cuvet contents and may alter the equilibrium temperature
of the cuvet contents.
With the optical kinetic procedure described here one can
investigate the temperature performance of a spectrophotometric instrument and then accurately calibrate the temperature of the liquid in the optical path of the instrument.
3. Lasky FD. Rate of achieving thermal equilibrium
in two models
of centrifugal analyzers. Clin Chem 1979;25:1667-8.
4. Price CP, SpencerK, Welch RP. Temperature measurement and
the centrifugal analyser. In: Price CP, Spencer K, eds. Centrifugal
analysers in clinical chemistry. New York: Praeger, 1980:381-94.
5. McDowell TL, Passey RB, Fowler MW. An evaluation of the
temperature control system of the CentrifiChem 400. Clin Chim
Acts 1981;11O:35-43.
References
1. Oliver RWA, Stott A. An optical thermometer:
the design,
construction and calibration of a photometric temperature scale for
use with a thermometric solution. J Physics E 1974;7:275-80.
2. Bowie L, Esters F, Bolin J, Gochman N. Development of an
aqueous temperature indicating technique and its application to
clinical laboratory instrumentation. Clin Chem 1976;22:449-55.
Miller WG, Smith DR. The temperature control system of the
Cobas-Biocentrifugal analyzer. Am J Clin Pathol 1983;80:867-70.
7. Adams PA, Berman MC. A kinetic standard for precise calibration of spectrophotometer cell temperature. Clin Chem
1981;27:753-5.
8. Harris RC, Huitman E. Kinetic methods that are independent of
the rate of reaction. Clin Chem 1983;29:2079-81.
6.
CLIN. CHEM. 33/10, 1895-1 897 (1987)
Glycated Proteins in Serum: Effect of Their Relative Proportions on Their Alkaline Reducing
Activity in the Fructosamine Test
Francesco Zoppl,1AndreaMosca,2SlmonettaGranata,1 and NorbertoMontalbetti’
The fructosamine test is considered clinically useful for
assessing short-term integrated control of blood glucose, but
there are few published data to support this hypothesis. We
fractionated glycated and nonglycated proteins by affinity
chromatography on phenylboronate columns and, with specific immunochemical methods, determined in the eluted
fractions the following proteins, selected according to their
biological half-lives and relative concentrations in serum:
albumin, IgA, IgG, 1gM, apolipoprotein B, haptoglobin, transferrin, a1-antitrypsin, and a2-macroglobulin. We found the
following correlations between fructosamine (mmol/L) and,.
respectively,
glycated albumin, lgG, and (albumin + lgG)
(each in grams per liter): r = 0.901,0.702,
0.878. 1gM had the
highest percentage of glycated molecules (range 11.137.5%, mean 22.4%), haptoglobin and a1-antitrypsin
the
least. This result was almost independent of the proteins’
molecular masses and fractional catabolic rate. Albumin
evidently contributes most to results of the fructosamine test,
confirming conclusions obtained in different ways by others.
Recently Johnson et al. (1) proposed a colorimetric assay
for determining glycated proteins in serum (2). The test and
its further modifications, especially as mechanized (3-5),
are based on the ability of ketoamine (formed in the
Amadori
rearrangement
of the condensation
products
be-
tween glucose and the amino groups of proteins) to reduce
Nitroblue tetrazolium in alkaline solution. Owing to its
simplicity, this method has gained wide popularity.
Its
correlations with other well-codified measures of diabetic
status (mainly glycated hemoglobin, Hb A1) have been
evaluated, the coefficient of correlation with Hb A1 being
‘Laboratorio
di Biochimica Climea ed Ematologia, Ospedale
Niguarda Ca’ Granda, I 20162 Milano, Italy.
2Universita’ degli Studi di Milano, Dipartimento di Scienze e
Tecnologie Biomediche, I 20132 Milano.
Received March 25, 1987; acceptedJune 23, 1987.
variously reported as 0.70 to 0.87 (3-6). Some authors (3)
claim that albumin influences the measurement, a feature
that has been further stressed recently (7-10).
We tried to correlate results with this new test with other
methods for determining glycated proteins. Here we present
correlations between fructosamine and several glycated
proteins (mainly albumin and immunoglobulins)
and Hb
A1.
Materials and Methods
We used a Rotochem CFA 2000 centrifugal analyzer
(‘Fravenol Laboratories Inc., Instr. Div., Savage, MA 20863)
for all immunoturbidimetric determinations; some of the
analytical protocols have been previously published (11).
For nephelometry we used a Behring Laser Nephelometer
and appropriate Behring antisera (Behringwerke, Marburg,
F.R.G.). Hb A1 was determined with an automated “highpressure” liquid chromatograph (Diamat-HPLC; Bio-Rad,
Richmond, CA 94804). We separated proteins with an AES
200 automated electrophoresis system (Olympus
Optical
Co., Ltd., Tokyo, Japan).
We performed the “fructosamine test” (Roche Diagnostica,
Milano, Italy) in the Rotochem CFA 2000, following the
protocol of Lloyd and Marples (3) except that the samplediluent volume was about three times the sample volume
and the measurement wavelength was 520 nm.
To separate glycated proteins from the total proteins, we
used columns ifiled with m-aminophenylboronate
gel (“Glycogel”; Pierce Chemical Co., Rockford, IL 61105) according
to the manufacturer’s instructions. The glycated and nonglycated proteins involved were albumin (Aib), a1-antitrypsin (a1AT), haptoglobin (Hp), apolipoprotein B (Apo B), a2macroglobulin (a2M), transferrin (Ti-f), and immunoglob3Nonstandard abbreviations:Hb A1, glycatedhemoglobin,Aib,
albumin; a2M, a2-macroglobulin;a1AT, a1-antitrypsin; Hp, haptoglobin;Ape B, apolipoproteinB; Ti-f, transferrin; and IgA, IgG, 1gM,
irnmunoglobulins
A, G, and M.
CLINICALCHEMISTRY, Vol. 33, No. 10, 1987 1895
ulins A, G, and M (IgA, IgG, 1gM). We quantified the
glycated and the nonglycated total protein colorimetrically
with Coomassie Brilliant Blue G 205 in phosphoric acid
solution (Bio-Rad).
We determined Alb, Trf, Ape B, and Hp immunoturbidimetrically, using antisera and standards purchased from
Orion Diagnostica, Helsinki, Finland. Immunonephelometry was used to quantifr a1AT and a2M. We determined
IgG, IgA, and 1gM with a modified kinetic latex immunoturbidimetric method involving antibody-coated latexes (Eiken
Chemical Co., Ltd., Tokyo, Japan). We also determined IgG
immunoturbidimetrically with Orion antisera and IgA and
1gM also immunonephelometrically with antisera from Beh-
B
A
//
/
34
‘4
‘4-
I
I
-
I
1234
Aib (gfL)
Glycotad
Glyajted I3 (g/L)
D
C
ringwerke.
3
+
1
Results
234
Ghwted (Aib
+
1ti3)
(9/LI
1
We found a low correlation between fructosamineexpressed in mmol/L or in mmol per gram of total
protein (12)-and
Hb A1 (r = 0.38, P = 0.013 and r = 0.40,
P = 0.007, respectively).
Table 1 summarizes for each protein the average (and
range) percentage of the fraction bound by phenylboronate
(with respect to the overall concentration of that protein)
and also lists some data from the literature (13, 14).
Independently of the proteins’ half-lives in the blood stream,
but probably depending on their amino acid composition (if
not on the carbohydrate chain, when present), this percentage differs for the various proteins. 1gM shows the highest
percentage of glycated molecules, Apo B the lowest, while
a1AT and Hp seem not to be glycated at all.
Figure 1 illustrates the relationships we found between
fructosamine
and the following determined or calculated
quantities: bound fractions (g/L) of Aib, IgG, Alb + IgG, and
Alb + IgG + IgA + 1gM + Ti-f + Ape B + a2M. The positive
intercept is probably ascribable
to the alkaline reducing
activity of substances other than glycated proteins.
234
Glycatad(Aib +
+tgA+
It+
I+
+ Iii’ +
AsoR)
(9/1)
whether
We correlated
4+4. +
+4-’.
Total protein was determined colorimetrically
in a Hitachi 705 analyzer (Boehringer Biochemia, Milano, Italy),
with reagents provided by Wako Pure Chemical
Industry
Ltd., Osaka, Japan (“total protein-HA”).
We analyzed 60 serum samples for all the above-mentioned analytes, from both insulin-dependent and non-insulin-dependent diabetic patients (types I and II).
Fig. 1. Relation between fructosamine (mmol/L) in the serum of 60
diabetic patients and (A) glycated albumin, g/L; ( glycated IgG, g/L;
(C) glycated (AIb + lgG),g/L;and (C) glycated (AIb + a2M + Tn + IgA
+ 1gM + lgG + ApoB), g/L
Regressionparametersare listed in Table 2.The numeralsin B referto resultsicr
certain patients(seetext)
by Sagniez et al. (12), the calculated ratio of fructosamine to
total protein (mmollg) and the ratio of various glycated
proteins to total specific protein (Table 2). The correlation
between fructosamine and glycated IgG fraction was unaffected by the presence of a monoclonal component (three
patients with IgG/X type). In contrast, we verified that a
disproportion in the concentrations of albumin and IgG
could produce outliers in the comparison between fructosamine and glycated IgG. Our cases included four persons
with increased IgG concentration and low albumin concentration and three with increased albumin concentration
and
low IgG concentration (respectively, cases 14, 18, 34, 50 and
cases 7, 55, 57 in Figure IB).
also, according to the procedure suggested
Table 1. Some Chemical and Biological Properties of Proteins as Measured In Our Study or Reported In the
Literature
Meanconcn
ProteIn
t,,, days
Albumin
19
a,AT
a2M
3.9
7.8
P4 (10
3.44
U
1.08
U
51.0
12.2
8-11
19.3
1.92
2,11
1.42
U
0.23-4.25
U
5.9
2.22
0.23-2.51
1.32
1.8
0.92
0.10-4.55
0.89
1.08-4.83
11.09-37.55
8.5
100.0
79.6
Apo B
3.2
265 (B48)
2-4
Mean
1.04-5.83
0
725.0
32.01
% glycatlon
Range
66.5
Trf
Hp 1-1
% carbohydratea
in our population
(n = 60), gIL
459 (B100)
IgA (monomer)
1gM
lgG
5-6.5
162
7.5
3.30
5.1
900
12.0
1.32
23.0
150
2.9
12.20
5.28-17.47
6506b
1.20-4.85
Total protein
Datafrom references 13, 14.
btemined
with biuretreaction.
CDete,.,jnedwithCoomassieBrilliantBlue.
U, undetectable.
1896
CLINICAL CHEMISTRY, Vol. 33, No. 10, 1987
2.69
22.40
11.10
3.06#{176}
Table 2. RegressIon Analysis between Fructosamlne (y) and Various Measured and Calculated Analytes (gIL), and
between Fructosamine/Total Protein (y) and the (Glycated/Total) Ratios for AlbumIn, Summation of Seven Protein
Fractions, and Total Protein
in
Fructosamine (mmol/L) vs:
Glycatedalbumin
Glycated lgG
Glycated (AIb + lgG)
Glycated (Aib#{247}
Trf + a2M
+ IgA + lgG + 1gM)
Glycated total protein8
+
Intercept
Slope
r
1.266
2.009
1.11
1.33
2.45
3.02
3.02
3.02
1.60
0.35
1.36
S,.
n
0.901
0.26
60
0.702
0.42
60
0.677
0.878
0.28
60
ApoB
Fructosamine (mmol)/fota! protein (g) vs:
Glycated albumin/total albumin
Glycated (AIb + a2M + Trf + ApoB
+ igA + lgG + lgM)/total (AIb+
a2M + Tn + ApoB + IgA + IgG
+ 1gM)
2.89
3.02
1.41
0.28
60
3.02
1.23
0.554
0.856
0.879
2.09
0.923
0.23
60
0.0344
0.0464
0.0244
0.640
0.913
0.0037
60
0.0464
0.0464
0.0200
0.537
0.861
0.849
0.897
0.0046
0.0040
60
0.0204
0.0491
Glycated total protein/total protein a
0.0306
8()tem,i
by Coomassie Brilliant Blue dye-binding method.
60
Discussion
References
The correlation we found between fructosamine and Hb
A1 is not consistent with those already published (3-6).
Indeed, a good correlation is not to be expected, because
glycated proteins and glycated hemoglobin represent blood
glucose concentrations integrated over two different time
1. Johnson
RN, Metcalf PA, Baker JR. Fructosamine: a new
approach to the estimation of serum glycoprotein. An index of
diabetic control. Clin Chim Acts 1983;127:87-95.
2. Johnson RN, Baker JR. The alkaline reducing activity of glycated proteins and its relevance to diabetes mellitus. Clin Chem
intervals.
Some authors (8-10), using a different approach, have
suggested that albumin is the major component influencing
the fructosamine test. We have reached the same conclusion.
Some authors (15) claim that gamma-globulins
are “more
readily glycated than other globulins.” We suspect this
might be explained by our finding that 1gM is the more
glycated protein fraction, regardless of its half-life in blood.
However, our results confirming the prevalence of an effect
of albumin on the fructosamine test also suggest that all the
major proteins, when glycated, may contribute to this reaction, as expressed in their overall effect. This is supported
both by the correlation between fructosamine and glycated
proteins (Aib + IgG + IgA + 1gM + a2M + Trf + Ape B)
and by the correlation between fructosamine (per gram of
total protein) and the above-mentioned ratio of glycation.
Although the fructosamine
test is now commercially
available, complete with calibrators, this does not necessarily make it a suitable test for use outside the clinical
laboratory without any other information-e.g.,
total protein concentration and electrophoretic pattern of proteins.
Patients with an altered protein distribution could easily
give deceptively positive or negative results in the fructosamine test.
In our opinion the fructosamine test cannot substitute for
the well-codified Hb A1 test, but it is useful for giving a
more complete picture of a diabetic patient’s status, once the
biochemical profile (particularly of proteins) is known.
We prefer to use a more specific method, i.e., to determine
a well-defined protein according to the interval of glycemic
control one wants to monitor-for
example, 1gM, one week;
Alb, two weeks; IgG, four weeks. In contrast, the fructosamine test determines an assortment
of glycated proteins,
having a wide range of half-lives. We believe that the
concentration of each glycated protein should be defined
every time if determination of glycated proteins is to be
meaningful.
1986;32:368-70.
3. Lloyd
D, Marples
J. Simple colorimetry
of glycated serum
protein in a centrifugal analyzer. Clin Chem 1984;30:1687-8.
4. Hindle EJ, Rosthn GM, Gatt JA. The estimation of serum
fructosamine: an alternative measurement to glycated haemoglobin. Ann Clin Biochem 1985;22:84-9.
5. Baker JR, Metcalf PA, Johnson RN, Newman D, Rietz P. Use of
protein-basedstandards in automated colorimetric determination
of
fructosamine
in serum. Clin Chem 1985;31:1550-4.
6. Baker JR. O’ConnorJP, MetcalfPA, Lawson MR. Johnson RN.
Clinical usefulness of estimation of serum fructosamine concentration as a screening
test for diabetes mellitus.
Br Med J
1983;287:863-7.
7. Seng LY, Staley MJ. Plasma fructosamine is a measure of all
glycated proteins [Letter]. Clin Chem 1986;32:560.
8. Van Dieijen-Visser MP, Seynaeve C, Brombacher PJ. Influence
of variation in albumin or total-protein concentration on serum
fructosamine concentration [Letter]. Clin Chem 1986;32:1610.
9. Van Dieijen-Visser MP, Salemans T, Van Wersch JWJ, Shellekens LA, Brombacher PJ. Glycosylated serum proteins and glycosylated haemoglobin in normal pregnancy. Ann Clin Biochem
1986;23:661-6.
10. Brombacher PJ, Dieijen-Visser MP, Seynaeve C, Salemas THB,
Shellekens LA. Effect of changing albumin and total protein
concentrations on the glycosylated serum protein (fructosamine)
concentration [Abstract V-25]. Quim Clin 1986;5:224.
11. Zoppi F, Perlangeli V, Faglia E. Improvement of a commercial
affinity chromatography method for determining glycated albumin
Clin Chem 1985;31:1238-9.
12. Sagniez M, Krempf M, La Bastard L. Correlation entre le
dosage de l’hemoglobin glycosilee A,,, et celui de Ia fructosamine
chezdes diabetiques soumis a une visite de contr#{244}le
systematique.
[Letter].
Pathol Biol 1985;33:914-6.
13. Marki AA, Leutner C. Sangue-proteine del plasma. In: Leutner
C, ed. Tavole scientifiche Geigy, 8th ed. Basal: Ciba-Geigy Ltd.,
1984:136-9.
14. Peters T
Ji-. Proteins. In: Brown SS, Mitchell FL, Young DS,
eds. Chemica] diagnosis of disease. New York: Elsevier-North
Holland Inc., 1979:311-62.
15. Lemon M, Forrest ARW. Fructosamine activity
serum [Tech. Briefi. Clin Chem 1986;32:2101.
of proteins
in
CLINICAL CHEMISTRY, Vol.33,No. 10,1987 1897