CUN. CHEM. 40/7, 1317-1321 (1994)
#{149}
Endocrinology
Glycohemoglobin Measured by Automated Affinity HPLC Correlates with Both
Short-Term and Long-Term Antecedent Glycemia
William T. Cefalu,’ Zhong
Q. Wang,
Audrey Bell-Farrow,
We evaluated glycohemoglobin (GHb) and glycated
plasma protein (GPP) by automated affinity HPLC for their
ability to monitor both short-term and long-term antecedent glycemia in 70 diabetic subjects. We placed 30 subjects on an intervention protocol in which insulin and (or)
dietary changes were made twice weekly to acutely decrease glycemia. We monitored 40 subjects at 6-week
intervals; changes in the clinical regimen were made at
that time only. Despite weekly changes in mean blood
glucose in the subjects who received more intensive intervention, GHb concentrations correlated significantly
with the weekly (r = 0.66, P <0.001), 2-week (r 0.70, P
<0.001), 3-week (r = 0.72, P <0.001), and 6-week (r =
0.83, P <0.001) mean glucose concentrations. GPP correlated significantly with measured glycated albumin determined by boronate affinity columns (r = 0.83, P <.001)
and correlated best with the 1-week (r = 0.66, P <0.001),
2-week (r = 0.64, P <.001) and 3-week (r = 0.60, P
<0.001) mean antecedent glucose concentration. Thus,
GHb, traditionally considered a marker for only long-term
diabetic control, correlated significantly with both shortterm and long-term antecedent glycemia.
=
Indexing Terms: diabetes mellitus/blood
cated proteins
glucose monitoring/gly-
Recent evidence has conclusively
related the development of diabetic complications
to clinical hyperglycemia
and demonstrated
that control of blood glucose, with use
of intensive insulin therapy, is associated
with a significant reduction in complications
such as retinopathy,
nephropathy,
and neuropathy
(1-4). Even before this
association
was scientifically
validated,
the measurement of glycohemoglobin
(GHb), an objective marker for
antecedent
glycemic control, was an integral part of
clinical
diabetes
management
(58).2
Because of its
high level of precision,ion-exchange
HPLC has been
recommended
as the “gold standard”
for measurement
of GHb. However,
this method is still less than ideal
because of potential
interference
by Hb variants
with
Diabetes Research Laboratories and Clinics, Department of Internal Medicine, Bowman Gray School of Medicine, Winston-Salem, NC 27157-1047.
‘Address correspondence to this author at: Section on Endocrinology/Metabolism,
Department
of Internal Medicine, Bowman
Gray School of Medicine, Medical Center Blvd., Winston-Salem,
NC 27157-1047. Fax 910-716-5895.
Abstract presented at the 52nd Annual Meeting, American Diabetes Association, San Antonio, TX, June 20-23, 1992.
2NO
abbreviations:
Hb, hemoglobin; GHb, glycohemoglobin; GPP, glycated plasma protein; CBG, capillary blood glucose; NIDDM,
non-insulin-dependent
diabetes
meffitus;
and
IDDM, insulin-dependent diabetes mellitus.
Received August 25, 1993; accepted April 11, 1994.
Fran D. Kiger, and Camille Iziar
altered charge (e.g., HbS or HbC) by many glucose adducts comigrating
with HbA1 species (e.g., carbamylated Hb), and by problems arising from artifacts
induced under various storage conditions
(9-13).
The
boronate
affinity
chromatography
method involving
minicolumns
has been gaining wide acceptance
because
it is not affected by Hb variants,
slight temperature
changes, carbamylated
Hb, the labile fraction (aldimine
or Schiff base), or ionic strength of the buffer (9, 14-19).
However, in some laboratories,
precision has been a
problem with these small columns. The application
of
HPLC used with affinity methods has markedly
improved precision and run-time while maintaining
the
major advantage
of the boronate affinity method (1720): Its applicability
to Hb and to soluble blood proteins
in serum and plasma samples allows objective glycation
assessment for all blood proteinsof interest with a single
method. This may allow objectivedetermination of antecedentglycemiccontrol for the previous 2-3 months (e.g.,
via GHb) and short-term diabetic control for the preceding 1-2 weeks [e.g., via glycatedplasma proteins (GPP),
especially glycated albumin] in the same specimen.
To clinically
evaluate the utffity of measurements
of
Glib and GPP by affinity HPLC in assessing durations
of antecedent
glycemia, we studied a diabetic outpatient
population undergoing both intensive and conventional
clinical intervention.
We evaluated
these analytes
for
correlation with mean glucose concentrations
for shortterm (1-3 weeks) and long-term (6 weeks) periods.
Patientsand Methods
Subjects. We studied 278 subjects, 70 of whom were
diabetic. The 208 subjects evaluated to establish
control
ranges were considered nondiabetic on the basis of their
known medical history
and repeated
fasting blood-glucose analysis.
The diabetic subjects were grouped according to treatment intervention.
Group I consisted of
30 non-insulin-independent
diabetes mellitus (NIDDM)
patients (16 women, 14 men, mean age 53 ± 3 years,
range 34-69),considered to be moderately uncontrolled
(GHb
10%), for which clinical intervention
consisted of
dietary changes and (or) insulin adjustments to decrease
mean weekly glucose significantly
over 8 weeks. These
patients were monitored with a home glucose-monitoring system with electronic log books as performed with
Accu-Check
II glucose meters (Merlin Diabetes Data
Management System; Boebringer
Mannheim
Diagnostics, Indianapolis,
IN). In this aspect of study, patients
performed capillary blood glucose (CBG) determinations
before meals and at bedtime. Patients
reported to the
clinic weekly, at which time mean CBG was determined, and insulin and (or) dietary adjustments
were
CLINICAL CHEMISTRY, Vol. 40, No. 7, 1994 1317
made twice weekly (via clinic and phone contact) on the
basis of the glucose recorded at home to decrease the
previous weeks’ glucose. Insulin therapy consisted of
twice-daily
injections
of intermediate-acting
insulin
(e.g., NPH) with mixed short-acting insulin (e.g., Regular) involving sliding scale at breakfast
and supper. We
refer to the Group I patients as the “intensive clinical
intervention”
group for the purpose of this study only;
we do not intend to misrepresent
the subjects as having
intensive insulin therapy as administered
in the Diabetes Control and Complications
Trial (4). Group II consisted of 40 diabetic subjects (22 women, 18 men; 24
NIDDM, 16 IDDM; mean age 49 ± 3 years, range 2573). These patients also recorded CBG values before
meals and at bedtime
and reported
to the clinic at
6-week intervals for dietary and (or) medication adjustments. All group II subjects were on a split/mixed
NPH/
Regular
insulin
regimen, except for eight IDDM patients who were on either a long-acting
insulin (e.g.,
Ultralentebased) or insulin-pump
regimen. The correlations of GHb and GPP with the weekly, 2-, 3-, and
6-week glucose means were determined for Group I subjects and with 6-week glucose means for Group II subjects. All patients gave informed consent and the study
was approved by the Clinical Research Practices
Committee of the Bowman Gray School of Medicine.
Sample prepara#{252}on.Whole blood collected by venipuncture
with EDTA as anticoagulant
was centrifuged
at 1500g for 10 mm; the plasma was removed and diluted 10-fold with GPP diluent
(Primus, Kansas City,
MO); the packed erythrocytes were diluted with 200-fold
Glib diluent containing
an aqueous hemolysis reagent
and preservative
(Primus). The samples were diluted
directly in the vials used for the assay.
Instrumentation.
The Primus CLC33O high-performance liquid chromatograph
determines both GHb and
GPP by automated
HPLC boronate affinity methods.
Equipped with an ultraviolet/visible
light detector, autosampler, computing integrator, and two disk drives,
the system is controlled through automation
software
provided by Primus. Both the analytical
system and the
automation
software
are described
in detail elsewhere
(21).
HPLC analysis.
Sample vials were loaded into the
autosainpler,identified
as to sample type,and analyzed
by the automation
software. The analytical
sequence is
as follows: Whole-blood
hemolysate
(5 L) or diluted
plasma (10 L) is automatically
injected onto a boronate
affinity column (Primus). The 0.5 x 5 cm glass analyticalcolumn used for the separation of GHb from other
Hb species contains m-aminophenyl
boronic acid covalently bound to a 10-tm-thick
vinyl polymer solid support, producing
>1200
theoretical
plates
per column.
Glycated species are retained by the column while nonglycated species (peak 1) are eluted through the column
and detector with elution reagent 1 (mobile phase from
injection
to 0.2 miii): ammonium
acetate
buffer (250
mmol/L),
sodium chloride
(500 mmoIIL) magnesium
chloride (30 mmolIL), ethanol (45 mldL), methanol (2.5
milL), isopropanol (2.5 milL), pH 9. Elution reagent 2
1318 CLINICAL CHEMISTRY, Vol. 40, No. 7, 1994
(0.2-1.3mm) contains a competing polyol designed to
release the boronate-bound
glycated
protein fraction
(peak 2). The reagent contains sodium chloride (150
mmol/L), ethanol (45 mLfL), methanol (2.5 mL/L), isopropanol (2.5 mL/L), and mannitol (100 mmol!L). Elution reagent 1 is then added (1.3-2.0 min) to prepare for
the next sample sequence. There is -0.7-mm
lag time
between switching to each mobile phase and its arrival
at the detector. All reagents were supplied by Primus.
The flow rate was 1.8 mL/min; the column oven was
kept at 40#{176}C.
GHb was detected at 413 nm, GPP at 280
nm. Quantification
was by peak-area analysis with a
calibration curve derived from calibrator samples. The
computing integrator calculates
the percentage
of the
glycated species by the following formula, as previously
described (19): % glycated species = 100 [peak 1 area!
(peak 1 area + peak 2 area)]. Intra- and interassay CVs
were 1.2% and 2.1%, respectively
(n = 10, mean =
5.7%).
Calibration.
The Primus CLC33O uses a two-point
calibration
system. Calibrators are prepared from human whole blood (for GHb) or from human plasma (for
GPP) obtained from nonthabetic
and diabetic donors.
Samples
are lyophilized
and
stored refrigerated
in
sealed vials until use. Values for the calibrators were
determined
by the manufacturer.
Data obtained in a
calibration
run were used to correct for variations
in
assay conditions.
Glycated serum albumin
(inter- and intraassay
CVs
4.5% and 3.4%, respectively)
was determined
with boronate affinity chromatography
with minicolumns,
as
previously described (22, 23).
Data were analyzed with Pearson correlation coefficients and repeated-measures
analysis
of variance
where appropriate.
Results
The range of GHb results for the nondiabetics
(n =
208, mean ± SD fasting glucose 4.6 ± 1.2 mmol/L) was
4.0-8.0% (mean ± SD 5.55 ± 0.74%). Using the 95th
percentile as the upper limit, we determined the upper
range of normal GHb for this assay to be 7.0%. The
range of GPP results for nondiabetics
was 13.4-25.0%
(mean 19.2%; SD 2.2%), with the 95th percentile (upper
range of normal) at 23.5%.
Figure
1A represents
the mean weekly concentrations of CBG, GHb, and GPP in the intensive intervention group. With use of repeated-measures
analysis of
variance, significant
decreases
in Glib (P <0.0001),
GPP (P <0.003), and CBG (P <0.0001) were noted over
time. As demonstrated
in Fig. 1B, no major changes in
GPP, mean CBG, or GHb were seen in Group II subjects,
demonstrating
relatively
stable glycemia over the period of observation.
The mean weekly CBG decreases for the intensively
intervened subjects (Group I) at weeks 1-8 were 15.7%,
12.3%, 6.7%, 5%, 9.1%, 5.8%, 1.4%, and 0.8%, respectively. Because
most the change
in mean glucose
(=50%) occurred after only 2 weeks of intervention,
we
correlated
the absolute
changes
in mean
glucose
30
A
25
GPP
minicolumns
for all Group I and Group II patients at
baseline
and correlated significantly
with the GPP by
automated
affinity HPLC (r = 0.83, P <0.001). Glib
correlated significantly
with the GPP at baseline
in
Group I (r
0.61, P <0.001), but by week 8 the correlation was no longer significant (r
0.26). In Group II,
Glib maintained
good correlation
with GPP at each
6-week cycle(r = 0.65-0.72,
P <0.001).
=
20
=
15
10
GHb
5
CBG
0
0
1
2
3
4
5
6
7
8
9
25
GPP
20
15
10
GHb
5
CBG
0
4
6
8101214161820222426
Weeks
Fig. 1. (A) Weekly changes in CBG (mmol/L), GPP (%), and GHb
(%) in Group I (intensively treated), and (B) mean CBG, GPP, and
GHb at 6-week intervals in Group II (conventional intervention) patients.
Dataare mean ± SE. CBG concentrations
areexpressed as 0.5 x absolute
value.
(ACBG) concentrations
with the changes
in Glib
(iGHb) for this period to determine whether GHb could
detect short-term glycemic changes. iCBG at the end of
week 1 and week 2 correlated significantly
with iXGHb
at week 1 (r
0.45, P <0.02) and week 2 (r
0.57, P
<0.002), but not with iCBG from weeks 3-8.
To determine the ability of Glib and GPP to monitor
antecedent
diabetic
control, we correlated
weekly
means for CBG, Glib, and GPP from Group I subjects.
Glib correlated
significantly
with the weekly, 2-week
and 3-week mean blood glucose (Fig. 2A-C and Table 1),
and with the mean 6-week glucose concentration
in both
Group I (r = 0.83, P <0.001) and Group II (r = 0.84, P
<0.00 1) subjects
(Table 1), despite the weekly changes
in glucose seen in Group I. GPP concentrations
at each
week of study correlated significantly
with the weekly,
2-week, and 3-week mean CBGs in the Group I patients
(Fig. 2D-F and Table 1). The 6-week mean CBG for
Group I subjects also correlated significantly
with the
GPP (r = 0.49, P <0.001), although at a lower correlation than when compared with Group II (r = 0.64, P
<0.001) (Table 1). Glycated albumin was determined by
=
=
Discussion
As would be expected, GHb determined
by automated
affinity HPLC correlates well with long-term glucose control in diabetic outpatients monitored at 6-week intervals.
Despite weekly
changes
in mean blood glucose in our
intensive intervention subjects, Glib, traditionally considered a marker only for long-term diabetic control, correlated significantly
with the weekly, 2-week, 3-week, and
6-week mean glucose and demonstrated
an ability to detect recent changes in glycemia. GPP correlated significantly with measured
glycated albumin concentrations
determined
by boronate
affinity minicolwnns
and correlated best with the 1-3-week mean glucose, demonstrating validity as a measure of short-term glycemia.
Long-term prospective
studies (e.g., 4) have demonstrated that the measurement
of glycated blood protein
markers as objective markers of antecedent
glycemic
control can predict the development
of diabetic complications such as retinopathy,
nephropathy,
and neuropathy (1-4). Physicians are now encouraged to normalize
the patient’s glycemic profile while minimizing
the side
effects of intensive therapy (4). The reduction in glycation of blood proteins, secondary
to a lowered mean
blood glucose concentration,
is postulated
to decrease
tissue accumulation
of nonenzymatically
attached glucose, thereby preventing
the advanced glycation
(i.e.,
protein cross-linking)
thought to contribute to the tissue
dysfunction
(24-27).
Therefore, the measurement
of
Glib as an objective marker for antecedent
glycemic
control has received renewed clinical interest.
Our data demonstrate
that Glib determined
by automated affinity HPLC is an excellent index of antecedent
long-term diabetic control (i.e., 6 weeks), showing excellent correlation in both an intensively
intervened and a
relatively stable outpatient population at 6-week intervals. GPP primarily
showed validity
in monitoring
short-term control in the acute intervention
patients,
and had decreased
correlation
with mean CBG when
evaluated
at 6-week intervals in this group. This suggests that the most appropriate
antecedent
glycemic
window for GPP may be the preceding 1-3 weeks. However, in the Group II patients, GPP maintained
high
correlation with the mean outpatient glucose concentration over a 6-week interval. Group II did not have major
changes in mean blood glucose, so the GPP would be
expected
to maintain
excellent
correlation
with both
mean CBG and GHb. However, GHb correlated with 1-,
2-, and 3-week mean glycemia, and the change in GHb
concentrations,
especially during the first weeks of therapy, correlated
significantly
with the change in mean
glycemia. These findings cannot be attributed to interCLINICAL CHEMISTRY, Vol. 40, No. 7, 1994
1319
25
-J
A
0
0
.
.
E
E
S
E20
E
w
15
0
0
>. 10
> 10
.
00
a)
.5
.
a)
S
5
r=
r .66
.66
p.001
p s .001
0
0
5
10
15
20
25
25
-J
-J
0
0
10
15
20
25
30
10
15
20
25
30
35
E
E
E20
E
CD
m
0
CD15
0
#{149}
10
00
.
a,
00
a)
C’J
N
0
0
5
10
15
20
25
-J
-J
0
0
E
E
E
E
CD
m
CD
0
.
0
00
00
00
a,
C,,
C)
0
5
10
20
15
25
10
group
antecedent
glycemla, weilc*
Slope
Intercept
GHb, %
1
2
3
6
6
0.66
0.70
0.72
0.83
0.84
(0.39-0.82)
(0.47-0.84)
(0.49-0.85)
(0.68-0.91)
(0.72-0.91)
0.87
0.89
0.91
0.98
0.74
0.66
0.64
0.60
0.49
0.64
(0.39-0.82)
(0.33-0.81)
(0.31-0.79)
(0.16-0.72)
(0.41-0.79)
0.85
0.78
0.74
0.66
0.47
0.27
-0.31
-0.28
-0.36
1.51
GPP, %
1
2
3
6
6
‘All correlations significant at P <0.001.
“95% confIdence intervallistedIn parentheses.
20
25
30
-10.9
-9.2
7.9
5.4
2.1
short-term control if evaluated
at frequent
an intensive monitoring program.
intervals
in
The relation between
Glib concentrations
and recent
glycemic excursions
has been reported recently (28).
Furthermore,
two other studies support our findings
that Glib may reflect short-term
changes.
Lee et al.
(29), using boronate affinity methodology,
demonstrated
that Glib and GPP assays provide the most usefulcliiiical indicators of short-term changes in glucose control.
Mortensen
and V#{216}Lund
(30) concluded that weekly determination
of HbA1C could be used as a supplement
to
preprandial glucose concentrations
to evaluate the effectiveness
of intensive
treatment
regimens.
Our data
__________
1320 CLINICAL CHEMISTRY, Vol. 40, No. 7, 1994
35
Fig. 2. Correlationof GHb
(A-C) and GPP (D-F) with
the mean weekly (A, D),
2-week (B, E), and 3-week (C,
F) mean CBG in Group I (intensiveintervention)patients.
ference from the Schiff base reflecting recent or current
glycemia, because the Schiff base is not detected by this
method. The results suggest that GHb determined
by
automated
affinity
HPLC may serve as a marker of
Table 1. CorrelatIon of antecedent
glycemla with
glycated
blood protelns.a
Duration of
ClInIcal
15
GPP, %
GHb, %
II
35
agree with the observation
noted in previous reports
and suggest that the GHb assay we used may provide
the same clinical information as the GPP assay. Therefore, one may have little need to measure
GPP if a
highly precise Glib assay is available.
We note, however, that an acute decrease in glycemia in this report
was accompanied by significant changes in an increased
GHb (our patients were considered in poor control initially). Whether Glib at a normal or near-normal value
on initial measurements can detect glycemic excursions
as well as short-lived proteins do need to be addressed.
In conclusion, this boronate affinity HPLC method for
the determination
of Glib and GPP is rapid (run time
<2 mm) and precise (CV s2%), and provides a valid
objective measure to assess both short- and long-term
antecedent control.
We are especially
grateful
to Beth Kivett for editorial assistance
and manuscript preparation.
References
1. Brinchmann-Hansen
0, Dahl-Jorgensen
K, Sandvik L, Hanssen KY. Blood glucose concentrations and progression of diabetic
retinopathy: the seven year results of the Oslo study. Br Med J
1992;304:19-22.
2. Dahi-Jorgensen
K, Bioro T, KierulfP, Sandvik L, Bangstad HJ,
Hansaen KY. Long-term glycemic control and kidney function in
insulin..dependent diabetes mellitus. Kidney mt 1992;41:920-3.
3. Reichard P, Berglund B, Britz A, Cars I, Nilsson BY, Rosenquist U. Intensified conventional insulin treatment retards the
microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS)
after 5 years. J Intern Med 1991;230:101-8.
4. The Diabetes Control and Complications Trial Research Group.
The effect of intensive treatment of diabetes on the development
and progression of long-term complications in insulin-dependent
diabetes mellitus. N Engl J Med 1993;329:977-86.
5. Gabbay KR, Hasty K, BreslowJL,et al. Glycosylated hemoglobins and long-term blood glucose control in diabetes meffitus. J
Clin Endocrinol Metab 1977;44:859-64.
6. Gonnen B, Rubenstein AH, Rochman H, et al. Hemoglobin A1:
an indicator of metabolic control of diabetes patients. Lancet
1977;il:734-7.
7. Lester E. The clinical value of glycated hemoglobin and glycated plasma proteins. Ann Clin Biochem 1989;26:213-9.
8. Armbruster DA. Diabetes mellitus, protein glycation and laboratory assays for monitoring long-term glycemic control. Clin Lab
Set 1990;3:184-8.
9. Fl#{252}ckiger
R. Glycated haemoglobins. J Chromatogr 1988;429:
279-82.
10. Kilpatrick
ES, Small M, Rumley A, Dominiczak MH. Increased fetal hemoglobin in insulin-treated diabetes mellitus contributes to the imprecision of glycohemoglobin measurements.
Clin Chem 1993;39:833-5.
11. Aleyassine H. Low proportions
of glycosylated hemoglobin
associated with hemoglobin S and hemoglobin C. Clin Chem
1979;25:1484-6.
12. Fairbanks V. The incidence of hemoglobin variants and their
effect on glycated hemoglobin assay results. In: The role of glycated hemoglobin in the management of diabetes. Cleveland: Case
Western Reserve University, 1988;28-31.
13. Abraham EC, Stallings M, Cameron BF, Huisman THK.
Minor hemoglobins in sickle cell heterozygotes and homozygotes
with and without diabetes. Biochim Biophys Acta 1980;625:109.
14. Fairbanks VF. Measurement of glycosylated hemoglobin by
affinity chromatography. Mayo Clin Proc 1983;58:770-4.
15. Wiley DG, Rosenthal MA, Caldwell S. Glycosylated hemoglobin and plasma glycoprotein assays by affinity chromatography.
Diabetologia 1984;27:56-8.
16. Abraham EC, Perry RE, Stallings M. Application of affinity
chromatography for separation and quantitation of glycosylated
hemoglobins. J Lab Clin Med 1983;102:187-97.
17. Scott KS, Peterson CM. Evaluating glycated protein by highperformance
boronate affinity chromatography.
Diabetes Prof
1990;Summer:20-4.
18. Sullivan K, Raskin P. Evaluation of a new automated affinitychromatographic method for the measurement
of glycated hemoglobin. Diabetes Res Clin Pract 1991;13:103-6.
19. Herold CD, Andree K, Herold DA, Felder RA. Robotic chromatography: development and evaluation of automated
instrumentation for assay of glycohemoglobin. Clin Chem 1993;39:
143-7.
20. Little RR, Wiedmeyer H-M, England JD, Wilke AL, Rohifing
CL, Wians FH, et al. Interlaboratory standardization of measurements of glycohemoglobins.
Clin Chem 1992;38:2472-8.
21. US Patent no. 07/774646; patent pending.
22. Rendell M, Rasbold K, Nierenberg J, Krohn R, Hermanson G,
Klenk D, et al. Comparison and contrast of affinity chromatographic determinations of plasma glycated albumin and total
glycated plasma protein. Clin Biochem 1986;19:216-20.
23. Rendell M, Kao G, Mercherikunnel P, Petersen B, Duhaney R,
Nierenberg
J, et al. Use of aminophenylboronic
acid affinity
chromatography
to measure glycosylated albumin levels. J Lab
Clin Med 1985;105:63-9.
24. Cohen MP. In: Diabetes and protein glycosylation-measurement and biologic relevance. New York Springer-Verlag, 1986:5-
16.
25. Brownlee M, Viassara H, Ceranu A. Nonenzymatic glycosylation and the pathogenesis of diabetic complications.
Ann Intern
Med 1984;101:527-37.
26. Monnier M, Vishwanath V, Frank KE, Elmonts CA, Dauchot
P, Kohn RR. Relation between complications of type I diabetes
mellitus and collagen-linked fluorescence. N Engl J Med 1986;314:
403-8.
27. Cerami A, Viassara H, Brownlee M. Role of advanced glyco..
sylation products in complications of diabetes. Diabetes Care
1988;11(Suppl 1):73-9.
28. Tahara Y, Shima K The response of Glib to stepwise plasma
glucose change over time in diabetic patients. Diabetes Care
1993;16:1313-4.
29. Lee PDK, Sherman LD, O’Day MR, Rognerud CL, Ou C-N.
Comparisons of home blood glucose testing and glycated protein
measurements.
Diabetes Res Clin Pract 1992;16:53-62.
30. Mortensen HB, V#{248}Lund
Aa. Application of a biokinetic model
for prediction and assessment of glycated hemoglobins in diabetic
patients. Scand J Clin Lab Invest 1988;48:595-602.
CUNICAL CHEMISTRY, Vol. 40, No. 7, 1994
1321