Vol. 81, No. 5
Printed
Journal
of Clinical
Endocrinology
and Metabolism
Copyright
0 1996 by The Endocrine
Society
Body Composition
Derived
from Whole Body Counting
Potassium
in Growth
Hormone-Deficient
Adults: A
Possible Low Intracellular
Potassium
Concentration
J. S. DAVIES,
W. BELL,
W. EVANS,
R. J. VILLIS,
AND
in U.S.A.
of
M. F. SCANLON
Section of Endocrinology,
Metabolism,
and Diabetes (J.S.D., M.F.S.), Department
of Medicine,
University of Wales College of Medicine;
the Department
of Medical Physics, University Hospital
of
Wales (W.E., R.J.V.), Heath Park, Cardiff, Wales, United Kingdom
CF4 4XW; and Cardiff Institute
of
Education
(W.B.), Cyncoed, Cardiff; Wales, United Kingdom
CF2 6XD
ABSTRACT
for both sexes compared
to values determined
by DEXA.
These discrepancies
may be accounted
for by the lower calculated
potassium
concentrations
compared
with standard
values for both males (56.2
vs. 66.4 mmol; P < 0.001) and females (53.1 vs. 59.6 mmol; P < 0.001).
These observations
suggest
that caution
should be exercised
in the
interpretation
of TBK in GHD
adults,
and the reduced
potassium
concentrations
would alleviate
inaccuracies
in the estimation
of body
composition.
Secondly,
the decreased
intracellular
potassium
concentration
of GHD adults may account for the decreased
muscle strength
and ease of fatigueability
seen in GHD adults.
(J Clin Endocrinol
Metub 81: 1720-1723,
1996)
The validity
of total body potassium
(TBK) measurement
in estimating fat mass and fat-free
mass (FFM)
in GHD adults was assessed
by comparison
with the reference
technique
of dual energy
x-ray
absorptiometry
(DEXA).
The TBK and FFM values
determined
by
DEXA were used to calculate
the potassium
concentration
per kg FFM
in GH-deficient
(GHD)
adults and compared
with standard
values for
normal
subjects
of 59.6 mmol for females
and 66.4 mmol for males.
There were considerable
differences
between
predicted
and measured
TBK values
for both males (3972 us. 3577 mmol; P < 0.001) and
females (2526 us. 2277 mmol; P < 0.001). Similarly,
the estimation
of
FFM and fat mass by TBK measurement
was significantly
inaccurate
T
HE MEASUREMENT of total body potassium (TBK) is
a standard technique used in the prediction of body
composition. The procedure involves measurement of y
emissionsfrom the body arising from the naturally occurring
radio-isotope 40K, which is a constant fraction of total potassium (0.012%).From this measurement, TBK, fat-free mass
@FM), and fat mass (FM) can be calculated assuming particular biological constants (1). In general, the technique provides good estimates of body composition, but its accuracy
has been questioned in malnutrition (2) and obesity (3, 4).
Adults with GH deficiency (GHD) display marked alterations in body composition, with decreased FFM and increasedFM (5,6), together with a reduced exercise capacity
(7). Such abnormalities may contribute to the increased cardiovascular mortality in GHD (8). In consequence, the accurate measurement of body composition in adult GHD has
assumedconsiderable importance, and several studies have
used TBK to measure such changes (5,9-11). However the
increased EM associated with GHD may result in greater
attenuation of the y emission from 40K and consequently
produce underestimates of TBK (3,4). Also, establishedGHD
may well reduce the potassium content of FFM, causing
further underestimates of FFM calculated using the normally
derived biological constants.
In this study we have investigated the value of TBK as a
method of estimating FFM and FM in adults with GHD using
Received April 13,1995. Revision received October 27,1995. Accepted
November
20,1995.
Address
all correspondence
and requests
for reprints
to: Dr. J. S.
Davies, Department
of Medicine,
University
of Wales College of Medicine, Heath Park, Cardiff,
Wales, United
Kingdom
CF4 4XN.
dual energy x-ray absorptiometry (DEXA) as the criterion
technique. DEXA was adopted as the reference technique
because it provides accurate estimates of FM and FFM as
assessedby meat block phantoms (12) and is independent of
the assumptions of invariant relationships among body components that mar certain other techniques, measuring bone
and soft tissue through direct radiological principles. Secondly, we have calculated the K concentration per kg FFM
in GHD adults (GK-FFM) using FFM measured by DEXA
and TBK measured by 40Kcounting and compared this with
the standard figures (K-FFM) for males and females (1).
Subjects
and
Methods
Subjects
Thirty-eight
subjects (21 males and 17 females),
between 20-60 yr of
age, were st;died.
GHD was confirmed
by a plasma GH response of less
than 5 us/L to either insulin-induced
hvnoelvcemia
(0.15 U/kg:
nadir
glucose,“<2
mmol/L)
or GHRH
(1 pgTkg”BW,
iv) together
Gith an
insulin-like
growth
factor I concentration
of 15 nmol/L
or less. GHD was
associated
with partial or complete
hypopituitarism
in 33 adults (receiving
stable replacement
therapy)
and was isolated
in 5 cases. No
subjects were taking diuretics
or p-antagonists,
which could adversely
affect potassium
estimates.
All participants
gave prior consent to investigations,
which were approved
by the joint ethics committee
of the
University
of Wales College of Medicine.
TBK
Measurement
of TBK relies on the detection
of 1.46 MeV ‘y-ray emission from the naturally
occurring
radioisotope
40K (12). 40K is present as
a constant fraction
(0.012%) of total potassium,
and from the 40K spectrum,
TBK can be calculated.
Assuming
that potassium
is confined
almost entirely
to the FFM (131, relationships
may be developed
that
allow calculation
of FFM from TBK.
1720
TOTAL BODY POTASSIUM
MEASUREMENT
TBK was determined
using a recently
refurbished
shadow
shield
whole body counting
chamber.
The counter
consists of six NaI (Tl)
detectors
mounted
radially
on a large steel annulus
within
a shielded
chamber.
Subjects were positioned
supine on a couch, lying along the
axis of the counting
annulus,
wearing
lightweight
clothing
to reduce
radioactive
contamination.
The detectors
made two sweeps of the subject, and the counts were recorded
as a pulse height spectrum
using
commercial
y-spectroscopy
software.
Total scanning
time was 24 min.
The 40K count rate was then corrected
for background
count, and the
actual TBK calculated
by reference
to the count rate obtained
from an
anthropometric
phantom
(14) containing
a known
potassium
content
(12). Finally,
the TBK was corrected
for body attenuation
of the *OK
radiation
using the following
formula
based on the administration
of
42K: TBK X IO.8333 (W/H)0.5
+ 0.48831, where H is height (centimeters),
and W is weight (kilograms)
(15). FFM was then calculated
from corrected TBK using the potassium
concentration
per kg F’FM, i.e. 2.60 g
(66.4 mmol) for males and 2.33 g (59.7 mmol) for females (1). FM was
calculated
by subtraction
of FF’M from body weight. The precision of this
method
is approximately
2% for calibration
standards
and 4% for subjects. Body weight was measured
on a digital scale to the nearest 0.1 kg,
and height was measured
to the nearest 0.1 cm using a Harpenden
stadiometer.
DEXA
DEXA measurements
were made using a Hologic QDR 1000/W
total
body scanner (Hologic,
Waltham,
MA). This technique
has been described previously
(16); in brief, x-radiation
at two different
energies
passes through
the subject whilst he is lying supine. Through
a process
of internal and external
calibration,
the attenuation
characteristics
of the
transmitted
beams are analyzed
to provide
quantitative
data on fat, lean,
and bone mineral masses; the latter two components
constitute
FFM. The
total measurement
time was 15 min with a radiation
dose of 5 @v (19).
The precision
of the technique
is better than 2% (19).
Statistical
analyses
The values of FFM and FM derived
from TBK were compared
to the
results obtained
using DEXA. Statistical
analyses were performed
using
mean, SD, mean difference,
SE of the estimate,
total error, and correlation.
Differences
were analyzed
using Student’s paired t test. Predicted
values
of TBK were obtained
using published
data (19) and compared
with
measured
values in our GHD group using a paired f test. Calculations
of GK-FFM
for males and females were compared
to their respective
standard
values using a single sample t test. All calculations
w&e performed
using. the Minitab
statistical
package
(Minitab
Inc., Universitv
Park, PA). R&ults are expressed
as the’mean”?
SD. P < 0.05 was accepted
as significant.
Results
IN GHD ADULTS
1721
The values of TBK were compared to those predicted from
age, height, weight, and sex using equations published by
Boddy et al. (20). The values in GHD adults were significantly
lower than predicted normal values (Table 1). The serum
potassium concentrations were similar for both sexes,and all
values were within the normal range.
Estimates of body composition
from TBK
(FM and FFM)
derived
Values of FM and FFM derived from TBK were compared
to values obtained using DEXA (Table 2). TBK significantly
underestimates FFM [53.9 ZIS.63.8 (P < 0.005); 38.2 ZJS.43.1
(P < O.OOl)]and overestimates FM (33.7 DS.23.0 and 32.0 IIS.
26.1; P < 0.001) in both males and females, respectively. The
validation statistics (Table 3) demonstrate highly significant
linear relationships between the two techniques for FM and
FFM (ranging from 0.94-0.97; P < 0.001) and significant
differences between TBK and DEXA. The mean differences
for males (-10 kg) were higher than those for females
(-5 kg), corresponding to errors of roughly 16% and 12% of
the absolute FFM and 43% and 19% of the absolute FM,
respectively. Also, total errors in the estimates of FM and
FFM were high, corresponding to errors of approximately
13% (males) and 7% (females) of the body weight. The plots
of the differences between TBK and DEXA values of FFM for
each subject against the average FFM for males and females
are illustrated in Fig. 1. Both plots demonstrate the consistently lower estimates of TBK through all measuresof FFM.
Furthermore, the difference in FFM increases with average
FFM ,with a strong correlation between the two parameters
in both males (r = 0.65; P < 0.01) and females (r = 0.61;
P < 0.01).
Potassium
concentration
per kg FFM
The GK-FFM was calculated using the TBK obtained from
40K counting and the DEXA value of FFM. Values of 56.2 +
2.9 and 53.1 ? 2.8 mmol were calculated for males and
females, respectively. These values were significantly lower
(P < 0.001) than the standard K-FFM values reported by
Womersley et al. (1) of 66.4 (males) and 59.7 mmol (females).
Descriptive data of the study group are listed in Table 1.
The body mass indexes (BMIs) indicate that both sexes are
overweight, but not obese. The DEXA percent body fat of
females (36.9 t 6.5%) is significantly greater than that of
males (26.2 ? 4.6%), although both sexeshave similar BMIs.
GH has major effects on growth and metabolism, such that
deficiency in childhood results in profound growth retardation as well as characteristic increases in FM and a more
TABLE
together
TABLE
2. Comparison
mass (FM) and fat-free
1. Anthropometric
with BMI, IGF-I,
data (mean
and TBK
+ SD) of the
Males (n = 21)
Age(yr)
Ht (cm)
Wt (kg)
IGF-I
(nmol/L)
BMI (kg/m2)
TBK measured
(mmol)
TBK predicted
(mmol)’
Serum K+ cont. (mmol/L)
44.8
175.8
87.5
10.1
28.2
3577
3972
4.4
a TBK predicted
from age, height,
’ P < 0.001 vs. predicted.
5
t
?
?
t
2
2
?
11.1
9.0
16.1
3.4
4.1
546”
669
0.4
sex, and weight
study
Females
45.4
161.8
70.1
8.5
26.7
2277
2526
4.2
(19).
group
Discussion
of TBK-derived
values
mass (FFM)
with those
(n = 17)
-t
?
?
t
5
2
2
i
10.1
7.7
14.6
3.4
4.8
341”
303
0.3
DXA
Males
FM (kg)
FFM (kg)
%BF
Females
FM (kg)
FFM (kg)
%BF
a P < 0.001.
b P < 0.005.
(mean
? SD) of fat
of DXA
TBK
23.0
63.8
26.2
2 7.1
t 10.7
2 4.6
33.7 2 9.9”
53.9 + 8.2b
37.9 2 5.gb
26.1
43.1
36.9
-c 8.9
!I 7.3
2 6.5
32.0 -c 10.2”
38.2 1- 5.7”
44.6 t 6.7”
DAVIES
TABLE
from
TBK
Males
TBK-FM
DXA-FM
TBK-FFM
DXA-FFM
Females
TBK-FM
DXA-FM
TBK-FFM
DXA-FFM
3. Validation
statistics
against
DXA
33.7
23.0
53.9
63.8
9.9
7.1
8.2
10.7
32.0
26.1
38.2
43.1
10.2
8.9
5.7
7.3
for the prediction
of FM
ET AL.
JCE & M . 1996
Vol81
. No 5
and FFM
10.6
2.4
11.3
0.94”
10.0
3.3
10.7
0.95”
5.8
2.0
6.3
0.97”
5.0
2.3
5.6
0.95”
-20 ’
SEE, SE of estimate.
a P < 0.001.
central fat distribution (20). The early treatment of GHD in
childhood using pituitary-derived GH resulted in increased
growth velocity (21) and improvement in body composition
(22). After the attainment of final height, therapy with pituitary-derived GH was usually stopped, principally because
of limited supplies (23). With the advent of recombinant
human GH (241,research into GHD gathered considerable
momentum. Previous studies on GHD in childhood had
indicated the adverse changes in body composition that occur after the cessation of GH treatment. Such data together
with the knowledge that GH is secreted well into adult life
(25) provided the impetus for the study of adult GHD. Increased FM, decreased FFM (5, 61, decreased extracellular
fluid volume (lo), abnormal lipid profile (26), reduced muscle strength (11,27), and decreased exercise capacity (7) are
all recognized complications of adult GHD. Such abnormalities may contribute to the increased cardiovascular mortality found in hypopituitary adults (8). Indeed, a beneficial
change in many of the above parameters is seen after treatment with recombinant human GH (5, 28), although data
concerning life expectancy are still awaited.
A number of studies have used TBK counting to assess
body composition in adults with GHD (5, 9-11). However,
the increased FM in adult GHD may cause increased y-ray
absorption and scattering, leading to underestimates of TBK
and, hence, FFM, as has been described in obesity (3, 4).
Furthermore, the assumption of a constant potassium concentration per kg FFM (K-FFM) is inappropriate in both
obesity (3,4) and malnutrition (2) and may also be inappropriate in GHD, as GH is known to affect the Na+/K+ ATPdependent pump of the cell membrane (29). This pump
maintains the resting potential of excitable cells, and a reduced resting potential may causemuscle weaknessand ease
of fatigueability (30). Landin et al. (31) reported increased
K-FFM in acromegalics and a reduction after surgical cure.
Thus, GHD may reduce the K-FFM and, hence, produce
underestimates of FFM from TBK counting and contribute to
the decreasedmuscle strength and easeof fatigueability seen
in GHD adults. Therefore, we investigated the accuracy of
TBK in predicting body composition in GHD adults compared with DEXA, which measuresFFM and FM without the
need for assumptions of biological constancy. DEXA-FFM
was also used to quantify GK-FFM in our group using the
measured values for TBK. This value was then compared
FIG.
FFM
25
30
1. Plots of the
for each subject
35
40
45
50
55 60
Average FFM (kg)
•~ females -) males
differences
between
against the average
65
70
75
60
TBK and DEXA values
of
FFM for males and females.
with the accepted standards of 66.4 mmol for males and 59.7
mm01 for females (1).
Measured TBK was significantly lower (P < 0.001)than the
predicted TBK for both males and females. Estimatesof FFM
based on TBK were significantly lower than DEXA values in
each sex, with correspondingly higher FM values. The degree
of difference translates into large errors of around 13%
(males) and 7% (females) of total body weight with respect
to estimated values of FM and FFM derived from TBK counting. These data indicate that TBK provides inaccurate estimates of body composition in adults with GHD compared
with those obtained by DEXA.
Are these discrepancies due to increased adiposity,
counter inaccuracy, or decreased intracellular potassium
concentration? If the inaccuracies were related solely to fatness,then greater errors for females might be expected due
to their higher percent body fat. This is not the case,however,
becausemales show the greatest disparities. Also, inaccuracies of 40K counting have been attributed to obesity, and
although a higher number of male patients have BMIs within
the obese range, we have attempted to reduce errors due to
y-ray absorption by correction for anthropometric differences (15). Hence, increased ‘y-ray attenuation by fat is an
unlikely explanation for the errors, although this may still be
a minor contributor to the reduced TBK values. Counter
accuracy was assessedby comparing the anthropometric
phantom counts of our instrument with those from an instrument at another center. No differences were found that
could implicate instrument error. The calculated GK-FFM
was significantly lower than the standard K-FFM for both
sexes,and this probably accounts for the differences between
measured and estimated values of FFM and FM derived from
TBK. Supportive data are provided by the increasing errors
with increasing FFM for both sexes. Thus, the actual TBK
count is probably accurate, but it is the assumption of normal
K-FFMs that causesthe errors in calculations of FFM and FM.
As skeletal muscle contains 75% of the TBK (32) and plasma
potassium concentrations are within normal ranges, then it
is reasonable to assumethat the low GK-FFM reflects a low
muscle potassium concentration, although a muscle biopsy
would be required to provide absolute evidence (33). This
may well contribute to the relative muscle weaknessin adult
GHD.
In conclusion, TBK counting has been found to be an
TOTAL BODY POTASSIUM
MEASUREMENT
inaccurate predictor of FM and FFM in GHD adults. The
data indicate that the inaccuracies are probably a consequence of a true decrease in K-FFM. To avoid such discrepancies in body composition
assessment, a revised KFFM should be adopted for both sexes in adult GHD.
Finally, low muscle potassium
concentrations
in adult
GHD may contribute to the muscular weakness associated
with this condition, and further investigation
of this hypothesis is now required.
References
1. Womersley
J, Boddy
K, King P, Dumin
JVGA. 1972 A comparison
of the fat
free mass of young adults estimated
by anthropometry,
body density
and total
body potassium
content.
Clin Sci. 43:469-475.
2. Royal1 D, Greenberg
GR, Allard
JP, Baker JP, Harrison
JE, Jeejeebhoy
KN.
1994 Critical
assessment
of body composition
measurements
in malnourished
subjects with Crohns
disease: the role of BIA. Am J Clin Nutr. 59:325-330.
3. Gray DS, Bray GA, Gemayel
N, Kaplin
K. 1989 Effect of obesity
on bioimpedance.
Am J Clin Nutr. 50:255-260.
4. Colt EWD, Wang J, Stallone
F, van Itallie
TB, Pierson
RN. 1981 A possible
low intracellular
potassium
in obesity.
Am J Clin Nutr. 34367-372.
5. Salomon
F, Cuneo
RC, Hesp R, Sonksen
PH. 1992 Effects of treatment
with
recombinant
human growth
hormone
on body composition
and metabolism
in adults with growth
hormone
deficiency.
N Engl J Med. 321:1797-1803.
6. Cuneo R, Salomon
F, McGauley
GA, Sonksen
PH. 1992 The growth hormone
deficiency
syndrome
in adults.
Clin Endocrinol
(0x0. 37:387-397.
7. Cuneo
RC, Salomon
F, McGauley
GA, Sonksen
PH. 1991 Growth
hormone
treatment
in growth
hormone
deficient
adults.
II. Effects on exercise
performance. J Appl Physiol.
70:695700.
B-A. 1990 Premature
cardiovascular
mortality
in hypopi8. Rosen T, Bengtsson
tuitarism-a
study of 333 consecutive
patients.
Lancet.
336285-288.
9. Orme SM, Sebastian
JP, Oldroyd
B, et al. 1992 Comparison
of measures
of
body composition
in a trial of low dose growth
hormone
replacement
therapy.
Clin Endocrinol
(OxB. 37:453-459.
10. Rosen T, Bosaeus
I, Tolli J, Lindstedt
G, Bengtsson
B-A. 1993 Increased
body
fat mass and decreased
extracellular
fluid
volume
in adults
with growth
hormone
deficiency.
Clin Endocrinol
(Ox@ 38:63-71.
11. Rutherford
OM, Salem AB, Johnston
DG. 1984 Quadriceps
strength
before
and after growth
hormone
replacement
in hypopituitary
adults: relationships
to changes in lean body mass and IGF-1. Endocrinol
Metab.
1:41-47.
12. Jensen MD, Kanaly
JA, Roust LR, et al. 1993 Assessment
of body composition
with use of dual energy x-ray absorptiometry:
evaluation
and comparison
with
other techniques.
Mayo Clin Proc. 68:867-873.
13. Boddy
K, King PC, Tothill
P, Strong JA. 1971 Measurement
of total body
potassium
with a shadow
shield whole body counter:
calibration
and errors.
Physiol
Med Biol. 16:275-282.
IN GHD ADULTS
1723
14. Spungen
AM, Bauman
WA, Wang
J, Pierson
RN. 1993 The relationship
between
total body potassium
and resting
energy expenditure
in individuals
with paraplegia.
Arch Whys Rehab. 74965-968.
15. Bush F. 1946 Energy
absorption
in radium
therapy.
Br J Radiol.
1914-21.
16. Thomson
WH. 1981 Calibration
and clinical
use of a scanning
whole body
counter.
PhD Thesis, University
of Wales College
of Medicine.
17. Bell W, Davies
JS, Evans W, Scanlon
MF. 1995 The validity
of estimating
total
body fat and fat free mass from skin-fold
thickness
in adults with growth
hormone
deficiency.
J Clin Endocrinol
Metab.
80:630-636.
18. Deleted
in proof.
19. Tothill
P, Avenell
A, Love J, Reid DM. 1994 Comparisons
between
Hologic,
lunar and Norland
dual energy
x-ray absorptiometers
and other techniques
used for whole body soft tissue measurements.
Eur J Clin Nutr. 48x781-794.
20. Boddy
K, King PC, Hume R, Weyers
E. 1972 The relationship
of total body
potassium
to height, weight and age in normal
adults. J Clin Pathol. 25:512-517.
21. Hindmarsh
PC, Brook
CGD. 1992 Disorders
of stature.
In: Grossman
A, ed.
Clinical
endocrinology,
London:
Blackwell;
810-836.
22. Raben MS. 1958 Treatment
of a pituitary
dwarf with human growth hormone.
J Clin Endocrinol
Metab.
l&901-903.
23. Collipp
PJ, Thomas
J, Curti
V, Sharma
RK, Meddiah
VT, Cohn SH. 1973
Body composition
changes
in children
receiving
human
growth
hormone.
Metabolism.
22589695.
24. Jorgensen
JOL. 1991 Human
growth
hormone
replacement
therapy.
Pharmacological
and clinical
aspects.
Endocr
Rev. 12189-207.
25. Goeddel
DV, Heynekee
HL, Hozumi
T, et al. 1979 Direct expression
in E. coli
of a DNA sequence
coding
for human growth
hormone.
Nature.
281:544-548.
26. Finkelstein
JW, Roffwarg
HP, Boyar RM, Kream
J, Hellman
L. 1972 Age
related changes in the 24 hr spontaneous
secretion
of growth hormone.
J Clin
Endocrinol
Metab.
35665-670.
27. Merimee
IJ, Hollander
W, Fineberg
SE. 1972 Studies
of hyperlipidaemia
in
the human
growth
hormone
deficient
state. Metabolism.
11:1053-1061.
28. Cuneo RC, Salomon
F, Wiles CM, Sonksen
PH. Skeletal
muscle performance
in adults with growth
hormone
deficiency.
Horm Res. 33fSuppl4):55-60.
29. Binnerts
A, Swart GR, Wilson
JHP, et al. 1992 The effect of growth hormone
administration
in the growth
hormone
deficient
adults on bone, protein,
carbohydrate
and lipid
homeostasis
as well as on body composition.
Clin Endocrinol
(Oxf). 3779-87.
30. Shimomura
Y, Lee M, Oku J, Bray GA, Glick
Z. 1982 Sodium
potassium
ATPase in hypophysectomised
rats: response
to growth hormone,
triiodothyronine
and cortisone.
Metabolism.
31:213-216.
31. Dorup
I. 1994 Magnesium
and potassium
deficiency.
Acta Physiol
Stand.
618(Suppl):l-55.
32. Landin
K, Petmson
B, Jakobsson
K-E, Bengtsson
B-A. 1993 Skeletal
muscle
sodium
and potassium
changes
after successful
surgery
in acromegalics:
relation to body composition,
blood glucose, plasma
insulin
and blood pressure.
Acta Endocrinol
(Copenh).
128:418-422.
A, Kjeldsen
K. 1991 Interrelation
of hypokalaemia
and potassium
33. Norgaard
depletion
and its implications:
a re-evaluation
based on studies
of the skeletal
muscle
sodium,
potassium-pump.
Clin Sci. 81449-455.