Clinical Science (1985) 68, 455-462
455
Mass and composition of the fat-free tissues of patients with
weight-loss
L. BURKINSHAW’
AND
D. B. MORGAN’
Departments of ‘MedicalPhysics and ‘Chemical Pathology, University of Leeds, The General Infirmary, Leeds, U.K.
(Received I May112 October 1984; accepted 2 November 1984)
Summary
1. An estimate of the mass of fat-free tissue in
the body can be calculated from body weight and
skinfold thickness; this estimate is called the ‘fatfree mass’. Total body potassium and nitrogen are
alternative estimates. Factor analysis of data for
healthy subjects has defined relationships between
the true values of these three quantities and
estimated the random component of the variance
of each, i.e. the component independent of variations in the mass of fat-free tissue. The results
indicated that all three were reliable measures of
the mass of fat-free tissue. However, it is not
known whether these findings are valid for patients
who have lost weight.
2. We have measured the same three quantities
in 104 wasted patients with heart disease or disorders of the gastrointestinal tract. The patients’
mean values were significantly less than corresponding values for healthy volunteers. The patients
had a mean ratio of total body nitrogen to fat-free
mass similar to that of healthy subjects, but lower
mean ratios of potassium to fat-free mass and
nitrogen. These findings suggest that the potassium
content of the patients’ fat-free tissues was
abnormally low.
3. Factor analysis of the patients’ data gave
relationships between the true values of the three
quantities similar to those for healthy subjects;
however, total body potassium was 100-300 mmol
lower in patients than in healthy subjects with the
same fat-free mass or total body nitrogen.
4. Factor analysis also showed that the random
components of variance of fat-free mass and total
body nitrogen were similar to those in healthy
Correspondence: Dr L. Burkinshaw, Department
of Medical Physics, University of Leeds, The
General Infirmary, Leeds LS1 3EX, U.K.
subjects. Therefore, in the patients as in healthy
subjects, fat-free mass was as valid a measure of
fat-free tissue as the more complex measurement
of total body nitrogen. The random component
of total body potassium was twice as big as in
healthy subjects; however, it formed no greater
a proportion of total variance than did the random
components of the other two quantities.
5. Total body nitrogen, and hence body
protein, could be estimated from measured fat-free
mass with a standard error of approximately
136 g (compared with 139 g for healthy individuals), and from total body potassium with
a standard error of 129 g (compared with 91 g in
healthy subjects).
Key words: body composition, fat-free mass,
fat-free tissue, total body nitrogen, total body
potassium .
Introduction
Loss of weight is a feature of many diseases, and in
extreme cases obviously includes loss of fat-free
tissues. However, the frequency and extent of this
loss are uncertain, largely because there is no
agreed estimate of the mass of fat-free tissue.
The simplest estimate, the ‘fat-free mass’, is the
difference between body weight and body fat
calculated from skinfold thickness [l, 21. The
accuracy of this method, even for healthy subjects,
has been questioned [3], and the assumptions on
which it is based may not be valid for patients who
have lost weight.
An alternative estimate of the mass of fat-free
tissue is obtained from total body potassium,
which can be determined by measuring the radioactivity of the natural radiosotope of potassium,
4‘k[4]. However the potassium content of the
45 6
L. Burkinshaw and D. B. Mrgan
fat-free tissues may not be the same in patients as
in healthy subjects, as the patients may be depleted
of potassium, or may have lost a potassium-rich
tissue such as skeletal muscle.
Finally, it has recently become possible t o
measure total body nitrogen by neutron activation
analysis in vivo [4-91. Total body nitrogen is
a measure of total body protein and thus of the
mass of fat-free tissue.
In healthy volunteers we found that these three
estimates of the fat-free tissues were correlated
[ 101. Total body potassium and nitrogen could be
estimated from each other or from fat-free mass,
by linear regression, with standard errors comparable with the errors of the measurements. We
therefore concluded that the three quantities were
equally valid measures of the mass of fat-free
tissue in health. However, linear regression analysis
is strictly valid only if the independent variable is
measured without random error, otherwise the
slope of the relationship is underestimated [l 11
and the error of estimation of the dependent
variable is exaggerated.
Only part of the variability of each of the three
quantities is due to variations in the true mass of
fat-free tissue; the remainder, the random component, is due to variations in the composition
of the fat-free tissues and to random errors of
measurement. Factor analysis [ 121 gives unbiased
estimates of the slopes and intercepts of the
relationships between the ‘true’ values of the
quantities, i.e. mean values for a group of people
containing a given mass of fat-free tissue, as well
as estimates of the components of variance.
We therefore subjected the values of fat-free
mass, total body potassium and total body nitrogen
of the healthy subjects to factor analysis [13]. The
relationships of true potassium and nitrogen with
true fat-free mass had positive intercepts on the
axis of fat-free mass; the relationship of true
potassium to true nitrogen had a negligible intercept. We concluded that healthy men have a higher
ratio of total body potassium to fat-free mass than
healthy women because they are bigger, and not
because of the sex differences per se. The random
component of variance, expressed as a proportion
of the total variance, was smallest for total body
potassium; therefore potassium was the most
reliable measure of the mass of fat-free tissue.
It is not clear whether the composition of fatfree tissue in wasted patients is abnormal. Wasted
patients have been shown to have ratios of total
body nitrogen to potassium different from those
of healthy subjects [14-161. Regression equations
relating potassium to nitrogen have been calculated
for combined groups of patients and healthy
people [16-181, but the separate relationships for
patients and healthy subjects have not been
compared. We have now analysed data for fat-free
mass, total body potassium and total body nitrogen
in wasted patients by factor analysis and compared
the results with those found for healthy volunteers.
Our aims were (a) to compare relationships
between the measures in health and disease and, if
they were different, to deduce how the composition of the patients’ fat-free tissues had changed,
and (b) t o compare the reliability of the three
measures as estimates of the mass of fat-free tissue
in disease.
Patients and methods
Patients
Ten men and 13 women had severe chronic
heart disease but were free of oedema. Their
potassium and nitrogen measurements have been
published [19], but not the measurements of
fat-free mass. Forty-one men and 40 women had
gastrointestinal disease; 56 had inflammatory
bowel disease and the majority of the others had
cancer of the large bowel or stomach. The patients
were compared with 9 1 healthy volunteers: 29
women, 18 policemen and 44 civilian men [ l o ,
131.
Methods
The methods have been described before [lo,
13, 191. The patients were weighed and the results
corrected to nude weight. Skinfold thicknesses
were measured with Harpenden calipers; the
cardiac patients were measured at the biceps,
triceps, subscapular and abdominal sites, but the
surgical patients were measured at the first three
sites only. The results were used to cdculate
fat-free mass [l].A fat-free mass of 56.5 kg (equal
to that of ‘Reference Man’ [20]) can be estimated
by this technique with a standard error of 2 kg.
Total body potassium and nitrogen were
estimated from the natural radioactivity in the
patient’s body and that induced by whole body
irradiation with 0.5 mSv of 14 MeV neutrons. The
standard error of each estimate was 4% of body
content [lo]. Total body protein was estimated by
multiplying total body nitrogen by 6.25 [21].
The mass of the mineral compartment of the
body was also estimated. The activation and
counting procedure gave the sum of the body
contents of potassium, sodium, chlorine, phosphorus and calcium with a standard error of
approximately 0.5 kg [21]. To this sum was added
the estimated oxygen content of bone mineral.
Oxygen forms 4 1.4% of bone mineral and calcium
Fat-free tissues of patients with weight-loss
39.8% [20]; therefore oxygen content was found
by multiplying total body calcium by 1.04.
Total body water was measured, in the surgical
patients only, by dilution of a tracer dose of
7.4MBq (200pCi) of tritiated water [21]. The
standard error was estimated to be approximately
1 litre [22].
Statistical methods
Means, standard deviations and correlation
coefficients were computed by the Statistical
Package for the Social Sciences [23]. Group means
were compared by Student’s t-test. Null hypotheses
were rejected above a 5% level of probability.
Factor analysis was carried out by the method
of maximum likelihood with the BMDP Statistical
Software [24]. To compare the slopes b l and bz of
relationships found by factor analysis for two
groups, the asymptotic variances of b l and b z , ot
and o:, were first calculated as described by
Barnett [25]. The test statistic T = (bl -bz)/
d(o:+02) was then calculated. A value of T
greater than 1.96 was taken to indicate a significant difference.
Ethical considerations
The volunteers and patients consented to the
studies after the experimental procedures had been
explained to them. The studies were approved by
the Neutron Activation Panel of the Medical
Research Council (forerunner of the Administration
of Radioactive Substances Advisory Committee
of the Department of Health and Social Security)
and by the Research Ethics Committee of the
General Infirmary at Leeds.
457
Results
Table 1 shows the mean values and standard
deviations of age and measured quantities for the
two groups of patients and the healthy volunteers.
The patients were on average shorter than the
healthy subjects, but this difference is too small
to explain why they weighed less and had less fatfree tissue.
Fat-free tissue is made up almost entirely of
water, protein and minerals [26]; therefore the
sum of these components is the mass of the
fat-free tissues. The mean value of the sum in the
surgical patients (45.3, kg; standard error 1.O kg)
was not significantly different from their mean
fat-free mass (45.2 kg; standard error 1.0 kg).
Similar agreement was found for male,and female
patients separately.
Table 1 shows that the ratio of nitrogen to fatfree mass was similar in all groups and in both
sexes. In contrast, the ratios of potassium to
fat-free mass and potassium to nitrogen were
significantly lower in the patients than in healthy
subjects of the same sex. This might suggest
a preferential loss of potassium by the patients.
However, the ratio of potassium to fat-free mass is
lower in smaller people, even when the tissues have
normal composition [13], so that the lower ratio
observed in the patients could be merely a consequence of the loss of fat-free tissue.
We therefore calculated, as described in the
appendix to our previous paper [13], the expected
mean ratios for a group of healthy subjects with
measured fat-free mass and nitrogen equal to the
mean values for the patients. The expected values
were 57.45 mmol of potassium and 31.68 g of
nitrogen/kg of fat-free mass, and 1.84 mmol of
TABLE1. Mean values and standard deviations of measured and derived quantities
The symbols FFM, TBK and TBN denote fat-free mass, total body potassium and total body nitrogen
respectively.
Diagnosticgroup
Sex
Healthy volunteers M
Patients with
heart disease
Surgicalpatients
No. of
Age
subjects (years)
62
F
29
M
10
F
13
M
41
F
40
42.53
110.86
43.84
k11.75
50.44
k8.60
50.59
k7.00
57.20
k13.23
55.63
k16.41
Height Weight
(cm)
(kg)
FFM TBK TBN TBK/FFM TBK/TBN TBN/FFM
(kg) (mmol) (g) (mmol/kd (mmol/g)
(g/kg)
77.27 62.33 3129 2046
177.4
26.9
29.85 k6.63 k454 k291
61.33 41.85 2334 1310
162.8
k6.6 k11.39 k5.23 k337 k213
172.0 63.48 51.29 2711 1646
26.2 28.66 k6.15 k455 k215
161.3
55.51 39.33 1966 1279
k4.1 i7.43 k4.12 k215 k99
63.90 51.55 2764 1664
170.0
k6.7 k14.10 k8.09 5609 k265
159.0 55.19 38.60 1947 1235
k6.7 k10.69 k5.39 k332 k183
59.87
k3.97
55.75
k4.04
52.91
k6.71
50.08
i3.49
53.49
k8.11
50.55
t6.13
1.83
kO.11
1.79
k0.13
1.65
k0.21
1.54
i0.15
1.65
i0.23
1.58
k0.18
32.79
k2.69
31.26
k2.88
32.15
i2.43
32.78
23.58
32.42
k3.24
32.18
i3.86
L. Burkinshaw and D.B. Morgan
458
potassium/g of nitrogen. The corresponding
observed values for the whole group of patients
were 51.88 mmol/kg, 32.35 g/kg and 1.61 mmol/g.
Thus the decrease in the ratios of potassium to
fat-free mass and nitrogen was not simply a consequence of the patients’ smaller size.
To establish unbiased relationships between the
measured quantities, and to determine the random
component of variance of each, we analysed the
data by factor analysis. Relationships did not
differ significantly either between the sexes or
between disease groups. Therefore data from all
the patients were combined.
The obtained relationships between the ‘true’
values of fat-free mass (FFM,, kg), total body
potassium (TBK,, mmol) and total body nitrogen
(TBN,, g) were:
TBKt =-704.96
+ 67.76 FFMt
TBKt = -577.39
+ 2.02 TBNt
TBNt =-63.25
+ 33.60 FFMt
TBKt = -1.69
TBNt =-341.58
+ 70.00 FFMt
+ 1.81 TBNt
+ 38.58 FFMt
1
2
4000
v
J
0
20
40
60
80
100
Fat-free mass (kg)
3000 ( b )
1
(1)
2500.
(2)
h
M
v
5
(3)
These relationships are shown in Fig. 1. Points
representing mean values for men and women
lie close to the lines, an observation consistent
with the use of a single set of relationships. The
slopes are, as expected, greater than those of the
corresponding regression equations (57.3, 1.79 and
28.4 respectively).
The relationships found by factor analysis for
healthy subjects [13] were:
TBKt = -621.39
5000 ( a )
3
2000-
M
8
c
‘E
1500.
a
P
3
1000-
g
50 1
0.
20
40
60
80
roo
Fat-free mass (kg)
(4)
9
(5)
(6)
These slopes also are greater than those of corresponding regression equations (64.77, 1.41 and
35.70 respectively) [lo].
Fig. l(a) shows that the patients contained less
potassium than healthy subjects with the same
measured fat-free mass; at the mean fat-free mass
the difference was 184 mmol. However, neither
slopes nor intercepts are significantly different
(T = 0.47 and 0.36 respectively). Fig. l(c) shows
that patients had less potassium than healthy
subjects with equal measured nitrogen content; at
the mean total body nitrogenvalue for the patients
the difference was 282 mmol. The slopes of the
relationships are not significantly different (T =
1.60), but the intercepts are ( T = 2.9). The
differences between -the slopes and intercepts of
the relationships of nitrogen with fat-free mass
.$E
E
50001 ( c )
/
4000
g
l
o
0
500 1000 1500 2000 2500 3000
Total body nitrogen (g)
FIG. 1. Relationships between fat-free mass, total
body potassium and total body nitrogen estimated
by factor analysis for patients (broken lines) and
healthy subjects (solid lines). The lengths of the
lines denote the ranges of the data. The mean
values for healthy men ( 0 ) and women (0), and
for male (m) and female (0)patients, are shown.
Fat-flee tissues of patients with weight-loss
(Fig. lb) approach statistical significance. For the
slopes the value of T is 2.0, for the intercepts it is
2.2.
The standard deviations of the random components of fat-free mass and nitrogen (3.58 kg and
99 g respectively) were very similar to those for
healthy subjects (3.12 kg and 103 g) [13], but the
value for potassium (226 mmol) was approximately
twice that in healthy subjects (108 mmol) [13].
When the random components of variance of
fat-free mass, potassium and nitrogen were each
expressed as a percentage of total variance, the
values found for the patients, 16%’ 14% and 11%
respectively, were almost equal. The corresponding
values for the healthy subjects, 7.5%, 1.9% and
5.676, were not equal; potassium had a smaller
proportion of random variance than fat-free mass
or nitrogen.
Discussion
The two groups of patients had strikingly similar
body composition. The mean values of the
measured quantities, their ratios and the relationships between them were all very similar. This
similarity, which presumably occurred by chance,
allowed us to combine the groups.
If the results are to be used to estimate the
mass of fat-free tissue remaining in wasted patients,
they must be accurate, and, in particular, any
systematic errors should be similar in patients and
in healthy subjects. This is probably the case for
the measurements of potassium, nitrogen and
minerals. The technique was calibrated by measuring a series of realistic anthropomorphic phantoms
simulating the weights and dimensions of patients
referred to us for measurement [27-291, and there
is no reason to suppose that the measurements are
more or less accurate for patients than for healthy
volunteers. The same can be said of the measurement of body water by dilution.
The estimate of fat-free mass from skinfold
thickness was originally standardized by relating
measured skinfold thickness to fat content,
estimated from measured body density, in a group
of healthy volunteers [ 11. No similar standardization has been carried out in wasted patients, where
error could arise because of changes in the proportion of total fat that is subcutaneous, or in the
composition of the fat-free tissues due to an excess
of fluid (as oedema). Errors do occur in some
individual estimates; two of our surgical patients
had values of fat-free mass greater than body
weight. However, the good agreement between
the mean values of fat-free mass and the sum of
water, protein and minerals shows that the method
is valid for groups of patients.
459
It is common practice to assess the composition
of the fat-free tissue by examining the ratios
between its constituents [ 14-16]. Our patients
had lower mean ratios of potassium to fat-free
mass and to nitrogen than the healthy subjects.
The differences were not accounted for by the
smaller size of the patients, supporting the conclusion that the patients’ tissues were relatively
low in potassium.
Factor analysis gave similar relationships
between fat-free mass and nitrogen in patients
and in healthy subjects but, at given nitrogen and
fat-free mass, the patients contained 100-300
mmol less potassium than the healthy subjects.
This small difference explains the patients’ relatively low ratios of potassium to nitrogen and
fat-free mass. In a previous paper [19] we concluded that the difference could result from
cellular depletion of potassium, from loss of
a potassium-rich tissue such as skeletal muscle or
from a combination of the two. Another possibility, recently proposed by James et al. [30], is
that the patients had lost protein mainly from
the cellular compartment, leaving them with
a relatively high proportion of potassium-free
extracellular collagen. Our present analysis cannot
distinguish between these possibilities.
In our earlier paper [13] we pointed out that
relationships (4) and (6) above, for healthy
subjects, have intercepts on the axis of fat-free mass
of 8.9 kg, each with an estimated standard error of
approximately 2 kg. We therefore suggested that
lean tissue had a non-cellular component of about
9 kg, which contained no potassium or nitrogen,
and a variable component containing potassium
and nitrogen in the ratio of 1.81 mmol/g. Transposing eqns. (1) and (3) for the patients we find:
+
FFM = 10.40 TBK/67.76
FFM = 1.88
(7)
+ TBN/33.60
Compared with the corresponding values for
healthy subjects, the intercept of eqn. (7) is not
significantly different, whereas that of eqn. (8) is
marginally so (T= 2.2). If these differences are
ignored, the equations can be replaced by others
defining relationships which have the intercepts
found in healthy subjects, but which pass through
the patients’ mean values:
+ TBK/64.94
FFM = 8.9 + TBN/40.12
FFM = 8.9
(9)
(10)
If these equations are valid, then the patients had
the same mass of non-cellular tissue as the healthy
L. Burkinshaw and D. B. Morgan
460
subjects, but had cellular tissue containing 1.62
mmol of potassium (rather than 1.81 mmol)/g of
nitrogen, i.e. approximately 10% less than the
healthy subjects. This explanation is not entirely
satisfactory because it requires the relationship
between potassium and nitrogen (eqn. 2) to pass
through the origin, whereas we find a significant
intercept. Thus we cannot confirm unequivocally
that the patients had the same mass of noncellular tissue as the healthy subjects.
In healthy subjects, potassium had the lowest
proportion of random variance, and thus can be
considered the best measure of the mass of fat-free
tissue in health [13]. However, in patients, the
random component of total body potassium was
twice as big as in healthy subjects, although
random variance, expressed as a percentage of
total, was similar for all three quantities.
The activation analysis technique is not widely
available, and therefore it is useful to consider how
accurately total body nitrogen, and hence protein,
can be estimated from the more easily measured
fat-free mass or total body potassium.
Two estimates are of interest. One is the
expected mean nitrogen content of a group of
people, each with the same mass of fat-free tissue
as a subject whose potassium content or fat-free
mass has been measured; the other is the actual
total body nitrogen content of the same individual.
Numerically these estimates are identical, being
calculated from eqns. ( 5 ) and (6) for healthy
subjects, or eqns. (2) and (3) for patients, but their
standard errors are different. The errors, calculated
as set out in the Appendix, are given in Table 2.
For healthy subjects, total body potassium is
clearly a better estimator of nitrogen than is
fat-free mass; the error with which the actual
nitrogen content of an individual can be estimated
from measured potassium (91 g) is comparable
with the estimated analytical error of 76 g [lo].
For patients, the errors of estimation of nitrogen
TABLE 2 . Standard errors (g) o f estimating the
nitrogen content of an individual, or the mean f o r
a group, f r o m measured total b o d y potassium or
fat-free mass
Measured
quantity
Fat-free
mass
Totalbody
potassium
Healthy subjects
Patients
Mean of Individual
a group
Mean of Individual
a group
120
139
120
136
60
91
112
129
from fat-free mass are hardly changed; the errors
of estimation from potassium are increased, but
do not exceed those from fat-free mass.
Acknowledgments
The surgical patients were under the care of Mr
G. L. Hill at the time of measurement. We are
grateful to Miss D. W. Krupowicz and Mr K.
Brooks for making the measurements on the
patients and volunteers. Development and application of neutron activation analysis in vivo were
supported by a series of grants from the Medical
Research Council. We are grateful to the referees
for helpful comments and particularly for pointing
out the oxygen content of bone mineral.
APPENDIX
Let M be the true mass of fat-free tissue in the
body. Let FFM,, TBK, and TBN, be the 'true'
values of the three measured quantities, i.e. the
mean values for a large group of subjects with
a given value of M. Assume that the true values
are each made up of a constant term (which may
be negative) and a term proportional to M, i.e.:
TBNt = (YN
+ PNM
(A31
where CYF,
OF, CYK,OK, CXN and PN are constants.
The quantity M may be written:
M=p+6
('44)
where p is the mean value of M for the population
sampled and 6 is a random variable, of mean zero
and variance u2, representing variations in the true
mass of fat-free tissue between individuals. If L =
6/u, L has mean zero and unit variance, and
M=p+uL
('45)
Substituting into eqns. (Al) to (A3):
('47)
where aF = (YF + &p etc. and bF = PFU etc. Thus
a F ,aK and aN are the mean values of the measured
Fat-fiee tissues of pa,tientswith weight-loss
variables in the population sampled, and bF, bK
and b~ are the standard deviations of those variations in the measured quantities that result from
variations in the true mass of fat-free tissue.
Let FFM,, TBK, and TBN, be the ‘actual’
values of the measured quantities for an individual,
i.e. the values that would be found if there were
no random errors of measurement. Actual values
differ from true values by variable amounts
because of variations in the composition of the
tissues. Thus:
FFM, = UF
~ F -kL EFb
649)
+
N +L EN b
(A10)
TBK, = a K 4- ~ K LEKb
~
TBN, = UN
factor analysis partitions the total variance of each
measured quantity into a component due to
Variations in the true mass of fat-free tissue ( b 2
etc.) and a component independent of such
variations (02etc.). If the variances of the errors
of measurement, uF: etc., are known, then the
‘biological’ variances, OF: etc., can be found from
equations such as:
Substituting numerical values of the parameters
into eqns. (A6) to (A8), and eliminating L between
pairs of equations, gives three equations relating
the true values of pairs of variables. Let the
equation relating TBN, to TBKt be:
(A1 1)
where EFb, EKb and ENb are random variables,
with means of zero and variances OF:, OK: and
ON:,
representing biological variations between
individuals.
Finally, let FFM,, TBK, and TBN, be the
measured values of the quantities. Then:
46 1
TBNt = CK + mK TBKt
(A20)
Expressing TBK, in terms of measured total body
potassium, TBK,:
TBNt = CK
+ mK TBK,
-mKEK
(A21)
(A12)
Thus the best estimate of TBNt from measured
total body potassium is
TBK, = aK
(A13)
TBNt = CK mK TBK,
TBN,
(A14)
The standard error of the estimate is mKoK.
The actual nitrogen content of an individual is
FFM, = U F
~ F L f F b 4-
E F ~
+ ~ K 4-LEKb 4- e~~
= a N + ~ N +LENb + EN^
where E F ~E, K and
~
EN^ are random variables, with
means of zero and variances OF:, OK: and ON:,
representing errors of measurement. Combining
the two random variations into a single term, we
have :
FFM,
TBK,
+ ~ F +LEF
=aK + bKL + E K
=UF
(A15)
(A16)
Here EF etc. are random variables of mean zero,
and variances uF’ OK’ and ON’, which represent
the combined effects of (a) variations in the
composition of fat-free tissue and (b) random
errors of measurement. The variances OF’ etc. are
given by equations such as:
02 = OF: + OF:
(A181
The values of aF, aK and a N are estimated by the
mean values for the subjects measured. Assuming
that L , E ~ eK
, and EN are all normally distributed,
factor analysis by the method of maximum likelihood estimates b ~b K, , bN, OF’, OK’ and ON’. Thus
+
TBN, = TBNt -mKEK 4-
(A22)
ENb
(A23)
Thus the estimate of TBN, is TBNt, as before, but
its standard error is d(mKzOK’ + ON:).
Similar
reasoning gives the errors of estimating body
nitrogen from measured fat-free mass. When
calculating the errors given in Table 2, the value of
uNe was taken to be 7 6 g [lo]. Readers of our
earlier paper [13] should note an error in eqn.
(A9). The term u F ~ on
F the right-hand side should
read a&.
References
1. Durnin, J.V.G.A. & Womersley, J. (1974) Body fat
assessed from total body density and its estimation
from skinfold thickness: measurements on 481 men
and women aged from 16 to 7 2 years. British Journal
of Nutrition, 32,71-91.
2. Lohman, T.G.(1981) Skinfolds and body density and
their relation to body fatness: a review. Human
Biology, 53, 181-225.
3. Cohn, S.H.,Ellis, K.J., Vartsky, D., Sawitsky, A.,
Gartenhaus, W., Yasumura, S. & Vaswani, A.N. (1981)
Comparison of methods of estimating body fat in
462
L. Burkinshaw and D. B. Morgan
normal subjects and cancer patients. American Journal
of Clinical Nutrition, 34, 2839-2847.
4. Burkinshaw, L. (1978) In vivo analysis and the wholebody radiation counter. In: Liquid Scintillation
Counfing,vol. 5 , p. 111. Eds. Crook, M.A. &Johnson,
P. Heyden, London.
5. Oxby, C.B., Appleby, D.B., Brooks, K., Burkinshaw,
L., Krupowicz, D.W., McCarthy, I.D., Oldroyd, B.,
Ellis, R.E., Collins, J.P. & Hill, G.L. (1978) A technique for measuring total-body nitrogen in clinical
investigations using the I4N(n, 2n) "N reaction. International Journal of Applied Radiation and Isotopes,
29,205-211.
6. Williams, E.D., Boddy, K., Harvey, I. & Haywood,
J.K. (1978) Calibration and evaluation of a system for
total body in vivo activation analysis using 14 MeV
neutrons. Physics in Medicine and Biology, 23, 405415.
7. Harvey, T.C., Jain, S . , Dykes, P.W., James, H., Chen,
N.S., Chettle, D.R., Ettinger, K.V., Fremlin, J.H. &
Thomas, B.J. (1973) Measurement of whole-body
nitrogen by neutron-activation analysis. Lancet, ii,
395-399.
8 . Mernagh, J.R., Harrison, J.E. & McNeill, K.G. (1971)
In vivo determination of nitrogen using Pu-Be sources.
Physics in Medicine and Biology, 22,831-835.
9. Vartsky, D., Ellis, K.J. & Cohn, S.H. (1979) In vivo
measurement of body nitrogen by analysis of prompt
gammas from neutron capture. Journal of Nuclear
Medicine, 20, 1158-1165.
10. Burkinshaw, L., Morgan, D.B., Silverton, N.P. &
Thomas, R.D. (1981) Total body nitrogen and its
relation to body potassium and fat-free mass in
healthy subjects. Clinical Science, 61,457-462.
11. Armitage, P. (1971) Statistical Methods in Medical
Research, 1st edn., p. 312. Blackwell Scientific
Publications, Oxford.
12. Lawley, D.N. & Maxwell, A.E. (1971) Factor Analysis
as a Statistical Method, 2nd edn. Butterworths,
London.
13. Morgan, D.B. & Burkinshaw, L. (1983) Estimation of
non-fat tissues from measurements of skinfold thickness, total body potassium and total body nitrogen.
Clinical Science, 65, 407-414.
14. Dabek, J.T., Vartsky, D., Dykes, P.W., Hardwicke, J.,
Thomas, B.J., Fremlin, J.H. & James, H.M. (1977)
Prompt gamma neutron activation analysis to measure
whole body nitrogen absolutely: its application to
studies of in vivo changes in body composition in
health and disease. Journal of Radioanalytical
Chemistry, 31, 325-331.
15. Cohn, S.H., Gartenhaus, W., Sawitsky, A., Rai, K.,
Zanzi, A, Vaswani, A, Ellis, K.J., Yasumura, S.,
Cortes, E. & Vartsky, D. (1981) Compartmental body
composition of cancer patients by measurement of
total body nitrogen, potassium and water. Metabolism,
30, 222-229.
16. McNeill, K.G., Mernagh, J.R., Jeejeebhoy, K.N.,
Wolman, S.L. & Harrison, J.E. (1979) In vivo measurement of body protein based on the determination of
nitrogen by prompt y analysis. American Journal of
Clinical Nutrition. 32, 1955-1961.
17. Ellis, K.J., Yasumura, S . , Vartsky, D., Vaswani, A.N.
& Cohn, S.H. (1982) Total body nitrogen in health
and disease: effects of age, weight, height and sex.
Journal of Lclboratory and Clinical Medicine, 99,
917-926.
18. Hill, G.L., Bradley, J.A., Collins, J.P., McCarthy, I.,
Oxby, C.B. & Burkinshaw, L. (1978) Fat-free body
mass from skinfold thickness: a close relationship
with total body nitrogen. British Journal o f Nutrition,
39,403-405.
19. Thomas, R.D., Silverton, N.P., Burkinshaw, L. &
Morgan, D.B. (1979) Potassium depletion and tissue
loss in chronic heart-disease. Lancet, ii, 9-11.
20. Snyder, W.S., Cook, M.J., Nasset, E.S., Karhausen,
L.R., Parry Howells, G. & Tipton, I.H. (1975) Report
of the Task Group on Reference Man. 1.C.R.P.Publication no. 23. Pergamon Press, Oxford.
21. Hill, G.L., McCarthy, I.D., Collins, J.P. & Smith, A.H.
(1978) A new method for the rapid measurement of
body composition in critically ill surgical patients.
British Journal of Surgery, 65,732-735.
22. Hill, G.L., Bradley, J.A., Smith, R.C., Smith, A.H.,
McCarthy, I.D., Oxby, C.B., Burkinshaw, L. & Morgan,
D.B. (1979) Changes in body weight and body protein
with intravenous nutrition. Journal of Paren teral and
En teral Nutrition, 3, 2 15-2 18.
23. Nie, N.H., Bent, D.H. & Hull, C.H. (1970) Statistical
Package for the Social Sciences. McGraw Hill, New
York.
24. Dixon, W.J. (ed.) (1981) BMDP Statistical Software
1981. University of California Press, Berkeley.
25. Barnett, V.D. (1969) Simultaneous pairwise linear
structural relationships. Biometries, 25, 129-142.
26. Siri, W.E. (1956) In: Advances in Biological and
Medical Physics, vol. 4, pp. 239-280. Ed. Lawrence,
J.H. &Tobias, C.A. Academic Press, New York.
27. Oxby, C.B. & Brooks K. (1979) A versatile variable
anthropomorphic phantom. Physics in Medicine and
Biology, 24,440-442.
28. McCarthy, I.D., Sharafi, A., Oxby, C.B. & Burkinshaw,
L. (1980) The accuracy of total body nitrogen determined by neutron activation analysis. Physics in
Medicine and Biology, 25,849-863.
29. Sharafi, A., Pearson, D., Oxby, C.B., Oldroyd, B.,
Krupowicz, D.W., Brooks, K. & Ellis, R.E. (1983)
Multi-element analysis of the human body using
neutron activation. Physics in Medicine and Biology,
28,203-214.
30. James, H.M., Dabek, J.T., Chettle, D.R., Dykes, P.W.,
Fremlin, J.H., Hardwicke, J., Thomas, B.J. & Vartsky,
D. (1984) Whole body cellular and collagen nitrogen
in healthy and wasted man. Clinical Science, 67,
73-82.