Albumin Quantitation by Dye Binding
and Salt FractionationTechniques
Don S. Miyada, Vernon Baysinger, Solomon Notrica, and R. M. Nakamura
and hemolyzed
Two dye-binding
reagents,
2-(4’.hyd roxybenzeneazo)benzoic
acid (HABA) and bromcresol
green
(BCG), for the quantitation
of serum albumin have
been compared
with the biuret colorimetric
procedure after salt fractionation.
The linear regres-
sion equations for the biuret-BCG and the biuretHABA comparisons were y = 0.43 + O.9lx and y =
-0.45 + l.l5x, respectively. Their corresponding
correlation coefficients were 0.93 and 0.83, respectively. The effects of icterus, lipemia, and hemolysis on the dye.binding procedures were generally
greater in the case of the HABA procedure.
A
comparison of results with BCG and HABA for 20
sera having A/G ratios of less than 1, indicates
them to be equally reliable.
Additional Keyphrases
green method
HABA
#{149}
effects of icterus,
Serum
albumin
quantitation
“protein-error”
of indicators
method
lipemia,
#{149}
bromcresol
hemolysis
based upon the
was introduced
in
(1). Since then, there
1953 by Bracken and Klotz
has been an increasing number of dye methods for
the determination
of serum albumin,
with various
attendant
modifications.
The dyes in widest use
have been the azo dyes such as methyl orange and
2-(4 ‘-hydroxybenzeneazo)
benzoic
acid
(HABA).
More recently
phthalein
dyes such as bromcresol
purple (2) and bromcresol
green (BeG) (3-10) have
been used for this purpose.
The BCG method,
initiated
by Rodkey
(3, 4),
has undergone
modifications,
primarily
a change of
reagent pH from 7.0 to 4.0 and the inclusion of a
nonionic
surfactant.
These changes have resulted
in absorbance
readings
that are directly
proportional to albumin concentration
and reagent blanks
with relatively
low absorbances.
Here, we compare
the merits
of the
BCG
and
methods
with the biuret colorimetric
technique after salt fractionation.
The BCG and HABA
methods
were compared
in the analysis
of sera
with
normal
and
abnormal
albumin/globulin
ratios. In addition
the effects of icteric, lipemic,
HABA
ing methods
serum
specimens
on the dye-bind-
were studied.
Materials and Methods
For all
used as
0814108;
Lambert
07950).
methods a commercial
control serum was
the reference
base (“Versatol,”
lot No.
General
Diagnostics
Division,
WarnerPharmaceutical
Co., Morris Plains, N. J.
Bromcresol
Green (BCG) Method (11)
Stock reagent
consisted
of 26 ml of sodium
hydroxide
(100 g/liter),
30 ml of lactic acid, 500
mg of BCG, and 10 ml of “Tween 20,” all diluted to
1 liter with distilled water, mixed, and the pH adjusted to 4.0 at room temperature
(“AlbuStrate,”
General Diagnostics).
The stock reagent
was diluted
five-fold
with
distilled water before use. Ten microliters
of serum
was added to 5 ml of diluted reagent, mixed, and
read at 630 nm vs. the reagent blank.
HABA Dye Method
The procedure was performed
(Beckman
92634) according
560
Instruments,
to Beckman
on the Model DSA
Fullerton,
Calif.
Procedure
No. 83917,
a modification of the method described by Martinek (12). In the modified HABA method,
HABA is
used at a concentration
of 1.5 mmol/liter
in a
phosphate
buffer of pH 6.4, the detergent,
“Brij35,” is incorporated
and the instrument
is usually
set on a serum blank. Serum blanks were prepared
by substituting
HABA
reagent
with phosphate
buffer. All absorbances
were read at 510 nm. The
manual HABA method
was identical
to the automated
procedure
except
for a proportional
increase in reagent
volumes;
where HABA reagent
blanks were used, serum was replaced by water.
Biuret Colorimetric
Method
After Salt Fractionation
From the Orange County Medical Center, Orange, Calif. 92668
(I). S. M., V. B., H. M. N.), University
of California
at Irvine,
Irvine, Calif. (1). S. M., H. M. N.), and St. Johns Hospital,
Santa
Monica
(S. N., presently
with
Laboratory
Procedures,
Inc.,
Woodland
Hills, Calif.).
Heceived Dec. 4, 1970; accepted Sept. 27, 1971.
52
CLINICAL CHEMISTRY,Vol. 18, No. 1,1972
The salt fractionation
technique
was performed
according
to the method of Gornall et al. (13). A
sodium sulfate-sulfite
mixture
(270 g/liter)
was
used for fractionation
of the albumin and globulins.
Methods
Results and Discussion
of Investigation
The three methods
were first compared
by analysis of 120 random
sera from hospital
patients.
All analyses were carried out within a 24-h petiod
from the time of collection.
Icteric, hemolytic,
or
lipemic sera were noted.
Twenty
sera having a
normal appearance
but abnormal
albumin/globulin
ratios were analyzed
for albumin
by the three
procedures
under study.
The effects
of bilirubin
and hemoglobin
on
serum albumin
determinations
were investigated
by use of the following solutions:
Spectral
absorbance
curves
for the BCG and
methods are illustrated
in Figure 1. Note that
whereas BCG measurements
made at 630 nm correspond to its spectral
absorbance
maximum,
the
wavelength
of 510 nm used for the HABA method
does not correspond
to its blank-corrected
spectral
absorbance
maximum,
which is about 485 nm.
Measurements
at 510 nm are less sensitive because
of the lower absorptivity.
Comparative
sensitivities
of the BCG, HABA and biuret methods
are shown
in Table 1. The BCG method is perhaps 20 times as
sensitive as the HABA method.
HABA
Bilirubin.
Bilirubin
solutions
were prepared
in
20 .tg/100 ml of sodium carbonate
in concentrations
ranging
to 50 mg/100
ml (bilirubin;
Harleco,
Philadelphia,
Pa. 19143). Vials of lyophilized
control serum were reconstituted
with 5 ml of the
above bilirubin solutions before analysis.
Hemoglobin. A concentrated
hemoglobin
solution
was freshly prepared
by hemolyzing
washed red
cells with water.
The hemolysate
was analyzed
and appropriately
diluted with saline to give concentrations
ranging
up to 500 mg/100
ml. Vials
of lyophilized
control
serum were reconstituted
with 5 ml of the above hemoglobin
solutions before
analysis.
lii
U
z
0
U,
d
The effects of bilirubin
and hemoglobin
on the
and HABA dye methods
were further
investigated by using an aqueous solution of purified human serum albumin
(Human
4X recryst.
serum
albumin;
Nutritional
Biochemical
Corp., Cleveland, Ohio 44128). As described
above, bilirubin
and hemoglobin
solutions
were prepared
and
added in appropriate
amounts to aqueous solutions
of purified serum albumin.
The resulting
solutions
were analyzed
by both BCG and HABA methods;
however, in the latter series of experiments,
spectral absorbance
patterns
were obtained
instead of
absorbance
measurements
at a fixed wavelength.
All spectral patterns
were obtained
with a Bausch
& Lomb Spectrophotometer,
Model 600, coupled
to a Sargent SRL Recorder.
BCG
WAVELENGTH,
nm
Fig. 1. Spectral absorbance curves obtained on an
albumin solutionby HABA and BCG methods
Composition
of solutions used:
Albumin,
Designation
g/100 ml
Dye
Blank
A
4
BCG
Water
B
0
C
4
BCG
HABA
Water
Water
D
0
HABA
Water
Table 1. Comparative Data for Albumin Determinations by HABA, BCG, and Biuret
Methods, with Purified Human Serum Albumin
Method
HABA
Biuret
Albustrate
Total volume
reagent
sample, ml
Sample
sIze, ml
Sample to
total volume
3.80
0.050
0.200
20.00
5.01
0.010
ratio
Cuvet
sIze,
mm
Absorbance#{176}
76
10
0.15
100
19
10
0.14
0.44
501
Sensltlvltyb
11.4
0.7
220
o All absorbance
measurements were read on a Coleman Jr. Spectrophotometer,
Model C. The albumin analyzed was 4 X crystallized human serum albumin prepared at a concentration of 4 g/100 ml. All samples were read against appropriate reagent
blanks.
(A)(sample:vol.
ratio)where A is absorbance
Sensitivity
=
d
,
.
and d iscuvetsize.
CLINICAL CHEMISTRY, Vol.18,No. 1,1972 53
Comparisons
of the BCG and HABA methods with
the biuret method
were made with scattergram
plots (Figures
2 and 3). Both visually
and statistically,
BCG measurements
correlate better than
the corresponding
HABA
measurements
with those
obtained
by biuret after salt fractionation;
this
is especially
true in cases of icteric and lipemic
sera, but this difference is negligible in the case of
hemolytic
sera.
The effect of lipemia was found to be greater
with the HABA procedure.
An explanation
for this
difference
can probably
be based on the greater
sensitivity
of the BCG procedure,
which requires less
serum and therefore
less lipemia-causing
lipids.
The quantitative
effects of bilirubin
and hemoglobin on the BCG and the HABA methods
for the
determination
of albumin were investigated
(Table
2). With the HABA method, the results indicate increased absorbances
with increased concentrations
for both bilirubin
and hemoglobin
when uncorrected with a serum blank. This increase amounts
to approximately
1% per milligram
of bilirubin
and an 18% increase
in absorbance
at a hemo5-
0
4
0
0
03
2
1
2
3
4
BIURET
5
Fig. 2. Correlation plots of serum assayed
g/100 ml, by the BCG and biu ret methods
for albumin,
in
Serum appearance; 0, normal; o, hemolyzed; #{149},
lipemic; +,
icteric. Regression line: BCG = 0.43 + 0.91 biuret; correlation
coefficient, 0.93
globin concentration
of 500 mg/100
blank-corrected
absorbance
readings
differences)
are compared
Table 2. Effect of Bilirubin and Hemoglobin on
the HABA and BCG Dye Methods for the
Determination of Serum Albumin
Absorbancesb
Concen
tratlon
mg/lOOml
Bilirubin
.
0
S
4
a
040
do
0
o
0
Hemoglobin
44.
1
2
4
3
4
5
BIU RET
Fig. 3. Correlation plots of serum assayed for albumin,
in g/100 ml, by the HABA and biuret methods
Serum appearance; 0, normal; 0, hemolyzed; #{149},
lipemic; -I-,
icteric. Regression line: HABA = -0.45 + 1.15 biuret; correlation coefficient, 0.83
54 CLINICAL CHEMISTRY, Vol.18,No. 1,1972
blank
Test
0.160
0.010
0.150
0.440
0.025
0.155
0.445
20
0.185
0.140
0.450
30
0.205
0.230
0.045
0.062
0.100
0.143
0.130
0.450
0.450
0.010
0.015
0.016
0.150
0.440
0.150
0.020
0.154
0.163
0.154
0.440
0.434
0.449
0.449
0
100
+
BCG
Test
0.170
50
00
2
HABA
Serum
-_________________________
0
10
50
0
instead of total absorb-
ance readings,
then the albumin values are found
to be decreased
with icteric sera. This apparent
depression
of albumin
values, amounting
to approximately
0.4% per milligram
of bilirubin,
can
probably
be ascribed to an “over-correction”
imposed by reading the test against a serum blank
(see Table 2 and Figure 4). With hemoglobin,
a
serum blank adequately
corrects for added hemoglobin. In both cases, the blank-corrected
results
are more acceptable
than the corresponding
noncorrected ones.
On the other
hand,
with the BCG method,
bilirubin and hemoglobin
interferences
were minimal, amounting
to 0.04% per milligram of bilirubin
and 3% at a hemoglobin
concentration
of 500 mg/
100 ml.
The effects of bilirubin
and hemoglobin
on BCG
and HABA methods were further investigated
with
use of purified serum albumin (Figures 4, 5, and 6).
The spectral absorbance
curves obtained
with the
HABA
method
indicate
a shift in the maxima
to
lower wavelengths
with increased
bilirubin
concentration
for both test and serum blank series.
This effect is more marked with the test series and
can probably
be explained,
in part, as the additive
superimposition
of the spectral patterns
for HABAalbumin and bilirubin.
A comparison
of the series
of curves obtained
with icteric sera indicates
a
small but significant
difference in the spread of the
absorbance
measurements
at 510 nm between the
serum blank and test series. The greater spread in
the absorbances
of the serum blank series with
5
4
ml. If serum
(absorbance
200
300
500
0.160
0.165
0.165
0.174
0.185
0.188
0.022
0.034
0.149
0.454
Concentration of bilirubin and hemoglobin added to a commercial, lyophilized control sera for which the bilirubin concentration was 0.8 mg/100 ml and hemoglobin was undetectable
by the Porter method (15).
Absorbance readings obtained on the Coleman Jr. Spectrophotometer, Model C, estimated to the third place.
0.4
HABA TEST
a4
I&I
C)
z4
m
0.2
0
U,
02
400
600
500
WAVELENGTH,nm
600
WAVELENGTH, nm
Fig. 4. Spectral absorbance scans showing the effect of
Fig. 5. Spectral absorbance scans showing the effect of
hemoglobin in concentrations of 0-500 mg/100 ml on the
HABA test. Scans were made against a HABA reagent
blank
bilirubin on the BCGand HABAmethods
Further composition of solutions used:
Designation
Bilirubin
Dye
A
B
C
Blank
0
8
None
Water
None
Water
40
0
8
40
None
HABA
HABA
HABA
0
BCG
Water
Serum blank
Serum blank
Serum blank
BCG blank
C
20
BCG
BCGblank
C
G
40
BCG
BCG blank
100
BCG
BCG blank
0
E
F
G
icteric sera would “overcorrect”
for the bilirubin
present, and the results confirm previous findings
(Table 2).
The
difficulty
encountered
with
the HABA
method
for icteric sera is made clear for those
utilizing
this procedure
on some of the more
popular
instruments
for automated
analysis.
Almost invariably,
when an albumin
result by the
HABA.
method
performed
on an icteric serum is
suspected
of being falsely low, and is compared
with that obtained
by the BCG method,
the result
by the latter method
will be found to be substantially
higher. These differences
have not been
found to be directly related to the bilirubin
concentration
of serum and, as shown by the investigations of Aryan and Ritz (14), the apparent
depression of albumin values obtained
by the HABA
method could be related to the conjugated
fraction
of bilirubin-free
bilirubin
had little effect. Their
findings are not inconsistent
with ours and offer a
logical explanation
for the inconsistency
in albumin
values by the HABA method in icteric sera. A more
definitive
study will require a better knowledge
of
the true nature of the chemical
entities
involved
and of their interactions,
information
not now
clear.
The effect of hemoglobin
on the HABA method is
shown in Figures
5 and 6. The results
are not
0.4
HABA
BLANK
0.2
510
WAVELENGTH,
nm
Fig. 6. Spectral absorbance
scans showing the effect of
hemoglobin
in concentrations
of 0-500 mg/100 ml on the
HABA serum blank. Scans were made against water
particularly
remarkable;
there is no shift in the
absorbance
maximum
with increased
hemoglobin
concentrations
up to 500 mg/100 ml, there appears
to be little interference
from the 540 nm peak of
oxyhemoglobin,
and the difference in the spread of
absorbance
at 510 nm for the blank and the test
series appears
insignificant.
The last observation
confirms the findings of Table 2, and indicates
that
the HABA method with serum blank correction yields
results that are uninfluenced
by hemoglobin.
Two properties
of the BCG method stand out: (a)
the greater
sensitivity
and specificity
of 13Cc1for
albumin
quantitation
minimize
interference
by
other
foreign
serum
chromogens,
and (b) the
greater
difference
in the wavelength
of measurement for the BCG method (630 nm) in comparison
to that for the HABA method (510 nm) minimizes
interference
from substances
such as bilirubin
(460
nm) and oxyhemoglobin
(418, 542, and 582 nm).
The BCG and HABA methods for determination
of
serum albumin were compared
for 20 sera having
A/G ratios of less than 1 (Table 3). The 20 samples
were considered
to have an abnormal
A/G ratio
CLINICAL CHEMISTRY,
Vol. 18, No. 1, 1972 55
Table 3. Comparison of Serum Albumin Measurements by BCG, HABA, and Biuret Methods
on “Normal” Appearing Serum Having
Abnormal A/G Ratios
ToUl
protein
Albustrate
HABA
AlbumIn (A/G)
Bluret
5. Dow, D., and Pinto, P. U. C., Determination of serum albumin
on the SMA-12/30
(Hospital
Model)
using bromcresol
green.
CLIN. CHEM. 15, 1006 (1969).
6. Hernandez,
0., Murray,
L.,
determination
of serum
albumin
CHEM. 13, 701 (1967).
and Doumas,
B., Automated
with bromcresol
green.
CLIN.
1
7.3
2.5(0.5)
2.2(0.4)
2.2(0.4)
2
3
6.7
3.3(0.9)
3.3(0.9)
3.1(0.9)
7. Harding,
J. R., and Keyser,
J. W., Bromocresol
green as a
reagent
for serum albumin.
Proc. Ass. Clin. Biochem.
5, 51(1968).
8. Bartholomew,
R. J., and Delaney,
A. M., Proc. Aust. A88.
Clin.Biochem. 1, 214 (1966).
5.1
1.1(0.3)
0.8(0.2)
0.8(0.2)
5
8.7
8.0
3.5(0.7)
3.5(0.8)
3.7(0.7)
3.3(0.7)
3.2(0.6)
3.2(0.7)
9. Northam,
4
6
SpecImen
Bull. No.
7.1
3.3(0.9)
3.4(0.9)
3.3(0.9)
7.6
7.7
7.3
7.9
4.1(1.2)
4.0(1.1)
3.0(0.6)
4.2(1.2)
4.0(1.1)
2.9(0.7)
2.6(0.5)
3.9(1.1)
3.8(1.0)
2.5(0.5)
2.3(0.4)
11
12
13
14
15
7.7
9.3
7.4
7.8
7.6
3.1(0.7)
3.3(0.6)
3.1(0.7)
3.6(0.9)
3.1(0.7)
2.8(0.6)
3.3(0.6)
3.2(0.8)
3.4(0.8)
3.5(0.9)
2.6(0.5)
3.0(0.5)
3.0(0.7)
3.3(0.7)
3.1(0.7)
16
17
18
19
20
7.6
7.7
7.8
6.5
6.5
3.5(0.9)
3.4(0.8)
3.7(0.9)
3.0(0.9)
3.6(0.9)
3.6(0.9)
3.6(0.9)
3.2(1.0)
3.6(0.9)
3.3(0.8)
3.4(0.8)
3.0(0.9)
3.1(0.9)
3.2(1.0)
3.0(0.9)
References
1. Bracken,
J. S., and Klotz, I. M., A simple method for the rapid
determination of serum albumin. Amer. J. Clin. Pathol. 23, 1055
(1953).
2. Louderback, A., Mealey, E. H., and Taylor, N. A., A new
dye-binding
technic using bromcresol
purple for determination of
albumin in serum. CLIN. CHEM. 14, 793 (1968).
4. Rodkey,
F. L., Direct
albumin
in human
serum.
spectrophotometric
CLIN.
56 CLINICAL CHEMISTRY,
CHEM.
determination
11, 478 (1965).
Vol. 18, No. 1, 1972
and Widdowson,
Biochem.
(1967).
G.
11. Babson,
A. L., see Addendum
12. Martinek,
determination
R. G., Evaluation
of serum albumin.
M.,
Albumin.
Tech.
to this paper.
of a dye-binding
CLIN.
CHEM.
serum
of
method
for
441 (1965).
11,
13. Gornall,
A. G., Bardawill,
C. J., and David,
M. M., Determination
of serum protein
by means
of the biuret
reaction.
J.
Biol. Chem. 177, 751 (1949).
14. Aryan,
D. A., and Ritz, A., Measurement
of serum albumin
by the HABA dye technique:
A study
of the effect of free and
conjugated
bilirubin,
of bile and of certain
drugs.
Gun. Chim.
Acta 26, 505 (1969).
15. Porter,
C. A., Spectrophotometric
method
for quantitative
plasma
hemoglobin
resulting
from acute hemolysis.
J. Lab. Clin.
Med. 60, 339 (1962).
Addendum
The bromcresol
green albumin
scribed in the preceding publication
in our laboratories
and submitted
and were selected
because
of no demonstrable
evidence
of lipemia,
hemolysis,
or icterus.
The
two methods
measure
albumin
equally
reliably.
In view of these results and findings of an occasional unreliable
albumin
result
by the HABA
procedure,
for sera with abnormal
A/G ratios, it
appears that these discrepancies
might result from
the analysis of icteric, lipemic, or hemolytic
(uncorrected
for serum blank) sera. Or possibly there
might be interference
from compounds
that competitively
bind albumin
such as fatty acids, sulfonamides,
salicylates,
glucuronides,
and phenobarbital.
We are currently
investigating
the effect
of these compounds
on both the HABA and the
BCG methods.
3. Rodkey,
F. L., Binding
of bromcresol
green by human
albumin.
Arch. Biochem.
Biophys.
108, 510 (1964).
F.,
Ass. Clin,
10. I)oumas, B. T., Watson, W. A., and Biggs, H.-G., Albumin
standards and the measurement
of serum albumin with bromcresol green. Clin. Chim. Ada 31, 87 (1971).
7
8
9
10
2.9(0.7)
B.
11,
CHEMISTRY
for
publication.
In
procedure
dewas developed
to CLINICAL
the
interim
an
essentially
identical
procedure
was published
by
Doumas et al. (1). These authors used 105 mg of
bromcresol
green per liter of succinate
buffer (75
mmol/liter,
pH 4.2) containing
“Brij-35”
(1.2 ml/liter), while our assay uses 100 mg of bromcresol
green per liter of lactate buffer (0.1 mol/liter,
pH
4.0) containing
2 ml of “Tween-20.”
Doumas et al.
also used 25 j.il of serum in 4 ml of reagent, while
our method calls for 10 uzl of serum in 5 ml of reagent. We have compared
the two methods,
and
they give similar results.
The advantages
of an acid-buffered
solution of
bromcresol
green containing
a nonionic surf actant
as a reagent for serum albumin
were thoroughly
documented
by Doumas
et al. We could see no
justification
for burdening
the literature
with substantially
the same information,
and reluctantly
withdrew our paper from consideration
by CLINICAL
CHEMISTRY.
ARTHUR
Department of
Diagnostics
L.
BABSON
Research
The Warner-Lambert
Research
N. J. 07950
Institute
Morris Plains,
I. Doumas,
B. T., Watson, W. A., and Biggs, H. G., Albumin
standards and the measurement
of serum albumin with bromcresolgreen. Clin.Chim. Acta 31, 87 (1971).