Perspectives
Adjustment
in
Nutrition
for Size of
B. Wnii,
A
Possible
Misuse
WILLIAM
Ratios
JR.,
MI).4
It is hoped that pret it moderately
lation to nutritional
Perspectives in Nutrition and present a spectrum of
research applied to
will
ideas
review
that will problenis.
the
literature
serve as
selectively,
a continual
interstimu
medical
In assessment
One the is, the
miiany
technic, determination
phases of body
which of
of
nutrition, composition
is external used retention
the is
balance, of extensively,
indirect essential.
is that a inherent
iiias
indicated
.
by
the
transversal
(line
of
dashes)
However, group
iii
it
and
is
clear
is,
that
instead,
groupi
similar
differs from to group II.
subThe more
Downloaded from www.ajcn.org by guest on June 6, 2011
calculation
of
the
specific
problem
concerns
the
stance
measured
by this
subtracting
intake.
measured
Many
excretion
problems
from use
of
of
different
the
ratio
size.
to
It
“equate”
is commonly
data such a surface
on
assumed
persons
that
in
methodology

have
purpose consider a of
been
of statistical ratios5 or, this
discussed
cornproblem
measurements be related the on
made a weight this three
are whether the
2
on
group area
should basis.
previously.’ munication concerning
The is the to misuse
or
Obviously more unless
previously knowing by
produces general
met. they There will this an
an improper ratio criteria mentioned
is be each animals no met time. experiment6 fed high way without of
specifically, to relate “correcting” presenting meter
To
the data for values per the
about
common practice on persons of this per calories of
two
of attempting varying size for per for : or (2) the should
variation, kilogram, metabolized. a ratio
quantities,
example, examining square Figure
showing
data from hydration
in
way actual of
is
and
justify
100 use
the
expressing
these
and
three in
low
animals
protein
of
diets
different
(HP,
LP)
size. or
which
of
resulted
food and
information
Subsequently,
conditions
items considered
should
be should
satisfied increase
(1)
the decrease
together line for
at 0, amid
in a linear the two
(3) the
fashion, variables
of
were deprived of food twothey water. Since the data on the control animals met the criteria regression intercept a ratio, they could be expressed should a ratio (gm. [FFDSJ) of 1120 or regression. per (per 100 cent The
hydration for the
equally gin. H20) use of starved and suggest in the a fatfree or
of the use of
well by ratio groups thirsted incoras dry an
variance
the
variable
increase the
An
with
example
the
of the
increasing
problem
magnitude
is illustrated
of solids
expression
variables. Figure
correct
.
to
describe
would
the
and comparable,
hydration
make the and related The use the two would to
of
two starved
the
experimental
in
the
1 The
relationship
graph
for
on
the
the
use
left
of
shows
a
animals
ratio. comparable,
The be
analysis.
data expressed
graphed
The use
on correctly
of
the
a to
right, only
ratio righthand that
however, by regression
would figure) of School
Cleveland,
cangroups
rectly the allowed sions, make of
that
there water.
is a difference of
hydration they were expres
animals on
whether regression
the data for group i (in appear to be comparable
From
cine,
*
the
group of MediOhio.
the
other is not
water,
hand,
but
shows the
whether
correctly animals
they were
the
Department
Reserve Professor Journal of
of Pediatrics,
University, of Pediatrics. Clinical Nutrition
that
were
the
difference fed high
in whether protein
Vol. 11,
Western Assistant
allowed
previously
249
or
low
diets.
September 1962
Anzerican
25()
Weil /
I
20 Ili
MEAN
MEAN
20 ‘II
RATIO1 (333
I
40
<
60
(22.59
I OOGM.
,ooct)
M(AN
RATIO1
MEAN SD,
RATIOS
SD1> 1>111
M#{128}ANATIO R
>
3D,,,
Raimo1
= SD
1>111
R&TO (iii.
MEAN
ii
00
GM)
kAN

M#{128}a RAno
(22 I
00
GM.)
FIG.
1.
and
Theoretical The omi the figure
the resultimig
considerations omi the left relationships deniomistrates
in which the
concernimig shows a the between situatiomiin
(lata should
the necessary the
which be
use of three
the
ratios
amid regresfor data.
a ratio
siomis.
ratio
conditiomis grupof
use of conipared
the The
would
use
of
a
figure
right
ami(l
l)e
mfliSlea(limlg,
regm ession by a
Downloaded from www.ajcn.org by guest on June 6, 2011
techmiic.
80
I 70 60

.
0
U
LP LP HP HP
St. S&T St
ONLY
50 40 130
‘/)
ONLY
a
sar
20
z

$00
w
90
80
70
w
3
60
0
U)
50
4.
30 20
I0
0 0 4 8
FFDS
2
IN of
6
GRAMS rat
20
24
28
32
36
40
44
48
ABSOLUTE
Fm;.
2.
were
The
water
content
carcasses.
Comitrol
mnals
of
the
was
fed 24 per cemit protein carcasses is expressed in
dry starved
dI1(l vater
(HP) or
grams expressed vater (St.
12 amid
per in
cemit grams.
and (LP)
experimiiental diets. with the group
The weight of rats
aniwater of on the each
compared unilamid another )
fatfree (liet ,,f food
solids ‘itli
(S
( FFI)S) access
& T).
also to
One
group
‘as (leprived
This nutrition
problemii when
may the
occur in retemition
the to
field
and
of
the
retention for inthe be
iii
by effects seen Figures
weight of
iii
will weight.
“correct” That
both this In side
retentiomi
figures is SO not
take The tion be
of usual ler assumed
a
substance is
are to examine intake per
be this kg. as It imitake
related. may reten shown might the and a
the
:,
hypothetical
4 and righthand
neither
examples
Figure illustrate
mior
technic kg. \rersus that
5.
3,
graphs
in
on
the
which
dividing
both
situation
Misuse
of Ratios
25 1
weight
RETENTION RETENTION 4
is
tendency
when
and to
compared
retention, l)roduce a more
intake
iith
the positiYe
the
greater ‘, ill
variatiomi
he correlation
in the
I
intake
2 6 8
(0 INTAKE (2 (4
234567
WEIGHT
between
Figure
physiologic
intake
4
data
per
may retention rthermore ith
kg.
be
since
amid retention
representative
increases
per
ith
kg.
of to does ever, per
4
weight,
3 RETENT3ON
PEA KG
and Fr change relating
produces a
shovs
,
no retentiomi intake.
relation Hoi to intake
\
2 1
.
eight.
absolute per
correlation.
INTAKE
not again

increasing retention
positive
::
‘
0
I
kg. an
2
3 INTAKE PER KS
45
I
234567 WEIGHT
kg.
Finally,
relation (,f Imi this intake the var\case la’twceii nutrient with left
iii
Figure absolute intake
increasimig
show 5
interesting decreases retention
per kg.
I’I(;.
3.
A theoretical
ami(l al)solute versus
poilits
intake
imi(liVidUalS
a of
of
miutriemit (lifferelit
ti()tI
retemitiomi
Paradox % ith
kg.
in which increasing
increases with
retention w hereas
intake
per the
weights. ilor retemitiumi
iieither weight
absolute (),right so a group by a intake ) gives I)ividing
reten
that
of
intake with
( upper
no correlation.
This larger absolutely
kg.
results persons but relatively on an smaller
fromii ingest less less absolute
a on
situation more a in per the basis, Retention of
in a kg. is larger the
which substance basis absolutely persons. correlation
overlapping
Downloaded from www.ajcn.org by guest on June 6, 2011
weight correlation
for
each betweemi
person, rltellti()n
however, per kg.
produces and
positive per
thami
the
persons.
intake not vary
varies
with
weight.
Further,
they
amid (10 Thus,
with in the
each
respect upper when
divided
to the
each
other, portion and a
as of
can reten
be figure.
tions
seen
are
lefthand intakes
by the and
betv the
een
on
retention
a relative the they are
and
basis, examples presented
intake
the given to which of like data. intake, the by multiple
is
negative
is be attention result Actually, retention ex
while
l)OsitiVe.
correlation may call may
However, of the
betaeen
appropriate
Although treme, is to serious
weight
correlatiomi
person
intake
involved, from taking
positive unrelated
retention
milisinterpretations adjustment variables are analyzed involved,
produced.
data of nal to for values two the
correlation
from uncritical two variables and dividing each pair when three by a third variable. Then the origiand weight variables are likely to show a positive should be when each is expressed as a ratio technics. third. The greater the variation in
This
arises
data
usually regression
RETENTION
RETENTION RETENTION
5$ 5
.
RETENTION
.
(0 INTAKE
5
‘4
2, WEIGHT
#{149} 7
2, S S 6 7
0 INTAKE
It
4 WEIGHT
RETENTION ,ssKs
:
57*53
‘0
S
I I
I K 3 WEIGHT 4 5 S 7
INTAKE
:
‘0
5 S
RETENTION
‘K, Ks
0 I
‘ 5
‘ 3 4 5
S
I
‘
INTAKE
0
I
.
‘
‘ 4 5
K
S
3 INTAKE
2
3
4
5
6
7
WEIGHT
FIG. to
4.
Theoretical
balance retentiomi example clues
data. intake not.
This
is comparable
FIG.
Figure
3 but imi this
whereas
weight
increases with There is a positive
5. is
Theoretical correlation made positive
balance between when
data. absolute both
Imi this are
example and
a
negative
retention expressed ‘cor‘
correlation
expressed
between
per kg.
these
variables
unIv
if they
are
both intake rected”
for
weight.
252
REFERENCES
Weil 4. \V.
FOMON,
S. J.
1.
WALLACE,
The upon In:
4, 2. p.
WALLAcE,
B., JR. and TAYLOR, A. effect of variable protein and mineral intake the body comnposition of the growing animal. Ciba Foundation Colloquia on Ageing, vol.
W. M.,
WElL,
bolic body
1962.
and OWEN, G. M. Comment balance studies as a method composition of infants. Pediatrics,
on
meta
of estimating 29: 495,
and relation perto spu
5.
TANNER,
J.
area
M,
Fallacy
standards J Appl.
of and
perweight their Physiol.,
116,
Boston,
1958.
Little,
Brown
&
Co.
surface rious
and
Fed. 3.
WALKER,
its
Proc.,
W. M. relation A. R. P.
Nitrogen to retention
1959.
content and
in the
of the body loss nitrogen. of
interpretabalance studies.
correlations. L. L.,
2: 1, 1949.
6.
HILL, HETRIcK,
18: 1125, validity
Uncertainties
of longterm 10:95,
Effect
studies
WElL, W. B., JR., BENNETT, M., M. , MILLER, I. and WALLACE, W. M. of prior protein intake on stressed rats:
tion
Am.
and J. Clin.
of
survival
and
body
composition.
In
Nutrition,
1962.
press.
Downloaded from www.ajcn.org by guest on June 6, 2011