close

Вход

Забыли?

вход по аккаунту

?

Comparison of body composition in middle-aged and elderly males using computed tomography.

код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 66289-295 (1985)
Comparison of Body Composition in Middle-aged and Elderly
Males Using Computed Tomography
GARY A. BORKAN, DAVID E. HULTS, STEPHEN G. GERZOF, AND
ALAN H. ROBBINS
Normative Aging Study, VA Outpatient Clinic, Boston, Massachusetts
02108 (G.A.B., D.E.H.) and Radiology Service, VA Medical Center, Boston,
Massachusetts 02130 (S.G.G., A.H.R.)
KEY WORDS
Body composition, Computed tomography, Obesity,
Aging
ABSTRACT
Computed tomography (CT) scans were taken of 21 middleaged men (mean age 46.3 years) and 20 older men (mean age 69.4 years) to
measure differences in body composition with age. Overall, the older men
weighed 8.2 kg less than the middle-aged men, and this difference was primarily the result of their having less lean tissue. Although fat mass (by whole
body potassium counting) was only slightly less in older men, there were
distributional differencesin fat between the age groups. Total abdomen adipose
tissue area (from CT) was similar in both groups, although the subcutaneous
portion of the abdomen adipose tissue was less in the older men, and they had
correspondingly more adipose tissue within the abdominal cavity. Muscle
areas of the leg and arm were significantly less in the older men, as were all
lean tissues of the abdomen and chest. When these data were corrected for
differences in body weight with age, the results were still significant, suggesting a centripetalization and internalization of fat with age. Causes of this
apparent fat redistribution and decrease of lean tissue with age were not
revealed by this study and are presently unknown.
The changes in physique and body compo- there have been no studies of site-specific
sition in adulthood have been less exten- muscle area or volume with age because of
sively studied than comparable changes in the lack of any practical technique t o meachildhood. Declines in muscle and organ sure these parameters.
We have been exploring the applications of
mass with age have been documented in
computed
tomography (CT) scanning to body
cross-sectional and longitudinal studies, using anthropometric techniques and more so- composition research and have found that
phisticated measures of lean body mass CT allows measurements heretofore imposCBrozek, 1952; Forbes, 1976; Norris et al., sible with other techniques. CT scans are
1963;Rossman, 1978). During late adulthood computer-reconstructedradiographic images
the fat mass of the body has been reported to which can be made of any body cross section
remain relatively constant or to increase and depict adipose and muscle tissues with
slightly (Borkan and Norris, 1977;Friis-Han- great clarity. Computer software available
sen, 1965; Krzywicki and Chinn, 1965).Thus, with modern CT scanners allows measurethe percentage of the body that is fat typi- ment of the total areas of adipose or lean
tissue in an image, measurement of linear
cally increases with age.
Site-specific external anthropometry has distance between two points, and determinarevealed that extremities are smaller in cir- tion of density of selected tissues.
Our previous investigations have shown
cumference in the elderly, while trunk meathat
tissue areas of a single abdomen cross
surements are larger (Borkan and Norris,
1977; Brozek, 1952; Garn and Young, 1956; section are highly predictive of the volume of
Skerlj, 1959). The site-specific skinfold mea- adipose tissue in the abdomen overall @orsurements confirm that this reflects de- kan et al., 1982). Other studies by ourselves
creased extremity adipose tissue and slightly (Borkan et al., 1983a) have shown that CT
increased trunk adipose tissue. These tech- measurements are better predictors of total
Received January 9, 1984; revised September 27, 1984; acniques do not measure intermuscular or internal abdomen adipose tissue. Furthermore, cepted October 4,1984.
0 1985 ALAN R. LISS, INC.
290
G.A. BORKAN ET AL
fat and fat free weight than estimates using
other body composition techniques.
The present investigation used CT scanning to measure differences in body composition in middle-aged and older men. In
particular, we were interested in age differences in adipose and muscle area and density
and in age-related differences in their relative distribution throughout the body.
MATERIALS AND METHODS
The study population comprised 21 men
between 41 and 52 years (mean of 46.3 years)
and 20 men between 59 and 76 years (mean
of 69.4 years). All were participants in the
Veterans Administration Normative Aging
Study (Bosse et al., 1984; Damon et al., 1972).
All men were of European ancestry and were
community living. Subjects had no known
significant or debilitating illnesses a t the
time of measurement. None were using diuretic medication a t the time of testing.
An Ohio Nuclear 2010 CT scanner was utilized. Three separate CT images were taken
of each participant during a single visit: at
the middle of the upper leg halfway between
the patella and pubic symphysis; a t the abdomen at umbilicus; and at the chest at the
nipples including the middle portion of the
upper arms. Sample CT images for the three
sites for two men are shown in Figures 1-6.
For all CT scans the subject was supine.
Scans were generated with a 50 mA, 120 kV
beam for a duration of 2 seconds. A scan
diameter of 40.3 cm was used for leg and
abdomen scans, and 50.3 cm was used for
chest scans. Scan thickness was constant at
1cm. This scanner derives a n image from a
256 x 256 matrix of picture elements (pixels). CT attenuation scores range from - 1000
(air) to +lo00 (dense bone), with zero representing the attenuation of water. Daily calibration assures a maximal variation of +/5 CT units per day. The average reported
dose in any cross-sectional image was 1.6 R,
with minimal scatter to surrounding tissues.
The study protocols were approved by the
Human Studies Committees of the VA Outpatient Clinic and the Boston VA Medical
Center, and fully informed consent was obtained from each subject.
Software available with the CT scanner
enabled us to encircle all or part of a n image
and calculate the adipose or lean area in
square centimeters by specifying the appropriate density ranges. Fat was calculated as
being all pixels with scores between -250
and -50 units; lean tissue was between -49
and +lo0 units. The rationale behind the
selection of these density ranges is discussed
in Borkan et al. (1982). It should be noted
that different CT attenuation ranges would
lead to slightly different estimates of adipose
and lean tissue, but relationships between
individuals and age groups should be constant. Because of the need to limit radiation
to subjects, we were not able to obtain replicate CT scans of the same individuals to determine reliability of measurement. However, remeasurement of the same CT images
for areas of tissues was perfect, if the same
range of tissue attenuation scores (CT numbers) was specified.
A potassium-40 scan was performed in a
whole body counter to determine lean body
weight and (by subtraction) fat weight (Tyson
et al., 1970). The low level counting facility
a t the VA Medical Center (Boston) consists of
two matched 4 x 8 inch sodium iodide crystal detectors mounted in a scanning geometry in a n 8 x 8 x 8 foot room with 9-inch
steel walls. The subject lay supine on a bed
between the two detectors for two 20-minute
traverses of the counter assembly. The
counter accumulates 10 to 15 counts per minute per nanoCurie of potassium, based on
calibration. The 40-minute scan period was
determined to obtain a SE of measurement
of 5%.Total milliequivalents of potassium-40
were converted to total body potassium, and
lean body weight was calculated using the
equation of Forbes (1963). The reader should
be aware that the potassium-40 technique
derives a n estimate of ether-extractable fat
(lipid), and by subtraction from total weight
produces a n estimate of fat-free weight. CT
cross-sectional areas are estimates of adipose
tissue, which contains some connective tissue, water, blood vessels, etc. However, for
our purposes, these measurements are comparable, because adipose tissue is primarily
composed of ether-extractable lipid.
RESULTS
Results from the anthropometry and whole
body counting demonstrate variation in body
composition between men of different ages
(Table 1).The significant difference in weight
between the middle-aged and older groups
(8.2 kg) was largely attributed to significant
differences in fat-free weight (6.6 kg) rather
than fat (1.5 kg). When corrected for differences in body size (weight) by analysis of
covariance, it is evident that fat is relatively
increased in the elderly men, and fat-free
291
COMPARISON OF BODY COMPOSITION USING CT
TABLE 1. Comparison of middle-aged (N = 21) and older (N = 20) men on
mean body composition parameters by analysis of variance and analysis of
covariance (correcting for differences in body weight)
Mean from
analysis of variance
MiddleF
aged
Old
Parameters
Age
Weight (kg)
Fat weight (kg)
Lean body weight (kg)
Height (cm)
46.3
82.0
25.0
56.2
174.1
69.4
73.8
23.5
49.6
170.8
Adjusted mean from
analysis of covariance
Middleaged
Old
F
8.2*
0.9
18.8*
2.7
22.6
54.6
172.7
26.0
51.3
172.5
4.0**
4.0**
0.1
Fat weight, percent fat, and lean body weight were derived from whole body potassium
counting.
*P < 0.01.
**P < 0.05.
TABLE 2. Comparison of middle-aged and old men in mean CT-derived tissue areas using analysis of
variance and analysis of covariance (controlling for weight)
Means from
analysis of variance
Middleaged
Old
Adipose
Upper leg
Abdomen (total)
Subcutaneousabd
Internal abd
Internalkotal abd
Chest
Upper arm
Lean
Upper leg
Abdomen
Chest
Upper arm
Total
Upper leg
Abdomen
Chest
Upper arm
F
Adjusted means from
analysis of covariance
Middleaged
Old
F
68.9
300.1
184.1
115.7
38.9
174.9
22.0
49.2
362.7
194.3
168.4
46.4
218.7
26.1
6.0**
5.5**
0.3
9.0*
4.0**
4.5**
1.9
0.1
6.4**
145.1
249.3
247.7
53.3
131.4
235.8
256.3
49.5
7.8**
1.5
0.5
1.9
33.8*
0.3
0.3
4.0**
224.7
576.2
476.6
83.7
189.6
641.5
518.1
83.2
17.5*
7.2**
3.8
0.0
71.0
331.2
205.7
125.6
38.5
193.7
23.5
47.1
330.0
171.1
158.3
46.7
198.9
24.5
10.7*
147.3
254.3
254.0
54.6
129.1
230.7
249.7
48.2
16.8*
229.3
613.7
501.1
86.5
184.8
602.2
491.0
80.3
0.0
3.8
4.1**
5.7**
0.1
0.2
5.8**
*p < 0.01.
**p < 0.05
tissue is significantly decreased. The slightly
shorter stature of the older men is consistent
with the combined effects of secular (generational) trends in stature and age-related
shrinkage. Statistical analysis indicated that
the weight and body composition differences
between middle-aged and older men are not
explained by the observed stature differences.
Tissue area measurements from the CT
scans (Table 2) indicate age-related differences in adipose and lean tissue distribution
by site. The middle columns show differences
between the age groups compared by analysis of variance. Adipose tissue in the upper
leg was significantly less in the older men
than in the middle-aged. Although total ab-
dominal adipose was similar between middle-aged and older men, there was less
subcutaneous adipose tissue and significantly more internal abdominal adipose tissue in the older men. Thus, the percentage of
total abdominal adipose tissue that is internal was significantly higher in the older
group. Both chest and arm adipose areas
were nonsignificantly greater in the older
men.
On the right side of Table 2 are the comparisons of tissue areas by analysis of covariance, which correct for differences in overall
body weight between middle-aged and older
men. By holding weight stable, it is possible
to obtain more definite evidence of body
292
G.A. BORKAN ET AL.
shape difference with age. Weight-corrected
means show that while leg adipose tissue
was significantly less in the old, abdominal
adipose actually increased significantly,
largely owing to a highly significant increase
in internal abdominal adipose tissue. Chest
adipose tissue was significantly greater in
the old, and the upper arm adipose difference
was still nonsignificant.
Lean tissue areas, which include both muscle and organs (Table 2, left), were significantly lower in the leg, abdomen, and upper
arm in the elderly. After correction for weight
by analysis of covariance (Table 2, right), this
trend remained significant only for the leg.
For the total area of tissue in each cross
section (adipose, muscle, and bone), the upper
leg was significantly decreased with age (the
largest F-statistic of any in Table 2). The
upper arm also showed a significant decrease
in total tissue with age. In the weight-corrected data (Table 2, right), upper leg remained significantly smaller in the old and
young. When corrected for differences in body
weight, the abdomen total area was significantly larger in the elderly as well.
The analyses in Table 2 compare two age
categories, whereas the analyses in Table 3
evaluate age as a continuous variable using
correlation analysis. The simple correlations
between the overall body composition measures and age show that while age was significantly and negatively correlated with
weight, this was largely due to its association
with lean body weight rather than with fat
weight.
Significant bivariate correlations between
fat areas and age were found for upper leg
(negative) and for internal abdominal adipose tissue (positive). The ratio of internal
abdomen adipose to total abdomen adipose
was also significant and positively correlated
with age. When we partialed out the effect of
weight (to provide a n indication of body
shape) the upper leg and internal abdominal
adipose areas and ratios retained their significance, and total abdominal adipose and
chest adipose also became significant as well.
Lean areas of the leg (abdomen and upper
arm) showed significant negative correlations with age. When corrected €or weight,
only upper leg remained significantly correlated with age. For total area of the cross
sections, upper leg was highly negatively correlated with age, and upper arm was negatively correlated as well. When corrected for
age by partial correlation analysis, upper leg
remained highly negative, and abdomen and
TABLE 3. Correlation of CT areas with age and partial
correlations of CT areas with age, controlling for weight
Simple
Corr
Overall
Weight
Fat weight
Lean body weight
Body mass index
Fat
Upper leg
Abdomen (total)
Subcut abd
Internal abd
IntiTotal abd
Chest
Upper arm
Lean
Upper leg
Abdomen
Chest
Upper arm
Total tissue
Upper leg
Abdomen
Chest
Upper arm
Partial
Corr
-0.36"
-0.04
-0.47"
-0.19
-0.47"
0.05
0.16
0.31**
0.36"
0.06
0.06
-0.43"
0.38"
0.15
0.43"
0.36"
0.31""
0.19
-0.08
-0.36"
-0.43"
-0.21
0.06
-0.22
-0.66"
0.01
-0.03
-0.26**
-0.60"
0.45*
0.28'"
-0.05
-0.53*
-0.33""
*P < 001
**P < 0 05
chest had significant positive correlations
with age.
IIISCUSSION
Computed tomography provides the first
practical opportunity to measure directly the
areas of fat and muscle in selected anatomical cross sections of living humans. To apply
Figs. 1-6. Two sets of CT scans taken using the body
composition protocol developed in our study. The left
hand series is from a 49-year-old man; the right hand
series is from a 70-year-oldman. The men differ in weight
by only 2 pounds and in height by 0.1 inch. On the CT
images black is air, dark gray is fat, light gray is muscle
and organs, and white is bone.
Figs. 1,2. Images of the upper legs at their midpoint.
Subcutaneous fat, individual muscles, and femur cross
section are visible. Fat infiltration is considerable in the
older man's leg, a characteristic seen in the elderly.
Figs. 3 , 4. Images of the abdomen at the umbilicus.
Internal abdomen fat is that portion inside the abdominal wall musculature. Subcutaneous fat is outside the
abdominal wall musculature. Intestines may be differentiated and intestinal gas as well. A greater amount of
internal abdomen fat exists in the abdomen of the older
man.
Figs. 5, 6. Images of the chest a t the nipples with
arms crossed (at midpoint of the upper arm). Visible are
the lungs, heart, fat, and muscle distribution and cross
sections of ribs. In the arms, fat, muscle, and bone are
visible.
COMPARISON O F BODY COMPOSITION USING CT
293
294
G.A. BORKAN ET AL.
CT scanning to questions of aging seems particularly appropriate because of the likelihood that many body composition changes
with age occur internally and are not measurable with other techniques.
Our findings suggest the dynamic nature
of adipose and lean tissues with advancing
age. Although adipose tissue in younger
adults is primarily subcutaneous, there appears to be increasing “centripetalization”
and “internalization” of fat with age. This is
most clear when the weight-corrected data in
Table 2 are considered. Perhaps of greatest
interest is the abdomen, where the significant relative increase in adipose tissue with
age is largely accumulated within the abdominal cavity, rather than the subcutaneous adipose depots. There was less agerelated difference in adipose tissue in the
arm than the leg (also true for lean tissue).
Indeed, the age trends were nonsignificant
and positive for the arm and significantly
negative for the leg. These results suggest
that aging in the upper arm may be more
related to the patterns seen in the trunk
than are leg age changes.
In our other studies, fat infiltration within
and between muscles was significantly
greater in the elderly at all sites examined
(Borkan et al., 1982). Histological studies by
other investigators (Frantzell and Ingelmark, 1951) have documented an increase in
fat content of dry human gastrocnemius
muscle from 10%at age 10 to 30%at age 80.
Age-related changes of similar magnitude
were found for deposits of fat beneath muscle
fascia and between muscles. Overall, these
findings suggest a progressive internalization of fat with age, both within the abdominal cavity and within and between muscle
tissues.
Furthermore, our study demonstrates that
the decrease in lean tissue area between middle-aged and older men is relatively consistent between the four cross-sectional sites
examined. However, when corrected for
weight, it appears that old men are disproportionately leaner in the upper leg than
middle-aged men. Because the upper leg is
the site of substantial decrease in both fat
and muscle with age, the total area of the leg
is highly significantly decreased in the elderly men. Thus, in ranking the various anatomical cross sections in terms of apparent
change with age, the upper leg area is the
most divergent of any variable we tested.
Since the total leg area is well approximated
by leg circumference, this latter measure
might be very useful as a biological age
measure in field studies.
Many of the age-related differences in body
composition shown in Tables 1-3 may be observed through comparing the CT scans of
the two men in Figures 1-6. Although these
men are almost identical in weight and
height, the older man has less leg tissue
overall, fat infiltration in and between leg
muscles, more internal abdomen fat, and fat
infiltration in the deep back muscles. While
these results do not themselves have statistical significance, the population analyses in
Tables 1-3 do show that these are significant
differences in our sample.
In growth data that are not longitudinal,
there is always the possibility that the age
differences reflect differences in sample selection, survivorship, or secular trends. Especially of concern in gerontologic data is
whether the older group is a subset of survivors with unusual physical or genetic characteristics. It is possible that results such as
ours could exist if there were premature
deaths of obese or mesomorphic individuals.
However, it has recently been noted (Andres,
1980) that in the large majority of studies,
moderately overweight individuals do not
have greater morbidity than underweight
ones. Further, the mortality in the Normative Aging Study sample has been very low
and thus far is unrelated to body weight (Borkan et al., 1981). Furthermore, longitudinal
data from the Normative Aging Study and
other longitudinal studies documents a decline in weight past age 55 within individuals morkan et al., 1983). The correction of
the data for body size (Tables 1-3) tends to
ensure that we are truly looking at body
shape differences, rather than the by-product
of body size differences. Our results support
the view that the CT adipose area findings
represent fat redistribution rather than being
an artifact of overall body size differences.
However, longitudinal studies must be made
to obtain definitive proof of actual fat redistribution taking place in individuals.
Computed tomography studies of body composition also demonstrated the remarkable
variation in fat distribution between individuals (Borkan et al., 1982). We have found
individuals who externally appear very lean
and have low skinfold thickness but whose
CT scans reveal extensive internal abdominal fat. It will be important to investigate
whether there are physiological differences
in internal abdominal and subcutaneous fat,
which may have bearing on health and dis-
295
COMPARISON OF BODY COMPOSITION USING CT
ease risk. Of particular interest is whether
internal abdominal fatness is correlated with
glucose intolerance, given recent findings
that adult onset diabetics have high trunk to
extremity fat ratios. Since both internal abdominal fat deposition and glucose intolerance are features of human aging, the
interrelationship needs investigation. Of especial interest is whether internal abdominal fat accumulation may precede or follow
glucose intolerance, or whether they are only
indirectly related by some third, presently
unidentified factor.
The promising results using CT scanning,
which provide accurate and reproducible
measures of body composition, including important parameters previously unmeasurable, suggests that future investigations in
body composition may find this technique of
great value.
ACKNOWLEDGMENTS
This research was supported by the Medical Research Service of the Veterans Administration. We thank Drs. J.E. Silbert and B.A.
Burrows for their role in making this study
possible. We thank M. Hourihan, E. Barrett,
B. Eblan, and J. Zozula for producing the CT
scans and J. Cardarelli for assistance with
the whole body counting.
LITERATURE CITED
Andres, R (1980)Effect of obesity on total mortality. Int.
J. Obesity 4:381-386.
Borkan, GA, Gerzof, SG, Robbins, AH, H u h , DE, Silbert, CK, and Silbert, J E (1982)Assessment of abdominal fat content by computed tomography. Am. J. Clin.
Nutr. 36:172-177.
Borkan, GA, Hults, DE, Gerzof, SG, Burrows, BA, and
Robbins, AH (1983a)Relationship between computed
tomography tissue areas, thicknesses and total body
composition. Ann. Hum. Biol. 10:537-546.
Borkan, GA, Hults, DE, and Glynn, R J (1981) Anthro-
pometric prediction of mortality in the Normative Aging Study. Am. J. Phys. Anthropol. 54:203.
Borkan, GA, Hults, DE, and Glynn, R J (198313) Role of
longitudinal change and secular trend in age differences in male body dimensions. Hum. Biol. 55.629641.
Borkan, GA, and Norris, AH (1977) Fat redistribution
and the changing body dimensions of the adult male.
Hum. Biol. 49:495-514.
Bosse, R, Ekerdt, DJ, and Silbert, JE (1984) The Veterans Administration Normative Aging Study. In SA
Mednick and M. Harvey (eds): Longitudinal Research
in the United States. Boston: Martinue Nijhoff (in
press).
Brozek. J (1952) Changes in bodv comDosition in man
during maturity and their nutritional implications.
Fed. Proc. 11:784-793.
Damon, A, Seltzer, CC, Stoudt, HW, and Bell, B (1972)
Age and physique in healthy white veterans at Boston.
J. Gerontol. 27:202-208.
Forbes, GB (1976) The adult decline in lean body mass.
Hum. Biol. 48:161-173.
Forbes, GB (1963)Age and sex trends in lean body mass
calculated from K40 measurements: With a note on
the theoretical basis for the procedure. Ann. N.Y. Acad.
Sci. 110:255-263.
Frantzell, A, and Ingelmark, BE (1951) Occurrence and
distribution of fat in human muscles at various age
levels. Acta Soc. Med. Upsalien 56:59-87.
Friis-Hansen, B (1965)Hygrometry of growth and aging.
Symp. Soc. Hum. Biol. 7:191-209.
Garn, SM, and Young, RW (1956) Concurrent fat loss
and gain. Am. J. Phys. Anthropol. 14:497-504.
Krzywicki, HJ, and Chinn, KSK (1965) Human body
density of an adult male population as measured by
water displacement. Am. J. Clin. Nutr. 16:305-310.
Norris, AH, Lundy, T, and Shock, NW (1963) Trends in
selected indices of body composition in men between
ages 30 and 80 years. Ann. N.Y. Acad. Sci. 110:623639.
Rossman, I (1978) Anatomic and body composition
changes with aging. In: C Finch and L Hayflick (eds):
Handbook of the Biology of Aging. New York: Van
Nostrand Reinhold.
Skerlj, B (1959) Age changes in fat distribution in the
female body. Acta Anat. 3856-63.
Tyson, JS, Genna, S, Jones, RL, Bikerman, V, and Burrows, BA (1970)Body potassium measurements with a
total body counter. J. Nuc. Med. llr255-259.
-
"
I
Документ
Категория
Без категории
Просмотров
0
Размер файла
621 Кб
Теги
using, malen, tomography, body, aged, comparison, composition, computer, middle, elderly
1/--страниц
Пожаловаться на содержимое документа