ArticlePDF Available

Energy content of diets of variable amino acid composition

Authors:

Abstract and Figures

Variation in the distribution of dietary nitrogen among the different amino acids is one factor that can modify the calorie equivalent per weight of amino acid or protein. This is important to consider when experimental diets with different amino acid compositions are compared and when indirect calorimetry is used to determine substrate oxidation rates. We developed a computer program to compute the energy content, oxygen equivalent, and respiratory quotient for arbitrary mixtures of amino acids and representative carbohydrates and fats. The calorie content of individual free amino acids was calculated by correction of the heat of combustion for the incomplete oxidation of amino acids characteristic of humans. Although these computations were presented before, we not report the limit of applicability of published values and the availability of the computer program to do these calculations.
Content may be subject to copyright.
770 Am iC/in Nutr l990;52:770-6. Printed in USA. © 1990 American Society for Clinical Nutrition
Original Research Communications--general
Energy content of diets of variable amino acid composition13
Michael E May and James 0 Hill
ABSTRACT variation in the distribution ofdietary nitro-
gen among the different amino acids is one factor that can mod-
ify the calorie equivalent per weight of amino acid or protein.
This is important to consider when experimental diets with
different amino acid compositions are compared and when in-
direct calorimetry is used to determine substrate oxidation
rates. We developed a computer program to compute the en-
ergy content, oxygen equivalent, and respiratory quotient for
arbitrary mixtures of amino acids and representative carbohy-
drates and fats. The calorie content of individual free amino
acids was calculated by correction of the heat of combustion
for the incomplete oxidation of amino acids characteristic of
humans. Although these computations were presented before,
we now report the limit ofapplicability ofpublished values and
the availability of the computer program to do these
calculations. Am J C/in Nutr 1990;52:770-6.
KEY WORDS Amino acids, branched-chain amino acids,
energy metabolism, indirect calorimetry
Introduction
The energy content ofa foodstuffis determined by multiple
factors including the efficiency of intestinal absorption, the
fraction ofnutrient or its metabolites lost in the urine, the frac-
tion of loss into feces of intestinal secretion induced by inges-
tion ofthe foodstuff, and the biochemical pathways of metabo-
lism. The energy content of absorbed nutrients is the heat of
combustion minus the heat of combustion of products of me-
tabolism that are not fully oxidized. These principles are well
known, and the usual factors for caloric content of foods have
traditionally been determined from the measured heat of corn-
bustion minus the heat of combustion of fecal material and
urine (1). Simple sugars, such as glucose, and short-chain fatty
acids, such as acetic acid, are oxidized nearly completely to car-
bon dioxide and water after absorption, and the energy content
is the heat ofcombustion. Amino acids are oxidized to carbon
dioxide, water, and several urinary metabolites containing ni-
trogen; these latter compounds account for the difference be-
tween biological energy content of protein and the heat of com-
bustion of protein. It seems intuitive that the biological energy
content will be greatest when the weight proportion of nitrogen
is least, and thus each amino acid may have a different caloric
value.
The purpose of this article is to review the chemical and
arithmetic bases for computation of the energy value of defined
food chemicals, to outline the approach to computerization of
these calculations, and to apply the computerized calculations
to a variety ofmixtures ofamino acids and other foods.
Methods
The heats of combustion and heats of formation of amino
acids were obtained from standard references (2, 3). Heats of
combustion were calculated for lysine, ornithine, proline, and
histidine from the heats of formation of amino acid, carbon
dioxide, and water (see Appendix for step-by-step description
of calculations). The biologic energy content was obtained by
subtraction ofthe residual heat ofcombustion contained in bi-
ologic end products from the heat of combustion. The major
nitrogenous excretory products were assumed to be urea, creat-
mine, uric acid, and ammonia. These products were chosen
because of their abundance in urine of normal adults (4). Be-
cause the heats of combustion of these compounds per mole
nitrogen are different, the energy content of urinary nitrogen-
containing compounds depends on the distribution of the ni-
trogen among these compounds. The sulfur-containing amino
acids yield a small amount of urinary taurine along with urea,
ammonia, creatinine, and uric acid and produce additional
metabolic energy by the oxidation ofthe sulfur to sulfate. These
computations were summarized recently (5), but the tedious
task of computing the urinary metabolite energy content for
all naturally occurring amino acids was simplified in the past
by assuming that urine nitrogen is in only one form, such as
urea (5).
We have written a microcomputer program, METENERG
(ME May, JO Hill, vanderbilt University, Nashville, TN) that
computes the metabolizable energy ofthe amino acids after the
user enters the distribution ofurine nitrogen either as a fraction
or as weight of compound excreted per day. Because the uri-
nary nitrogenous metabolites contain carbon, hydrogen, and
oxygen in addition to nitrogen, the volume of oxygen con-
sumed in the catabolism ofamino acid sufficient to produce 1
g ofurine nitrogen (O:N), the metabolic energy equivalent per
liter of gaseous oxygen used in oxidation of an amino acid
IFrom the Departments of Medicine and Pediatrics, Vanderbilt
University, Nashville, TN.
2Supported in part by a pilot and feasibility grant from the Diabetes
Research and Training Center, NIH 5 P60 DK20593-lO, and by NIH
grants DK38088 and DK26657.
Address reprint requests to ME May, AS 105 MCN Vanderbilt Uni-
versity, Nashville, TN 37232-2230.
Received September 5. 1989.
Accepted for publication December 27, 1989.
by guest on January 4, 2012www.ajcn.orgDownloaded from
CALORIE CONTENT OF FREE AMINO ACIDS 771
O:N =X.gasconstant/14.0067.n5
(H02), and the ratio of gaseous carbon dioxide produced to
oxygen consumed (RQ) also depend on the distribution of
urine nitrogen among the possible metabolites. METENERG
computes these quantities for each amino acid after the urine-
nitrogen distribution is specified according to the following
computational algorithm:
1) The heat of combustion of urinary nitrogenous corn-
pounds is computed per mole urine nitrogen by the expression
: (F1,).(HJn1)
where F, is the fraction of urine nitrogen in compound i, H,
is the heat ofcombustion ofcompound i, and n is the number
of moles nitrogen per mole i; i represents any urinary corn-
pound containing nitrogen derived from amino acids (includ-
ing amino acids and proteins).
2) The maximum metabolizable energy ofeach amino acid
is computed as heat ofcombustion, amino acid, minus residual
heat in urinary nitrogenous compounds.
3) The number ofcarbons (ca), hydrogens (he), and oxygens
(on) in the urine per mole nitrogen are derived from analogous
formulas:
c =  (F1,) .(c1/n1)
h =>(F1,)-(h1/n1)
ou =>:(F1,)-(o1/n1)
where c, h, and o1 are the numbers ofmoles ofcarbon, hydro-
gen, and oxygen per mole of nitrogen in compound i, respec-
tively. In the calculation ofthe correction for energy content of
amino acids, only nitrogenous compounds in the urine will be
included rather than all urinary compounds.
4) The chemical reaction for biologic oxidation of an amino
acid can be written in general terms as
CcaHhaOoaNna + X02  CCUHhUOOUN + YCO2 + ZH2O
The molecular compositions of the natural amino acids are
known and the apparent molecular composition of the urinary
nitrogenous compounds depends on the distribution of urine
nitrogen as detailed in reference 3. The stoichiometric coeffi-
cients are computed as follows:
Y=ca -c#{149}na
Z=(ha -hufla)/2
X(2Y+Z+Oufla)/2
where Y is the CO2 production in mol/mol amino acid and X
is the 02 utilization in mol/mol amino acid.
5) The RQ for a diet is the ratio of carbon dioxide produced
to oxygen consumed in complete oxidation of the diet after
complete absorption:
RQ =Z/X
6) The heat equivalent of oxygen (kcal/L 02) is
Heq =(Metabolizable energy)/(X .gas constant)
7) The ratio of oxygen utilization to urine nitrogen produc-
tion (L 02:g N) is
8) Sulfur amino acids require additional algebraic terms to
account for energy release and oxygen utilization in oxidation
ofsulfur to sulfate and the small change in apparent molecular
composition ofunnary nitrogenous compounds occasioned by
the excretion of taurine.
The above algorithm is not original (5), but the algebraic for-
mulation of the chemical balance allows the computerization
of the calculations, with easy extension of the calculations to
any specified distribution of urinary nitrogenous compounds.
The ionization fraction of organic acids is required to obtain
correct RQ values (5), and the program includes input of the
ionization fraction and the RQ correction for five common
short-chain carboxylic acids.
METENERG was written in Turbo Basic 1.0 and compiled
for use on an IBM-PC#{174}--compatible microcomputer under
MSDOS. It requires a minimum of 128 kb RAM memory and
one disk drive. Printer output uses ASCII codes only and
should be compatible with most common printers. The pro-
gram files are METENERG.EXE, MOLECULE.FW. MOL-
ECULE.MW, AAFORWT.BAT, and AAMOLWT.BAT. The
two DOS batch files allow selection ofeither molecular weights
or acyl residue formula weights for the parameter file MOL-
ECULE.CON. The program searches only the default directory
for the latter file and puts output files in the same directory.
Report files are named RQDI??KA.OUT. The 7? in the name
of the report file is 01 for the first report and then increases
to 99 as additional reports are generated. This feature allows
multiple reports to be generated in one run of the program.
When the report counter reaches 99, all subsequent report files
are named RQDIO1KA.OUT until the report files are cleared.
It is the responsibility of the user to delete old report files by
using appropriate DOS commands.
The program is organized as follows: The first screen displays
the name, copyright notice, and an option to exit to DOS. The
initialization sections read the chemical formulas and heats of
combustion of amino acids and other substrates, and the de-
fault distribution of urine nitrogen from the file MOLE-
CULE.CON. All numbers are stored as double-precision reals
(8 bytes, 14-digit precision) to minimize rounding errors. The
main menu allows four choices or exit: enter fractional distri-
bution of urinary nitrogen, enter mass of urinary metabolites
with nitrogen, calculate amino acid metabolizable energy and
O:Nu, and calculate energy parameters ofa food mixture.
Program flow is directed by a SELECT CASE <choice) state-
ment, which forces return to the main menu before to exit from
the program. There are two functions and 32 procedures that
do the following operations: screen manipulation, 1 1 routines
to clear specified blocks, print messages, format numbers as
strings for display, and print successive screens; keyboard input
routines, 7 routines for entry ofspecified quantities; calculation
routines, 10 routines, each for a specific portion ofthe compu-
tations; output routines, 4 routines; and program flow, main
program plus 2 routines comprising groups ofthe input, calcu-
lation, and output routines.
The substrates for diets included in MOLECULE.CON are
22 amino acids, 5carbohydrates, 3 polyols, 5short-chain or-
ganic acids, ethanol, and 4 fats. Thermodynamic properties for
the latter 18 substrates were obtained from published literature
(2). The program is supplied with two parameter files; one con-
tains compositions and molecular weights of amino acids
(MOLECULE.MW) and the other contains compositions and
by guest on January 4, 2012www.ajcn.orgDownloaded from
772 MAY AND HILL
TABLE 1
Heats ofcombustion and biologic oxidation ofamino acids
Compound Heat ofcombustion
Metabolizable energy5
Set 1 Set 2 Set 3 Set 4
kca//g kca//g(kJ/g)
Alanine 4.34 1 3.425 ( 13.70) 3.549 ( 14.20) 3.494 ( 13.98) 4.369 (I8.28)
Arginine S. 129 3.254 ( 13.02) 3.508 ( 14.03) 3.396 ( 13.58) 3.784 (I 5.83)
Asparagine 3.488 2.252 (9.0 1) 2.4 19 (9.68) 2.345 (9.38) 2.7 18 ( 11.37)
Aspartic acid 2.875 2.26 1 (9.04) 2.344 (9.38) 2.307 (9.23) 2.689 ( 1 1.25)
Cysteine 3.256 4.3 12 ( 17.24) 4.402 ( 17.6 1)4.362 (17.44) 4.792 (20.05)
Cystine 3.0 15 4.093 ( 16.37) 4. 182 ( 16.73) 4. 143 ( 16.57) -
Glutamic acid 3.646 3.09 1 ( 12.36) 3. 166 ( 12.66) 3. 133 (12.53) 3.583 (14.99)
Glutamine 4.207 3.089 (12.36) 3.241 (12.96) 3.173 (12.69) 3.61 1 (15.11)
Glycine 3.097 2.01 1 (8.04) 2.158 (8.63) 2.093 (8.37) 2.741 (11.47)
Histidine 4.85 1 3.273 ( 13.09) 3.487 ( 13.94) 3.392 ( 13.57) 4.2 16 ( I 7.64)
Isoleucine 6.523 5.902 (23.6 1) 5.986 (23.94) 5.949 (23.80) 6.886(28.81)
Leucine 6.524 5.903 (23.6 1) 5.988 (23.95) 5.950 (23.80) 6.888 (28.82)
Lysine 6.038 4.92 1 (19.68) 5.072 (20.29) 5.005 (20.02) 5.686 (23.79)
Methionine 4.456 5.3 13 (2 1.25) 5.385 (2 1.54) 5.353 (2 1.4 1) 6.06 (25.34)
Ornithine 5.493 4.259(17.04) 4.426(17.71) 4.352(17.41) -
Phenylalanine 6.723 6.229 (24.92) 6.296 (25. 18) 6.266 (25.06) 7.036 (29.44)
Proline 5.681 4.970(19.88) 5.066(20.27) 5.024(20.10) 5.937(24.84)
Serine 3.308 2.532 ( 10. 13) 2.637 ( 10.54) 2.590 ( 10.36) 3. 107 (13.00)
Threonine 4.120 3.434(13.74) 3.527(14.11) 3.486(13.94) 4.101 (17.16)
Tryptophan 6.588 5.787(23.14) 5.896(23.58) 5.848(23.39) 6.419(26.86)
Tyrosine 5.859 5.409 (2 1.64) 5.470 (2 1.88) 5.443 (2 1.77) 6.039 (25.27)
Valine 5.963 5.264 (2 1.06) 5.358 (2 1.43) 5.3 16 (2 1.27) 6.276(26.26)
I 4.781 4.136(16.54) 4.253(17.01) 4.201 (16.80) 4.847 (20.28)
SD 1.324 1.342(5.370) 1.321 (5.283) 1.339(5.321) 1.530(6.402)
CV(%) 27.7 32.4 31.1 31.7
aFor four distributions ofurine nitrogen: set 1-90% urea, 3% ammonia, 5% creatinine, 2% uric acid; set 2-100% ammonia; set 3-l00% urea;
and set 4-data from Table 3 ofreference (2) for 100% urea.
formula weights ofthe acyl portion (dehydrated) ofthe amino
acids (MOLECULE.FW). Selection of parameter files is ac-
complished by the DOS batch files. The program as written is
appropriate for diets composed of modular components (pro-.
tein and/or amino acids, carbohydrate, fats), and extension to
natural diets would require modification or assumption that
the nonprotein nitrogen was negligible in quantity. The corn-
putations strictly apply to the portion of the diet that is ab-
sorbed or taken parenterally; digestibility is not specifically en-
tered in the program. Because the user is prompted for the diet
components by weight, known digestibility factors (amount ab-
sorbed/amount eaten) can be accounted for by entering Weight
xDigestibility instead ofWeight for each nutritive component
and entering  Weight X (1-Digestibility) plus the weight of
dietary minerals as Mineral for the program-this allows cor-
rect energy computations per weight of the original diet. (The
compiled program is available from the authors for $25 (US)
to cover reproduction and shipping.)
Results and discussion
Table 1 shows the heats of combustion and metabolizable
energy for amino acids. Metabolizable energy is presented for
four distributions of urine nitrogen. The right-hand column of
Table 1 is the metabolizable energy reported by Livesey and
Elia (5) under the assumption that urea is the sole urinary end
product. Minor differences would be expected from the use of
slightly different primary sources, but the large differences in
our calculations and those previously reported was bother-
some. We replicated the numbers of Livesey and Elia by use of
the compositions ofthe acyl portion for computation of nitro-
gen content and formula weights rather than the molecular
composition for each ofthe amino acids. That is, the previously
reported parameter values refer to amino acids in protein when
the amino acid data has been expressed as
Amount ofamino acid =fraction ofprotein nitrogen
in amino acid X acyl formula weight per mole nitrogen
Ifamino acid composition is expressed in these terms, the sum
of amino acid weights will be equivalent to the protein weight
and we can call this mode of expression protein equivalent.
Actually, the sum ofamino acid weights obtained by hydrolysis
of protein will be greater than the weight of the initial protein
by the amount ofwater added to hydrolyze the peptide bonds.
There is no convention stated for expression of amino acid
data in standard tables of food composition (6). Furthermore,
the protein-equivalent mode ofexpression ofamino acid corn-
position ofproteins is not applicable to diets composed of crys-
talline amino acids, as are often used in experiments and in
parenteral nutrition. The previously published values for me-
tabolizable energy of amino acids do not apply to diets corn-
by guest on January 4, 2012www.ajcn.orgDownloaded from
CALORIE CONTENT OF FREE AMINO ACIDS 773
SSets described in Table 1.
TABLE 2
Heat equivalent ofoxygen for amino acids
Amino acid
Heat equivalen tofoxygen consum ed in oxidation5
Set 1 Set 2 Set 3
kcal/L (ki/L)
Alanine
Arginine
Asparagine
Aspartic acid
Cysteine
Cystine
Glutamic acid
Glutamine
Glycine
Histidine
Isoleucine
Leucine
Lysine
Methionine
Ornithine
Phenylalanine
Proline
Serine
Threonine
Tryptophan
Tyrosine
Valine
j:
SD
CV (%)
4.627 (I 8.5 1 )
4.799 ( I 9. 19)
4.600 (1 8.40)
4.563 ( 18.25)
5.318 (21.27)
5.308 (2 1.23)
4.567 ( 18.27)
4.593 ( 18.37)
3.979 ( 1 5.9 1)
4.693 ( 18.77)
4.64 1 ( 18.56)
4.642 (18.57)
4.662 ( 18.65)
4.790 ( 19. 16)
4.663 ( 18.65)
4.6 17 ( 18.47)
4.69 1 ( 18.76)
4.859 ( 19.44)
4.629 ( 18.52)
4.631 (18.53)
4.63 1 (18.52)
4.630(18.52)
4.688 ( 18.75)
0.259 ( 1.04)
5.5
4.703 ( 18.8 1)
4.957 ( 19.83)
4.753 (19.0 1)
4.640 ( 18.56)
5.359 (21.44)
5.350 (2 1 .40)
4.6 18 ( 18.47)
4.696 ( 18.78)
4. 1 30 ( 16.52)
4.827 ( 19.3 1)
4.67 1 ( 18.68)
4.672 (18.69)
4.726 ( 18.90)
4.8 18 ( 19.27)
4.745 ( 18.98)
4.640 ( 18.56)
4.732 ( 18.93)
4.945 ( 19.78)
4.686 ( 18.74)
4.671 (18.68)
4.655 ( 18.62)
4.668(18.67)
4.757 ( 19.03)
0.248 (0.99)
5.2
4.630(18.52)
4.798 (19.19)
4.607(18.43)
4.567 (18.27)
5.31 1 (21.24)
.300 (21.20)
4.70(18.28)
4.98(18.39)
4.005 (16.02)
4.696(18.79)
4.642 (18.57)
4.643 (18.57)
4.664 ( I 8.65)
4.790(19.16)
4.666(18.66)
4.618(18.47)
4.692 (18.77)
4.858 (19.43)
4.63 1 (18.52)
4.633(18.53)
4.632 (18.S3)
4.631 (18.52)
4.690 ( I 8.76)
0.24(1.01)
5.4
SSets described in Table 1.
prising crystalline amino acids. We prepared two parameter
files containing either molecular weights ofamino acids or acyl
residue formula weights for amino acids, for use with actual
amino acid weights or protein-equivalent weights of amino
acids, respectively. There is over twofold variation in the values
for individual amino acids, and the essential amino acids have
higher metabolizable energy on average than do the nonessen-
tial amino acids.
Tables 2-4 present the values for free amino acids rather
than protein-equivalent amino acids. The heat equivalent of
oxygen varies little among the amino acids (CV <6%, Table 2),
but large variations among amino acids in the values of oxygen
equivalent of urine nitrogen (CV 57%) and respiratory coeffi-
cient (CV 18%) are shown in Tables 3 and 4. As demonstrated
in Table 3, the oxygen equivalent of urine nitrogen is the same
whether the urinary metabolite is urea or ammonia; this calcu-
lated result was verified as a general rule for any monoamino
monocarboxylic acid. The mixture of multiple amino acids
and addition of other caloric sources will tend to minimize
variation in energy content of natural diets, especially because
amino acids usually contribute only 10-20% of the energy in
normal human diets.
Between the time of first review and publication of this
manuscript, an Erratum was published stating that the metabo-
lizable energy density of individual amino acids published in
reference 5of this paper was expressed in units of energy per
g protein equivalent and not per g free amino acid (7). This
communication is in agreement with the findings ofthis paper.
These data are ofinterest for two reasons: the design of isoca-
loric, isonitrogenous research and therapeutic diets requires
these computations ifthe distribution ofamino acids is atypi-
cal, and the measurement ofenergy expenditure and substrate
oxidation rates by indirect calorimetry or measurement of en-
ergy expenditure in free-living subjects by doubly labeled water
require proper values of 0:N, H02, and RQ for accuracy.
Table 5 shows the effect ofapplying these values to the compu-
tation ofthe biologic energy content ofseveral amino acid mix-
tures. variation in amino acid composition results in different
energy contents despite similar weight percentages of total
amino acids. For example, solutions enriched in the branched-
chain amino acids have a higher caloric value than do formula-
tions enriched in alanine and glycine despite similar or lower
weight percentages oftotal amino acids. The essential or indis-
pensable amino acids have not only a high metabolizable en-
ergy but also a high oxygen equivalent ofurine nitrogen. Table
5 includes several enteral products that contain carbohydrate
in addition to amino acids, and the calculated RQ is corre-
spondingly higher for those products. O:N directly reflects the
amino acid composition ofeach diet.
It can be predicted that it would be difficult to balance both
calories and nitrogen among test diets of a few amino acids
chosen from the extremes ofcaloric density. The most extreme
TABLE 3
Oxygen equivalent ofurine nitrogen
Amino acid
Oxygen used in production
ofurine nitrogen5
Set 1 Set 2 Set 3
L 02/g N
Alanine 4.709 4.801 4.801
Arginine 2.108 2.200 2.200
Asparagine 2.308 2.400 2.400
Aspartic acid 4.709 4.801 4.801
Cysteine 7.015 7.105 7.105
Cystine 6.615 6.705 6.705
Glutamicacid 7.109 7.201 7.201
Glutamine 3.509 3.600 3.600
Glycine 2.709 2.800 2.800
Histidine 2.575 2.667 2.667
Isoleucine I 1.910 12.002 12.002
Leucine 11.910 12.002 12.002
Lysine 5.509 5.601 5.601
Methionine 11.816 11.906 11.906
Ornithine 4.309 4.401 4.401
Phenylalanine 15.9 10 16.002 16.002
Proline 8.709 8.801 8.801
Serine 3.909 4.001 4.001
Threonine 6.309 6.401 6.401
Tryptophan 9. 109 9.20 1 9.201
Tyrosine 15. 1 10 15.202 15.202
Valine 9.509 9.601 9.601
I 7.154 7.245 7.245
SD 4.117 4.117 4.117
CV (%) 57.5 56.8 56.8
by guest on January 4, 2012www.ajcn.orgDownloaded from
774 MAY AND HILL
TABLE 4
Respiratory quotient for amino acids
Aminoacid
CO2 produced per oxygen
consumed in catabolism5
Setl Set2 Set3
L/L
Alanine 0.835 1.000 0.833
Arginine 0.727 1.091 0.727
Asparagine 1.011 1.333 1.000
Asparticacid 1.175 1.333 1.167
Cysteine 0.554 0.667 0.556
Cystine 0.588 0.706 0.589
Glutamic acid 1.004 1 .1 1 1 1.000
Glutamine 0.893 1.1 1 1 0.889
Glycine 0.862 1.143 0.857
Histidine 0.906 1.200 0.900
Isoleucine 0.733 0.800 0.733
Leucine 0.733 0.800 0.733
Lysine 0.714 0.857 0.714
Methionine 0.600 0.667 0.601
Ornithine 0.727 0.909 0.727
Phenylalanine 0.85 1 0.900 0.850
Proline 0.819 0.909 0.818
Serine 1.006 1.200 1.000
Threonine 0.877 1.000 0.875
Tryptophan 0.871 0.957 0.870
Tyrosine 0.896 0.947 0.895
Valine 0.750 0.833 0.750
i 0.824 0.976 0.822
SD 0.151 0.196 0.148
CV(%) 18.3 20.1 18.0
5Sets described in Table I.
imbalance arises for comparisons ofa single amino acid among
diets. This is of great interest because the branched-chain
amino acids (leucine, isoleucine, and valine) may have benefi-
cial roles in the modulation of proteolysis in severe catabolic
states (9, 10), and the energy content of the branched-chain
amino acids is high. For example, the effects of five solutions
on nitrogen balance were studied in traumatized rats (1 1): each
solution contained (per L) 50 g glucose and, in addition, the
first had 15 g leucine, the second had 15 g isoleucine, the third
had 15 g valine, the fourth had 15 g alanine, and the fifth had
an additional I 5g glucose. The diets were thought to be bal-
anced for energy (assuming all amino acids and glucose had a
biologic energy content of4 kcal/g), but the actual energy con-
tents of the five solutions were 275, 275, 266, 238, and 242
kcal/L, respectively. The improvement in nitrogen balance
seen with the leucine solution compared with the alanine or
dextrose solutions was as likely to have been because ofthe 14%
greater energy content rather than special properties of leucine
(10). METENERG will allow such computations to be done
easily for many other special diets and will allow those compu-
tations to be done for different distributions ofnitrogen among
urinary metabolites to allow customization for genetic or
pathophysiologic states in which urine nitrogen is shunted
away from urea.
The proper interpretation ofthe measurement of respiratory
gases depends on several parameters computed by ME TEN-
ERG. The precision ofthe heat equivalent ofoxygen (Table 2)
supports its use as a measure ofcaloric expenditure. However,
the partitioning of total oxygen consumption into carbohy-
drate and fat oxidation requires correction of the oxygen utili-
zation by subtracting urine nitrogen X O:N .As seen in Table
3, O:N is highly variable among the natural amino acids. A
problematic but critical question is whether the mix of amino
acids being oxidized is the same under all conditions or among
all subjects in a research study. Computation of a correction
for urine nitrogen from the dietary amino acid composition
implies the assumption that urine nitrogen is derived from
amino acids in the same relative proportions as in the diet. This
assumption may be true in a steady state in adult humans or
nongrowing animals but is only one ofseveral possible assump-
tions in states of nitrogen accretion (growth) or negative nitro-
gen balance (starvation, postinjury or postoperative catabo-
lism). Other assumptions, such as that the mix of amino acids
being oxidized is distributed the same as the amino acid distri-
TABLES
Calorimetric variables ofsome amino acid mixtures5
Mixture ME RQ Heq 02:N
kcal/L
Travasol 8.5%t 495 0.776 4.669 5.126
AminoSyn 8.S%t 556 0.772 4.687 5.769
HepatAmine 483 0.772 4.689 6.151
FreAmine III 8.S% 344 0.780 4.697 5.876
Nephramine S.4% 282 0.744 4.708 9.372
TrophAmine6% 266 0.780 4.720 6.212
Indispensable-8.5% fl 442 0.787 4.705 8.460
Dispensable-8.S%i 27 1 0.864 4.800 4.054
RDA mixture-6.36%55 341 0.761 4.715 9.698
Nutrisource Aminoacidtt 4435ff 0.807 4.694 6.885
Nutrisource AA-HBCtt 47 18  0.780 4.706 7.206
HepaticAidII 1077 0.874 4.864 6.740
TravasorbRenalt 1170 0.945 4.921 6.471
StandardVivonex 764 0.977 4.950 5.368
VivonexHighN 758 0.961 4.920 5.636
Amin-Aid 1769 0.924 4.950 9.534
Traum-Aid HBC 997 0.909 4.917 7.633
TravasorbHepatict 1108 0.937 4.964 6.723
Caseclill 4605ff 0.822 5.397 6.544
Traum-Aid HN 949 0.925 4.912 7.0 14
Isocallill 1073 0.870 4.861 6.755
Ensure11J 1064 0.882 4.878 6.503
Vital HN1 41 16ff 0.940 4.943 6.077
SValues calculated for urine nitrogen: 90% urea, 3% ammonia, 5%
creatinine, 2% uric acid.
tTravenol, Inc, Deerfield, IL.
:t Abbott Laboratories, Inc. North Chicago, IL.
§American McGaw, Inc, Santa Ana, CA.
II His, lie, Leu, Lys, Met, Phe, Thr, Trp, Tyr, Val: 8.5 g/L each.
#{182}Ala, Arg, Asn, Asp, Cys, Glu, GIn, Gly, Pro, Ser: 8.5 g/L each.
55 Recommended dietary allowance (8). 3.49 g Cys, 8.38 g Ile, 1 1.17
g Leu, 8.38 g Lys, 3.49 g Met, 5.59 g Phe, 5.59 g Thr, 2. 10 g Trp, 5.59 g
Tyr, and 9.78 g Val in 1 L solution.
tt Sandoz Nutrition, Minneapolis.
if Dry powder (kcal/kg).
§ Norwich Eaton, Inc. Norwich, NY.
II II Mend Johnson Nutritionals, Evansville, IN.
#{182}11Ross Laboratories, Columbus, OH.
by guest on January 4, 2012www.ajcn.orgDownloaded from
tion
CALORIE CONTENT OF FREE AMINO ACIDS 775
6C+7H2+02+N2-.+6C02+7H2O+N2
bution ofwhole body protein or that dispensable amino acids
are oxidized in preference to indispensable amino acids, would
result in computation of different values for 0:N .Errors in
estimating protein oxidation from end products ofamino acid
metabolism can affect the estimates ofrates ofoxidation of car-
bohydrate and fat and of total energy expenditure obtained
from indirect calorimetry (5). Determination ofthe oxidation
rates of multiple amino acids simultaneously is technically
difficult because of the limited number of tracers for the corn-
mon end product carbon dioxide. METENERG allows corn-
putation ofbounds ofvalues ofO:N and Heq for differing as-
sumed proportions ofamino acids.
An alternative technique for measurement of energy expen-
diture in living subjects is determination ofthe decay of specific
activity ofdoubly labeled water, which estimates carbon diox-
ide production from the difference in dilution ofenrichment in
water of deuteriurn and oxygen- 18. The difference in dilution
curves results from the distribution ofoxygen-l 8 into the bicar-
bonate pool in addition to the water pool labeled by the deute-
num. Oxygen consumption is then estimated by assuming that
the subject’s RQ equals the RQ of the subject’s usual diet. Er-
rors in estimation of the diet RQ can lead to errors in estima-
tion ofoxygen consumption and thus to error in estimation of
total energy expenditure. As detailed in Table 4, there is wide
variation in RQ among amino acids as well as the known varia-
tion in RQ among fats and carbohydrates (5). The estimation
of energy expenditure by doubly labeled water decay curves
also usually includes the assumption that the subjects are in a
net zero-energy balance during the period under study, which
may be considered part of the assumption that RQ equals the
calculated dietary RQ (12). The RQ calculated for the diet will
be appropriate for the individual subject only ifthe distribution
ofoxidized substrates is the same as the distribution of dietary
substrates, and this assumption would be untenable for sub-
jects altering weight by deposition or loss ofbody tissue during
the period of study.
Urine nitrogen can also reflect nucleic acid catabolism and
excretion of many non-amino acid nitrogenous compounds
contained in natural diets. Another major assumption of the
above discussion is that urinary nitrogenous compounds are
derived from amino acids. This assumption is valid whenever
the nitrogen content ofdietary amino acids is overwhelmingly
large compared with the nitrogen content of non-amino acid
compounds. Foods or experimental diets that contain large
amounts of non-amino acid nitrogen cannot be treated by the
equations or program described above nor can they be allowed
in studies employing indirect calorimetry for computation of
substrate oxidation rates.
Finally, we feel that proper design of amino acid-modified
research diets will require application of the individual energy
parameters rather than assumption of constant values for all
amino acids. We are making METENERG available for this
purpose. El
References
I. Widdowson EM. Note on the calculation of the energy value of
foods and of diets. In: Paul AA, Southgate DA, eds. The composi-
tion of foods. 4th ed. New York: Elsevier/North-Holland Biomedi-
cal Press, 1978:322-6.
2. Hutchens JO. Heat of combustion, enthalpy and free energy of for-
mation ofamino acids and related compounds. In: Sober HA, ed.
Handbook ofbiochemistry. 2nd ed. Cleveland: CRC Press, 1970:
B62-4.
3. Dean JA, ed. Lange’s handbook ofchemistry. 13th ed. New York:
McGraw-Hill, 1985.
4. Diem K, Lentner C, eds. Scientific tables. 7th ed. Ardsley: Ciba-
Geigy Corp 1975:663-72.
S. Livesey G, Elia M. Estimation ofenergy expenditure, net carbohy-
drate utilization, and net fat oxidation and synthesis by indirect
calorimetry: evaluation of errors with special reference to the de-
tailed composition offuels. Am J Clin Nutr l988;47:608-28.
6. Consumer and Food Economic Institute. Composition of foods:
dairy and egg products. Agricultural handbook no. 8-1. Washing-
ton, DC: USGovernment PnntingOffice, 1976.
7. Erratum. Am J Clin Nutr l989;50:l475.
8. National Research Council. Recommended dietary allowances.
10th ed. Washington, DC: National Academy Press, 1989.
9. Brennan MF, Cerra F, Daly JM, et al. Report ofa research work-
shop: branched chain amino acids in stress and injury. JPEN
1986; 10:446-52.
10. May ME, Buse MG. Effects ofbranched chain amino acids on pro-
tein turnover. Diabetes Metab Rev l989;S:227-45.
1 1. Freund H, Yoshimura N, Fischer JE. The role of alanine in the
nitrogen conserving quality ofthe branched chain amino acids in
the post injury state. J Surg Res 1980; 29:23-30.
12. Schoeller DA. Measurement of energy expenditure in free-living
humans by using doubly labeled water. J Nutr 1988; 1 18:1278-89.
APPENDIX
Sample computations for an individual amino acid
1) Urine nitrogen will be assumed to be 90% urea, 5% ammo-
nia, and 5% creatinine. The molecular formulas and heats
of combustion of these compounds are
Urea C1H4O1N2 l5l.Okcal/mol
Ammonia C0H3O0N1 70.6 kcal/mol
Creatinine C4H701N3 558.1 kcal/mol
2) The heat of combustion of urinary nitrogenous corn-
pounds is (0.9 x151.0/2) + (0.05 x70.6/1) + (0.05 X
558. 1/3) =80.78 kcal/(mol urine N).
3) The C, H, and 0 lost in urinary nitrogenous compounds
arec =(0.9 X l/2)+(0.05 X0/l)+(0.05 X4/3)= 0.5167
(mol/mol N).
h =(0.9 x4/2) + (0.05 x3/ 1 )+ (0.05 x7/3) =2.0667
(mol/mol N).
ou =(0.9 X 1/2) + (0.05 x0/1) + (0.05 x1/3) =0.4667
(mol/mol N).
4) The remaining calculations will be illustrated with lysine.
The molecular formula of lysine is C6 14 N202 .The heat
ofcombustion oflysine is not tabulated but can be calcu-
lated from the heat of formation. The heat offormation is
the enthalpy change for the reaction
6C+ 7 H2 +02 + N2-C6H14N202
The heat ofcombustion is the enthalpy change for the reac-
C6H 14N202 +02 #{248}6 CO2 + 7 H2O + N2
The summation reaction
by guest on January 4, 2012www.ajcn.orgDownloaded from
776 MAY AND HILL
has an enthalpy that is a sum of heats of formation of the
products, ie,
(Heat of formation)L, + (heat of combustion)L5
=6 X (heat of formation)co2 +7X (heat of formation)H20
Substituting the known values for heats of formation, we
get
(Heat of combustion)L, =[6 X94.38 + 7 X 68.38
-162.2] kcal/mol =882.7 kcal/mol
5) The metabolizable energy is the heat of combustion minus
the heat of combustion of urinary products, ie,
882.7 kcal/mol Lys -(80.78 kcal/mol N)
X (2 mol N/mol Lys) =721.14 kcal/mol Lys
Because food items are usually specified by weight, we con-
vert this result by division by the molecular weight of lysine
(146.19) to 4.933 kcal/g.
6) The CO2 produced by oxidation of lysine is
(6 -0.5167 X 2) mol/mol LYS =4.9667 mol/mol LYS
7) The H2O produced by oxidation of lysine is
(14 -2.0667 X 2)/2 mol/mol LYS =4.9333 mol/mol LYS
8) The oxygen consumed in catabolism of lysine is
(2 X 4.9667 + 4.9333 + 0.4667 x2
-2)/2 mol/mol LYS =6.9000 mol/mol LYS.
9) RQ =4.9667/6.9000 =0.720 L C02/L 02
10) Heq =721.14/6.9000 X 22.414 =43.663 kcal/L 02
11) O:N =6.900 X 22.414/14.0067 X 2 =5.521 LO2/gN
by guest on January 4, 2012www.ajcn.orgDownloaded from
... It is found that sodium chloride increases the solubility of AAs, so "the solubility of AAs is highly important in determining biological action of bio-macromolecules such as proteins" [3]. The combustion of free AAs inside human body gives (1.3-6.5 Kcal/g) as heat of combustion (enthalpy change for AAs oxidation) [4]. In another side "ascorbic acid (Vitamin C, AA) is one of the most important vitamins with antioxidant properties that widely exist in various fruits and vegetables and widely used in pharmaceutical formulations and cosmetic applications. ...
Conference Paper
Full-text available
Molar conductivity for hydrogen bonds ion association of ascorbic acid (AA) with Cysteine (Cys) and Methionine (Met) at concentration (2-5×10-3 , 2×10-3 and 2×10-3 M respectively) in water and in (2×10-3 M) sodium chloride (NaCl) solutions was examined at temperatures range (298 K to 313 K). The conductance equation of Shedlovsky was used to calculate the association constants (KA) and the limiting molar conductance (Λo). The association heat, the Gibbs free energy, the change of entropy and the activation energy (ΔH o , ΔG o , ΔS o , and ΔES respectively) also calculated. All results are computed. The data showed increasing of Λo, with increases in temperature and the same trend in decreasing for KA. The data showed negative values for ΔH o and ΔG o parameters. This refers to the exothermic association processes, and values increased with increasing in AA concentration, while ΔS o and ΔES parameters show a positive value related to the higher mobility of the ions and decreasing of ion association solvation in water and NaCl solutions. The ΔEs and ΔH o of a different solution in present sodium chloride show decreasing in values as compared with water solution at different concentration of AA, while the opposite trend values appear for ΔG o and ΔS o with some exception.
... The sulfur coefficient is lower than in either equation and is based on matching the trends in heats of combustion reported for amino acids. 5 Sulfur content is so low in food, however, that it has a negligible effect on the final result. Eq. 3 was also compared to the heat of combustion of twenty organic compounds of varying C, H, N, and O composition, ranging from normal alkanes to alcohols to amino acids, taken from the 63 rd CRC Handbook. ...
Technical Report
Full-text available
This report presents improved correlations for relating the heats of combustion and formation to the elemental composition, moisture content, and ash content. The correlations are also able to calculate heats of combustion of carbohydrates, proteins, and lipids individually, including how they depend on elemental composition. The starting point for these correlations are relationships commonly used to estimate the heat of combustion of fossil fuels, and they have been modified slightly to agree better with the ranges of chemical structures found in foodstuffs and biomass.
... Study of the chelating ability of glycine peptides (Gp), in particular diglycine (GG) and triglycine (GGG), with some divalent metal ions (Co 2? , Ni 2? and Cu 2? ) has shown that glycine peptides are potential metal chelators via their amine and carboxyl functional groups [11]. For GG and GGG, the occurrence of an amide bond in its backbone may also effect the stability of the formed complex. ...
Article
Full-text available
The stability of binary and mixed-ligand complexes among trivalent transition metal ions (chromium and iron), glycine peptides (glycylglycine and glycylglycylglycine) and phenolates (ferulic acid and gallic acid) were studied by using pH-potentiometric titration in aqueous solution at 298.15 K and ionic strength of 0.15 mol·dm-3 NaNO3. The complexation model for each system was obtained by processing the potentiometric titration data using the HYPERQUAD2008 program. The stability constant trend of complexes in both systems and the contributions of deprotonated or protonated amide peptides to the stability of the complexes is discussed. The stability of the mixed-ligand complexes relative to their corresponding binary complexes was also investigated by calculating the log10 K parameter of each system. In addition, the Gibbs energies of reaction (Δr G) obtained from the Gaussian modeling program with B3LYP/6-31+G(d) basis set were used to verify the contributing binding sites of the ligands and to predict the structures of the M-L complexes.
... In general, hunger is produced in animals by depriving them of carbohydrates. We estimate that the AA solutions provided 10% more calories than sucrose alone (May and Hill, 1990), so it is possible that this population was less hungry than those fed sucrose alone. However, our data may also indicate that the same neuronal mechanisms that regulate and respond to carbohydrates also respond to AAs. ...
Article
Full-text available
Obtaining the correct balance of nutrients requires that the brain integrates information about the body's nutritional state with sensory information from food to guide feeding behaviour. Learning is a mechanism that allows animals to identify cues associated with nutrients so that they can be located quickly when required. Feedback about nutritional state is essential for nutrient balancing and could influence learning. How specific this feedback is to individual nutrients has not often been examined. Here, we tested how the honeybee's nutritional state influenced the likelihood it would feed on and learn sucrose solutions containing single amino acids. Nutritional state was manipulated by pre-feeding bees with either 1 M sucrose or 1 M sucrose containing 100 mM of isoleucine, proline, phenylalanine, or methionine 24 h prior to olfactory conditioning of the proboscis extension response. We found that bees pre-fed sucrose solution consumed less of solutions containing amino acids and were also less likely to learn to associate amino acid solutions with odours. Unexpectedly, bees pre-fed solutions containing an amino acid were also less likely to learn to associate odours with sucrose the next day. Furthermore, they consumed more of and were more likely to learn when rewarded with an amino acid solution if they were pre-fed isoleucine and proline. Our data indicate that single amino acids at relatively high concentrations inhibit feeding on sucrose solutions containing them, and they can act as appetitive reinforcers during learning. Our data also suggest that select amino acids interact with mechanisms that signal nutritional sufficiency to reduce hunger. Based on these experiments, we predict that nutrient balancing for essential amino acids during learning requires integration of information about several amino acids experienced simultaneously.
... g. Lysine (Sigma) was supplemented as 1.8% solution, a concentration equicaloric with the leucine solution [38]. ...
Article
Full-text available
Brain pathways, including those in hypothalamus and nucleus of the solitary tract, influence food intake, nutrient preferences, metabolism and development of obesity in ways that often differ between males and females. Branched chain amino acids, including leucine, can suppress food intake, alter metabolism and change vulnerability to obesity. The SLC6A15 (v7-3) gene encodes a sodium-dependent transporter of leucine and other branched chain amino acids that is expressed by neurons in hypothalamus and nucleus of the solitary tract. We now report that SLC6A15 knockout attenuates leucine's abilities to reduce both: a) intake of normal chow and b) weight gain produced by access to a high fat diet in gender-selective fashions. We identify SNPs in the human SLC6A15 that are associated with body mass index and insulin resistance in males. These observations in mice and humans support a novel, gender-selective role for brain amino acid compartmentalization mediated by SLC6A15 in diet and obesity-associated phenotypes.
... We therefore used a BAL mixture of 90 g in which the quantities of all amino acids were reduced while maintaining equivalent proportions with respect to the TYR/ PHE-free mixture (see Table 1). We used the method of May and Hill [24] to confirm that the amino acid mixtures were of equivalent metabolisable energy content (TYR/PHE-free = 480 kcal; BAL = 487 kcal). The mixtures were prepared by suspending the amino acids in tap water. ...
... The concentration of essential amino acids in R+E food was 7-40 % higher than in R food, depending on the concentration of each amino acid in brewer's yeast (see Bass et al., 2007, Supplementary Table 2). Assuming the energy content of amino acids from May & Hill (1990), caloric content of R+M and R+E food was higher than that in the R food due to the addition of amino acids by 0 . 28% and by 3 . ...
Article
Full-text available
We measure genetic variation in lifespan and fecundity at two food levels in 34 core lines of the Drosophila Genetic Reference Panel collection. Lines were significantly different from each other in lifespan and fecundity at both restricted and full food. There was a strong food-by-line interaction for the slope of age-specific mortality, fecundity and proportion of fertilized eggs, indicating the presence of genetic variation for the strength of the dietary restriction effect, likely to represent standing genetic variation in a natural population from which the lines used have originated. No trade-off between fecundity and lifespan manifested in life-history variation among inbred lines. Our data partially corroborate the recent proposition that availability of nutrient-free water eliminates the apparent dietary restriction at least in some conditions. Although flies on full food with water added had lifespan slightly higher than those without a water source, it was still significantly lower than that in flies on restricted food, with no indication of interaction. We fully corroborate the recently discovered effect of addition of essential amino acids to the medium: addition of 1.5 mM methionine to restricted food significantly increased fecundity without a measurable decrease in lifespan; addition of each of 10 essential amino acids increased fecundity and decreased females lifespan to the levels observed on full food, again with no evidence of line-by-food interactions. We propose a mechanistic hypothesis explaining the observed data, based on the assumption that food consumption by flies is adjusted according to flies' saturation in water and methionine.
Article
Children with phenylketonuric (PKU) are at risk for fractures. This study used a PKU murine model (PAHenu‐2) to evaluate effects of moderate dietary protein restriction and elevated plasma phenylalanine concentration impact upon bone status. Fifty‐four male weanling PKU and control mice were assigned to either an elemental phenylalanine (Phe)‐restricted diet (treated) or Phe‐unrestricted diet (untreated) with low or normal protein levels for 56 days. Untreated mice and control mice received equal amounts of dietary Phe; treated mice consumed prescribed dietary Phe to maintain plasma Phe concentrations between 120 and 480 µmol/L. Plasma Phe, osteocalcin, and urine deoxypyridinoline (DPD)/creatinine were analysed at baseline and at days 28 and 56. Femur strength, bone mineral density (BMD) and bone mineral content (BMC) were analysed at day 56. Moderate protein restriction did not significantly affect bone status. Mean plasma Phe concentrations were significantly greater in untreated vs treated and control mice (p<0.0001). Total body weight was significantly less in untreated vs control mice (p<0.01). Mean femur weight was reduced in untreated mice vs both treated and control mice (p<0.03). Untreated mice had smaller mean femur length than control mice (p<0.002). Femur strength was greater in treated mice compared to control mice (p<0.01) but not compared to untreated mice. No significant difference among groups was found in BMD and BMC. At day 56 there was a statistical trend (p<0.056) towards higher urine DPD/creatinine excretion in untreated mice than in treated mice. Plasma Phe concentration was positively correlated with urine DPD/creatinine. These data suggested that hyperphenylalaninaemia may adversely affect bone status in PKU mice.
Chapter
Glycine is a nonessential small neutral amino acid containing 18.7% nitrogen. The nonessential amino acid glycine (Gly) is an inhibitory neurotransmitter. It is needed for the synthesis of peptides and proteins, creatine, glutathione, porphyrins, and purines, and for the conjugation of bile acids and xenobiotics. Gly breakdown requires thiamin, riboflavin, niacin, vitamin B6, folate, vitamin B12, pantothenate, lipoate, ubiquinone, iron, and magnesium. Adequate amounts are consumed when total protein intakes meet recommendations. Dietary supplements containing crystalline Gly are commercially available. With adequate total protein intake, enough Gly is available directly and from the conversion of serine and threonine. A small percentage is converted to oxalate that may increase the risk of kidney stones. Protein in food is denatured by gastric acid and the action of gastric, pancreatic, and enteric enzymes, many of them cleaving peptide bonds between specific amino acids and Gly. Filtered free Gly is taken up into proximal renal tubules mainly by the sodium-amino acid cotransport system B° , di- and tripeptides via PepT1 and PepT2. Ala is then exported across the basolateral membrane via the sodium-dependent systems N (SN2), A (ATA2), and ASC (ASCT1). As a result of very efficient reabsorption, the loss of Gly into urine is minimal in healthy people. Losses into feces are negligible while gastrointestinal function is normal. Nitrogen from metabolized Gly is excreted into urine as urea or as uric acid. Nearly a quarter of the nitrogen in uric acid comes from Gly.
Chapter
Acidosis is common in patients with AKI in the ICU and often associated with acidemia. It is typically secondary to the accumulation of lactate, chloride and unmeasured anions. Its correction appears desirable and can be more reliably and safely achieved with CRRT. Use of bicarbonate-based fluids is safest as the initial approach. However, lactate- and citrate-buffered fluids can also correct acidosis if appropriately metabolized by the liver and other key organs. CRRT can also be used to correct extreme acidosis in the absence of a major degree of renal impairment. As CRRT controls volume status easily, it would additionally enable bicarbonate infusion to occur for more rapid correction of acidemia.
Note on the calculation of the energy value of foods and of diets The composition of foods
  • I References
  • Em Widdowson
References I. Widdowson EM. Note on the calculation of the energy value of foods and of diets. In: Paul AA, Southgate DA, eds. The composition of foods. 4th ed. New York: Elsevier/North-Holland Biomedical Press, 1978:322-6.