Jun 6, 1995 - quality-control survey on glycohemoglobin. (GHb), monitored in France by the. Soci#{233}t#{233}Fran#{231}aise de Biologie. Clinique.
CLIN.
CHEM.
41/11,
1644-1648
Glycohemoglobin Philippe
We
the
Dominique
results
Evaluated
Labb#{233},2Glues
of a national
glycohemoglobin
authority
of the
lyophilized
“Agence
hemolysate
du
in France
and
and
sent
utilization by clinicians (QC)
the
laboratories.
Results were obtained from 2770 laboratories. HbA1C, HbA1, and total GHb were measured by 50%, 24%, and 26% of the participants, respectively. Of these measurements, 79% of the HbA1C results and 76% of the total GHb results, but only 48% of the HbA1 results, were within the ± 20% limits of the indicated target values. Mean values for the hemolysate ranged from 8% to 11 % for
HbA1C,
from
13% for total 3% to 20%, used
according
GHb
for
ples,
7% to 12% for HbA1, GHb. The interlaboratory assay,
which
very
different
exhibit
Nonetheless, techniques
to
and from 11% to CV5 ranged from used. So, methods
method are
based
on various
analytical
this large-scale study indicates that can support transferability of results
laboratory
surveys
some from
to laboratory.
We report GHb assays
source
Terms:
comparison/variation,
Materials
QC material. The material used lyophilized hemolysate obtained
and
after 15 mm This material Before
its
of glycohemoglobin
major component index of metabolic
(GHb),
and
partic-
HbA1C, is considered a control in diabetes melli-
tus.3 GHb values are used as routine retrospective markers of glycemic status, reflecting diabetic balance of 4 to 8 weeks before blood collection (1). The optimal use of these biological markers in public health implies the use of reliable, well-controlled methods (2, 3). However, many techniques are used in clinical biology laboratories, some of them measuring all glycohemoglobins, other measuring specifically
or the bulk
of HbA1.
These
methods
are based
on
various principles that influence the analytical performances of the technique: ion-exchange chromatography, electrophoresis, immunology, and affinity chromatography (4). In the absence of reference methods and (or) reference materials for GHb, there may be a large discrepancy of results obtained from laboratory to laboratory. The only way to evaluate the respective analytical performances of the different methods and their actual ‘Laboratoire Central de Biochimie, H#{244}pital Robert Debr#{233}, CHU Reims, Avenue du G#{233}ndral Koenig, F-51092 Reims Cedex, France (address for correspondence). Fax 33.26.78.85.39. 2 Laboratoire de Biochimie A and Socidt#{233} Francaise de Biologie Clinique, H#{244}pital Necker-Enfants Malades, Paris, France. Nonstandard abbreviations: SFBC, Societe Francaise de Biologie Clinique; Hb, hemoglobin; GHb, glycohemoglobin; and QC, quality control. Received June 6, 1995; accepted August 4, 1995. 1644
CLINICAL
CHEMISTRY,
Vol. 41, No. 11, 1995
were
pool The
HbA1C and
concentration
to help
prevent
specific
materials
determined added of the
a
desired
were are
of a
having the
was
denaturation used
he-
portions
obtaining
lyoprotectants
donor before
erythrocyte
hemolysate until
a SA
with 0.5 foaming,
isolated
various
of erythrocyte
CryoThe
with
concentration,
HPLC. material
and
a normal
supplemented
targets.
bin.
washed
lyophilization,
was
was
processed as a patient’s sample. was prepared from multiple
which
HbA1C
for QC survey from Bio-Rad
France), to be reconstituted water, gently mixed without
concentrated
Measurement
HbAIC
and Methods
molysate
of
auspice Labora involve over th referenc pro
cedure.
high
ularly valuable
the results of a national QC survey o monitored in France by the Soci#{233}t#{233} Fran
erythrocytes,
diabetes/intermethod
quality-contro
#{231}aisede Biologie Clinique (SFBC), under the of the Agence du M#{233}dicament (Direction des toires et des Contr#{244}les). This large-scale study >2700 laboratories of clinical biology all country. Because of the lack of international method, this survey was run as an experimental
lysing.
Indexing
is to undertake
(5).
(Ivry-sur-Seine, mL of distilled
princi-
performances.
Survey
Vassault2
by
on behalf of the A sample of
to 3109
Anne
Quality-Control
survey
M#{233}dicament.”
was
Utilization
in a Large-Scale
Dumont,2
quality-control
(GHb), monitored Soci#{233}t#{233} Fran#{231}aise de Biologie Clinique on
Management
#{149}Laboratory
Assays
Gillery,’
report
(1995)
by to
the
hemoglo-
proprietary.
Before distribution, the technical characteristics o the QC material were verified to be suitable for use with the various methods of GHb determination (i.e., ion-exchange chromatography, electrophoresis, immunology, affinity techniques), as described in a previous study of the Glycated Proteins Committee of SFBC (6). A pathological value for GHb was chosen for this study, the expected HbA1C value, as determined by the HPLC separation method, being 10% of total Hb. Target values for the various glycated Hb subspecies were checked with different methods in reference laboratories (Table 1). In the absence of approved reference methods, we chose the values obtained by HPLC as indicative target values for the purpose of the study. The values obtained for HbA1C and HbA1 were respectively 10.0% and 11.0% of total Hb, whereas a target value for total GHb could not be determined because o
high case,
variability we decided
value
of
all
results
(11.7%). The stability checked, able control
to avoid to
transport material
between to use
the methods as the target
obtained
of QC
material
problems
during
the
during
storage
of interpretations
conditions. under
tested. In this value the mean
three
GHb different
QC
survey was
alse
attributwas
assayed conditions:
in on
Table
1. Setting
target
values
Mean
Method
truncation discarded the retained results, between were processed for further
for the QC material. SD
±
Manulacturer
Target
value
% of total Hb
HbAIC
Diamat
(Lab.
Bio-Rad
10.0
±
0.1
Bio-Rad
10.5
±
0.2
±
0.4
1)
HPLC Diamat (Lab.
2)
Immunological method (Tinaquant)
Boehringer Mannheim
9.8
Of the 3109 laboratories that received QC material, 2770 sent results (89% response). Each individual result was evaluated in comparison with (a) the indicated target value and (b) the peer-group method mean value. With respect to (a), the results of each laboratory were judged acceptable if within ± 20% of the target
11.0 Diamat
Total GHb Automatic affinity method (IMx) Affinity chromatography (glycaffin minicolumn) a
n
=
5 each.
b
Mean
value
Bio-Rad
11.0
±
0.3
Abbott
12.2
±
0.9
Eurobio (Isolab)
10.4
±
1.3
11.7”
value.
Only
tained results
results varied
Hb during
QC survey
(see text).
but
With
day
of reconstitution
(basal
conditions),
on
day
7
only
regard
2 shows
the
the
results
proved
method
and
anonymous personal number. An updated list of methods commercially available and currently in use was enclosed with the report form. Each technique used was identified with two characters: The first one indicated the principle of the method (e.g., cation-exchange chromatography); the second one was specific for the method. Results could be expressed as percentages of HbA1C, HbA1 or total GHb, according to the method used. Protocol of QC survey and data collection. A vial of QC material was sent by mail in July 1993 together with the QC survey form and technical information to
liquid
survey tory, result.
report
form
coding
of
The
form
the was
as
to
identification
method,
and
returned
by
of transcription mail
within
the
laboraof
the
5 days
to
the Agence du M#{233}dicament, where results were collected. Statistical calculations were performed by the SFBC center. Statistical calculations. The peer-group mean values were calculated from was complete information according to the following
the
results for which there about the method used, procedures: A nonparametric
obtained
the was
all
The
and
Many
or
were used
absence
of
eliminated. for
component
laboratories
Table
laboratories.
of results
2092.
acceptable. means,
of error
coding
number
were
method
because
analyte
1056
results,
from
unusable
calculations in
of HbAI peer-group
(or)
Eventually, HbA1C
48% to the
data
after reconstitution and storage at 4 #{176}C (extreme conditions of storage in the laboratory), and on the day of reconstitution of a lyophilized sample stored 7 days at 37 #{176}C (extreme conditions of transport duration and temperature). These storage conditions did not alter GHb values, as determined by HPLC, immunoassay, or affinity methods (data not shown). GHb survey form. A survey form was sent with the QC material. Each laboratory was identified by an
the 3109 participating laboratories. The participating laboratories were determined on the basis of the answer to a questionnaire investigating the actual measurement of GHb by each laboratory, sent during a previous QC survey. Each laboratory performed the GHb assay by its usual method and filled the QC
71.5% of participating laboratories obin this range. The number of acceptable according to the glycated component of 79% of HbA1C results and 76% of GHb
assayed:
results, obtained
of the mean The imprethe between-
Results
HbAI HPLC
outlying data. The and 0.975 fractiles, calculation: num-
ber of results, SD, and standard estimate (S), obtained after recurrent truncations, cision of the methods was defined as laboratory CV for each method.
10.0
HPLC
1% most the 0.025 statistical
statistical
measured
(50.4%
of
was
laboratories),
HbA1 in 501 laboratories (23.9%), and total GHb in 537 laboratories (25.7%). Various methods were used, principally for the HbA1C assays. Cation-exchange chromatography represented 57.8% of the methods used for determination of HbA1C (58% HPLC or low-pressure chromatography,
phoresis
3.1%,
mated sults)
affinity
method
29.0% or
measured or
its
columns
minicolumns), assays
(yielding
was
chromatography by
electrophoresis
by methods derivatives, (75.6%),
by
affinity
or
by
and
measured (63.5%
(36.5%).
based
electro-
10.1%,
calculated
HbA1
of methods.
cation-exchange ries)
42%
immunological
of
Total
on affinity
re-
either
by
laboratoGHb
was
for boronic
acid
chromatography automated
auto-
HbA1C
methods
on
mini-
(24.4%).
Chemical methods were not frequently used (n 11). The accuracy and the imprecision of the different methods used were highly variable, even for a single group of techniques based on the same principle (Fig. 1). For HbA1C assays, mean values ranged from 8% to 11% of total Hb, and CVs ranged from 2% to 18%. Among the methods involving cation-exchange chromatography, the automated methods, HPLC and lowpressure liquid chromatography (code 2), showed the best precision (CVs generally between 2% and 5%). The precision depended on the system, the difference of means obtained by different HPLC methods varying in a range of 20%. Minicolumns (codes CX) yielded generally results lower than the indicated target value, and showed a much greater imprecision, with CVs ranging from 12% to 19%. Results obtained with immunological methods (codes FIX) showed mean values close to the target value (±5%) and CVs ranging between 2% and 8%, whereas agarose gel electrophore=
CLINICAL
CHEMISTRY,
Vol. 41, No. 11, 1995
1645
Table
2. Results
from
all the participating
laboratories
for HbA1C,
HbA1,
and total
% of total Methoda
Code
n
Mean
SD
Median
GHb.
Hb 0.5 percentile
99.5 percentile
HbA1 Overall
mean
1442
Helena Hb glyquee Sebia HbA1 Electro Other
REP (1) (2)
electrophoresis
Menarini Ciba
HA 8110
Corning
Merck Bio-Rad Bio-Rad Other Eurobio
(7)
Clevenot
Abbott
(6)
system Hb
glyquee
(5)
IMx (8)
Other affinity chromatography Bio-Rad HbA,C macromethod Bio-Rad
HbA1C micromethod
Bio-Rad
HbA1 mini-col
Realef
Chembio
Helena
Hb glyquee
Fumouze
(6)
(6)
Glyco-Sep
A1
(9)
(1)
ion-exchange
Other
(6)
chromatography
A1 (10)
2000
Bayer-Ames
DCA
Boehringer
Tinaquant
Biotrol
(12)
Hb glycosylee
Unidentified
(11) A1
(9)
4.6
18.8
5.1
23.6
11.8
7.6 8.4
16.3
32 18
8.3 8.7
8.3
0.4
6.8
9.6
8.7
0.2
8.2
9.2
2H
17
7.7
7.5
0.7
2K 2M
84 81
10.1 9.9
10.1
0.3
7.0 9.2
10.0
0.4
8.9
2X
123
9.5
9.6
1.2
6.5
16.9
9.3 10.7
Al
46
13.1
13.0
2.4
6.7
11.1 10.9 18.8
AM AT
25 310
10.8 8.8
10.8
3.3
3.9
22.5
8.9
1.1
3.5
16.0
AX
26
9.9
9.3
2.8
5.0
16.1
CD CK
58 166
8.5
8.6
1.5
6.7
14.5
8.8
1.5
5.6
CO
11
8.8 8.7
8.6
0.8
7.7
20.0 10.1
CR CS
36 17
8.1 10.6
7.9 10.5
1.6
5.2
11.9
1.6
7.0
14.2
CX
45
8.6
43
6.8
2.0 1.9
1.1 4.2
13.6
CZ
8.6 6.9
HB
37
10.3
0.3
9.7
HO PF
62 ii
10.3 9.6
9.8 10.0
0.8 6.2
6.8 2.0
11.0
11.4 9.5
9.7
2.0
2.2
17.1
18.6
XX
technique
1.9 3.0 2.1 2.3
(4)
(5)
9.3 11.1
9.8
2D 2G
HPLC
9.2
10.9 12.1 9.7
34 13
(3)
Glycaffin
Merck
iS 1X
Diamat (6) modular system HPLC
35
system
Glycomat
Clevenot
1H
112
13.1 10.8 19.5
HbA1 592
9.7
10.1
2.6
3.7
62 112
11.4 12.3
11.3
1.5
7.1
16.3
12.2
2.2
4.5
20.0
1X 2X
14 56
12.3 10.2
12.1
10.7
2.1 1.6
7.7 3.2
15.5 12.6
(7)
Al
17
12.7
Boehringer A1 (12) Eurobio HbA1 rapides
CB
19
10.6
12.8 10.3
2.2 1.0
8.3 8.0
18.6 11.7
13.0
Overall
mean
Helena
Hb glyquee
Sebia HbA1 Electro Other electrophoresis HLPC systems Eurobio
Glycaffin
Other
1H
(2)
iS system
(7)
ion-exchange
Fumouze
chromatography
A1 (10)
Unidentified
Total
REP (1)
technique
CI
11
9.0
CX
41
9.9
9.3 10.0
1.5 1.8
7.5 6.8
CZ
202
7.2
7.2
1.4
3.7
12.7
XX
58
10.3
10.7
2.0
5.8
15.0
GHb
Overall
mean
Helena
Hb
Sebia
HbA,
Eurobio
2.5
4.5
22.2
11.6
11.6
2.3
5.4
17.0
(2)
iS
28
11.7
11.5
2.1
9.1
18.7
11(7)
AH Al
47 197
10.9 13.1
10.9
7.4 1.3
22.2
13.1
2.5 2.2
2H
127
11.1
11.1
2.0
4.7
23.0
AT AX
132 29
11.3 12.8
11.2
1.3
7.3
14.6
13.0
4.5 8.8
19.6
REP
Electro
Abbott
IMx
Other
affinity
Hb
Other
glyquee
(5)
(8) chromatography
Hb glyquee
Helena
(1)
(7)
Clevenot
Sigma
11.6
36
736
Glycaffin
Merck
11.7
1H
glyquee
Glycotest
Eurobio
totale
(13)
Hb glyqu#{233}e(1) ion-exchange
Fumouze
chromatography
A1 (10)
Unidentified a Addresses
technique of suppliers
Cergy-Pontoise,
(all in France)
(5) Nogent-sur-Marne,
corresponding (6) lvry-sur-Seine,
CLINICAL
CHEMISTRY,
22.0
A6
20
13.4
13.7
2.7 1.8
CS CX
11 15
9.1 10.3
9.2 10.0
1.7 2.3
7.4 6.8
13.3
CZ
33
7.3
7.2
1.5
4.5
11.2
XX
61
3.0
5.2
to the numbers (7) Les Ulis,
Fallavier.
1646
15.6
Vol. 41, No. 11, 1995
10.6 in parentheses:
(8) Rungis,
10.8 (1) Saint-Leu
(9) Paris,
Ia For#{234}t. (2) Issy-les-Moulineaux,
(10) Asni#{232}res,(ii)
Puteaux,
(12) Meylan,
15.5
13.9
23.0 (3) Chevilly-Larue, and
(13) Saint-Ouentin
(4)
Mean GHb(%
of total Hb)
A
14
B
C
#{149}
#{149}
12 #{149}
S
#{149}
10
#{149} S
#{149}‘
8
#{149}
#{149}#{149} #{149}
6
AT
2M 2K
20
of vanation
Coefficient
2H
cK
01
2(3
CD
HB 1H
2K HO
Cl
iS
CB
CZ
J
AT
1H
AH
AS AM
Fig. 1. Accuracy and imprecision of GHb assays: mean values (A, B, C) and CVs (D, E, F) for the various methods evaluating HbA1C (A, D), HbA1 (B, E), and total GHb (C, F).
(%)
20
16
The horizontal lines represent the target values. The methods are identified as in Table 2. Al, AH, AM, and A6 (C, F) are affinity methods (minicolumns); AT (A, C, 0, F) is an affinity method (automated); 2K, 2M, 20, and 2H (A, B, 0, E) are HPLC ion-exchange methods; 2G (A, 0) is a low-pressure ion-exchange liquid chromatographic method; CK, CD, CR, CB, Cl, and CZ (A, B, 0. E) are ion-exchange chromatographic minicolumns; 1H and iS (A, B, 0, E) are electro-
12 8. 4.
0. CK
AT2M2H
2K
20
20
CD
CRHB 1H
2K
ce
HO
a
a
is
P
AT
1H
AH
(code 1H) gave higher mean values and CVs (13%). The results of calculated HbA1C obtained by an automated affinity technique (code AT) gave a mean value 10% lower than that obtained by HPLC or immunology, with a CV of 10%. HbA1 assays were characterized by a high number of unacceptable results: Only 48% were within ± 20% of the target value. After elimination of the extreme aberrant values, electrophoretic methods (codes 1X) had mean values near target value but CVs >15%. Results obtained with minicolumns (codes CX) varied according to the kit; they were characterized by high CVs (20%) and, for technique CZ, a very low mean value. All GHb assays were based on the same principle of
text,
affinity.
conditions
sis
and
Their were
obtained
mean
values
compatible by
the
ranged
with
specific
the automated method umns (13% to 18%).
the
methods.
AT
from
11%
HbA1C CVs
(11%)
to
mean were
than
lower
with
13.5% values with
minicol-
46
the
large
far,
Thus materials lack
of the
disease.
The
availability
generated long-term of reliable
by diabetes complications assay
methods
for measuring biological markers of diabetes is of major importance, because these assays allow diabetologists to better monitor the glycemic balance of their patients, thus delaying the onset of complications. Since the introduction of the first chromatographic methods into clinical laboratories -20 years ago, many techniques have been developed, based on various principles and equipment. This report provides an overview of the actual situation in July 1993. Moreover, the principle of the methods used affected the glycated Hb components actually assayed. In this con-
general
agreement for
of standardization,
materials, The
only
an
way
of their
quality The
major
results
obtained
matter
of fact,
the
of
lower
scale,
our
study,
the but
of the
isfactory
ries, The
those
results
for
the
pointed
out
obtained
by
during
The ferent
the as
analytical
methods
huge
such
better
particularly incompatible
clinical
use
CLINICAL
a In
only
the
a particular
to
the
effective
provide
in many
unsat-
laboratois a critical
methods which by
generates biologists,
of aberrant
operated
as
answers
exhibit
Obviously, precision
minicolumns
imprecision
its
case at
study.
techniques,
and
the
not to
can
number
performances. as
a
routine
not
include
attributable
routinely
provide
niques
As
the
laboratories.
comprehension
this
methods
is
of results,
of method
to report
undertaken
linked
transferability
a confusion
in
are
if it is not sufficiently robust. large number of available
factor
was
which
Any technique implemented
when
the
surveys.
measured
obtained
technique.
re-
and
of laboratories.
intrinsically also
QC
specialized
results
of variation
of
used
study
surveys between
calibration
dispersion
laboratory,
e.g.,
The
the
methods
present
was
on reference
for
number
GHb
any
interlaboratory
exists
is to perform of the
value
measurements.
the
in a large
when
for
unavoidable
results
between
determination monitoring blood
HbA1C
GHb
to control
interest
(A, 0) are
HO
found
particularly
induces
sults.
and
semiological
though index
techniques
no and
HB
results
effective
of some GHb assays, even is considered the reference glucose equilibrium (7).
use
problems of public health are associated with the
of the
and
methods.
variability
decreases
technique,
The meffitus
methods;
immunological
laboratories
factors
Discussion
phoresis
frJ
than
or the with
(3, 8). We CHEMISTRY,
very
manual
tech-
electrophoresis.
HbA1 the
suggest
Some
methods, goals
dif-
automated
show of the
that
an
method
techniques
Vol. 41, No. 11, 1995
1647
evaluating any
HbA1
more,
as
because
a whole (a)
should
their
not
general
be
performed
performances
HbA1 fraction. On the basis of the results obtained, however, some of methods evaluated in this study seem able to support transferability of results from laboratory to laboratory. Confirmation of the present results by future
but
rather
a target
based
value
the results obtained by HPLC and confirmed with different available methods. The choice of HPLC values was dictated by the fact that HPLC techniques allow the specific evaluation of HbA1 and are considered good candidates for reference techniques (4, 5, 9). However, each result was compared not only with the HPLC target value but also with the peer-group method mean calculated for every group of techniques, provided the method used was correctly identified on the survey form. Mean values in this study were widely dispersed, even in a single group of techniques, such as HPLC. This betweenmethod
dispersion
cannot
be
substantially
improved
un-
til an international standardization of GHb assays is achieved. As a temporary solution, a systematic calibration of techniques by convenient materials could be used (9-12). This is supported by the fact that methods based on very different principles, such as HPLC and immunology for HbA1C assays, provide comparable results if the immunological method is calibrated with a HPLC-titrated that
calibrator.
HbA1C calibration
Moreover,
of the
methods
others
have
improves
study
was
showed methods, (6).
not
exactly
comparable
interlaboratory
in particular
Moreover,
experimental
the
CLINICAL
to
the
in same
material
conditions
and mailing). Nevertheless, ferred to fresh biological particularly for accuracy Finally, this study (15) 1648
different properties from explain some differences the material used in this
identical behavior
CHEMISTRY,
patients’
samples,
several
of the
(high
well number
conclusions
samples studies. confirms
without the
it
various
chromatographic was
pattern
adapted
to
of
participants
cannot
be
our
trans-
modification, absolute
Vol. 41, No. 11, 1995
we
neces-
of
here,
techniques,
as
statistical
and
for
recommend only
or
that
in
should
well
as
general
good
not
GHb
assays
technical
be
and
should financial
performed.
This work was monitored by SFBC on behalf of the authority of the Agence du M#{233}dicament (Direction des Laboratoires et des Contr#{244}les) and was made possible by complementary grants of Bioforma. We thank A. Nicolas for technical assistance and S. Etienne
for
typing
the
manuscript.
References 1. Larsen ML, H#{248}rder M, Mogensen monitoring of glycosylated hemoglobin dent diabetes mellitus. N Engl J Med 2. Goldstein DE, Little RR, Wiedmeyer ing CL, Wilke AL. Is glycohemoglobin mellitus? Lessons from the diabetes trial. Clin Chem 1994;40:1637-40. 3. Kolatkar
NS,
Intensive
Cembrowski
diabetes
GS,
management
glycohemoglobin 4. Gillery P,
[Letter].
EF. Effect levels in 1990;323:1021-5.
HM,
Clin
Chem
England
JD,
Rohlf-
useful in diabetes and complications
testing control
Callahan
requires
of long-term insuim-depen-
PL, very
Etzwiler
precise
DD.
testing
of
1994;40:1608-9.
Guillemin C, Delpech M. Hemoglobine m#{233}thodes de dosage et probl#{232}mes de standardisation Ann Biol Clin 1994;52:157-63.
glyquee: [Review).
5.
Goldstein
Little
DE. lege
RR,
Wiedmeyer
Interlaboratory of American
HM,
England
comparison Pathologists
JD,
Naito
HK,
of glycohemoglobin survey data. Clin
results: ColChem 1991;37:
1725-9. 6. Guillemin C, Gillery P, Hue G, Bordas-Fonfr#{232}de M, Dumon MF, P#{233}rierC, et al. Probl#{232}mes poses par le contr#{244}lede qualite des dosages d’hemoglobine glyquee. Inform Sci Biol 1993;N19:6-11. 7. Steffes NW. can we measure Lytken-Larsen
for glycated 9. Bodor Nahm
goal
Little
setting
Standardization
in the clinical 1994;38:2414-8.
but
Blaabjerg
haemoglobin.
MH.
it,
0, Hyltoft Petersen P, Hansen H, prior to selection of a method Scand J Clin Lab Invest 1990;50:715-51. RR, Garrett N, Brown W, Goldstein DE,
M,
GS,
we can define 1995;41:180-1.
Glycemic control: perhaps it [Editorial]? Clin Chem
H#{248}rden M. Analytical
CVs (12) and is the best way at present to ensure standardization of results (9-13). On the other hand, we found in this study that calculations for obtaining HbA1C values from the assay of total Glib, a different glycated component with a completely different method, were actually possible, provided that a systematic calibration was performed (14). However, the correction used in some affinity methods to provide a calculated HbA1C value may be biased when used with control preparations showing patients’ samples-which may of accuracy (13). Even though
rule,
performed
conditions,
8.
reported
as
data at the level of every laboratory. The withinlaboratory QC is an integral part of good laboratory practices and is central point for the quality of the results. The implementation of still nonrobust methods such as Glib assay methods in a laboratory may induce higher costs of reagent and (or) equipment and higher control necessities (16). Every analyst should be aware of these causalities before choosing a method. As a be
on
QC
studies,
general
studies could permit the recommendation of selected methods relevant for clinical use. The problem of the accuracy of the methods remains. In the absence of reference methods and (or) materials, we could not use in this study a definitive target value for our preparation,
intensive
an
large-scale
not satisfactory; (b) the principles used for HbA1 also allow HbA1C measurements; and (c) they constitute a factor of confusion for clinicians. Furthermore, HbA1C is generally considered as the reference marker and is less subject to interferences than is the heterogeneous
control
of
sity are
laboratory:
of glycohemoglobin three
years
determinations
of experience.
10. Little R, Erhart hemoglobin
RR, England JD, Wiedmeyer HM, McKenzie PM, et al. Interlaboratory standardization determinations. Clin Chem 1986;32:358-60.
11. Little CL, Wians
RR, Wiedmeyer HM, England JD, Wilke FH, et al. Interlaboratory standardization
Clin
EM,
Chem
Mitra
of glycated
AL, Rohifing of measure-
ments of glycohemoglobins. Clin Chem 1992;38:2472-8. 12. Weikamp CW, Penders TJ, Muskiet FAJ, van der Slik W. Effect of calibration on dispersion of glycohemoglobin values determined by 111 laboratories using 21 methods. Clin Chem 1994;40:138-44. 13. Weykamp CW, Penders TJ, Miedema K, Muskiet FAJ, van der Slik W. Standardization of glycohemoglobin results and reference values in whole blood studied in 103 laboratories using 20 methods. Clin Chem 1995;41:82-6. 14. Turpeinen glycohemoglobin
U,
Karjalainen compared.
U, Stenman Clin Chem
U-H. Three 1995;41:191-5.
assays
for
15. Gillery P. Labb#{233}D, Maurice-Estepa L, Dumont G, Vassault A. Hemoglobines glyqu#{233}es. Ann Contr#{244}le Qual Nati Biochim 1993;13:11-28. 16. Gillery P, Endler of HbA1C on routine 39:1076-9.
AT, Rubi K, Scheurmann clinical chemistry analyzers.
T. Determination Clin Lab 1993;
