Download Energy requirements and physical activity in free

Energy requirements and physical activity in free-living
older women and men: a doubly labeled water study
RAYMOND D. STARLING, MICHAEL J. TOTH, WILLIAM H. CARPENTER,
DWIGHT E. MATTHEWS, AND ERIC T. POEHLMAN
Division of Clinical Pharmacology and Metabolic Research, Department of Medicine,
University of Vermont, Burlington, Vermont 05405
total daily energy expenditure; elderly; aerobic capacity;
physical activity
whether energy needs in the elderly are
higher or lower than current recommendations (10, 28).
Moreover, factors that predict daily energy requirements in free-living elderly are poorly defined. Traditionally, energy intake methods have been used to determine energy requirements. However, the accuracy of
these techniques are questionable in the elderly because of misreporting of habitual energy intake (12,
31). To circumvent this problem, the assessment of
daily total energy expenditure (TEE) is used to estimate individual energy needs. When an individual is in
energy balance, the assessment of daily energy expenditure becomes a proxy measure of energy needs. Doubly
labeled water (24, 32) allows an integrated measurement of daily energy expenditure and, therefore, provides a valuable tool to examine daily energy needs in
free-living elderly.
IT IS UNCLEAR
http://www.jap.org
Present energy requirement recommendations, based
on the physical activity level (PAL) ratio [PAL 5
TEE/resting metabolic rate (RMR)], are estimated to be
1.51 3 RMR (10). Recent results (28) based on the
pooling of TEE data (12, 17, 26, 29, 30) suggest that this
recommendation may underestimate energy requirements of older individuals. Although the PAL ratio may
be a useful tool to estimate group energy needs, research should be directed at predicting individual
energy requirements.
Another method to estimate daily energy requirements is multiple-regression analysis. This approach
predicts individual daily energy requirements by using
parameters related to energy needs (i.e., anthropometric, physiological, and physical activity indexes). Because energy expenditure of physical activity is the
most variable component of TEE among the elderly (11,
12, 17, 26), measures of physical activity and/or aerobic
fitness may allow a more accurate prediction of individual energy requirements. Evidence for this assertion is based on previous data which show a strong
association between aerobic capacity and daily energy
expenditure measured via questionnaire (3) and doubly
labeled water (12). However, the use of multipleregression analysis to estimate energy requirements
has been limited by small numbers of subjects and the
lack of cross-validation procedures to assess the accuracy of these equations in independent populations (2,
4, 15, 27, 33).
Thus the primary aim of this study was to develop
and crossvalidate an equation to predict daily energy
requirements from anthropometric, physiological, and
physical activity indexes associated with TEE in a
relatively large sample of healthy, elderly women and
men. A secondary aim was to identify factors which correlate with physical activity energy expenditure (PAEE)
estimated from doubly labeled water and indirect calorimetry. Because low physical activity levels are predictive of cardiovascular and metabolic disease risk (18a,
22), an understanding of factors modulating physical
activity has important public health implications.
MATERIALS AND METHODS
Subjects
Subjects were 99 healthy, elderly Caucasians (51 women
and 48 men, ages 56–90 yr) recruited from the Burlington,
VT, area via advertisements in local newspapers. A subset of
these individuals were control subjects in previous studies
from our laboratory examining energy metabolism in various diseased populations (7, 23, 36, 38). All participants
were healthy and had no history or evidence on physical
examination of 1) coronary heart disease (e.g., S-T segment
8750-7587/98 $5.00 Copyright r 1998 the American Physiological Society
1063
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
Starling, Raymond D., Michael J. Toth, William H.
Carpenter, Dwight E. Matthews, and Eric T. Poehlman.
Energy requirements and physical activity in free-living older
women and men: a doubly labeled water study. J. Appl.
Physiol. 85(3): 1063–1069, 1998.—Determinants of daily
energy needs and physical activity are unknown in free-living
elderly. This study examined determinants of daily total
energy expenditure (TEE) and free-living physical activity in
older women (n 5 51; age 5 67 6 6 yr) and men (n 5 48; age 5
70 6 7 yr) by using doubly labeled water and indirect
calorimetry. Using multiple-regression analyses, we predicted TEE by using anthropometric, physiological, and physical activity indexes. Data were collected on resting metabolic
rate (RMR), body composition, peak oxygen consumption
(V̇O2 peak), leisure time activity, and plasma thyroid hormone.
Data adjusted for body composition were not different between older women and men, respectively (in kcal/day): TEE,
2,306 6 647 vs. 2,456 6 666; RMR, 1,463 6 244 vs. 1,378 6
249; and physical activity energy expenditure, 612 6 570 vs.
832 6 581. In a subgroup of 70 women and men, RMR and
V̇O2 peak explained approximately two-thirds of the variance in
TEE (R2 5 0.62; standard error of the estimate 5 6348
kcal/day). Crossvalidation of this equation in the remaining
29 women and men was successful, with no difference between predicted and measured TEE (2,364 6 398 and 2,406 6
571 kcal/day, respectively). The strongest predictors of physical activity energy expenditure (P , 0.05) for women and men
were V̇O2 peak (r 5 0.43), fat-free mass (r 5 0.39), and body
mass (r 5 0.34). In summary, RMR and V̇O2 peak are important
independent predictors of energy requirements in the elderly.
Furthermore, cardiovascular fitness and fat-free mass are
moderate predictors of physical activity in free-living elderly.
1064
ENERGY REQUIREMENTS IN THE ELDERLY
depression .1 mm at rest or exercise); 2) hypertension
(resting blood pressure .140/90); 3) medications that could
affect cardiovascular function or metabolism; 4) diabetes; 5)
body mass fluctuation of .2 kg in the past yr; 6) exerciselimiting noncardiac disease (arthritis, peripheral vascular
disease, cerebral vascular disease); 7) smoking; or 8) hormone
replacement therapy. No subject was regularly engaging in
aerobic or resistance training (i.e., ,2 days/wk). Each subject
signed a consent form approved by the Institutional Review
Board of the University of Vermont before participating in the
study.
Testing Protocol
N 5 1/2 (NO/cO 1 NH/c8H)
where cO and c8H are the sizes of the exchangeable oxygen and
hydrogen pool sizes relative to TBW. Pool size based on the
average of the cO and c8H pool spaces reduces error in TBW by
,1/2, which has been previously discussed (16). 18O dilution
space relative to TBW, cO, was assumed to be 1.01 (24, 25, 34).
2H dilution space relative to TBW, c8 , was assumed to be
H
1.053 (13). This value includes all analytical methodology
effects of hydrogen sample preparation and is consistent in
our hands where c8H/cO ratio is 1.046 6 0.002, averaging the
results of .250 studies.
Rate of CO2 production (rCO2, mol/day) was calculated by
using Eq. 3 of Speakman et al. (34)
rCO2 5 N/2.196 · (cOkO 2 cHkH)
where cH is the true hydrogen dilution space parameter (i.e.,
does not contain any analytical-method-induced component).
Following the recommendation of Racette et al. (25), a value
of 1.041 was used for cH. If a value of 1.053 had been used
instead for cH, the cOkO 2 cHkH difference would have been
decreased by ,1%, and the calculation of rCO2 would have
been decreased by #5% (16). Assuming a respiratory quotient
of the food consumed of 0.85 (5), total CO2 production was
converted to oxygen consumed or daily TEE (in kcal/day) by
using the Weir formula (39)
TEE 5 (3.044/0.85 1 1.104)rCO2
DETERMINATION OF PAEE. PAEE was calculated by using
the following equation as previously described (12)
PAEE (kcal/day) 5 (0.90 3 TEE) 2 RMR
This approach assumes that the thermic effect of feeding is
10% in the elderly (20).
Body composition. Fat and fat-free mass were measured by
dual-energy X-ray absorptiometry with the use of a Lunar
DPX-L densitometer (Lunar Radiation, Madison, WI). A total
body scan was completed in 40 min and provided measurements of fat-free mass, fat mass, and percent body fat. In six
older women from our laboratory, the coefficient of variation
for body fat was 1.7% during test-retest on two occasions
within 1 wk.
Aerobic capacity and leisure time activity. Peak oxygen
consumption (V̇O2 peak) was determined during an incremental
cycling test to voluntary exhaustion. Cycling cadence was 50
rpm, with a workload during the first 3 min of 25 and 50 W for
the women and men, respectively. Workload was increased 25
W every 2 min until voluntary exhaustion. V̇O2 peak (l/min) was
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
All subjects were tested in the morning after an overnight
visit to the General Clinical Research Center (GCRC) at the
University of Vermont. Subjects provided a baseline urine
sample on the evening of admission (at 1600–1800) and were
administered a mixed dose of doubly labeled water to measure daily TEE. After a 12-h overnight fast, each subject was
awakened at 0630 for a measurement of RMR and collection
of a venous blood sample and two urine samples for doubly
labeled water analysis. Body composition was assessed by
using dual-energy X-ray absorptiometry, and a Minnesota
Leisure Time Physical Activity questionnaire was administered to each subject. Each subject returned 10 days later to
provide two urine samples for doubly labeled water analysis
and to complete a cycling aerobic capacity test. Specific
details about data collection are provided below.
RMR. RMR was measured by indirect calorimetry by using
the ventilated hood technique (19). Respiratory gas analysis
was performed with the use of a Deltatrac metabolic cart
(Sensormedics, Yorba Linda, CA). RMR (kcal/day) was calculated from the equation of Weir (39). The intraclass correlation and coefficient of variation for RMR, as determined by
using test-retest in 17 volunteers from our laboratory, are
0.90 and 4.3%, respectively. RMR was also estimated from
body mass and height by using gender-specific equations (10).
TEE. TEE was measured over a 10-day period by using the
doubly labeled water (2H218O) method of Schoeller and van
Santen (32). Subjects arrived at the GCRC on day 0, and a
urine sample was acquired for measurement of baseline 2H
and 18O enrichment. Between 1600 and 1800, a premixed
dose containing ,0.12 g of 2H2O and 0.15 g of H218O/kg of
estimated total body water (TBW) was given to each subject
to drink (,70 ml). Two urine samples were collected the next
morning (day 1), and another two urine samples were collected on the return visit to the GCRC on day 10. These urine
samples were obtained between 0800 and 1200. Aliquots of
the urine samples were stored frozen at 220°C in vacutainers
until later analysis by isotope-ratio mass spectrometry (IRMS).
For measurement of 18O, duplicate 1-ml aliquots of urine
were placed in vacutainers, which were then filled with a low
pressure of CO2 and shaken at room temperature overnight.
CO2 oxygen equilibrates with the water oxygen, and the 18O
isotopic enrichment of the water was then measured by
IRMS. For measurement of 2H enrichment, triplicate 5-µl
aliquots of urine were placed in stopcock-sealed reaction
vessels containing 100 mg of zinc catalyst, following the
method of Wong et al. (40). Vessels were sealed under
nitrogen gas and then evacuated with the water frozen. The
water was reduced to hydrogen gas by heating the vessels in a
block at 500°C for 30 min. 2H enrichments of the hydrogen
samples were measured by IRMS on the same day they were
prepared. Aliquots of the 2H218O dose were also measured for
2H and 18O enrichment after being quantitatively diluted
with unlabeled water. Urine sample 2H and 18O enrichments
were calculated by using the diluted dose 2H and 18O enrichments as calibrants for analysis in each study.
DETERMINATION OF TEE. After the dose of 2H218O is consumed, the 2H and 18O tracers equilibrate in the TBW pool. 2H
and 18O tracers are then lost at an exponential rate from the
body over the following 10 days. Rate of CO2 production,
integrated over the 10-day period, was determined by subtracting the difference between the rate of 18O tracer loss (water
turnover 1 CO2 production) and 2H tracer loss (water turnover alone) (32). Rates of 18O and 2H loss (kO and kH,
respectively) from body water and the 18O and 2H dilution
spaces (NO and NH, respectively) were determined from the
semilogarithmic regression of the 18O and 2H enrichments vs.
time. The slopes define the rates of tracer disappearance, and
the intercepts define the dilution spaces (liters of water).
TBW (N) in liters was calculated by taking the average of the
dilution space of the 18O and 2H tracers
1065
ENERGY REQUIREMENTS IN THE ELDERLY
considered to be achieved with a respiratory exchange ratio
.1.1 or a heart rate at or above the age-related predicted
maximum (220 2 age). Test-retest conditions (within 1 wk)
for V̇O2 peak on nine older individuals in our laboratory have
yielded an intraclass correlation of 0.94 and a coefficient of
variation of 3.8%. Leisure time physical activity was measured by a structured questionnaire and interview (35). This
questionnaire assessed each individual’s physical activity
level over the past year.
Plasma thyroid hormones. Plasma thyroxine (T4 ), free T4,
and triiodothyronine (T3 ) concentrations were measured by
using commercially available enzymatic methods (Baxter,
Cambridge, MA), whereas free T3 concentration was determined via an analog assay (Diagnostics Products, Los Angeles, CA).
Statistical Analyses
RESULTS
There were no differences in TEE, RMR, and PAEE
between men and women after adjustment for body
composition differences. Measured PAL was also not
significantly different between women and men. Pooled
gender data for physical characteristics and energy
expenditure are presented in Tables 1 and 2, respec-
Variable
Value
n (women/men)
Age, yr
Height, cm
Body mass, kg
Body mass index, kg/m2
Body fat, %body mass
Fat mass, kg
Fat-free mass, kg
V̇O2 peak , l/min
LTA, kcal/day
Total T3 , ng/dl
Free T3 , pg/dl
Total T4 , µg/dl
Free T4 , ng/dl
51/48
69 6 8
168 6 9
71 6 12
24.9 6 3.8
31 6 10
22 6 9
49 6 10
1.72 6 0.56
415 6 320
116 6 25
3.1 6 0.7
6.9 6 1.6
1.1 6 0.6
Values are means 6 SD. V̇O2 peak , peak aerobic capacity; LTA,
leisure time activity; T3 , triiodothyronine; T4 , plasma thyroxine.
tively. Pearson correlations between TEE and other
selected dependent variables for all subjects (n 5 99)
are displayed in Table 3. Measured and predicted RMR,
fat-free mass, V̇O2 peak, body mass, and percent body fat
were significantly correlated with TEE.
There were no differences in physical characteristics
between validation and cross-validation groups, and a
Levene’s test demonstrated equal variances between
groups for all data (see Table 4). Stepwise regression
analysis was performed on a randomly selected sample
of 70 women and men, and predicted RMR, V̇O2 peak,
body mass, percent body fat, fat-free mass, and gender
were entered into the model based on their strong
associations with TEE from simple correlations. The
following equation accounted for the largest amount of
variance (R2 5 0.62) in TEE
TEE(kcal/day) 5 [1.95 · predicted RMR(kcal/day)]
1 [217.3 · V̇O2 peak (l/min)] 2 825.5
The standard error of the estimate (SEE) for this
equation was 6348 kcal/day. No difference was detected between women and men for the mean residuals
(i.e., measured minus predicted TEE); this demonstrates that this equation predicts TEE accurately for
both genders. Cross-validation procedures on this equation demonstrated no significant difference between
predicted and measured TEE for the 29 women and
men (see Fig. 1). However, the correlation coefficient
between the measured and predicted TEE in the crossTable 2. Daily total energy expenditure, resting
metabolic rate, physical activity energy expenditure,
and physical activity level ratio for 99 women and men
Variable
Value
TEE, kcal/day
RMR, kcal/day
PAEE, kcal/day
PAL
2,379 6 556
1,422 6 255
719 6 377
1.68 6 0.28
Values are means 6 SD. TEE, daily total energy expenditure;
RMR, resting metabolic rate; PAEE, physical activity energy expenditure; PAL, physical activity level measured as TEE/RMR.
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
All data are expressed as means 6 SD. Potential differences between women and men for daily TEE and its components were examined by using independent t-tests. Comparison of V̇O2 peak between women and men was completed after
normalizing for fat-free mass by using analysis of covariance
(37). Significance was accepted at the P , 0.05 level. Data for
women and men were pooled for all subsequent analyses if no
differences were detected for daily TEE and its components
after adjustment for body composition.
Pearson product-moment correlations were calculated to
examine relationships among TEE, PAEE, and other selected
independent variables for all 99 women and men. To account
for the influence of body composition on PAEE, correlations
between PAEE and various independent variables were performed accordingly. 1) PAEE was correlated with various
independent variables by using partial correlation analyses
to account for the influence of fat and fat-free mass, as
previously suggested (8). 2) PAEE was indexed as the ratio of
measured TEE to RMR (i.e., PAL ratio) and correlated with
various independent variables to account for the influence of
body weight.
By using stepwise regression analysis (9), an equation was
developed to determine the relative contribution of selected
independent variables to the variation in TEE. Independent
variables introduced into the analysis 1) needed a physiological rationale to be included, and 2) correlated significantly
with TEE from Pearson product tests. Additionally, a moderate effect size (R2 5 0.50–0.75) was hypothesized, and a
subject-to-independent variable ratio of 6:1 was maintained
during stepwise regression analyses, as suggested previously
(14). This prediction equation was generated on a randomly
selected two-thirds of women (n 5 36) and men (n 5 34). The
remaining one-third of women (n 5 15) and men (n 5 14)
served as a cross-validation group. The accuracy of the
prediction equation was determined by comparing 1) predicted and measured TEE in the cross-validation group by
using a paired t-test; and 2) the correlation coefficient (r) of
predicted and measured TEE vs. the multiple R obtained
from the prediction equation by an independent z-test (18).
The reliability of our prediction was also assessed by calculating an intraclass correlation coefficient.
Table 1. Descriptive characteristics of subjects
1066
ENERGY REQUIREMENTS IN THE ELDERLY
Table 3. Pearson product correlations between TEE
and other independent variables for 99 women
and men
Variable
TEE
Predicted RMR, kcal/day
Fat-free mass, kg
Measured RMR, kcal/day
Body mass, kg
V̇O2 peak , l/min
Body fat, %body mass
LTA, kcal/day
Fat mass, kg
Age, yr
Total T4 , µg/dl
Total T3 , ng/dl
Free T4 , ng/dl
Free T3 , pg/ml
0.70*
0.69*
0.68*
0.62*
0.61*
20.23*
0.17
0.12
20.11
20.23
20.18
0.09
0.04
Predicted RMR estimated from FAO/WHO/UNU equations (10).
* P , 0.05.
DISCUSSION
This study was prompted by the paucity of data
regarding energy requirements and physical activity in
free-living elderly. Our relatively large numbers of
subjects allowed for the determination of TEE by using
several independent predictors while maintaining a
favorable subject-to-variable ratio for multiple-regression analysis. Regression data demonstrate that predicted RMR and V̇O2 peak explain 62% of the variance in
TEE. Furthermore, this equation accurately predicts
mean TEE in an independent group of women and men
with a SEE of 6460 kcal/day, although prediction on an
Table 4. Descriptive characteristics for validation and
cross-validation groups
Variable
Validation
Crossvalidation
n (women/men)
Age, yr
Height, cm
Body mass, kg
Body mass index, kg/m2
Body fat, %body mass
Fat mass, kg
Fat-free mass, kg
V̇O2 peak , l/min
LTA, kcal/day
70 (36/34)
68 6 6
167 6 9
71 6 13
25.3 6 3.8
32 6 9
22 6 8
49 6 10
1.72 6 0.54
368 6 255
29 (15/14)
69 6 8
169 6 8
69 6 11
24.0 6 3.6
29 6 11
20 6 10
49 6 10
1.72 6 0.64
529 6 426
Values are expressed as means 6 SD. No significant differences
were present between groups (P . 0.05).
Fig. 1. Relationship between predicted and measured daily total
energy expenditure (TEE) for cross-validation group of 15 women and
14 men. Values are group means 6 SD. SEE, standard error of
estimate. No significant difference was present between predicted
and measured TEE (P . 0.05).
individual basis was less precise. Moreover, cardiovascular fitness and fat-free mass are moderate predictors
of PAEE in free-living elderly.
Prediction and Validation of TEE Equation
An understanding of factors regulating TEE and
energy requirements in the elderly are important public health concerns. Energy requirements of the elderly
are estimated at 1.51 times RMR on the basis of the
factorial approach (10). It is suggested that present
energy intake guidelines underestimate the energy
needs of the elderly and that a PAL ratio of 1.62 to 1.68
may be more appropriate (28). PAL results for women
(1.63 6 0.27) and men (1.73 6 0.28) in the present study
confirm these findings. However, indexing energy requirements by using the PAL ratio assumes PAL is a
constant function of an individual’s RMR. Failure to
account for variation in physical activity among individuals may introduce significant error into the prediction of energy requirements.
Table 5. Pearson product correlations between PAEE
and various independent variables for 99 women
and men
Variable
Age, yr
Body mass, kg
Body fat, %
Fat mass, kg
Fat-free mass, kg
LTA, kcal/day
V̇O2 peak , l/min
Unadjusted
Adjusted
Measured
PAEE, kcal/day PAEE, kcal/day PAL Ratio, kcal/day
0.10
0.34*
20.10
0.04
0.39*
0.21*
0.43*
20.09
0.09
20.14
0.20*
0.31*
20.06
0.18*
20.16
0.01
0.12
0.18*
0.23*
Adjusted PAEE is covaried for fat and fat-free mass, as previously
suggested (8). * P , 0.05.
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
validation group (r 5 0.61) was significantly different
from the multiple R of the regression equation developed from the validation group (R 5 0.79); this indicates that the equation was less precise in this independent cohort. Moreover, the intraclass correlation
coefficient was 0.80 for our prediction equation with a
SEE of 6460 kcal/day in the cross-validation group.
Simple correlations between PAEE and various dependent variables for all subjects are shown in Table 5.
Independent of the method for indexing PAEE, low to
moderate correlations were demonstrated with fat-free
mass, body mass, V̇O2 peak, and leisure time activity
score.
ENERGY REQUIREMENTS IN THE ELDERLY
between the multiple R of the original equation and the
r value of measured vs. predicted TEE in the crossvalidation group. Nevertheless, cross-validation results demonstrate that our equation is internally consistent, but the equation may not be universally applicable
because it was not developed in a random group of
elderly individuals selected from the general population.
Determinants of PAEE
A secondary aim of this study was to examine potential determinants of PAEE in free-living elderly. This is
the most understudied component of TEE, which is
normally estimated from activity diaries and motion
detectors but can be more objectively quantified from
doubly labeled water and indirect calorimetry data.
Low levels of physical activity are most predictive of
cardiovascular and metabolic disease risk (18a, 22);
therefore, identifying determinants of PAEE has important public health implications.
PAEE is both a behavioral characteristic and a
physiological attribute. It includes volitional activity
during structured exercise, as well as nonstructured
movement such as fidgeting. Because PAEE is influenced by body mass and composition, correlations are
presented in Table 5 on an absolute and adjusted basis.
Because there is no consensus on how to adjust PAEE
for differences in metabolic size, we present several
adjustment procedures that are currently used in other
studies. Briefly, we adjusted PAEE 1) for fat and
fat-free mass by using partial correlation procedures,
as previously suggested (8), and 2) by using the PAL
ratio which divides TEE by RMR.
Absolute levels of PAEE are related to fat-free mass,
body mass, V̇O2 peak, and leisure time activity in the
present study. As with TEE, it is logical to postulate a
close association between aerobic fitness and volitional
participation in physical activity. Previous work (12) in
13 elderly individuals shows that PAEE is strongly
correlated with V̇O2 peak (r 5 0.52) and leisure time
activity score (r 5 0.83). Moderate correlations between
PAEE and V̇O2 peak in the present study are probably
reflective of differences between a predominantly behavioral characteristic like PAEE and a physiological attribute such as V̇O2 peak. That is, factors other than
physiological indexes (e.g., socioeconomic status, availability of recreational facilities, seasonality) may be
related to PAEE. We also show a low correlation
between PAEE and the Minnesota Leisure Time Activity Survey (r 5 0.21), which demonstrates that this
structured questionnaire does not accurately reflect
PAEE in the elderly, as previously suggested (12). After
adjusting PAEE for body weight and composition, correlations were similar or slightly attenuated. At this
point, we conclude that our ability to predict PAEE in
free-living elderly is modest. Given the importance of
physical activity in the determination of cardiovascular
and metabolic disease risk, future research is needed to
determine other factors associated with PAEE in older
women and men.
Summary. The primary aim of this study was to
attempt the first large-scale prediction of daily TEE
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
To address this problem, we used multiple-regression
analyses to predict individual energy requirements
from anthropometric and physiological indexes including measures of physical activity. Previous studies
examining energy needs in the elderly have relied on
small sample sizes (6, 13, 17, 26, 29, 30). Moreover,
investigators who have used regression analyses to
predict energy requirements have not crossvalidated
their equations in independent cohorts (2, 4, 15, 27, 33).
We attempted to account for these experimental drawbacks in the present design.
Our results show that 62% of the variance in TEE is
explained by predicted RMR (10) and V̇O2 peak, with a
SEE of 6348 kcal/day. Others demonstrate that 74% of
the variance in TEE, as measured by doubly labeled
water, is explained by V̇O2 peak and leisure time physical
activity, with a SEE of 6217 kcal/day (12). Moreover,
after correcting for fat-free mass by using regression
procedures, a significant partial correlation persists
between TEE and V̇O2 peak (r 5 0.24, P , 0.05) in the
present study. This finding provides additional evidence that aerobic fitness level is an important predictor of TEE, independent of its covariance with fat-free
mass. Collectively, these studies demonstrate that
V̇O2 peak and other indexes of physical activity are important predictors of TEE.
It is logical to assume that an individual may be more
physically active because of a greater fitness level,
although being more physically active may improve an
individual’s aerobic fitness. Nevertheless, an individual
with a higher absolute V̇O2 peak would work at a lower
percentage of individual maximum aerobic capacity for
a similar quantity of physical activity compared with
an individual with a lower V̇O2 peak. Theoretically, an
elderly person with a higher aerobic fitness should be
able to complete a greater quantity of work throughout
the day, resulting in greater energy expenditure and
requirements. Because maximal aerobic capacity cannot be easily measured in all of the elderly, predicted
aerobic capacity from submaximal exercise data and
other more direct measures of physical activity (i.e.,
axial/triaxial accelerometers, pedometers, and agespecific activity questionnaires) may be useful in quantifying the contribution of physical activity to variation
in TEE. Overall, identifying other physiological (e.g.,
sympathetic nervous system activity), environmental
(e.g., socioeconomic status, education level), and physical activity indexes may help explain a greater amount
of variance in TEE and increase the precision of this
predictive model (e.g., 6100 kcal/day).
To our knowledge, the present study is the first to
assess the accuracy of a TEE prediction equation in an
independent sample of the elderly the (i.e., crossvalidation) who had TEE measured by doubly labeled water.
Results from the cross-validation group demonstrate
that our equation accurately predicts TEE on a group
basis, as evidenced by the absence of a difference
between predicted and measured TEE (see Fig. 1). The
precision is less than desired on an individual basis, as
reflected by a SEE of 460 kcal/day in the crossvalidation group and by the significant difference
1067
1068
ENERGY REQUIREMENTS IN THE ELDERLY
from various anthropometric, physiological, and physical activity indexes in a relatively large group of
healthy, elderly women and men. Our equation explains 62% of the variance in daily TEE by using
predicted RMR and peak aerobic capacity with a SEE of
,350 kcal/day. Crossvalidation demonstrates that this
equation works accurately on a group mean basis in a
randomly selected, independent group of healthy elderly individuals; however, this equation is less precise
on an individual basis. Future studies that use other
physiological, environmental, and physical activity indexes are needed to determine what factors may explain more variation in daily TEE and improve the
prediction of energy needs within 6100 kcal/day. Moreover, free-living physical activity, determined from
doubly labeled water and indirect calorimetry, is moderately associated with cardiovascular fitness and fatfree mass in the elderly.
Received 22 September 1997; accepted in final form 14 May 1998.
REFERENCES
2. Astrup, A., G. Throbek, J. Lind, and B. Isaksson. Prediction
of 24-h energy expenditure and its components from physical
characteristics and body composition in normal-weight humans.
Am. J. Clin. Nutr. 52: 777–783, 1990.
3. Berthouze, S. E., P. M. Minaire, J. Castells, T. Busso, L.
Vico, and J. R. Lacour. Relationship between mean habitual
daily energy expenditure and maximal oxygen uptake. Med. Sci.
Sports Exerc. 27: 1170–1179, 1995.
4. Black, A. E., W. A. Coward, T. J. Cole, and A. M. Prentice.
Human energy expenditure in affluent societies: an analysis of
574 doubly-labelled water measurements. Eur. J. Clin. Nutr. 50:
72–92, 1996.
5. Black, A. E., A. M. Prentice, and W. A. Coward. Use of food
quotients to predict respiratory quotients for the doubly-labelled
water method of measuring energy expenditure. Hum. Nutr.
Clin. Nutr. 40C: 381–391, 1986.
6. Campbell, W. C., D. Cyr-Campbell, J. A. Weaver, and W. J.
Evans. Energy requirements for long-term body weight maintenance in older women. Metabolism 46: 884–889, 1997.
7. Carpenter, W. H., T. Fonong, M. J. Toth, P. A. Ades, J.
Calles-Escandon, J. Walston, and E. T. Poehlman. Total
daily energy expenditure in free-living older African-Americans
and Caucasians. Am. J. Physiol. 274 (Endocrinol. Metab. 37):
E96–E101, 1998.
8. Carpenter, W. H., E. T. Poehlman, M. O. O’Connell, and M. I.
Goran. Influence of body composition and resting metabolic rate
on variation in total energy expenditure: a meta-analysis. Am. J.
Clin. Nutr. 61: 4–10, 1995.
9. Draper, N., and H. Smith. Applied Regression Analysis. New
York: Wiley, 1981.
10. Food and Agricultural Organization-World Health Organization-United Nations University. Energy and Protein Requirements. Geneva: World Health Organization, 1985.
11. Fuller, N. J., M. B. Sawyer, W. A. Coward, P. A. Paxton, and
M. Elia. Components of total daily energy expenditure in freeliving elderly men (over 75 years of age): measurement, predictability and relationship to quality-of-life indices. Br. J. Nutr. 75:
161–173, 1996.
12. Goran, M. I., and E. T. Poehlman. Total energy expenditure
and energy requirements in healthy elderly persons. Metabolism
41: 744–752, 1992.
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017
This study was supported by National Institutes of Health Grants
AG-07857, AG-00564, DK-52752, F32 AG-05791 (to R.D. Starling),
and RR-00109 (to the University of Vermont General Clinical Research Center) and by American Association of Retired Persons
grants (to E. T. Poehlman).
Address for reprint requests: E. T. Poehlman, Univ. of Vermont,
Dept. of Medicine, Given Bldg. C-247, Burlington, VT 05405 (E-mail:
[email protected]),
13. Goran, M. I., E. T. Poehlman, K. S. Nair, and E. Danforth,
Jr. Effect of gender, body composition, and equilibration time on
the 2H-to-18O dilution space. Am. J. Physiol. 263 (Endocrinol.
Metab. 26): E1119–E1124, 1992.
14. Green, S. B. How many subjects does it take to do a regression
analysis? Mult. Behav. Res. 26: 499–510, 1991.
15. Klausen, B., S. Toubro, and A. Astrup. Age and sex effects on
energy expenditure. Am. J. Clin. Nutr. 65: 895–907, 1997.
16. Matthews, D. E., and C. D. Gilker. Impact of 2H and 18O pool
size determinants on the calculation of total daily energy expenditure. Obesity Res. 3: 21–29, 1995.
17. Pannemans, D. L. E., and K. R. Westerterp. Energy expenditure, physical activity and basal metabolic rate of elderly subjects. Br. J. Nutr. 73: 571–581, 1995.
18. Pedhauzer, E. J. Multiple Regression in Behavioral Research.
Fort Worth, TX: Holt, Rinehart & Winston, 1982.
18a.National Institutes of Health. Physical, Activity, and Cardiovascular Health. NIH Consens. Statement 13: 1–33, 1995.
19. Poehlman, E. T., T. L. McAuliffe, D. R. Van Houten, and E.
Danforth, Jr. Influence of age and endurance training on
metabolic rate and hormones in healthy men. Am. J. Physiol. 259
(Endocrinol. Metab. 22): E66–E72, 1990.
20. Poehlman, E. T., C. L. Melby, and S. F. Badylack. Relation of
age and physical exercise status on metabolic rate in younger
and older healthy men. J. Gerontol. A Biol. Sci. Med. Sci. 46:
B54–B58, 1991.
21. Poehlman, E. T., and M. J. Toth. Mathematical ratios lead to
spurious conclusions regarding age and gender-related differences in resting metabolic rate. Am. J. Clin. Nutr. 61: 482–485,
1995.
22. Poehlman, E. T., M. J. Toth, L. B. Bunyard, A. W. Gardner,
K. E. Donaldson, E. Colman, T. Fonong, and P. A. Ades.
Physiological predictors of increasing total and central adiposity
in aging men and women. Arch. Intern. Med. 155: 2443–2448,
1995.
23. Poehlman, E. T., M. J. Toth, M. I. Goran, W. H. Carpenter, P.
Newhouse, and C. J. Rosen. Daily energy expenditure in
free-living non-institutionalized Alzheimer’s patients. Neurology
48: 997–1002, 1997.
24. Prentice, A. The Doubly Labeled Water Method for Measuring
Energy Expenditure: Technical Recommendations for Use in
Humans. Vienna: International Atomic Energy Agency, 1990.
25. Racette, S. B., D. A. Schoeller, A. H. Luke, K. Shay, J.
Hnilicka, and R. F. Kushner. Relative dilution spaces of 2H to
18O-labeled water in humans. Am. J. Physiol. 267 (Endocrinol.
Metab. 30): E585–E590, 1994.
26. Reilly, J. J., A. Lord, V. W. Bunker, A. M. Prentice, W. A.
Coward, A. J. Thomas, and R. S. Briggs. Energy balance in
healthy elderly women. Br. J. Nutr. 69: 21–27, 1993.
27. Rising, R., I. T. Harper, A. M. Fontvielle, R. T. Ferraro, M.
Spraul, and E. Ravussin. Determinants of total daily energy
expenditure: variability in physical activity. Am. J. Clin. Nutr.
59: 800–804, 1994.
28. Roberts, S. B. Energy requirements of older individuals. Eur. J.
Clin. Nutr. 50, Suppl.: S112–S118, 1996.
29. Roberts, S. B., V. R. Young, P. Fuss, M. A. Fiatarone, B.
Richard, H. Rasmussen, D. Wagner, L. Joseph, E. Holehouse, and W. J. Evans. What are the dietary energy needs of
elderly adults? Int. J. Obes. 16: 969–974, 1992.
30. Sawaya, A. L., E. Saltzman, P. Fuss, V. R. Young, and S. B.
Roberts. Dietary energy requirements of young and older
women determined by using the doubly labeled water method.
Am. J. Clin. Nutr. 62: 338–344, 1995.
31. Sawaya, A. L., K. Tucker, W. Willet, E. Saltzman, G. E.
Dallal, and S. B. Roberts. Evaluation of four methods for
determining energy intake in young and older women: comparison with doubly labeled water measurements of total energy
expenditure. Am. J. Clin. Nutr. 63: 491–499, 1996.
32. Schoeller, D. A., and E. van Santen. Measurement of energy
expenditure in humans by doubly labeled water. J. Appl. Physiol.
53: 955–959, 1982.
ENERGY REQUIREMENTS IN THE ELDERLY
33. Schulz, L. O., and D. A. Schoeller. A compilation of total daily
energy expenditures and body weights in healthy adults. Am. J.
Clin. Nutr. 60: 676–681, 1994.
34. Speakman, J. R., K. S. Nair, and M. I. Goran. Revised
equations for calculating CO2 production from doubly labeled
water in humans. Am. J. Physiol. 264 (Endocrinol. Metab. 27):
E912–E917, 1993.
35. Taylor, H. L., D. R. Jacobs, B. Schucker, J. Knudsen, A. S.
Leon, and G. Debacker. A questionnaire for the assessment of
leisure time physical activities. J. Chron. Dis. 31: 741–755, 1978.
36. Toth, M. J., P. S. Fishman, and E. T. Poehlman. Free-living
daily energy expenditure in patients with Parkinson’s disease.
Neurology 48: 88–91, 1997.
1069
37. Toth, M. J., M. I. Goran, P. A. Ades, D. B. Howard, and E. T.
Poehlman. Examination of data normalization procedures for
expressing peak V̇O2 data. J. Appl. Physiol. 75: 2288–2292, 1993.
38. Toth, M. J., S. S. Gottlieb, M. I. Goran, M. L. Fisher, and E. T.
Poehlman. Daily energy expenditure in free-living heart failure
patients. Am. J. Physiol. 272 (Endocrinol. Metab. 35): E469–
E475, 1997.
39. Weir, J. B. New methods for calculating metabolic rate with
special reference to protein metabolism. J. Physiol. (Lond.) 109:
1–9, 1949.
40. Wong, W. W., L. S. Lee, and P. D. Klein. Deuterium and
oxygen-18 measurements on microliter samples of urine, plasma,
saliva, and human milk. Am. J. Clin. Nutr. 45: 905–913, 1987.
Downloaded from http://jap.physiology.org/ by 10.220.33.3 on August 1, 2017