, differences were considered significant at P
All subjects were weight stable during the 1-mo period
preceding the study [? body weight (mean ? SD), 1.46 ? 0.35
kg of starting weight] and were weight stable during the study
period (? body weight, 4.2 ? 26 g/d). Physical characteristics
of the subjects except for body weight after fasting and BMI
were similar among the tertiles of BMI in these extremely
obese women (Table 1), and values of all biochemical vari-
ables were within normal range for all subjects. Glucose was
significantly higher in fasting subjects in the highest BMI
tertile compared with the middle tertile (P ? 0.05), and leptin
was elevated in fasting subjects in the middle BMI tertile (P
? 0.05) compared with the lowest tertile.
Energy expenditure variables by tertile of BMI are shown
unadjusted for body composition (Table 2) and therefore
illustrate the effect of increasing BMI on energy expenditure.
Both TEE and REE were significantly higher in the highest
BMI tertile compared with the lowest BMI tertile. However,
when energy expenditure variables such as TEE and REE were
adjusted for FFM, there were no significant differences among
tertiles. This indicated that the differences in energy expen-
diture between the highest and lowest tertiles were due to
differences in FFM.
TEE values for all BMI tertiles were significantly higher
than the mean TEE (9.51–10.21 MJ/d; P ? 0.001) in non-
obese women from a published summary (3). Measured TEE
was also significantly higher (14.3 MJ/d) than the mean U.S.
DRIs estimates of dietary energy needs using the subjects’
actual weight in 1) equations developed in normal weight
women (11.69 MJ/d; P ? 0.001), 2) the combined equation for
normal, overweight and obese women (12.42 MJ/d; P
? 0.001), or 3) the equation for overweight and obese women
(12.77 MJ/d; P ? 0.001) (5).
No significant differences were observed among BMI ter-
tiles for AEE, TEF, physical activity level (PAL), or food
quotient (FQ). In addition, the respiratory quotient (RQ) in
fasting subjects was significantly higher in the lowest BMI
tertile compared with the highest BMI tertile, but postprandial
RQ (averaged over the 4 h TEF) did not differ significantly
Subject characteristics by tertiles of BMI in 30 weight stable,
extremely obese women1
36.17 ? 0.5 40.10 ? 0.535.38 ? 0.9
106.74 ? 0.2a
134.29 ? 0.3b
161.98 ? 0.6c
161.78 ? 0.4 164.94 ? 0.3163.73 ? 0.6
40.84 ? 0.71a
49.47 ? 1.0b
60.30 ? 2.2c
0.90 ? 0.060.90 ? 0.040.92 ? 0.05
0.30 ? 0.02
2.63 ? 0.49
0.31 ? 0.04
2.45 ? 0.31
0.28 ? 0.03
2.48 ? 0.42
5.29 ? 0.27ab
5.12 ? 0.14a
5.95 ? 0.20b
143.64 ? 30.22a
152.51 ? 19.56b
195.50 ? 39.81ab
42.11 ? 3.50a
59.14 ? 4.6b
55.73 ? 6.62ab
4.51 ? 1.025.10 ? 0.777.69 ? 1.76
1Values are means ? SEM. Means in a row with superscripts
without a common letter differ, P ? 0.05. Symbols indicate a significant
main effect of BMI, * P ? 0.001 and ** P ? 0.05.
2Abbreviations: rT3, reverse triiodothyronine, TSH, thyroid-stimu-
3HOMA IR, insulin resistance by homeostasis model assessment
ENERGY EXPENDITURE IN EXTREME OBESITY
at TUFTS UNIVERSITY on July 24, 2012
among BMI tertiles. There were no significant differences
among BMI tertiles for self-reported physical activity when
expressed in min/d or MJ/d (data for the latter not shown) or
for any of the activity categories (data also not shown). In
addition, there was no significant relation between measured
AEE and energy expenditure calculated from LTPA and oc-
cupational activities (r ? 0.04; P ? 0.8). PAL was also not
predicted by LTPA, occupational, or total activity in the 3
activity categories i.e., light, moderate, or heavy activity (P
Regression analyses (Table 3) showed that TEE was best
predicted by FFM (adjusted R2? 0.56; P ? 0.001). Although
FFM alone significantly predicted REE (adjusted R2? 0.63; P
? 0.001), the addition of fat mass to the regression model
substantially improved the prediction for REE (adjusted R2?
0.71; P ? 0.001). AEE was also significantly predicted by FFM
(adjusted R2? 0.12; P ? 0.05).
TEE and PAL (mean ? SD), obtained in our study were
contrasted with values from published energy expenditure data
compiled by Prentice et al. (3) (Fig. 1). TEE increased with
increasing BMI (P ? 0.0001) in the Prentice et al. data set (3).
In our own data set, mean TEE in the highest BMI tertile was
significantly higher than mean TEE in the lowest tertile of
BMI. In contrast, values for PAL did not differ with increasing
BMI in the women in the Prentice et al. study (3) or in our
extremely obese women.
Measured REE was compared with REE predicted using
several published equations, i.e., the WHO prediction equa-
tion (6), the Harris-Benedict equation (7) for normal weight
individuals, and 2 equations developed for use in the obese
(8,9) (Fig. 2). Both the WHO equation (?0.12 MJ/d; adjusted
P ? 0.97) and the Harris-Benedict equation (0.23 MJ/d;
adjusted P ? 0.79) showed nonsignificant mean differences
from measured REE. The Harris-Benedict equation underpre-
dicted, and the WHO equation overpredicted REE by only
Summary of energy expenditure data by tertiles of BMI in 30 weight stable, extremely obese women1
TEF, % of test meal energy
Self-reported physical activity, min/d
12.80 ? 0.5a
7.87 ? 0.2a
3.7 ? 0.4
6.85 ? 0.71
1.61 ? 0.06
0.79 ? 0.01a
0.81 ? 0.01
0.85 ? 0.01
14.7 ? 0.5a,b
8.78 ? 0.3a,b
4.7 ? 0.3
8.18 ? 1.0
1.67 ? 0.04
0.77 ? 0.01b
0.80 ? 0.01
0.84 ? 0.01
16.1 ? 0.9b
9.94 ? 0.6b
4.9 ? 0.6
10.46 ? 2.2
1.62 ? 0.05
0.77 ? 0.01b
0.80 ? 0.01
0.84 ? 0.01
1Values are means ? SEM. Means in a row with superscripts without a common letter differ, P ? 0.05. Symbols indicate significant main effects
of BMI, * P ? 0.001,†P ? 0.01, and#P ? 0.01. TEE values in all tertiles were also significantly higher than the current U.S. DRI (5) for nonobese
persons, P ? 0.001.
Regression models showing predictors of TEE, REE, and AEE
in 30 weight stable, extremely obese women1
?-Coefficient ? SEE2
P-value Adjusted R2(P)
3.364 ? 1.806
0.177 ? 0.029
2.183 ? 0.930
0.106 ? 0.015
2.872 ? 0.856
0.029 ? 0.01
0.062 ? 0.020
?0.032 ? 1.505
0.071 ? 0.024
0.006 0.21 (0.006)
normal weight, overweight, obese, and extremely obese BMI groups
[Prentice et al. (3)] and in 30 weight stable, extremely obese women in
this study according to BMI tertiles.
TEE (A) and PAL (B) (mean ? SD) by BMI (kg/m2) in
DAS ET AL.
at TUFTS UNIVERSITY on July 24, 2012
3%. The sum of squared differences between measured and
predicted REE was larger for the WHO equation (69 vs. 53)
than for the Harris-Benedict equation, indicating that the
Harris-Benedict equation was more precise. When equations
developed in obese populations were used, measured REE was
significantly different from that predicted using the Ireton-
Jones equation (?3.30 MJ/d; adjusted P ? 0.001) and the
Bernstein equation (1.87 MJ/d; adjusted P ? 0.001). The
Ireton-Jones equation overpredicted REE by 40% and the
Bernstein equation underpredicted REE by 18%. In addition,
the sum of squared differences between measured and pre-
dicted REE was very large for both equations (122, for the
Bernstein equation and 458, for the Ireton-Jones equation),
indicating that the obese-specific equations were poor predic-
tors of REE.
The major finding of this study was that the TEE of our
extremely obese subjects was very high and increased with
increasing BMI from (mean ? SEE) 12.79 ? 0.5 MJ/d in the
lowest BMI tertile [(mean ? SD), BMI, 40.8 ? 2.1kg/m2] to
16.61 ? 0.76 MJ/d in the highest BMI tertile (mean BMI, 61.2
? 8.8 kg/m2). We chose to examine and present data using
BMI tertiles because we wanted to determine whether energy
expenditure increased with increasing body mass over the wide
range of BMI in our study. The TEE value for the highest BMI
tertile was 41% higher on average than the mean TEE of
nonobese women published in a summary of doubly labeled
water data (3), and 38% higher than previous measurements of
TEE in nonobese women [(mean ? SD), BMI, 20.9 ? 1.9] by
our own research group (14), and the predicted TEE using the
equations of the U.S. DRIs (5) for nonobese women (BMI
? 30.0 kg/m2). These results indicate that extremely obese
individuals have high rather than low rates of TEE, and
therefore must consume very high levels of dietary energy to
maintain their excess weight. This conclusion is inconsistent
with numerous reports of low energy intake in obese individ-
uals (15–18), but is entirely consistent with previous TEE data
from smaller studies of extremely obese persons (34 subjects
total) (19–26). There are several aspects of our study design
that deserve mention. The use of doubly labeled water to
measure TEE and a multicompartment model for the measure-
ment of body composition was important. Doubly labeled
water is the only recognized accurate method for estimation of
TEE in free-living subjects (27,28) and is the only method that
is relatively independent of inaccurate dietary reporting (15–
18). Although 1 previous study that compared TEE measured
in a calorimetric chamber with that of TEE from doubly
labeled water measurements indicated that doubly labeled
water may underestimate TEE in obese subjects (20), the
degree of underestimation in that study was small (mean
? SD, ?4.4 ? 5.2‰). Furthermore, an underestimation of
TEE in obese individuals would only strengthen our findings of
high TEE in our population.
In addition to finding high TEE in this extremely obese
population, we also observed that PAL values (an index of the
proportion of TEE attributable to physical activity) were some-
what low compared with published summaries of PAL deter-
mined by doubly labeled water in nonobese populations (3).
Although differences in measurement errors between labora-
tories necessarily limit between-laboratory comparisons, all
methods were validated against reference techniques in our
laboratory and in the published data. It is also interesting to
note that the observed PAL values in our study, although
apparently substantially lower than those of nonobese individ-
uals (3), were actually higher than anticipated for sedentary
adults on the basis of current (5) U.S. DRIs.
In addition to the finding of high levels of TEE, varying
degrees of accuracy were found in the prediction of REE in our
population with the equations recommended for use in the
nonobese and obese (8,9). Contrary to expectation, the widely
used WHO equation and Harris-Benedict equations (6,7,29)
developed in normal weight individuals provided the closest
values to measured REE. The Harris-Benedict equation was
the most accurate of all. The 2 obese-specific equations sub-
stantially under- or overpredicted REE, indicating that the
standard equations provide better predictions of REE in ex-
tremely obese individuals.
In conclusion, absolute values for TEE and REE in ex-
tremely obese individuals are high and increase with increas-
ing body mass. In addition, TEE and REE in our extremely
obese subjects were higher than published values for nonobese
individuals (3), and also higher than the U.S. DRI estimates of
dietary energy requirements in nonobese individuals (5,6,29).
These observations strongly suggest that high energy intakes
contribute to the maintenance of excess weight in extremely
obese individuals. Finally, the most widely used Harris-Bene-
dict equations and WHO equations provided the closest esti-
mate for predicted REE, and the obese-specific equations were
inaccurate for use in our extremely obese individuals
We thank all of the subjects who participated in the study,
Marshall Otter for his assistance in the doubly labeled water analysis,
and the nursing staff at the General Clinical Research Center, New
England Medical Center, for their expert patient care and support.
1. Flegal, K. M., Carroll, M. D., Ogden, C. L. & Johnson, C. L.
Prevalence and trends in obesity among US adults, 1999–2000. J. Am. Med.
Assoc. 288: 1723–1727.
2. Black, A. E., Coward, W. A., Cole, T. J. & Prentice, A. M.
using the WHO equation (6) and the Harris-Benedict equation (7) for
normal weight individuals and the obese-specific Bernstein (8) and
Ireton-Jones (9) equations.
Relation between measured REE and REE predicted
ENERGY EXPENDITURE IN EXTREME OBESITY
at TUFTS UNIVERSITY on July 24, 2012
energy expenditure in affluent societies: an analysis of 574 doubly-labelled water
measurements. Eur. J. Clin. Nutr. 50: 72–92.
3. Prentice, A. M., Black, A. E., Coward, W. A. & Cole, T. J.
expenditure in overweight and obese adults in affluent societies: an analysis of
319 doubly-labelled water measurements. Eur. J. Clin. Nutr. 50: 93–97.
4. Schoeller, D. A. (1999) Recent advances from application of doubly
labeled water to measurement of human energy expenditure. J. Nutr. 129: 1765–
5. Institute of Medicine(2002) Dietary Reference Intakes: Energy, Car-
bohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids. Na-
tional Academy Press, Washington, DC.
6. National Research Council (1989)
10th ed. National Academy Press, Washington, DC.
7. Harris, J. A. & Benedict, F. G.
Metabolism in Man. Carnegie Institute of Washington, Washington, DC.
8. Bernstein, R. S., Thornton, J. C., Yang, M. U., Wang, J., Redmond, A. M.,
Pierson, R. N., Pi-Sunyer, F. X. & Van Itallie, T. B.
resting metabolic rate in obese patients. Am. J. Clin. Nutr. 37: 595–602.
9. Ireton-Jones, C. S. (1989)Evaluation of energy expenditure in obese
patients. Nutr. Clin. Pract. 4: 127–129.
10. Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher,
D. F. & Turner, R. C.(1985) Homeostasis model assessment: insulin resistance
and ?-cell function from fasting plasma glucose and insulin concentrations in
man. Diabetologia 28: 412–419.
11. Bonora, E., Targher, G., Alberiche, M., Bonadonna, R. C., Saggiani, F.,
Zenere, M. B., Monauni, T. & Muggeo, M.
ment closely mirrors the glucose clamp technique in the assessment of insulin
sensitivity. Diabetes Care 23: 57–63.
12. Das, S. K., Roberts, S. B., Kehayias, J. J., Wang, J., Hsu, G.L.K., Shikora,
S. A., Saltzman, E. & McCrory, M. A.(2003)
extreme obesity and after massive weight loss induced by gastric bypass surgery.
Am. J. Physiol. 284: E1080–E1088.
13. Das, S. K., Roberts, S. B., McCrory, M. A., Hsu, G.L.K., Shikora, S. A.,
Kehayias, J. J., Dallal, G. E. & Saltzman, E.
energy expenditure and body composition after massive weight loss induced by
gastric bypass surgery. Am. J. Clin. Nutr. 78: 22–30.
14. Sawaya, A. L., Saltzman, E., Fuss, P., Young, V. R. & Roberts, S. B.
(1995) Dietary energy requirements of young and older women determined by
using the doubly labeled water method. Am. J. Clin. Nutr. 62: 338–344.
15. Schoeller, D. A. (1990) How accurate is self-reported dietary energy
intake? Nutr. Rev. 48: 373–379.
16. Schoeller, D. A., Bandini, L. G. & Dietz, W. H.
Recommended Dietary Allowances,
(1919) Biometric Studies of Basal
(1983)Prediction of the
(2000)Homeostasis model assess-
Body composition assessment in
(2003) Long-term changes in
self-reported intake identified by comparison with the doubly labelled water
method. Can. J. Physiol. Pharmacol. 68: 941–949.
17. Westerterp, K. R., Verboeket-van de Venne, W.P.H.G., Meijer, G.A.L. &
Ten Hoor, F.(1991) Self-reported intake as a measure for energy intake. A
validation against doubly labeled water. Obes. Eur. 91: 17–22.
18. Heitmann, B. L. & Lissner, L.(1995)
individuals—is it specific or non-specific? Br. Med. J. 311: 986–989.
19. Prentice, A. M., Black, A. E., Coward, W. A., Davies, A. L. & Goldberg,
G. R.(1986)High levels of energy expenditure in obese women. Br. Med. J.
20. Ravussin, E., Harper, I., Rising, R. & Bogardus, C.
expenditure by doubly-labeled water: validation in lean and obese subjects.
Am. J. Physiol. 261: E402–E409.
21. Westerterp, K. R., Meijer, G.A.L., Saris, W.H.M., Soeters, P. B., Winantis,
Y. & Ten Hoor, F.(1991)Physical activity and sleeping metabolic rate. Med.
Sci. Sports Exerc. 23: 166–170.
22. Westerterp, K. R., Saris, W.H.M., Soeters, P. B. & ten Hoor, F.
Determinants of weight loss after vertical banded gastroplasty. Int. J. Obes. Relat.
Metab. Disord. 15: 529–534.
23. Buhl, K. M., Gallagher, D., Hoy, K., Matthews, D. E. & Heymsfield, S. B.
(1995)Unexplained disturbance in body weight regulation: diagnostic outcome
assessed by doubly labeled water and body composition analyses in obese
patients reporting low energy intakes. J. Am. Diet Assoc. 95: 1393–1400.
24. Su, W., Jones, P.J.H., Cleator, I.G.M., Phang, P. T. & Birmingham, C. L.
(1996) Determinants of weight loss following ileogastronomy. Int. J. Obes. Relat.
Metab. Disord. 20: 481–487.
25. van Gemert, W. G., Westerterp, K. R., Van Acker, B.A.C., Wagenmakers,
A.J.M., Halliday, D., Greve, J. M. & Soeters, P. B.
protein metabolism in morbid obesity before, during and after massive weight
loss. Int. J. Obes. Relat. Metab. Disord. 24: 711–718.
26. Gibney, E. R., Murgatroyd, P., Wright, A., Jebb, S. & Elida, M.
Measurement of total energy expenditure in grossly obese women: comparison of
the bicarbonate-urea method with whole-body calorimetry and free living doubly
labeled water. Int. J. Obes. and Relat. Metab. Disord. 27: 641–647.
27. Schoeller, D. A.(1988)Measurement of energy expenditure in free-
living humans by using doubly labeled water. J. Nutr. 118: 1278–1289.
28. Roberts, S. B. (1989)Use of the doubly labeled water method for
measurement of energy expenditure, total body water, water intake, and metab-
olizable energy intake in humans and small animals. Can. J. Physiol. Pharm. 67:
29. World Health Organization(1985)
Report of a Joint FAO/WHO/UNU Expert Consultation, p. 206. WHO, Geneva,
Dietary underreporting by obese
(2000)Energy substrate and
Energy and Protein Requirements.
DAS ET AL.
at TUFTS UNIVERSITY on July 24, 2012