![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Epidemiology, University of Michigan, Ann Arbor, MI 48109-2029
1To whom correspondence should be addressed.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
58 y, with the modeled relationship including interactions with
waist circumference and height. These models accounted for 70% of
observed variance in lean mass. Age is associated with body composition
but explains <10% of variation. When measures of height and
circumferences are available, amounts of lean and fat mass are highly
predictable. This is particularly important for lean mass because no
other surrogate measures exist for lean mass, whereas there are
surrogates for fat mass, including body mass index.
KEY WORDS: • fat mass • lean mass • dual X-ray densitometry • body composition • body circumference measures
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, Kissebah and Krakower 1994
).
Although these surrogate measures of body composition and body topology
are sometimes portrayed as interchangeable, they represent different,
though related underlying constructs. Body composition measures are
estimates of the amount of mass of a specific body compartment (i.e.,
fat and lean), whereas topology measures are proxies for the locational
distribution of gluteal and visceral adipose tissue. In epidemiologic
research, questions remain concerning whether the two types of measures
are independent predictors of chronic disease risk or whether there are
combined effects of body composition and body topology that affect the
ability of either to predict disease risk independently. Because age is
a major marker for most chronic diseases, an understanding of the
interaction of body composition and body topology with age could
advance our understanding of the interpretation of studies in which
surrogate measures of body composition are used. Some effects of aging
on body composition or body topology, notably loss of muscle mass in
the elderly (Rosenberg and Roubenoff 1995
) or more
central adiposity after menopause (Kotani et al. 1994
,
Zamboni et al. 1997
) are relatively well documented.
However, few studies have addressed the effect of chronological age on
the association between actual measures of body composition and
topology.
The purpose of this analysis was to evaluate the relationship between age and the size and distribution of the fat and lean tissue compartments in a population-based sample of heterogeneously aged women. The following three questions were addressed: 1) What is the proportion of measures of body composition and body topology that can be explained by age? 2) If the sizes of the compartments can be described as a function of age, what are the shapes of those functions? 3) Does age affect the predictive relationship between body topology measures and the sizes of the fat and lean compartments?
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Data were collected in 1993–1994 from 1300 women, aged 18–94 y, who
were residents of three Iowa communities, selected for the unique
mineral characteristics in the community water supply. Each community
is similar with respect to size (<2000 residents per community),
ethnic characteristics (primarily Anglo-Saxon and Germanic),
including the number of foreign born, mean income and primary
occupations. In 1993–1994, a census of community residents was taken.
All women 18 y or older who were ambulatory (capable of climbing
three stairs without assistance) and capable of providing informed
consent were asked to participate in the study without any additional
selection criteria. The study design, methodology and populations have
been described previously (Sowers et al. 1998
). The
participation rates among age-eligible women were >77%. A second
survey was undertaken in the same communities in 1997, and 91% of the
1993–1994 participants were reevaluated. Studies were approved by the
Institutional Review Boards of the Universities of Michigan and Iowa.
Data collection.
Body composition measures.
Body composition, including fat mass and lean mass (excluding bone
mineral content) compartments, was measured using dual X-ray
absorptiometry (DXA-Lunar Corporation, Madison, WI, DPX-L, software
version 1.3y). Two certified technicians scanned each enrollee for
total body bone mineral density, a measure that was accompanied by body
composition information including lean mass, fat mass and bone mineral
content (Elowsson et al. 1998
, Kamel et al. 1999
, Lukaski et al. 1999
). Scans with technical
problems were either sent to the manufacturer for corroborating
analysis (n = 3) or the woman was rescanned
(n = 2). System calibration was performed daily and
a lumbar spine phantom was scanned weekly because calibration phantoms
were not available to evaluate the precision of body composition
measures by DXA. However, the CV for bone mineral density by DXA was
< 1.0% for the femoral neck and lumbar spine sites. DXA provided
data to evaluate fat mass (kg) and fat-free mass (including lean
mass and bone mineral content). The foci of this report are fat mass
(kg) and lean mass (kg).
Height was measured using an anthropometric plane and scale to the nearest 0.1 cm and weight was taken to the nearest 0.1 kg, after preimplementation training. Women had not been fasting, but urinated before measurement for weight. From these two measures, body mass index (BMI, kg/m2) was calculated. There were 18 subjects whose height measures differed by 3 cm or more between the 1993–1994 wave and the 1997–1998 data collections (7 increased, 11 decreased). For these subjects, height measures were set to missing. The average age of subjects with recorded height differences >3 cm was 57.7 ± 4.5 y (mean ± SEM) compared with 54.9 ± 0.5 y for subjects whose measured height changed <3 cm (P = 0.52).
Waist circumference (cm) was measured as the smallest location of the midsection and was undertaken following a forced expiration. Hip circumference (cm) was measured at the location of the greatest gluteal mass. All measures were taken using a nonstretching tape over a single layer of light clothing.
Because of the well-known, if not well-understood effects of
smoking on body weight and metabolism (Grunberg et al. 1992
, Wack and Rodin 1982
), only data from
participants who reported never smoking were used in these analyses. Of
the 1300 baseline participants, 876 were never-smokers and thus
were eligible for inclusion in the analyses. The average age of current
smokers (45 ± 1.3 y) was about a decade younger than never
and former smokers (57 ± 0.6 and 56 ± 1.2 y,
respectively), (P < 0.0001). Women who had ever
used oral contraceptive agents (n = 54) or hormone
replacement therapy (n = 19) were included because
preliminary investigation indicated that the use of these preparations
was not associated with variation in the body composition measures.
Not all age-eligible women had all body composition measurements
taken. One woman refused DXA and anthropometric measurement and was not
included in the analyses. In addition, there were 68 participants
without DXA data whose anthropometric data were included in descriptive
statistics but not in the regression models. The average weight of
participants without DXA scans was 92.6 ± 2.8 kg compared with
72.2 ± 0.5 kg for those who were scanned (P
< 0.0001). The participants without scans also had significantly
greater mean hip and waist circumferences (P < 0.0001); they were 
4.5 y older than the group with DXA scans
(P = 0.07).
Statistics.
The body composition data were modeled as two compartments, fat mass
and lean mass, including body fluids and muscle mass (but excluding
bone mineral content). The data were evaluated for distribution
characteristics and did not require the use of transformations to
address the assumptions of an underlying normal distribution. Linear
regression analyses were used to model body compartment sizes as a
function of age. Age was included in these models as a continuous
variable. Because curvilinear models were evaluated, all models
initially included an age2 as well as the age term. For
models including age2, mean age was subtracted from age
before the calculation of age2 to remove the colinearity
between the age and age2 terms. The hip and waist
circumferences and height were evaluated as main effect terms and were
included in interaction terms with age in the regression models. The
significance level was defined as 
< 0.05. The range of
effects of many independent variables on an outcome, as characterized
by the multiple variable regression models, was depicted with
conditioning display methodology (Becker et al. 1996,
Cleveland 1993).
Weight was not included in the regression models for this investigation
because weight was conceptualized as the sum of outcomes (fat mass +
lean mass without bone + bone mass) rather than as an independent
predictor of compartment size. Further, this paper addresses the
absolute lean and fat body composition compartments, rather than BMI,
because BMI is a surrogate for body fat, but not for percentage of body
fat as shown in Figure 1A and B
.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The physical characteristics of the study sample are shown stratified
by age groupings in Table 1
. The data are shown according to percentiles within four age groups to
facilitate comparison with distributions of values in other
populations, most notably the 85th percentile value used by the
National Center for Health Statistics to describe obesity with data
from the National Health and Nutrition Examination Surveys
(Kuczmarski et al. 1994
).
|
|
). Percentage of body fat, BMI and fat mass (kg) were not
interchangeable in their association with each other (Fig. 1A and B
). Therefore, this report focuses on fat
mass (kg). There were lower correlations between lean mass and weight
and lean mass and BMI (Table 2)
. There were high correlations between
fat mass and waist and hip circumferences, whereas there were lower
correlations between lean mass and waist and hip circumferences.
|
4
5
6
represent the relation of age to measures of body composition (fat mass
and lean mass) and to measures of body topology (waist circumference
and hip circumference). The relationship between fat mass and age and
the circumference measures and age could be summarized with quadratic
models. Figures 3
4
5
depict the fitted age + age2
polynomial model for each measure accompanied by the 95% prediction
intervals for the fitted line. According to these single variable
models, fat mass increased gradually with increasing age until 56 y
at which time the mass of the fat compartment started to decrease.
Waist circumference increased steadily with age throughout young
adulthood and middle age and then declined slowly after 67 y. Hip
circumference values also increased in young adulthood and middle age,
reaching their highest average value at age 56 y. The relationship
between lean mass and age (Fig. 6)
, however, was best
described using two separate regression lines, one for subjects <58 y
and one for those 58 y.
|
|
|
|
The quadratic model of fat mass predicted by age in Figure 3
accounted for 8% of the variance (P < 0.05) in the
data. When measures of height and body topology were also considered in
the modeling, in addition to age, the terms for waist circumference,
hip circumference, height and two interaction terms (age x waist
circumference and age x height) were significant. This model
accounted for 91% of the variation in fat mass; the terms in the
final model are presented in Table 3
.
|
show fat mass predicted by the multiple variable fat mass model over
the nonextreme ranges (5th–95th interquartile range) of observed age
and waist circumferences and at the 33rd, 50th and 85th percentile
values of height and hip circumference. By depicting the model in this
dynamic fashion, it is possible to demonstrate the predicted value of
fat mass over many different sets of predictors simultaneously and to
gain insight from the model. The relationships shown in Figure 7
were
consistent with Figures 3
4
5
6
, which indicate that both the predictors
and outcomes changed with age.
|
The two piecewise regression models predicting lean mass from age 18 to
57 y and 58 to 94 y are presented in Table 3
. In the single
variable model, the use of age did not predict lean mass in the younger
(<58 y) age group; however, when body topology and height were
included in the multiple variable model, age was a significant
predictor of lean mass. The final multiple variable model for lean mass
in subjects < 58 y old accounted for 71% of the observed
variance in the model. Above age 57 y, the modeled relationship
between age and lean mass was more complex; the final model for lean
mass in subjects >58 y old included interaction terms for age x waist circumference and age x height in addition to the
anthropometric measures of lean mass. This portion of the piecewise
regression accounted for 69% of the variance in lean mass. Similar
relationships were observed when fat-free mass (lean mass plus bone
mineral content) was evaluated.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
,
Sowers et al. 1992
) have shown that lean mass is more
strongly associated with bone mineral density than is fat mass.
Further, it is in the muscle mass, rather than the fat mass, that
efficient carbohydrate metabolism takes place.
The simple regression models considering age and
age2 in relation to body topology fat mass had
highly significant overall P-values. It could be surmised
that measures of body composition and body topology change with age in
a curvilinear manner; this generalization, however, did not hold for
lean mass, which was best modeled as a two-piece line, flat and
then declining. The concave-shaped relationship between fat mass
and age in women seen in Figure 2A
has been reported by
several other authors who used four-compartment elemental analysis
(Aloia et al. 1996
, Ellis 1990
),
two-compartment bioelectrical impedance (Silver et al. 1993
) and two-compartment DXA analyses (Mautalen et al. 1996
, Mott et al. 1999
).
Although age was related to both fat mass and body topology, the age at
which the apex of the curves was observed was not the same for each of
the measures. This result suggested that even if these constructs were
correlated biologically, they had different trajectories and were not
behaving similarly. The data presented were consistent with loss of
muscle mass of the elderly, but the data did not provide evidence of
the "excess" muscle loss of the elderly identified as sarcopenia
(Rosenberg and Roubenoff, 1995
). These data also
suggested that the time interval between 40 and 55 y reflected
substantial transition with respect to the ratio of lean-to-fat tissue.
The second aspect of the analysis evaluated the effect of age as well as body topology measures to predict body composition measures effectively. Age was a predictor of lean mass in women < 58 y only after inclusion of body topology predictors in the multiple variable regression models. The absence of an interaction term between age and the circumference measures, as it related to lean mass in the younger age group, was not surprising given that the waist and hip circumferences are usually considered surrogate markers for gluteal or abdominal fat depots.
Among the older age group in the second segment of the multiple variable lean mass model, we had hypothesized that an age x height interaction would reflect two groups. The first group would have included those who were shorter in the young adult and middle-aged group. A second group would have included those who had greater height as young adults and middle-aged women, but who had lost height with aging and become "short" because of spinal compression and/or vertebral crush fractures (height is a recognized risk factor for osteoporosis). We speculated that we might observe that young adult and middle-aged shorter women had more muscle per unit height than the taller women, but were more likely to lose that greater muscle mass per unit height in older age. This hypothesis, however, was not substantiated by the data (not shown).
In contrast to observations in lean mass, age modified the effect of both waist circumference and height in the multiple variable regression model of fat mass. An age2 term was included in the fitted fat mass model as well, indicating that even after accounting for the other predictors, the relationship between age and fat mass was curvilinear.
In the fat mass model, the slope of the waist effect became less steep
as age increased. The two variable regression models (Figs. 3
, 4)
provide one possible explanation for these results. Although fat mass
(Fig. 3)
reached its peak and began to decline at age 56 y, waist
circumference (Fig. 4)
peaked at age 67 y and declined at a
negligible rate after that. Therefore, the outcome (fat mass) and
predictor (waist circumference) were not changing at the same rate.
Changing abdominal tissue distribution with increasing age may be one
possible explanation for this finding. An additional possible
explanation hypothesized by Spencer and Clive (1996)
is
"senescent convergence," i.e., extreme values of body size,
chemistry or function converge with increasing age.
Considering sources of measurement error.
Caution should be exercised in extrapolating the aforementioned
relationships in obese or morbidly obese populations, although DXA is
considered an accurate means of assessing body composition (precision
error < 2%) (Clasey, 1999
, Hansen, 1999
) and is highly correlated with hydrodensitometry
(underwater weighing) values. Dual X-ray densitometry cannot be
undertaken in the morbidly obese because the scan table is mechanically
stable only to weights between 118 and 132 kg (260–290 pounds)
(depending upon the manufacturer). It deforms with loads beyond those
weights, jeopardizing the entire system. Among the subjects without
scans in this investigation, seven women exceeded the 120 kg (260
pound) limit for the DXA machine used. Second, an underlying assumption
of DXA use is that measurements are not affected by the anteroposterior
thickness of the body. However, studies have shown consistently that
body thickness > 25 cm does have an effect on evaluating the
energy signal, typically leading to overestimates of the fat mass.
Measuring fat and lean compartment size with DXA is based on the
assumption that fat-free mass hydration remains constant over all
ages. The testing of this assumption is particularly important because
these data suggest that the fat and lean compartments do change in mass
as a function of age. Although Pietrobeilli et al. (1996
and 1998)
provided theoretical and experimental evidence that DXA
fat mass estimates are sensitive to fat-free mass hydration, it is
reassuring that the size of the error is extremely small, <1% with
hydration changes between 1 and 5%. Schoeller (1989)
concluded that total body water decreases with age, but hydration of
fat-free mass remains relatively constant. In their review of the
literature, Wang et al. (1999)
concluded that the
assumption of constancy of fat-free mass hydration can be assumed
only for nonelderly adults; however, any difference in fat-free
mass hydration with senescence is likely to be small.
Last, the estimation of body composition is a function of the 40–50% of the pixels that do not contain bone, and the fewer the pixels available to provide data, the larger the measurement error is likely to be. Thus, if investigators are more interested in the measurement of regions of the body, including the thorax and arm that may have relatively fewer pixels without bone, they should anticipate that body composition data from those regions are more likely to be prone to measurement error.
There were potential sources of error in the measurement of body
topology. Although standardized protocols were used to measure the
waist and hip circumferences used in this analysis, not all sources of
variability (some of them age dependent) could be eliminated when the
data were being collected. The location of the waist moves up and down
with changes in weight and muscle tone (Baumgartner et al. 1988
); typically, therefore, long-term reproducibility
could be a problem. Both waist and hip circumferences included fat,
lean, bone, skin and, in the case of waist circumference, organ tissues
as well. Changes in any of these tissue compartments with age would be
included in the circumference measurements but would not necessarily be
reflective of changes in the specific composition compartment being
predicted. Hip circumference might have been affected by measurement
modifications due to pelvic size and shape.
Inferences with biological relevance.
The results of these analyses were based on cross-sectional data; thus, no assumptions about rates of change over age can be made. However, if cohort effects can be ruled out as a reason for different compartment sizes in different ages among the study subjects, then the analysis of the data suggests that in this population of women, both the fat and lean compartments change in size over the age span.
In the case of the fat compartment, the age x waist interaction
term in the model suggested that fat may be distributed differently in
different age groups. Recent reports have characterized individuals
whose fat is concentrated mostly in the abdomen (android obesity) as
more likely to develop many of the health risks associated with obesity
than in those with gynoid obesity. Abdominal fat can be further
subdivided into visceral and subcutaneous adipose tissue with quantity
of visceral fat being a better predictor of some adverse health
conditions than subcutaneous fat (Jensen 1997
).
The estimate of total body adipose tissue provided by DXA did not
discriminate between the two types of fat; however, the interaction
term suggested that both a measure of body composition and location
would be more optimal in evaluating disease risk.
Ultimately, questions about change in body composition and topology over age can be answered definitively only with population-based longitudinal studies that measure within-subject rates of change over the course of increasing age. The need for easy-to-apply cross-sectional measures will remain, however, for clinical and public health screening and epidemiologic studies with large samples. The use of more complex measures of body composition, such as underwater weighing or computed tomography, in large studies has typically been precluded by logistical demands including the number of persons to be evaluated, lack of available facilities and time required to complete the measurements. With better understanding of the relationships between more readily gathered measures (e.g., circumferences), more valid and reliable inferences about the underlying human biology can be made from those observations. Furthermore, an understanding of the interactions among these measures may help in determining definitions of compartment sizes at specific chronological ages that constitute a health risk. It also may help explain the discrepant observations reported relative to the importance of composition and topology and health risk among populations on the basis of age, sex or ethnicity.
Manuscript received October 12, 1999. Initial review completed December 15, 1999. Revision accepted April 12, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. Aloia J. F., Vaswani A., Ma R., Flaster E. Aging in women—the four-compartment model of body composition. Metabolism 1996;45:43-48[Medline]
2.
Baumgartner R. N., Heymsfield S. B., Roche A. F., Bernardino M. Abdominal composition quantified by computed tomography. Am. J. Clin. Nutr. 1988;48:936-945
3.
Clasey J. L., Kanaley J. A., Wideman L., Heymsfield S. B., Teates C. D., Gutgesell M. E., Thorner M. O., Hartman M. L., Weltman A. Validity of methods of body composition assessment in young and older men and women. J. Appl. Physiol. 1999;86:1728-1738
4. Ellis K. J. Reference man and woman more fully characterized. Variations on the basis of body size, age, sex, and race. Biol. Trace Elem. Res. 1990;26–27:385-400
5.
Elowsson P., Forslund A. H., Mallmin H., Feuk U., Hansson I., Carlsten J. An evaluation of dual-energy X-ray absorptiometry and underwater weighing to estimate body composition by means of carcass analysis in piglets. J. Nutr. 1998;128:1543-1549
6. Grunberg N. E., Greenwood M. R., Collins F., Epstein L. H., Hatsukami D., Niaura R., O’Connell K., Pomerleau O. F., Ravussin E., Rolls B. J., Audrain J., Coday M. National working conference on smoking and body weight. Task force 1: mechanisms relevant to the relations between cigarette smoking and body weight. Health Psychol 1992;11:4-9
7.
Hansen R. D., Raja C., Aslani A., Smith R. C., Allen B. J. Determination of skeletal muscle and fat-free mass by nuclear and dual-energy X-ray absorptiometry methods in men and women aged 51–84. Am. J. Clin. Nutr. 1999;70:228-233
8. Jensen M. D. Health consequences of fat distribution. Horm. Res. 1997;48:88-92
9. Kamel E. G., McNeill G., Han T. S., Smith F. W., Avenell A., Davidson L., Tothill P. Measurement of abdominal fat by magnetic resonance imaging, dual energy X-ray absorptiometry and anthropometry in non-obese men and women. Int. J. Obes. 1999;23:686-692
10.
Kissebah A. H., Krakower G. R. Regional adiposity and morbidity. Physiol. Rev. 1994;74:761-811
11. Kotani K., Tokunaga K., Fujioka S., Kobatake T., Keno Y., Yoshida S., Shimomura I., Tarui S., Matsuzawa Y. Sexual dimorphism of age-related changes in whole-body fat distribution in the obese. Int. J. Obes. Relat. Metab. Disord. 1994;18:207-212[Medline]
12.
Kuczmarski R. J., Flegal K. M., Campbell S. M., Johnson C. L. Increasing prevalence of overweight among U.S. adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. J. Am. Med. Assoc. 1994;272:205-211
13. Lukaski H. C., Marchello M. J., Hall C. V., Schafer D. M., Siders W. A. Soft tissue composition of pigs measured with dual X-ray absorptiometry: comparison with chemical analyses and effects of carcass thicknesses. Nutrition 1999;15:697-703[Medline]
14. Mautalen C., Bagur A., Vega E., Gonzalez D. Body composition in normal and osteoporotic women. Medicina 1996;56:29-34
15.
Mott J. W., Wang J., Thornton J. C., Allison D. B., Heymsfield S. B., Pierson R. N. Relation between body fat and age in four ethnic groups. Am. J. Clin. Nutr. 1999;69:1007-1013
16.
Pietrobelli A., Formica C., Wang Z., Heymsfield S. B. Dual-energy X-ray absorptiometry body composition model: review of physical concepts. Am. J. Physiol. 1996;271:E941-E951
17.
Pietrobelli A., Wang Z., Formica C., Heymsfield S. B. Dual-energy X-ray absorptiometry: fat estimation errors due to variation in soft tissue hydration. Am. J. Physiol. 1998;274:E808-E816
18.
Rosenberg I. H., Roubenoff R. Stalking sarcopenia. Ann. Intern. Med. 1995;123:727-728
19. Salamone L. M., Glynn N. W., Black D. M., Ferrell R. E., Palermo L., Epstein R. S., Kuller L. H., Cauley J. A. Determinants of premenopausal bone mineral density: the interplay of genetic and lifestyle factors. J. Bone Miner. Res. 1996;11:1557-1565[Medline]
20. Schoeller D. A. Changes in total body water with age. Am. J. Clin. Nutr. 1989;50:1176-1181
21. Silver A. J., Guillen C. P., Kahl M. J., Morley J. E. Effect of aging on body fat. J. Am. Geriatr. Soc. 1993;41:211-213[Medline]
22.
Sowers M. F., Jannausch M., Wood C., Pope S. K., Lachance L. L., Peterson B. Prevalence of renal stones in a population-based study with dietary calcium, oxalate and medication exposures. Am. J. Epidemiol. 1998;147:914-920
23.
Sowers M. F., Kshirsagar A., Crutchfield M. M., Updike S. Joint influence of fat and lean body composition compartments on femoral bone mineral density in premenopausal women. Am. J. Epidemiol. 1992;136:257-265
24. Spencer R. P., Clive J. M. Evidence of senescent convergence during human aging. Mech. Ageing Dev. 1996;91:185-194[Medline]
25.
Wack J. T., Rodin J. Smoking and its effects on body weight and the systems of caloric regulation. Am. J. Clin. Nutr. 1982;35:366-380
26.
Wang Z. M., Deurenberg P., Wang W., Pietrobelli A., Baumgartner R. N., Heymsfield S. B. Hydration of fat-free body mass: review and critique of a classic body-composition constant. Am. J. Clin. Nutr. 1999;69:833-841
27.
Zamboni M., Armellini F., Harris T., Turcato E., Micciolo R., Bergamo-Andreis I. A., Bosello O. Effects of age on body fat distribution and cardiovascular risk factors in women. Am. J. Clin. Nutr. 1997;66:111-115
This article has been cited by other articles:
|
|
 |
|
  E. Atlantis, S. A Martin, M. T Haren, A. W Taylor, G. A Wittert, and for the Florey Adelaide Male Aging Study Lifestyle factors associated with age-related differences in body composition: the Florey Adelaide Male Aging Study Am. J. Clinical Nutrition, July 1, 2008; 88(1): 95 - 104. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. R. Sowers, M. Crutchfield, K. Richards, M. K. Wilkin, A. Furniss, M. Jannausch, D. Zhang, and M. Gross Sarcopenia Is Related to Physical Functioning and Leg Strength in Middle-Aged Women J. Gerontol. A Biol. Sci. Med. Sci., April 1, 2005; 60(4): 486 - 490. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
32, percentage
body fat = -48.82 + 5.11 BMI - 0.07 BMI2; for BMI > 32, percentage of body fat = 37.14 + 0.28 BMI. (B)
Scatter plot showing the relationship between BMI and body fat.
