Journal of Gerontology: MEDICAL SCIENCES
1996, Vol. 5IA. No. 2. M7I-M73
Copyright 1996 by The Gerontological Society of America
Effect of Age on Body Water
and Resting Metabolic Rate
Naomi K. Fukagawa,123 Linda G. Bandini,3 William H. Dietz,3 and James B. Young1
'The Charles A. Dana Research Institute and the Harvard Thorndike Laboratory,
Department of Medicine, Beth Israel Hospital.
department of Pediatrics and Division on Aging, Harvard Medical School.
'Clinical Research Center, Massachusetts Institute of Technology.
Background. We previously reported that differences in fat-free mass (FFM) estimated by isotope dilution of 18Olabeled water could not fully account for lower resting metabolic rates (RMR) in old men and women compared to RMR
in young men.
Methods. Since age-related changes in the distribution of water between extracellular and intracellular spaces could
lead to overestimation of FFM in the old, we reanalyzed our data using estimates for total body and intracellular water
(TBW and ICW, respectively) derived from published equations and included data from adolescent boys and girls studied
under similar conditions.
Results. In both sexes, the age-related reduction in RMR remained significant after adjustment for estimated body
water compartments (p < .05). While adjusted RMR differed in boys and girls (p < .0001), it did not in old men and
women (p — .15).
Conclusion. We conclude that aging per se reduces RMR in lean tissue, a difference which cannot be fully explained
by changes in body water or its distribution. Investigators should be cautious when selecting models and equations to
estimate body water compartments.
T^UBLISHED equations are frequently used in studies
•1 requiring an estimate of body water compartments (1,2).
We previously reported that differences in fat-free mass
(FFM) estimated by l8O-labeled water could not fully account for the lower resting metabolic rate (RMR) in old men
and women compared to the RMR in young men (3). A
frequent question raised about the original report was
whether our observed differences could be due to a shift in
body water compartments rather than a decrease in the
metabolic activity of FFM. To examine whether changes in
body fluid distribution would alter the interpretation of our
previous findings (3), we calculated intracellular water
(ICW) and extracellular water (ECW) in our young and old
subjects using frequently applied published equations. In
addition, we included adolescents studied under similar
conditions (4) in our reanalysis to determine the effect of
aging on metabolic rate over a longer age span.
MATERIALS AND METHODS
Subjects. — Data from five groups of volunteers are
included in this report. Sixty-eight young, 18-33 yrs, and
elderly men and women, 67-89 yrs (3), and 25 nonobese
adolescent boys and girls, 12-16 yrs (4), were studied
under similar conditions. In all subjects, total body water
(TBW) was estimated using isotope dilution of 18O-labeled
water (5). Resting metabolic rate was measured using an
open-circuit ventilated hood indirect calorimetry system
interfaced to a desktop computer (3,4).
Calculations.
— Estimates of TBW, ECW, and ICW
were calculated from commonly accepted published equations (1,2,6,7) and are included in the appendix. It should be
noted that Cheek et al. (6,7) assumed that the deuterium
space exceeded actual body water by 2%, whereas Moore et
al. (1) used a correction factor of 4%. In all cases, ICW was
derived by subtracting ECW from TBW.
Statistical analysis. — The relationships between FFM
and RMR in all groups were examined by analysis of
covariance (ANCOVA). RMR data among subjects or between groups are often compared by expressing RMR per
kilogram of FFM, i.e., by dividing RMR by an estimate of
FFM (8). However, the use of the ratio of RMR/FFM is only
appropriate when the mathematical equation relating the two
parameters is one in which RMR is in constant proportion to
FFM with an intercept equal to zero (8,9). Analysis of
covariance permits derivation of adjusted RMR values according to a specific relationship between RMR and FFM
and avoids the assumption of a zero intercept. In this analysis, the tests for the equality of adjusted means and zero
slopes are based on the assumption that the slopes of the
dependent variable (RMR) and the covariate (FFM, TBW,
or ICW) are the same in every group. If the slopes differ
significantly, the results of the comparisons of adjusted
means should be interpreted with caution. In this report, age
was used as the class variable and FFM, TBW, or ICW as the
covariates to examine the relationships between RMR (dependent variable) and the various estimates of composition
M71
FUKAGAWA ET AL.
M72
Table 1. Body Water (kg) for Young and Elderly Men, Elderly Women,
and Boys and Girls From Published Equations^
Young Men
(18-33 yrs)
TBW (18O) [Measured]
TBW (Moore) [Age, Weight]
ICW (Moore) [TBW]
TBW (Pierson) [Age, Weight]
ICW (Pierson) [Age, Weight]
TBW (Cheek) [Weight]
ICW (Cheek) IWeightJ
40.4
42.3
22.9
48.0
31.8
± 1.4
± 0.8*
± 0.9
± 1.4t
± 0.9
NA
NA
Elderly Men
(67-89 yrs)
34 .8
37 .0
17.8
41 .4
26 .6
Elderly Women
(67-89 yrs)
± 0.7
±
±
±
±
25.8
28.5
12.7
29.9
17.9
0.5t
0.4
0.9t
0.6
± 0.6
±0.7t
± 0.3
± 1.11± 0.7
NA
NA
NA
NA
Boys
(12-16 yrs)
Girls
(12-16 yrs)
34.0 ± 1.8
NA
NA
NA
NA
34.1 ± 1.7
19.7 ± 0.9
30.0 ± 1.2
NA
NA
NA
NA
:S2.0 ± 1.6
19.3 ± 1.0
Notes: [ ] lists the variables used in the predictive equations; NA = not applicable.
tCheek and Graystone, 1968 (7); Moore et al., 1963 (1); Pierson et al., 1982(2).
*p < .05, ~\p < .001 compared to measured TBW.
(covariate). Comparisons between groups were made using
the Student's /-test. All analyses were done using BMDP
Statistical Software, Los Angeles, CA.
RESULTS
The young men were taller than the elderly men and boys
(177 ± 2, 172 ± 1, and 167 ± 3 cm, respectively;/? <
.05), but their body weight was comparable to the old men
(72 ± 2 and 73 ± 2 kg, respectively;/? = .41). The boys
weighed significantly less (55 ± 3 kg) than both adult
groups. In contrast, the old women and girls were of similar
height(159 ± 1 and 162 ± 2cm, respectively;/? = .10) and
weight (59 ± 2 and 56 ± 3 kg, respectively; p = .32).
Body mass index (BMI = weight in kg/height2 in m) was
lowest in the adolescent boys and girls and highest in the old
men. The data summarized in Table 1 reflect the application
of published equations to estimate TBW and ICW, which
were then used to compare RMR adjusted for ICW. Our
measured TBW was significantly less than that predicted by
both Moore's (1) and Pierson's (2) equations based on body
weight for different adult age groups (/? < .05), but the
correlation coefficients between measured TBW and both
estimates were identical: r = .94 (p < .001). In contrast,
prediction of TBW by Cheek et al. based on body weight for
boys and girls was similar to the measured TBW.
RMR adjusted for FFM estimated by 18O-water in the five
groups have been previously reported (3,4) and shown in
Table 2. The FFM common to all groups and used by the
ANCOVA was 46.0 kg. Within each gender, boys and girls
exhibited the highest adjusted RMRs. In contrast, adjusted
RMR did not differ between the old men and old women
(p= .15).
In the adult males and females, the age-related reduction
in RMR remained significant after estimation of total body
water using the equations of Moore et al. (1) and Pierson et
al. (2) (Table 2). We did not include the adolescents in the
analysis of RMR adjusted for body water estimates because
of the different set of regression equations available for the
boys and girls. In adult males, the adjusted RMRs for the
different estimates of body water compartments in the old
men were all significantly different from the young (Table
2). In females, RMRs adjusted for measured TBW and for
ICW estimated by Moore's equation and for TBW estimated
by Pierson's equations differed significantly from that in
Table 2. RMR (kcal/min) Adjusted for Body Water
Estimates in Adults
Young Men
(18-33 yrs)
TBW (18O)
TBW (Moore)
ICW (Moore)
TBW (Pierson)
ICW (Pierson)
.128
.080
. 102
.115
.090
±
±
±
±
±
0.021
0.024
0.024
0.021
0.022
Elderly Men
(67-89 yrs)
.028
.022
.046
.022
.021
±
±
±
±
±
0.016*
0.016t
0.017*
0.016*
0.016
Elderly Women
(67-89 yrs)
0.986
1.051
0.994
1.008
1.041
it
it
it
dt
it
0.024*
0.029
0.026§
0.025*
0.027
Source: Moore et al., 1963 (1); Pierson et al., 1982 (2).
*p < .004; 1y? < .04; \p < .06; §/> < .02; \\p < .01 compared to
young men.
young men. In all cases, old women did not differ from old
men. However, in the analysis of TBW estimated by both
Moore's and Pierson's equations, the criteria for parallelism
were not met, i.e., the slope of the dependent variable on the
covariate was not similar in every group. In both cases, the
young had a steeper slope than the elderly, suggesting that
for a given change in TBW, the young had a greater rise in
RMR, providing additional support for our suggestion that
age has an independent effect on RMR.
DISCUSSION
The effect of age on the relationship between body composition and RMR remains controversial because the appropriate terms for expression of metabolic measurements such
as energy expenditure are disputed. As elegantly discussed
by Moore and Boyden in 1963 (10), the use of "lean body
mass" or "fat-free body" as reference standards for energy
turnover includes support tissues such as extracellular fluid,
plasma volume, cartilage, skeleton, tendons, hair, teeth,
elastin, etc., which lack direct significance to oxidative
energy turnover. Because the metabolic activity of tissue is
related to ICW, and direct measurements of TBW and ECW
may be cumbersome, investigators have used published
regression equations to estimate the different compiirtments
of body water. Many of the well-accepted equations are
based on a small sample size and the assumption that the
individuals whose measures were used to generate the equations represent an unbiased population sample. The application of these equations may be appropriate and useful in
CHANGES IN BODY WATER AND RMR
clinical circumstances, but their utility from a research
standpoint is open to question.
We have found that RMR adjusted for differences in TB W
is inversely related to age (3). The present analyses using
established equations to estimate body water compartments
support our conclusion that energy metabolism is influenced
by age, independent of changes in TBW and its distribution.
This is further strengthened by our inclusion of data from
boys and girls. In both cases, the younger individuals had
higher RMR adjusted for FFM estimated from 18O-water
distribution and for estimated TBW and ICW than the adults.
These calculations also highlight the large differences between measured TBW and TBW estimates made with regression equations.
In summary, although we only measured TBW, the comparisons made with frequently published equations suggest
that it is unlikely that changes in body water distribution
could account for the lower metabolic rate in older individuals. The actual mechanisms responsible for age-related
changes in energy expenditure remain to be clarified. Moreover, caution should be exercised when utilizing published
equations to estimate body water compartments under different experimental conditions and among different groups of
volunteers.
ACKNOWLEDGMENTS
This research was supported by National Institutes of Health grants RR00088, RR-01032, AG-00599, HD-17696, and DK-26678.
The authors thank Ms. Dawn Griffiths and Ms. Allison Hawxhurst for
excellent secretarial assistance in the preparation of the manuscript, and the
staff of the Clinical Research Centers at the Massachusetts Institute of
Technology and Beth Israel Hospital for help in the conduct of the study.
The authors also gratefully acknowledge the contribution of Dr. Dale A.
Schoeller of the Clinical Nutrition Research Unit at the University of
Chicago.
M73
Dr. James B. Young is now located at the Department of Medicine,
Northwestern University Medical School, Chicago, IL 60611.
Dr. Naomi K. Fukagawa is now located at The University of Vermont
College of Medicine, Given Building Room C-207, Burlington, VT 054050068. Address correspondence and requests for reprints to Dr. Fukagawa at
this address.
REFERENCES
1. Moore FD, Olesen KH, McMurrey JD, Parker HV, Ball MR, Boyden
CM. The body cell mass and its supporting environment: body composition in health and disease. Philadelphia: W.B. Saunders, 1963.
2. Pierson RNJ, Wang J, Colt EW, Neumann P. Body composition
measurements in normal man: the potassium, sodium, sulfate and
tritium spaces in 58 adults. J Chron Dis 1982,35:419-28.
3. Fukagawa NK, Bandini LG, Young JB. Effect of age on body composition and resting metabolic rate. Am J Physiol 1990;259:E233-8.
4. Bandini LG, Schoeller DA, Dietz WE, Jr. Energy expenditure in obese
and nonobese adolescents. PediatrRes 1990;27:198-203.
5. Schoeller DA, Van Santen E, Peterson DW, Dietz W, Jaspan J, Klein
PD. Total body water measurement in humans with 18O- and 2Hlabelled water. Am J Clin Nutr 1980;33:2686-93.
6. Mellits ED, Cheek DB. Growth and body water. In: Cheek DB, ed.
Human growth: body composition, cell growth, energy, and intelligence. Philadelphia: Lea & Febiger, 1968:135-49.
7. Cheek DB, Graystone JE. Intracellular and extracellular volume [and
sodium], and exchangeable chloride in children. In: Cheek DB, ed.
Human growth: body composition, cell growth, energy, and intelligence. Philadelphia: Lea & Febiger, 1968:150-64.
8. Katch VL. Use of the oxygen/body weight ratio in correlational
analyses: spurious correlations and statistical considerations. Med Sci
Sports 1973,5:253-7.
9. Tanner JM. Fallacy of per-weight and per-surface area standards, and
their relation to spurious correlation. J Appl Physiol 1949;2:1-15.
10. Moore FD, Boyden CM. Body cell mass and limits of hydration of the
fat-free body: their relation to estimated skeletal weight. Ann NY Acad
Sci I963;l 10:62-71.
Received February 2, 1994
Accepted October 13, 1995
Appendix
Equations Used to Estimate Body Water Compartments
Source
Sex
n
Age
Equation:
Intracellular Water
Moore etal., 1963(1)
M&F
M&F
20
14
20-60 yrs
61-84 yrs
ICW = 0.633 (TBW) -2.67
ICW = 0.571 (TBW)-2.06
Pierson etal., 1982(2)
M
F
30
28
20-80 yrs
19-78 yrs
ICW = 0.01 (BW*) [47-0.14 Age]
ICW = 0.01 (BW) [45.1-0.2 Age]
Cheek and Graystone, 1968 (7)
M
F
18
20
5-16 yrs
5-16 yrs
ICW = 0.3236 (BW) + 1.773
ICW = 0.3288 (BW) + 0.930
Total Body Water
Moore etal., 1963(1)
M
M
M
F
63
56
13
34
16-30 yrs
31-60 yrs
61-90 yrs
31 -90 yrs
TBW
TBW
TBW
TBW
Pierson etal., 1982(2)
M
F
30
28
20-80 yrs
19-78 yrs
TBW = 0.01 (BW) [70.3-0.18 AgeJ
TBW = 0.01 (BW) [67.8-0.24 Age]
Boys
Girls
61
46
Cheek and Graystone, 1968 (7)
Derived equations for boys and girls
by combining two sets of data
*Where BW = body weight in kilograms.
1 month-16 yrs
1 month-16 yrs
=
=
=
=
13.26 + 0.404 (BW)
11.03 + 0.397 (BW)
12.02 + 0.343 (BW)
8.84 + 0.331 (BW)
TBW = 0.611 (BW) + 0.251
TBW = 0.551 (BW) + 1.244