![]()
12
| 3 | 2008
Klinisk Biokemi i Norden
(Fortsat fra side 10)
The present situation as reflected in inter compa-
rison studies
The overall purpose with the development of a
reference measurement procedure is to improve the
field methods by making HbA1c values from dif-
ferent procedures and from different time periods
more comparable. With the introduction of methods
where results are traceable to the IFCC procedure
we therefore expect to see an increased agreement
between results in external quality assessment (EQA)
schemes.
For the Nordic countries we have an excellent opp-
ortunity to monitor the effect of the change of stan-
dardization from the results of our EQA programs.
We have therefore collected HbA1c data reported
in the Labquality and EQUALIS HbA1c schemes for
2007. Both organizations distribute 10 fresh whole
EDTA-blood pools a year. EQUALIS distributes 1
sample in 10 surveys, while Labqality distributes 2
samples in 5 surveys. In Sweden all reported HbA1c
results are traceable to Mono S standardization (9).
For the ease of comparison of the results from the
two schemes, all results in EQUALIS scheme have
been converted to NGSP levels by the equation
“HbA1c(NGSP) = 0,96 x HbA1c(Mono S) + 1,18”
(this equation is derived from the two equations
above) prior to all calculations. Results outside mean
+/- 3SD calculated from the raw data for each distri-
bution in the two schemes are excluded as outliers.
The target value in the EQUALIS scheme is the mean
result from 5 Mono S expert laboratories. The mean
target value for the 10 distributions was 6.8% (range
5.0 – 8.4%). The target value for materials distributed
by Labquality is the NGSP value assigned by European
Reference Laboratory (ERL) in the Netherlands. The
mean target value for the 10 distributions was 7.5%
(range 5.3 – 10.2%). The deviation from the target
value for each reported result is used to calculate the
mean and 95% confidence limits for the mean for the
major method groups in the 10 distributions (figure
3 and 4).
Fig 3. Mean deviation from Mono S target during 2007 for the
major method groups in the Swedish EQA scheme fot HbA1c.
‘Afinion’ is Axis Shield Afinion AS100 Analyzer, ‘DCA 2000’ is
Siemens DCA 2000, ‘DCA Vantage’ is Siemens DCA Vantage,
‘Mono S (exp)’ is the group of Swedish Mono S HPLC field
methods from which the target is calculated, ‘Mono S’ are the
HPLC systems independet from the target forming systems,
‘Tq/Unimate’ is Roche Tinaquant and Unimate reagents on
Roch analytical platforms, ‘Tosoh’ is Tosoh G5 and G7 HPLC
system from Tosoh Bioscience, ‘Variant II’ is Variant II HPLC
from Bio Rad, ‘Variant Turbo’ is Variant Turbo HPLC from
Bio Rad. ‘N’ denotes total number of results (number of exclu-
ded outlier results).
Fig 4. Mean deviation from NGSP target for the major method
groups during 2007 in the Labquality EQA scheme for HbA1c.
‘Arch/Aero’ is Architecht and Aeroset from Abbot, ‘DCA2000/
Vantage’ is DCA2000 and DCA Vantage from Siemens,
‘TinaQuant’ is TinaQuant reagent on Roche platforms,
‘NycoCard’ is NycoCard HbA1c, Medinor, ‘Konelab’ is Konelab
HbA1c test, Thermo Electron, ‘Integra, wb’ is Roche whole
blood method on Integra, ‘Olympus’ is Olympus HbA1c on
Olympus platform, ‘Tosoh’ is Tosoh HPLC, Tosoh Bioscience,
‘Variant II’ is Variant II HPLC from Bio Rad, Advia is Advia
HbA1c, Siemens, ’D-10’ is D-10 HPLC, Bio Rad. ‘N’ denotes
total number of results (number of excluded outlier results).
vv
0,0
0,2
0,4
0,6
rget, HbA1c %-unit
±
95% Confidence interval
T
-0,6
-0,4
-0,2
Afinion
N=200 (0)
DCA 2000
N=3653 (17)
DCA Vantage
N=21 (0)
Mono S
N=45 (0)
Mono S (exp)
N=45 (0)
Tq/Unimate
N = 85 (4)
Tosoh
N=113 (1)
Variant II
N= 111 (2)
Variant Turbo
N=42 (5)
Deviation from ta
T
-0,6
-0,4
-0,2
0,0
0,2
0,4
0,6
Arch/Aero
N=40 (0)
DCA2000/Vantage
N=462 (1)
TinaQuant
N=60 (2)
NycoCard
N=32 (1)
Konelab
192 (4)
Integra, wb
N=252 (10)
Olympus
N=24 (0)
Tosoh
N=182 (1)
Variant II
N=110 (0)
Advia
N=73 (2)
D-10
N=32 (0)
Deviation from target, HbA1c %-unit
±
95% Confidence interval
