CCQM-K12.1: Subsequent Key Comparison on the …kcdb.bipm.org/appendixb/appBResults/CCQM-K12/CCQM-K12.1.pdf · CCQM-K12.1: Subsequent Key ... INTRODUCTION The accuracy and ... KRISS
Post on 14-Feb-2018
215 Views
Preview:
Transcript
1
CCQM-K12.1: Subsequent Key Comparison on the Determination of
Creatinine in Serum
Final Report
September 2006
Lian Hua Shi, Byung-Joo Kim, Hwa-Shim Lee, and Sang-Ryoul Park
Korea Research Institute of Standards and Science (KRISS)
Daejeon, Korea
Miryan Balderas Escamilla and Melina Pérez Urquiza
Centro Nacional de Metrología (CENAM)
Querétaro, México
Hisashi Kato
National Metrology Institute of Japan (NMIJ)
Tsukuba, Japan
Can Quan and Hongmei Li
National Research Center for Certified Reference Materials (NRCCRM)
Beijing, China
INTRODUCTION
The accuracy and traceability of routine clinical diagnostic tests has become a matter of
great concern. Inconsistency in test results due to the lack of high quality measurement
standards implemented in clinical practices often leads to confusion in making medical
decisions as well as unnecessary and costly repeats of tests. This problem has been
legally addressed by the European Union to implement of an In-vitro Diagnostic (IVD)
Directive. Other parts of the world are likely to follow the EU in requiring greater
traceability for IVD products. Upon this change, CIPM, IFCC, and several other related
international organizations have organized the Joint Committee on Traceability in
Laboratory Medicine (JCTM) to help implementation of measurement standards in the
field of laboratory medicine [1]. National metrology institutes (NMIs) are to take great
2
responsibility in development and dissemination of measurement standards of higher
order. Therefore, Key Comparisons of substances of great diagnostic values have been
actively organized and carried out under CCQM.
Creatinine is an important diagnostic marker for renal function, and is frequently
measured in various clinical situations. Routine clinical tests mostly based on enzymatic
reactions, however, are often subject to interferences from various materials coexisting
with creatinine in samples. Therefore, the accuracy of clinical tests is still of great
concern. CCQM Organic Analysis Working Group (OAWG) had performed a key
comparison of determination of creatinine in human serum (CCQM-K12) in which five
NMIs had participated. The results are available from the BIPM KCDB [2].
There were other NMIs that needed to demonstrate their measurement capability
through a form of KC. Reflecting upon the current situation, the chair of OAWG
arranged for a subsequent Key Comparison (CCQM-K12.1). Under the guidance of
OAWG, KRISS who demonstrated its measurement capability in CCQM-K11
coordinated the Subsequent Key Comparison.
CENAM (Mexico), NMIJ (Japan), and NRCCRM (China) participated in the
Subsequent Key Comparison, and two sets of test materials representing normal and
elevated creatinine levels in human serum were sent. Results were received from
CENAM and NRCCRM. NMIJ withdrew from the Key Comparison because of some
technical difficulties in measurement. Returned data were presented in OAWG fall
meeting of 2005 with minimum data processing, and no significant comments on the
results of the Subsequent Key Comparison were made by OAWG. In the draft
document of the final report, modifications of initially reported expanded uncertainties
were made as NRCCRM and KRISS found needs for correction in their assessments. As
not allowed, no modification was made in the initially reported mean values. The draft
document of the final report was submitted and discussed during the CCQM OAWG
meeting in April 2006.
The draft of the final report was submitted to the 2006 April meeting of OAWG,
OAWG decided to place the originally reported expanded uncertainties by participants
as the final values. This recommendation has been reflected in the revised report.
Corrections of some miscalculations in the previous draft repro were also made.
3
SUMMARY OF ORIGINAL CCQM-K12 STUDY
A Key Comparison on the determination of creatinine in human serum, CCQM-K12
was conducted in 2002 with National Institute of Standards and Technology (NIST) as
the coordinating laboratory. Five laboratories participated in this Key Comparison:
Country Institution
EU Institute of Reference Materials and Measurements, IRMM
Germany Physikalisch-Technische Bundesanstalt, PTB
Korea Korea Research Institute of Standards and Science, KRISS
The Netherlands NMi Van Swinden Laboratorium, NMi
UK Laboratory of the Government Chemist, LGC
USA National Institute of Standards and Technology, NIST (Pilot Lab)
The details of the study outcome can be found at the BIPM website [2]. The conclusion
of the study was that the participating NMIs could successfully measure serum
creatinine at normal and elevated levels, using ID/MS-based methods, with
interlaboratory expanded uncertainties of less than 0.8%. The essential part of the
results of CCQM-K11 is presented in Table 1.
CONDUCT OF THIS STUDY (CCQM-K12.1; CCQM-K12-Subsequent)
Participants
The following five countries participated in this study:
Country Institution
China National Research Centre for Certified Reference Materials, NRCCRM
Japan National Metrology Institute of Japan, NMIJ (data not returned)
Korea KRISS (Coordinating Laboratory)
Mexico National Center of Metrology, CENAM
NMIJ had received test materials, but experienced a technical difficulty in measurement,
and decide not to return data. KRISS is the only laboratory participated in CCQM-K12
and had demonstrated capability for accurate measurement of creatinie in human serum.
Therefore, KRISS, the coordinating laboratory was to provide a link to CCQM-K12 for
the results of this study.
4
Methods Used for the CCQM-K12-Subsequent Comparison
NIMJ measured creatinine in serum by ID-LC-MS method. The reproducibility of
measurements was very poor to report the results. NMIJ scientists detected unknown
peaks at m/z 114, and peak areas increased by the time spent on the defrosted the serum
sample in the refrigerator. For this problem, NMIJ decided to withdraw from this
Subsequent Key Comparison. No other laboratories reported a similar problem.
All participants including NMIJ used LC/ID-MS. KRISS used ultra-filtration with 3000
Da cut-off filters for removal of proteins. Other laboratories precipitated proteins with
acetonitrile or ethanol. KRISS used NIST SRM 914a as the high purity creatinie for
preparation of standard solutions. CENAM and NRCCRM used commercial chemical
products from Aldrich and Fluka, respectively. CENAM determined the purity of the
high purity creatinine by itself. NRCCRM took of 1 % of estimated uncertainty in purity
determination. CENAM and KRISS validated their measurement procedures with the
reference materials of CRM DMR-963a and NIST SRM 909b, respectively. Table below is
the summary of the methods used by each laboratory.
Laboratory Sample prep. &
Derivatization
Measurement Calibration
material
Validation
CENAM Acetonitrile precip.,
no derive.
LC/ID-MS Aldrich* CRM DMR-
963a
KRISS Ultra-filtration
(3000 Da), no derive.
LC/ID-MS NIST SRM 914a NIST SRM 909b
NMIJ NA (suffered from
poor reproducibility
NA
(LC/ID-MS)
NA NA
NRCCRM Ethanol precip.,
no derive.
LC/ID-MS Fluka, > 99%
* Purity was determined by CENAM.
Materials Used for the CCQM-K12-Subsequent Comparison
Candidate certified reference materials of human serum prepared by KRISS were used
as test materials in this study. The test materials were natural human serum to which no
preservative was added. Only antiseptic filtering and/or fortification were applied.
5
Homogeneities of creatinine were estimated 0.65 % and 0.75% RSD for Test Material
01B and Test Material 02B, respectively as presented in the figures below. Target
ranges were given as follows:
01B 02B
Creatinine 4 - 7 mg/kg 20 – 40 mg/kg
0 2 4 6 8 10
0.90
0.95
1.00
1.05
1.10
RSD = 0.65 %
De
via
tio
n
Vial Number
0 2 4 6 8 10
0.90
0.95
1.00
1.05
1.10
RSD = 0.75 %
De
via
tio
n
Vial Number
0 2 4 6 8 10
0.90
0.95
1.00
1.05
1.10
RSD = 0.65 %
De
via
tio
n
Vial Number
0 2 4 6 8 10
0.90
0.95
1.00
1.05
1.10
RSD = 0.75 %
De
via
tio
n
Vial Number
Results of homogeneity test of the test materials for creatinine measurement.
The test materials were to be kept frozen (either at -20 oC or at -70
oC ). However, no
visible sign of degradation was observed during storage at room temperature for several
weeks as far as the seals were intact. Each test material consisted of 4 vials of 3 mL of
human serum. For the set of 4 vials, one is for practice and/or for better determining the
target concentration whereas the rest 3 vials are for measurement. The test materials
were sent as packed with a sufficient amount of dry ice to keep them frozen for several
days. However, test materials sent to NRCCRM were exposed to ambient temperature
for about two weeks because of trouble at custom offices. Stability of creatinine in this
storage condition was not performed. Because of the absence of any assuring delivery
mean, NRCCRM performed the measurements without replacement of originally
received test materials.
6
Measurement Protocol and Calculation of Uncertainty
As requested, all participants except NMIJ produced two independent results from each
vial of a test material. Three vials of a test material were distributed. Therefore, 6
measurements were resulted from each test material. The results were to be reported on
an absolute basis (corrected for chemical purity of the calibration material used by the
participant) together with the expanded uncertainty. Space was provided at the end of
the data reporting sheets for inclusion of a full uncertainty budget, including definition
of terms and assessment of which components made significant contributions.
Linking the Subsequent Key Comparison to the Original Key Comparison
The results of CCQM-K12-Subsequent were initially treated as if an independent Key
Comparison was carried out (Table 2-1 & 2-2). Then, liking them to the results of the
original Key Comparison was made by applying proportionality factors obtained
through comparison of KRISS results from both Key Comparisons. KRISS results from
the Subsequent Key Comparison were assumed to have the same fractions of deviation
from KCRVs (virtual values) as determined in the original Key Comparison. From this
assumption, KCRVs of the Subsequent Comparison were assigned as noted as assigned
KCRVs in Table 3. Although concentrations of creatinine were not exactly matched, the
results of Test Material 01B of CCQM-K12.1 were related to the results of Material I of
CCQM-K12 as a normal creatinine level whereas Test Material 02B of CCQM-K12.1
was related to Material II of CCQM-12 as an elevated creatinine level.
Calculation of the assigned KCRVs is shown below.
Assignment of the KCRV for Test Material 01B of CCQM-K12.1:
Assigned KCRV = 5.939 mg/kg x (8.217/8.186) = 5.961 mg/kg
8.217 mg/kg: KCRV for Material I of CCQM-K12;
8.186 mg/kg: KRISS result for Material I of CCQM-K12;
5.939 mg/kg: KRISS result for Test Material 01B of CCQM-K12.1
Assignment of the KCRV for Test Material 02B of K12.1:
Assigned KCRV = 27.361 mg/kg x (18.645/18.539) = 27.517 mg/kg
18.645 mg/kg: KCRV for Material II of CCQM-K12;
18.539 mg/kg: KRISS result for Material II of CCQM-K12;
27.361 mg/kg: KRISS result for Test Material 02B of CCQM-K12.1
7
Expanded uncertainties were also linked to the results of original Key Comparison in
the same manner as applied for calculation of deviations of from KCRVs. Calculation
of adjusted uncertainties are shown below.
Adjustment of expanded uncertainties for Test Material 01B
Adjusted Uncertainties = Reported uncertainties x (1.9/0.92)
1.9 %: KRISS Rel. Uncertainty (%) for Material I of CCQM-K12;
0.92 %: KRISS Rel. Uncertainty (%) for Test Material 01B of CCQ-K12.1
Adjustment of expanded uncertainties for Test Material 02B
Adjusted Uncertainties = Reported uncertainties x (1.7/0.86)
1.7 %: KRISS Rel. Uncertainty (%) for Material II of CCQM-K12;
0.86 %: KRISS Rel. Uncertainty (%) for Test Material 02B of CCQ-K12.1
During the CCQM meeting in April 2006, Organic Analysis Working Group decided to
accept the originally reported expanded uncertainty by each participating laboratory as
the final values of the expanded uncertainty. Therefore, the values calculated above for
linking to the original study were not taken as the final values.
RESULTS
Results for the CCQM-K12-Subsequent Comparison are summarized in Table 2. The
uncertainty bars in the figures represent expanded uncertainties as reported by the
participating laboratories. Uncertainties in the results of NRCCRM were changed from
the values reported in OAWG fall meeting 2005 as requested by NRCCRM. KRISS
also refined its uncertainty budgeting procedure and made modifications of originally
reported uncertainties. Evaluation of uncertainty reported by each laboratory is
summarized in Table 5. The results of linking the results of this Subsequent Key
Comparison to the original Key Comparison are presented in Table 3. As decided by
OAWG, the uncertainties originally reported by participants were placed as the final
values regardless the results of linking to the results of CCQM K-11.
8
DISCUSSION
KRISS is the only laboratory participated in the original CCQM-K12 Key Comparison
[2]. KRISS, the coordinating laboratory, was to provide a link for this Subsequent Key
Comparison to the original Key Comparison. KRISS proved its capability for accurate
measurement of creatinine in human serum in the original Key Comparison study. In
addition, KRISS validated its measurement procedure by performing simultaneous
measurement of NIST SRM 909b as shown below.
NMIJ had experienced poor reproducibility problem and decided not to submit the
results. NMIJ scientists informed that they found an unknown peak of m/z 114, of
which area increased in a time course. If this phenomenon was responsible to the poor
reproducibility of measurement of creatinine, it may be related with cross conversion
between creatine and creatinine as described in the final report of CCQM-K11 [2].
NMIJ would have to solve this problem. No other participating laboratories reported a
similar problem.
50.389 ± 0.887KRISS measurement
50.317 ± 0.571Certified value
mg/kgLevel 2
6.384 ± 0.105KRISS measurement
6.363 ± 0.062Certified value
mg/kgLevel 1
NIST SRM 909b
(freeze dry serum)
50.389 ± 0.887KRISS measurement
50.317 ± 0.571Certified value
mg/kgLevel 2
6.384 ± 0.105KRISS measurement
6.363 ± 0.062Certified value
mg/kgLevel 1
NIST SRM 909b
(freeze dry serum)
NRCCRM reported relatively smaller mean values for both test materials. Especially for
Test Material 01B, scattering of data was noticeable (See a figure below). One possible
reason for this scattering and underestimations would be degradation of creatinine due
to exposure of test materials to ambient temperature for about two weeks. Stability of
creatinine in serum media against a long term exposure to ambient temperature was not
tested yet. Inhomogeneity of the test materials is ruled out from the possible causes of
poor agreements of the reported results. Homogeneity of the test materials was proved
9
by the results of the homogeneity test shown above as well as high precision of reported
results of this study except NRCCRM’s results for Test Material 01B.
5.2
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
6.1
Cre
atinin
e (
mg
/kg
)
25.0
25.5
26.0
26.5
27.0
27.5
28.0
28.5
Cre
atin
ine (
mg
/kg
)
CE
NA
M
KR
ISS
NR
CC
RM
CE
NA
M
KR
ISS
NR
CC
RM
Test Material 01B Test Material 02B
Mean values reported by CENAM were relatively close to KRISS values. However,
deviations were slightly greater than the expanded uncertainties.
In overall, the results of the Subsequent Key Comparison were poorer than the outcome
of the original Key Comparison in both deviations and uncertainties. The lack of
performing a pilot study prior to participating in this Subsequent Key Comparison
might be the major reason for the poor outcome.
10
CONCLUSIONS
Probably because of lacking of participation in an adequate pilot study, the deviations
and uncertainties of CCQM-12-Subsequent Key Comparison were substantially larger
than the average results of the original Key Comparison. CENAM’s results were
relatively close to the KCRVs, and would be within the expanded uncertainties if the
values calculated for linking to the original study were taken. Nevertheless, CENAM
might need to improve the level of accuracy to the levels other NMIs demonstrated in
the original Key Comparison. NRCCRM might have encountered partially degraded test
materials as they reported substantially underestimating results. The test materials were
stored at ambient temperature for about two weeks at a customs office. If not done yet,
it will be helpful for NRCCRM to validate its measurement procedure using reliable
reference materials. NMIJ withdrew from the Subsequent Key Comparison due to a
technical difficulty. As proved in this study, having sufficient experience through a Pilot
Study prior to participating in a Key Comparison is valuable to demonstrate satisfactory
equivalence through a Key Comparison.
REFERENCES
[1] http://www.bipm.org/en/committees/jc/jctlm/
[2] http://kcdb.bipm.org/AppendixB/appbresults/ccqm-k12/ccqm-k12_final_report.pdf
11
Table 1-1. Results of CCQM-K12 Creatinine in Human Serum: Material I
units: mg/kg
Participant Mean Standard Un
certainty
Degrees of f
reedom
k Expanded
uncertainty
IRMM 8.360 0.1060 60 2.000 0.212
KRISS 8.186 0.0796 69 1.995 0.159
LGC 8.193 0.0080 60 2.000 0.016
NIST 8.277 0.0319 3.9 3.182 0.102
PTB 8.211 0.0289 60 2.000 0.058
Mean 8.217 excluding IRMM
Range (%) 2.12
Std dev of mean 0.0208
Degrees of Freedom 3
K factor 3.182
U 0.063
U(rel) % 0.81
KCRV 8.217 mg/kg 0.066 mg/kg
Table 1-2. Results of CCQM-K12 Creatinine in Human Serum: Material II
units: mg/kg
Participant Mean Standard Un
certainty
Degrees of f
reedom
k Expanded
uncertainty
IRMM 18.720 0.2396 60 2.000 0.479
KRISS 18.539 0.1627 503 1.965 0.320
LGC 18.614 0.0316 60 2.000 0.063
NIST 18.708 0.0722 3.9 3.182 0.230
PTB 18.718 0.0650 34.7 2.032 0.132
Mean 18.645 excluding IRMM
Range (%) 0.97
Std dev of mean 0.0423
Degrees of Freedom 3
K factor: 3.182
U 0.14
U (rel) % 0.72
KCRV 18.65 mg/kg 0.14 mg/kg
12
Table 2-1. Results of CCQM-K12.1 Creatinie in Human Serum: Test
Material 01B
Units: mg/kg
Participant Mean Comb. Std.
Uncertainty
Degrees of
freedom
k Expended
uncertainty
CENAM 5.745 0.031 1000 1.96 0.061
KRISS 5.939 0.027 31 2.04 0.055
NMIJ N/A N/A N/A N/A N/A
NRCCRM 5.58 0.12 60 2.00 0.24
Mean 5.754
Range (%) 6.3
Std dev of mean 0.180
Degrees of Freedom 2
K factor 4.30 (95% confidence level)
U 0.77
Relative U (%) 13
KCRV 5.75 mg/kg 0.77 mg/kg
(treated as an independent study)
CE
NA
M
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
Cre
atin
ine
(mg/
kg)
KR
ISS
NR
CC
RM
CE
NA
M
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
Cre
atin
ine
(mg/
kg)
KR
ISS
NR
CC
RM
CE
NA
M
5.3
5.4
5.5
5.6
5.7
5.8
5.9
6.0
Cre
atin
ine
(mg/
kg)
KR
ISS
NR
CC
RM
13
Table 2-2. Results of CCQM-K12.1 Creatinine in Human Serum: Test
Material 02B
Units: mg/kg
Participant Mean Comb. Std.
Uncertainty
Degrees of
freedom
k Expanded
uncertainty
CENAM 28.26 0.13 1.96 0.25
KRISS 27.36 0.12 31 2.04 0.23
NMIJ N/A N/A N/A N/A N/A
NRCCRM 25.61 0.12 2.00 0.23
NMIJ data were excluded in the following calculations.
Mean 27.08
Range (%) 9.8
Std dev of mean 1.34
Degrees of Freedom 2
K factor 4.30 (95% confidence level)
U 5.8
Relative U (%) 21
KCRV 27.1 mg/kg 5.8 mg/kg
(treated as an independent study)
26
27
28
Cre
atin
ine
(m
g/k
g)
CE
NA
M
KR
ISS
NR
CC
RM
26
27
28
Cre
atin
ine
(m
g/k
g)
CE
NA
M
KR
ISS
NR
CC
RM
14
Table 3-1. Link to the original study: Test Material 01B (normal creatinine
level)
Units: mg/kg; Linked KCRV = 5.961 (See Page 6, 7)
Participant Mean of
K12.1
Difference
from
KCRV
% Difference
from KCRV
% Rel.
uncertainty,K12.1
(Linked to K12)***
Finalized
rel.
uncertainty
(%)
CENAM 5.745 -0.22 -3.6 1.1 (2.24) 1.1
KRISS 5.939 -0.022 -0.37 0.92 (1.94) 0.92
NRCCRM 5.579 -0.38 -6.4 4.3 (9.03) 4.3
IRMM 1.7* 2.5
**
LGC -0.29* 0.20
**
NIST 0.73* 1.2
**
PTB -0.073* 0.71
**
* % difference in CCQM-K12;
** relative expanded uncertainty in CCQM-K12;
*** % relative uncertainty
calculated as linked to the results of CCQM-K12 assuming the uncertainty of KRISS was exactly
maintained.
-12
-10
-8
-6
-4
-2
0
2
4
6
% D
iffe
ren
ce f
rom
KC
RV
CE
NA
M
KR
ISS
NR
CC
RM
PT
B
NIS
T
LG
C
KR
ISS
IRM
M
CCQM-K12 CCQM-K12.1
-12
-10
-8
-6
-4
-2
0
2
4
6
% D
iffe
ren
ce f
rom
KC
RV
CE
NA
M
KR
ISS
NR
CC
RM
PT
B
NIS
T
LG
C
KR
ISS
IRM
M
CCQM-K12 CCQM-K12.1
15
Table 3-2. Link to the original study: Test Material 02B (elevated
creatinine level)
Units: mg/kg; Linked KCRV = 27.517 (See Page 6, 7)
Participant Mean of
K12.1
Difference
from
KCRV
% Difference
from KCRV
% Rel.
uncertainty,K12.1
(Linked to K12)***
Finalized
rel.
uncertainty
(%)
CENAM 28.26 0.74 2.7 0.91 (1.8) 0.91
KRISS 27.36 -0.16 -0.57 0.85 (1.7) 0.85
NRCCRM 25.6 -1.9 -6.9 0.85 (1.7) 0.85
IRMM 0.40* 2.6
**
LGC -0.17* 0.34
**
NIST 0.34* 1.2
**
PTB 0.39* 0.71
**
* % difference in CCQM-K12;
** relative expanded uncertainty in CCQM-K12;
*** % relative uncertainty
calculated as linked to the results of CCQM-K12 assuming the uncertainty of KRISS was exactly
maintained.
-10
-8
-6
-4
-2
0
2
4
6
% D
iffe
ren
ce f
rom
KC
RV
CE
NA
M
KR
ISS
NR
CC
RM
PT
B
NIS
T
LG
C
KR
ISS
IRM
M
CCQM-K12 CCQM-K12.1
-10
-8
-6
-4
-2
0
2
4
6
% D
iffe
ren
ce f
rom
KC
RV
CE
NA
M
KR
ISS
NR
CC
RM
PT
B
NIS
T
LG
C
KR
ISS
IRM
M
CCQM-K12 CCQM-K12.1
16
Table 4-1-A. Uncertainty Reports: Test Material 01B (CENAM)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude) 01B
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( ) (mg
g-1
)
Degrees
of
freedom
(i)
Type Source of
data
Method
precision
Between batch
precision for
the method as a
whole (major
source)
5.7454 0.0064 ˜ 1 5 A Replicate
analysis of
sample
across tree
samples
Calibration
solution
Concentration
of the
calibration
solution
(corrected for
purity) (major
source)
5.9462 0.0013 ˜ 1 5 A Gravimetric
preparation
Weight of
sample
Balance
linearity (minor
source)
0.6354 0.00004 large B Balance
calibration
certificate
Weight of
calibration
compound
Balance
linearity (minor
source)
0.6411 0.00004 large B Balance
calibration
certificate
ARsample Area Ratio
sample
1.0544 0.0026 ˜ 1 5 A Replicate
analysis of
sample
across tree
samples
ARstd Area Ratio
standard
1.0997 0.0058 ˜ 1 2 A Replicate
analysis
17
Table 4-1-B. Uncertainty Reports: Test Material 02B (CENAM)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude) 02B
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( ) (mg
g-1
)
Degrees
of
freedom
(i)
Type Source of
data
Method
precision
Between batch
precision for
the method as
a whole (major
source)
28.2563 0.0658 ˜ 1 5 A Replicate
analysis of
sample
across tree
samples
Calibration
solution
Concentration
of the
calibration
solution
(corrected for
purity) (major
source)
27.6171 0.0062 ˜ 1 5 A Gravimetric
preparation
Weight of
sample
Balance
linearity
(minor source)
0.62270 0.00004 large B Balance
calibration
certificate
Weight of
calibration
compound
Balance
linearity
(minor source)
0.62507 0.00004 large B Balance
calibration
certificate
ARsample Area Ratio
sample
1.0130 0.0084 ˜ 1 5 A Replicate
analysis of
sample
across tree
samples
ARstd Area Ratio
standard
0.9927 0.0027 ˜ 1 2 A Replicate
analysis of
sample
across tree
samples
18
Table 4-2-A. Uncertainty Reports: Test Material 01B (KRISS)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude)
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( )
Degrees of
freedom
(i)
Type Source of
data
Method
precision
Between batch
precision for the
method as a
whole (major
source)
5.9391
mg/kg
0.00745
mg/kg
~ 1 5 A Replicate
analysis of
sample
across tree
samples
Calibration
solution
Purity correction
(major source)
99.7% 0.3% ~ 1 large B Certified
value from
SRM
Provider
Weight of cal.
compound
2.2352
mg
0.002
mg
~ 1 large B Calibration
certificate
Weight of cal.
solution
26
g
0.00005
g
~ 1 large B Calibration
certificate
Standard
Solution
Accuracy in prep.
of std soln.
0.796871 0.00328 ~ 1 3 A Std dev of
R.F. of 4
std solns
u
(Mis_sol,std/Ms_sol,std)
1.0020073
0.0001421
~ 1 large B Calibration
certificate
Sample
solution
u
(Mis_sol,std/Ws)
0.2989 0.00010 ~ 1 large B Calibration
certificate
Random components were counted in method precision.
ARstd Repeatability
of determination
0.9870
0.00520
~ 1 4 A Replicate
analysis of
std. soln.
ARsample Counted in method precision
19
Table 4-2-B. Uncertainty Reports: Test Material 02B (KRISS)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude)
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( )
Degrees of
freedom
(i)
Type Source of
data
Method
precision
Between batch
precision for the
method as a
whole (major
source)
27.3615
mg/kg
0.036176
mg/kg
~ 1 5 A Replicate
analysis of
sample
across tree
samples
Calibration
solution
Purity correction
(major source)
99.7% 0.3% ~ 1 large B Certified
value from
SRM
Provider
Weight of cal.
compound
2.2352
mg
0.002
mg
~ 1 large B Calibration
certificate
Weight of cal.
solution
26
g
0.00005
g
~ 1 large B Calibration
certificate
Standard
Solution
Accuracy in prep.
of std soln.
0.796871 0.00328 ~ 1 3 A Std dev of
R.F. of 4
std solns
u
(Mis_sol,std/Ms_sol,std)
1.00201
0.00014
~ 1 large B Calibration
certificate
Sample
solution
u
(Mis_sol,std/Ws)
1.412078
0.000496
~ 1 large B Calibration
certificate
Random components were counted in method precision.
ARstd Repeatability
of determination
0.9870
0.00235
~ 1 4 A Replicate
analysis of
std. soln.
ARsample Counted in method precision
20
Table 4-3-A. Uncertainty Reports: Test Material 01B (NRCCRM)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude)
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( )
Degrees
of
freedom
(i)
Type Source of
data
Method
precision
Between
batch
precision for
the method as
a whole
(major
source)
5.579
mg/kg
0.0856
mg/kg
0.0856
mg/kg
5 A Replicate
analysis of
sample
across two
batches
Calibration
solution
Concentration
of the
calibration
solution
(corrected for
purity) (major
source)
5.514mg/kg 0.0032
mg/kg
0.059 mg/kg large B Supplier’s
specification
Weight of
sample
Balance
linearity
(minor
source)
0.48590g 0.000029 g -0.000599g large B Balance
calibration
certificate
Weight of
calibration
compound
Balance
linearity
(minor
source)
0.42290 mg 0.0058 mg 0.012 mg large B Balance
calibration
certificate
Other
Other
Other
Other
21
Table 4-3-B. Uncertainty Reports: Test Material 02B (NRCCRM)
Uncertainty Budget (Sources of uncertainty, their type, evaluation and magnitude)
Parameter Source of
uncertainty
xi u(xi)
f
xu x
i
i
( )
Degrees
of
freedom
(i)
Type Source of
data
Method
precision
Between
batch
precision for
the method as
a whole
(major
source)
25.613
mg/kg
0.115mg/kg 0.115mg/kg 5 A Replicate
analysis of
sample
across two
batches
Calibration
solution
Concentration
of the
calibration
solution
(corrected for
purity) (major
source)
25.705mg/kg 0.000205
mg/kg
0.0180mg/kg large B Supplier’s
specification
Weight of
sample
Balance
linearity
(minor
source)
0.49690g 0.000029g -0.001526
μg
large B Balance
calibration
certificate
Weight of
calibration
compound
Balance
linearity
(minor
source)
0.28110 mg 0.058 mg 0.005 mg large B Balance
calibration
certificate
Other
Other
Other
Other
Other
top related