PROFICIENCY STUDY AQA 13-12 COCAINE - DECEMBER 2013
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
PROFICIENCY STUDY AQA 13-12 COCAINE DECEMBER 2013
ACKNOWLEDGMENTS
This study was conducted by the National Measurement Institute (NMI). Support funding was
provided by the Australian Government Department of Industry.
I would like to thank the management and staff of the participating laboratories for supporting
the study. It is only through widespread participation that we can provide an effective service
to laboratories.
The assistance of the following NMI staff members in the planning, conduct and reporting of
the study is acknowledged.
Raluca Iavetz
Paul Armishaw
Manager, Chemical Proficiency Testing
105 Delhi Rd, Riverside Corporate Park,
North Ryde NSW 2113
PO Box 138,
North Ryde NSW 1670
Phone: 61-2-9449 0149
Fax: 61-2-9449 0123
paul.armishaw@measurement.gov.au
Accredited for compliance with ISO/IEC 17043
iTHIS PAGE IS INTENTIONALLY BLANK
iiTABLE OF CONTENTS
SUMMARY 1
1 INTRODUCTION 2
1.1 NMI Proficiency Testing Program 2
1.2 Study Aims 2
1.3 Study Conduct 2
2 STUDY INFORMATION 3
2.1 Study Timetable 3
2.2 Participation 3
2.3 Test Material Specification 3
2.4 Laboratory Code 3
2.5 Test Sample Homogeneity 3
2.6 Sample Dispatch and Receipt 4
2.7 Instructions to Participants 4
2.8 Interim Report 4
3 PARTICIPANT LABORATORY INFORMATION 5
3.1 Test Method Summaries 5
3.2 Reported Basis of Participants’ Measurement Uncertainty Estimates 7
3.3 Details of Participant Reference Standard 9
3.4 Participants’ Comments 10
4 PRESENTATION OF RESULTS AND STATISTICAL ANALYSIS 11
4.1 Results Summary 11
4.2 Assigned Value 11
4.3 Between-Laboratory Coefficient of Variation 11
4.4 Target Standard Deviation 12
4.5 z-Score 12
4.6 En-Score 12
4.7 Traceability and Measurement Uncertainty 12
5 TABLES AND FIGURES 13
6 DISCUSSION OF RESULTS 20
6.1 Assigned Value 20
6.2 Measurement Uncertainty Reported by Participants 20
6.3 z-Score 20
6.4 En-Score 21
6.5 Identification of Cutting Agent 22
6.6 Participants’ Analytical Methods 23
6.7 Summary of participation and performance in Cocaine Studies 25
7 REFERENCES 26
APPENDIX 1 - PARTICIPANT LABORATORIES 27
APPENDIX 2 - HOMOGENEITY TESTING 28
APPENDIX 3 - MEASUREMENT UNCERTAINTY OF THE ASSIGNED VALUE 29
APPENDIX 4 - ACRONYMS AND ABBREVIATIONS 30
iiiSUMMARY AQA 13-12 was conducted in August 2013. Three test samples of cocaine hydrochloride were sent to twenty-nine laboratories. Twenty-eight laboratories submitted results by the due date. Two laboratories submitted two sets of results analysed independently by two analysts. Laboratory Performance Laboratory performance was assessed by z-score and En-score. Laboratories 1, 12, 15, 21 and 26 returned both a satisfactory z-score and En-score for all results. Laboratory 3 reported results for only two out of the three samples and both results returned satisfactory z and En scores. The remaining twenty-four participants returned at least one questionable score for one sample or more. z-Score A total of 70% scores were satisfactory (62 scores from a total of 89). Laboratories 1, 2, 6, 12, 15, 21, 26 and 28 returned a satisfactory score for all results. Laboratory 3 reported results for only two out of three samples, both returning satisfactory z- scores. The remaining twenty-one laboratories returned at least one questionable score for one sample or more. En-Score A total of 66% scores were satisfactory (59 scores from a total of 89). Laboratories 1, 8, 12, 13, 15, 16, 17, 21, 23 and 26 returned a satisfactory score for all results. Laboratory 3 reported results for only two out of three samples and both results returned satisfactory En-scores. Laboratories 4, 5 and 29 returned no satisfactory En-scores for any sample. The remaining sixteen laboratories returned at least one unsatisfactory score for one or more samples. Estimation of Measurement Uncertainty Eighty-three results (93%) were reported with an associated expanded uncertainty. Laboratories 4 and 28 did not report uncertainty. Both laboratories were not accredited. Laboratories 13, 17, 21, 23, 26 and 29 reported an identical uncertainty for samples which were of significantly different concentrations. The magnitude of reported uncertainties was within the range 2% to 68% relative. Identification of Cutting Agent Twenty-nine participants (97%) reported on the identity of the cutting agents in the test samples. All test samples were prepared using an illicit seizure of cocaine hydrochloride, approximately 73 % base (m/m) supplied by the Australian Federal Police. The following were added by the study coordinator: glucose to all test samples, procaine to Sample S2 and levamisol to Sample S3. Eleven participants identified correctly all the cutting agents added by the study coordinator to Samples S1, S2 and S3. Twenty-eight laboratories correctly identified procaine in Sample S2 and levamisole in Sample S3. 1 AQA 13-12 Cocaine
1 INTRODUCTION
1.1 NMI Proficiency Testing Program
The National Measurement Institute (NMI) is responsible for Australia’s national
measurement infrastructure, providing a range of services including a chemical proficiency
testing program.
Proficiency testing (PT) is: ‘evaluation of participant performance against pre-established
criteria by means of interlaboratory comparison.’1 NMI PT studies target chemical testing in
areas of high public significance such as trade, environment, law enforcement and food
safety. NMI offers studies in:
• pesticide residues in fruit and vegetables, soil and water;
• petroleum hydrocarbons in soil and water;
• metals in soil, water, food and pharmaceuticals;
• controlled drug assay and clandestine laboratory;
• allergens in food; and
• folic acid in flour.
1.2 Study Aims
The aims of the study were to:
• assess the accuracy of participants’ measurements of cocaine in samples typical of a
routine seizure;
• test the ability of participants to identify a cutting agent commonly found in controlled
drug preparation; and
• develop a practical application of traceability and measurement uncertainty and provide
participants with information that will assist uncertainty estimates.
The choice of the test method was left to the participating laboratories.
1.3 Study Conduct
NMI is accredited by the National Association of Testing Authorities, Australia (NATA) to
ISO170431 as a provider of proficiency testing schemes. This controlled drug proficiency test
is within the scope of NMI’s accreditation.
The conduct of NMI proficiency tests is described in the NMI Chemical Proficiency Testing
Study Protocol2. The statistical methods used are described in the NMI Chemical Proficiency
Statistical Manual3. These documents have been prepared with reference to ISO 17043 and
The International Harmonized Protocol for Proficiency Testing of (Chemical) Analytical
Laboratories4.
AQA 13-12 Cocaine 22 STUDY INFORMATION 2.1 Study Timetable The timetable of the study was: Invitation issued: 10 July 2013 Samples dispatched: 16 September 2013 Results due: 08 November 2013 Interim report issued: 15 November 2013 2.2 Participation A total of seventy-four international, national, state government and private laboratories were invited to participate. Twenty-nine laboratories agreed to participate and twenty-eight submitted results. These laboratories are listed in Appendix 1. Three laboratories requested two sets of each test samples in order to be analysed by different analysts. Only two laboratories submitted two sets of results. 2.3 Test Material Specification Three test samples were prepared in June 2013. The starting material was cocaine hydrochloride approximately 73% base (m/m) supplied by the Australian Federal Police. The cocaine was ground and sieved through a 180 µm sieve. The study coordinator purchased levamisole from Acros Organics, D-glucose anhydrous from Mallinckrodt and procaine from Sigma. These were used as cutting agents. Glucose was used to prepare Sample S1, glucose and procaine were used to prepare Sample S2 and glucose and levamisole were used to prepare Sample S3. The cutting agents were similarly processed as the cocaine powder. Each test sample was then prepared by mixing a known mass of sieved drug material with a known mass of sieved cutting agent in a tumbler for five hours. Portions of 150 mg of test samples were weighed into labelled glass vials. Sample S1 was prepared to contain approximately 30% cocaine base (m/m). Sample S2 was prepared to contain approximately 13% cocaine base (m/m). Sample S3 was prepared to contain approximately 18% cocaine base (m/m). 2.4 Laboratory Code Each participant was randomly assigned a confidential laboratory code between 1 and 32. 2.5 Test Sample Homogeneity The preparation of homogeneous test samples is an important part of a proficiency testing study. Given the small (150 mg) test portions normally used for controlled substances analysis the particle size must be sufficiently small and uniformly distributed to ensure minimal influence on analytical precision. The homogeneity testing is described in Appendix 2. The study samples were demonstrated to be sufficiently homogeneous. 3 AQA 13-12 Cocaine
2.6 Sample Dispatch and Receipt
A set of three test samples, each containing approximately 150 mg of test material, were
dispatched on 16 September 2013.
The following items were packaged with the samples:
• a covering letter with instructions for participants; and
• a faxback form for participants to confirm the receipt of the test samples.
An Excel spreadsheet for the electronic reporting of results was e-mailed to participants.
2.7 Instructions to Participants
Participants were asked to analyse the samples using their routine quantitative method and
return the following information:
• one result for each sample as % m/m cocaine base;
• an estimate of the expanded uncertainty associated with the result as % m/m cocaine
base at the 95% confidence level;
• brief detail on how the uncertainty was calculated e.g. uncertainty budget method;
• the identity of the cutting agent in all three samples, if part of routine analysis;
• origin and stated purity of the analytical reference standard used;
• brief summary of the quantitative method used;
• the completed results sheet by 08 November 2013, as late results cannot be included in the
report; and
• any other comment.
2.8 Interim Report
An interim report was emailed to all participants on 15 November 2013.
AQA 13-12 Cocaine 43 PARTICIPANT LABORATORY INFORMATION
3.1 Test Method Summaries
Reported participant method summary is presented for information in Table 1.
Table 1 Participant Summary of Test Methods
Lab. Extraction Internal Calibration Detecto
Technique Column
Code solvent standard points r
Acetonitrile:Water ODS Hypersil (C18, 5µm,
1 Diethylphthalate 3 HPLC PDA
(75:25) 250mm x4.6mm)
2 Methanol 7 HPLC DAD RP-18 ZORBAX
S1: chloroform Benzopinacolone 1 S1. GC FID HP1
3
S3: mobile phase None 4 S3. HPLC UV C18 ubondapak
4 Methanol Loxapine 5 HPLC DAD X TERRA C18
Acetic acid/
UV POROSHELL 120 EC-
5 acetonitrile/ water None 4 HPLC
DAD C18
(4/20/76 V/V/V).
6 Methanol None 5 HPLC DAD Phenomenex C18-XB
Acetonitrile/Metha Pholcodine
7 3 UPLC PDA Waters Acquity C-18
nol (95:5) 1mg/ml
Phenomenex C18 %um
8 Methanol None 5 HPLC DAD
Luna
Phenomex C8-Luna 3u
10 Water None 5 HPLC DAD
Narrow Bore 100mm
11 Methanol Diazepam 6 GC FID J&W 128-5512
Isopropyl
12 Ethyl acetate 5 GC FID HP5
cocaine
13 Ethanol Tribenzylamine 6 GC FID HP5
Methanol/
14 Triphenylamine 4 GC FID HP5 MS
chloroform 20:80
Agilent Lichrospher 60
15 Methanol Lidocaine 1 HPLC DAD
RP-Select B
External standard Phenomenex C8 Luna 3u
16 Methanol 5 HPLC DAD
calibration Narrow Bore 100mm
Acetonitrile/
17 None 3 HPLC UV Column C8
eau (80/20)
Water/acetonitrile/
Diode
18 n10 sulphuric None 3 HPLC shimpack XR-ODS
array
90:10:2
19 GC FID SGE 12 m x 0,22 mm
Methanol Tetracosane 4
(F/N054046)
10%
FID &
20 Dichloromethane Eicosane 3 GC HP1 (FID & MSD)
MSD
in Ethanol
5 AQA 13-12 CocaineLab. Extraction Internal Calibration Detecto
Technique Column
Code solvent standard points r
S1, S3: Procaine 5 GC FID DB 35
Butylacetate hydrochloride
(C6H12O2)
21
3,5-
S2: d6-DMSO dimetoxybens- NMR proton experiment
aldehyde
23 Acetonitril/water None 1 HPLC DAD Kromasil
24 Ethanol Tetracosane 3 GC FID BPX5
25 Methanol Bupivacaine 5 GC FID ZB-Semivolatiles
Sodiumphosphate UV- Zorbax Eclipse XBD-C8
26 None 4 HPLC
(pH 2.4) DAD (Agilent)
ACN, MeOH:H2O
27 None 5 HPLC DAD C-18 coloumn
(25:25:50)
28 Ethanol Propyl paraben 7 UPLC PDA BEH Shield RP 18
29 Ethanol Ethaverine 6 GC FID HP5MS
30 Chloroform C28 n-alkane 4 GC FID HP1
31 Methanol ----------------- 3 HPLC DAD ECLIPSE- XDB-C18
Acetonitrile/ Pholcodine Waters Acquity
32 3 UPLC PDA
Methanol (95:5) 1mg/ml C-18
AQA 13-12 Cocaine 63.2 Reported Basis of Participants’ Measurement Uncertainty Estimates
Participant returns as received are listed in Table 2.
Table 2 Reported Basis of Uncertainty Estimate
Lab. Approach to Estimating Information Sources for MU Estimation Guide Document for
Code MU Estimating MU
Precision Method Bias
Precision and estimates
of method bias and Control samples,
1 Standard Purity Eurachem/CITAC Guide
standard deviation of Duplicate analysis
replicate analysis
Standard Purity,
Control samples, Matrix Effects, Instrument
2 Validation data Eurachem/CITAC Guide
Duplicate analysis calibration, Masses and
Volumes, Homogeneity
3
4
Expanded uncertainty
ISO 5725-2 and ISO /TS
5 (k=2) Based of RSD Control samples, Standard Purity, Matrix Effects
21748
intermediate precision
Bottom Up (ISO/GUM,
Standard Purity, Matrix
6 fish bone/ cause and Control samples Eurachem/CITAC Guide
Effects, Instrument calibration
effect diagram)
Top Down - precision
Standard Purity,
and estimates of the Control samples,
7 Instrument calibration, Masses NATA Technical Note 33
method and laboratory Duplicate analysis
and Volumes, Homogeneity
bias
Top Down - precision
and estimates of the Control samples, Laboratory bias from PT
8 Eurachem/CITAC Guide
method and laboratory Duplicate analysis studies
bias
Standard Purity,
Control samples,
10 Top Down approach Instrument calibration, Masses NATA Technical Note 33
Duplicate analysis
and Volumes
Estimating Measurement
11 Uncertainty by black box ISO/GUM
by pairs of values
Top Down - precision Laboratory bias from PT
and estimates of the Control samples, studies, Standard Purity,
12 Eurachem/CITAC Guide
method and laboratory Duplicate analysis Instrument calibration, Masses
bias and Volumes
Top Down - precision
and estimates of the
13 Control samples Standard Purity
method and laboratory
bias
Top Down - precision
and estimates of the
14 Control samples Standard Purity, Homogeneity Eurachem/CITAC Guide
method and laboratory
bias
Top Down - precision
and estimates of the Laboratory bias from PT
15 Control samples Eurachem/CITAC Guide
method and laboratory studies, Standard Purity
bias
Top Down - precision
Standard Purity,
and estimates of the Control samples,
16 Instrument calibration, Masses
method and laboratory Duplicate analysis
and Volumes
bias
7 AQA 13-12 CocaineLab. Approach to Estimating Information Sources for MU Estimation Guide Document for
Code MU Estimating MU
Precision Method Bias
Top Down - precision
and estimates of the Laboratory bias from PT
17 Control samples
method and laboratory studies, Standard Purity
bias
Control samples, Standard Purity,
18 Professional judgement ISO/GUM
Duplicate analysis Instrument calibration
Laboratory bias from PT
Top Down - precision studies, Recoveries of Spiked
and estimates of the Samples, Standard Purity, EA-4/16: 2003 and ILAG
19 Control samples
method and laboratory Matrix Effects, Instrument G-17:2002
bias calibration, Masses and
Volumes, Homogeneity
Laboratory bias from PT
studies, Recoveries of Spiked
Bottom Up (ISO/GUM,
Control samples, Samples, Standard Purity,
20 fish bone/ cause and Eurachem/CITAC Guide
Duplicate analysis Matrix Effects, Instrument
effect diagram)
calibration, Masses and
Volumes, Homogeneity
Top Down - precision
Laboratory bias from PT
and estimates of the
21 Control samples studies Nordtest Report TR537
method and laboratory
bias
Standard deviation of
23 replicate analyses Control samples ISO/GUM
multiplied by 2 or 3
Standard Purity,
Uncertainty budget Control samples,
24 Instrument calibration, Masses In-house Methods Manual
method Duplicate analysis
and Volumes
Top Down - precision
and estimates of the Control samples,
25 Recoveries of Spiked Samples Internal validation
method and laboratory Duplicate analysis
bias
Standard deviation of Standard deviation from
Laboratory bias from PT
26 replicate analyses Control samples control samples and z-
studies
multiplied by 2 or 3 value from PT-studies.
Top Down - precision
and estimates of the
27 Control samples
method and laboratory
bias
28
Top Down - precision Internal quality online
and estimates of the Control samples document based on
29
method and laboratory Eurachem/CITAC,
bias ISO/GUM
Uncertainty Budget
Control samples, Laboratory bias from PT
30 (Ustandard, Umean, U ISO/GUM
Duplicate analysis studies
reproducibility)
Top Down - precision
and estimates of the Standard Purity, Instrument
31 Duplicate analysis Nata Technical Note 33
method and laboratory calibration
bias
Precision and estimates
Standard Purity,
of method bias and Control samples,
32 Instrument calibration, Masses Eurachem/CITAC Guide
standard deviation of Duplicate analysis
and Volumes, Homogeneity
replicate analysis
AQA 13-12 Cocaine 83.3 Details of Participant Calibration Standard
Participant returns as received are listed in Table 3.
Table 3 Participant Calibration Standard
Lab. Purity
Reference Standard* (%)
Code
1 NMI 99.6
2 LIPOMED 99.774
3 NMI
4 Lipomed 1 mg/ml >98
5 SIGMA 98.6
6 Lipomed >98.5
7 NMI 96.1
8 N/A 100
10 NMI 96.1
11 LIPOMED 99.6
12 NMI 96.1
13 Lipomed 99.4
14 Cerillinat 98.8
15 NMI 99.6
16 NMI 96.1 +/- 2.6
17 Lipomed 88.8
18 Sigma-aldrich 98.4
19 Merck
20 NMI 96.1
21 Apoteket AB (supplier of Medical drugs) 100
23 Lipomed 99.6
24 NMI 96.1
25 Sigma 98.4
26 Other* ca 100
27 Unikem 100
28 NMI >98
29 NMI 96.1
30 Siegfried 99.8
31 LGC STANDARDS 1,000±0,003 mg/mL(ACN)
32 NMI 96.1
*Some data has been edited to preserve confidentiality
9 AQA 13-12 Cocaine3.4 Participants’ Comments
The study manager welcomes comments or suggestions from participants as it provides
information which will improve future studies. All returns are listed as received in Table 4
along with the study manager response, where appropriate.
Table 4 Participant Comments
Lab.
Participant comments Study Manager’s response
Code
S1 and S2 failed quality standards Most participants use less than 50 mg
S1 was repeated using a less accurate method due to sample size for each analysis. For reasons of
3 restraints. security and accountability, NMI
S3 could not be repeated due to insufficient sample. conducts these PT’s using the
Please provide more sample in future studies minimum practical amount of drug.
6 Average of two determination
Tetramisole was not reported for Sample S3 as the laboratory does not
15
have the reference standard.
The name tetramisole has been used because the analysis cannot
20
discriminate between the stereoisomers, levamisole and dexamisole.
Sample 1 and 3 analysed by GC-FID
21
Sample 2 analysed by NMR
24 A greater quantity of powder for each sample would be beneficial. See response for laboratory 3
28 Cocaine is present as cocaine hydrochloride
AQA 13-12 Cocaine 104 PRESENTATION OF RESULTS AND STATISTICAL ANALYSIS
4.1 Results Summary
Participant results are listed in Tables 5 to 7 with resultant summary statistics: mean, median,
maximum, minimum, robust average, robust standard deviation (Robust SD) and robust
coefficient of variation (Robust CV).
Bar charts of results and performance scores are presented in Figures 2 to 5.
An example chart with interpretation guide is shown in Figure 1. Independent estimates of analyte
concentration with associated expanded
uncertainties (coverage factor is 2).
Distribution of results around the assigned Assigned value and associated
value as kernel density estimate expanded uncertainty (coverage Md = Median (of participants’ results)
(illustrates participant consensus). factor is 2). R.A. = Robust Average
H.V. = NMI Homogeneity Value
Uncertainties reported by
participants.
Figure 1 Guide to Presentation of Results
4.2 Assigned Value
The assigned value is defined as: ‘value attributed to a particular quantity and accepted,
sometimes by convention, as having an uncertainty appropriate for a given purpose’ 4.
For a proficiency test, the assigned value is the best available measurement of the true
concentration of an analyte in the test sample.
4.3 Between-Laboratory Coefficient of Variation
The between laboratory coefficient of variation is a measure of the between laboratory
variation that in the judgement of the study coordinator would be expected from participants
given the analyte concentration. It is important to note this is not the coefficient of variation of
participant results.
11 AQA 13-12 Cocaine4.4 Target Standard Deviation
The target standard deviation (σ) is the product of the assigned value (Χ) and the between-
laboratory coefficient of variation (CV) as presented in Equation 1. This value is used for
calculation of participant z-score.
σ = Χ * CV Equation 1
4.5 z-Score
For each participant result a z-score is calculated according to Equation 2 below:
(χ − X )
z= Equation 2
σ
where:
z is z-score
χ is participant result
Χ is the study assigned value
σ is the target standard deviation from equation 1
A z-score with absolute value (|z|):
• |z| ≤ 2 is satisfactory;
• |z| > 2 ≤ 3 is questionable;
• |z| > 3 is unsatisfactory.
4.6 En-Score
The En-score is complementary to the z-score in assessment of laboratory performance.
En-score includes measurement uncertainty and is calculated according to Equation 3 below:
(χ − X )
En = Equation 3
U χ2 + U X2
where:
En is En-score
χ is a participant’s result
Χ is the assigned value
Uχ is the expanded uncertainty of the participant’s result
UX is the expanded uncertainty of the assigned value
An En-score with absolute value (|En|):
• |En| ≤ 1 is satisfactory;
• |En| > 1 is unsatisfactory.
4.7 Traceability and Measurement Uncertainty
Laboratories accredited to ISO/IEC Standard 17025:20055 must establish and demonstrate the
traceability and measurement uncertainty associated with their test results. Guidelines for
quantifying uncertainty in analytical measurement are described in the Eurachem /CITAC
Guide. 6
AQA 13-12 Cocaine 125 TABLES AND FIGURES
Table 5
Sample Details
Sample No. S1
Matrix. Powder
Analyte. Cocaine
Units % base (m/m)
Participant Results
Lab Code Result Uncertainty z-Score En-Score
1 24.6 1.9 -1.43 -0.52
2 27 2.37 1.69 0.51
3 25.5 3.4 -0.26 -0.06
4 26.9 NR 1.56 1.33
5 22.4 1.5 -4.28 -1.89
6 27 2.5 1.69 0.49
7 24.5 1.5 -1.56 -0.69
8 24.0 3.5 -2.20 -0.47
10 25.7 2.57 0.00 0.00
11 31 2 6.87 2.42
12 25.4 0.9 -0.39 -0.24
13 24.6 6.4 -1.43 -0.17
14 28.3 0.8 3.37 2.16
15 27.1 1.8 1.82 0.70
16 24 2.4 -2.20 -0.66
17 23.3 2.5 -3.11 -0.90
18 26 1.95 0.39 0.14
19 25.8 1.3 0.13 0.06
20 30.0 2.0 5.58 1.96
21 25.2 10 -0.65 -0.05
23 28 4.8 2.98 0.47
24 21 1.0 -6.10 -3.49
25 26.5 1.45 1.04 0.47
26 24.3 2 -1.82 -0.64
27 25.2 2.5 -0.65 -0.19
28 25.3 NR -0.52 -0.44
29 29.2 2 4.54 1.60
30 27.3 0.6 2.08 1.48
31 24.29 2.44 -1.83 -0.54
32 25.4 1.5 -0.39 -0.17
Statistics
Assigned Value 25.7 0.9
Robust Average 25.7 0.9
Median 25.5 0.7
Mean 25.8
N 30
Max. 31
Min. 21
Robust SD 1.87
Robust CV 7.3%
13 AQA 13-12 CocaineFigure 2 AQA 13-12 Cocaine 14
Table 6
Sample Details
Sample No. S2
Matrix. Powder
Analyte. Cocaine
Units % base (m/m)
Participant Results
Lab Code Result Uncertainty z-Score En-Score
1 12.1 0.9 0.75 0.58
2 10.4 0.95 -1.37 -1.02
3 NR NR
4 10.2 NR -1.61 -2.60
5 10.6 0.7 -1.12 -1.05
6 13 1.2 1.86 1.15
7 9.3 1.2 -2.73 -1.69
8 11.6 1.7 0.12 0.06
10 9.5 0.95 -2.48 -1.86
11 13 1 1.86 1.34
12 11.5 0.4 0.00 0.00
13 11.4 6.4 -0.12 -0.02
14 11.6 0.3 0.12 0.17
15 11.2 0.8 -0.37 -0.32
16 11 1.1 -0.62 -0.41
17 12.9 2.5 1.74 0.55
18 12 0.9 0.62 0.49
19 12.7 0.7 1.49 1.39
20 12.2 0.8 0.87 0.74
21 10.5 1.1 -1.24 -0.83
23 13 4.8 1.86 0.31
24 11.5 0.5 0.00 0.00
25 11.9 1.39 0.50 0.27
26 11.2 2 -0.37 -0.15
27 10.9 1.1 -0.75 -0.50
28 11.7 NR 0.25 0.40
29 12.7 1 1.49 1.07
30 12 0.3 0.62 0.86
31 9.85 1.03 -2.05 -1.44
32 9.5 1.2 -2.48 -1.54
Statistics
Assigned Value 11.5 0.5
Robust Average 11.5 0.5
Median 11.5 0.4
Mean 11.4
N 29
Max. 13
Min. 9.3
Robust SD 1.14
Robust CV 9.9%
15 AQA 13-12 CocaineFigure 3 AQA 13-12 Cocaine 16
Table 7
Sample Details
Sample No. S3
Matrix. Powder
Analyte. Cocaine
Units % base (m/m)
Participant Results
Lab Code Result Uncertainty z-Score En-Score
1 16.6 1.5 1.15 0.51
2 16.4 1.57 0.89 0.39
3 14.8 0.7 -1.15 -0.79
4 12.1 NR -4.59 -4.00
5 13.8 0.9 -2.42 -1.49
6 16 1.5 0.38 0.17
7 15.0 1.3 -0.89 -0.44
8 16.6 2.4 1.15 0.35
10 13.1 1.31 -3.31 -1.64
11 15 1 -0.89 -0.52
12 14.7 0.5 -1.27 -0.97
13 13.6 6.4 -2.68 -0.32
14 14.9 0.4 -1.02 -0.81
15 15.5 1.0 -0.25 -0.15
16 17 1.7 1.66 0.68
17 17.3 2.5 2.04 0.60
18 18 1.35 2.93 1.42
19 17.4 0.9 2.17 1.34
20 15.8 1.0 0.13 0.07
21 14.6 10 -1.40 -0.11
23 19 4.8 4.20 0.68
24 13.5 0.6 -2.80 -2.03
25 19.3 2.04 4.59 1.61
26 15.6 2 -0.13 -0.05
27 13.2 1.3 -3.18 -1.58
28 17.1 NR 1.78 1.56
29 19 2 4.20 1.50
30 17 0.4 1.66 1.32
31 15.25 1.39 -0.57 -0.27
32 15.3 1.3 -0.51 -0.25
Statistics
Assigned Value 15.7 0.9
Robust Average 15.7 0.9
NMI Homogeneity 16.6 1.1
Median 15.6 0.7
Mean 15.7
N 30
Max. 19.3
Min. 12.1
Robust SD 1.86
Robust CV 12%
17 AQA 13-12 CocaineFigure 4 AQA 13-12 Cocaine 18
Table 8 Participants’ identification of cutting agents
Lab Cutting agents
Code S1 S2 S3
1 Procaine Levamisole
2 Procaine Levamisole
3 Glucose Procaine, glucose Levamisole, glucose
4 Procaine Levamisole
5 Glucose Glucose, procaine Glucose, levamisole
6 ND Procaine Levamisole
7 Procaine Levamisole
8 Not determined Not determined Not determined
10 Procaine Levamisole
11 Procaine Levamisole
12 Procaine Levamisole
13 Sucrose Procaïne Levamisol
14 Procaine Levamisole
15 Glucose Glucose, procaine Glucose, Tetramisol
16 Glucose Glucose, procaine Glucose, levamisole
17 Glucose Procaine, glucose Glucose, levamisole
18 None detected Procaine Levamisole
19 None Procaine Levamisole
20 Glucose Glucose and procaine Glucose and tetramisole*
21 None Procaine Tetramisol
23 Glucose Glucose - procaïne Glucose - levamisole
Levamisole and glucose
24 Glucose (indications) Procaine & glucose (indications)
(indications)
25 Glucose Procaine Glucose
26 None Procaine Levamisole/tetramisole
27 None Procaine Levamisole
Glucose 56.7% Glucose 63.1%
28 Glucose 64.2%
Procaine 23.8% Lévamisole 11.7%
29 Glucose Procaine, Glucose Levamisole, Glucose
30 Glucose Procaine Levamisole
31 Procaine Tetramisole
32 Procaine Levamisole
19 AQA 13-12 Cocaine6 DISCUSSION OF RESULTS
6.1 Assigned Value
The assigned value is the robust average of the results reported by the participants. The robust
average and associated expanded uncertainties were calculated using the procedure described
in ‘ISO13258:2005(E), Statistical methods for use in proficiency testing by interlaboratory
comparisons’.7 The calculation procedure for the expanded uncertainty is presented in
Appendix 3.
Traceability: The consensus of participants’ results is not traceable to any external reference,
so although expressed in SI units, metrological traceability has not been established.
6.2 Measurement Uncertainty Reported by Participants
Participants were asked to report an estimate of the expanded measurement uncertainty
associated with their results and the basis of this uncertainty estimate (Table 2).
It is a requirement of the ISO Standard 170255 that laboratories have procedures to estimate
the uncertainty of chemical measurements and to report this uncertainty in specific
circumstances, including ‘when the client’s instruction so requires.’ From 1 July 2012 this is
also a requirement of ASCLD/Lab-International accreditation program.
Eighty-three results (93%) were reported with an associated expanded uncertainty.
Laboratories 4 and 28 did not report uncertainty. Both laboratories were not accredited.
Laboratories 13, 17, 21, 23, 26 and 29 reported an identical uncertainty for samples which
were of significantly different concentrations.
The magnitude of reported uncertainties was within the range 2% to 68% relative.
Fifty-eight of eighty-three (70%) expanded uncertainties were between 3% and 10% relative
to the result. Laboratories reporting uncertainties smaller than 3% or larger than 10% relative
may wish to consider whether these estimates are fit for purpose.
Laboratories having a satisfactory z-score and an unsatisfactory En-score are likely to have
underestimated the expanded uncertainty associated with the result.
In some cases the results were reported with an inappropriate number of significant figures.
The recommended format is to write the uncertainty to no more than two significant figures
and then to write the result with the corresponding number of decimal places (for example
instead of 16.4 ± 1.57% the recommended format is 16.4 ± 1.6%).6
6.3 z-Score
Target standard deviations equivalent to 3%, 5% and 7% CV were used by the study
coordinator to calculate z-scores in order to achieve comparable target standard deviation
(SD) for all three samples. Target SDs, the between laboratory coefficient of variation
predicted by Thomson - Horwitz equation8 and between laboratories coefficient of variation
obtained in this study are presented in Table 9.
AQA 13-12 Cocaine 20Table 9 Target standard deviations, coefficient of variations from predictive model and
between laboratories
Target SD Target SDa Thompson Between
Assigned value
Sample Analyte (as CV) (% base m/m) Horwitz laboratories
(% base m/m)
CV CV
S1 Cocaine 25.7 3% 0.8 2.4% 7.3%
S2 Cocaine 11.5 7% 0.8 2.8% 9.9%
S3 Cocaine 15.7 5% 0.8 2.6% 11.8%
a
Calculated using equation 1.
A summary of z-scores by laboratory is presented in Figure 5.
10
8
6
4
2
z-Score
0
-2
-4
-6
-8
-10
1 2 3 4 5 6 7 8 10 11 12 13 14 15 16 17 18 19 20 21 23 24 25 26 27 28 29 30 31 32 All
Figure 5 Summary of participants’ z-score.
Sixty-two of eighty-nine numeric results (70%) returned a satisfactory z-score with |z| ≤ 2.
• Eight laboratories (27%) - 1, 2, 6, 12, 15, 21, 26 and 28 returned a satisfactory score for
all three samples. Laboratory 3 reported results for only two of the three samples, all
results returned satisfactory z-scores; and
• Twenty-one laboratories returned at least one questionable z-score.
6.4 En-Score
The dispersal of participants’ En-scores is graphically presented in Figure 6. Where a
laboratory did not report an expanded uncertainty with a result, an expanded uncertainty of
zero (0) was used to calculate the En-score.
21 AQA 13-12 Cocaine10
8
6
4
2
En-Score
0
-2
-4
-6
-8
-10
1 2 3 4 5 6 7 8 10 11 12 13 14 15 16 17 18 19 20 21 23 24 25 26 27 28 29 30 31 32 All
Figure 6 Summary of participants’ En-Score
Fifty-nine of eighty-nine numeric results (66%) returned a satisfactory En-score with |En| ≤ 1.
• Ten laboratories (33%) – 1, 8, 12, 13, 15, 16, 17, 21, 23 and 26 returned a satisfactory
scores for all three samples; Laboratory 3 reported results for only two of the three
samples, all results returned satisfactory En –scores;
• Sixteen laboratories returned at least one questionable En-score; and
• Laboratories 4, 5, and 29 returned |En| > 1 for all test samples. This indicates that the
reported results did not agree with the assigned value and that the uncertainties associated
with the results have been understated.
6.5 Identification of Cutting Agent
Twenty-nine participants (97%) reported on the identity of the cutting agents and the findings
are presented in Table 8.
All test samples were prepared using an illicit seizure of cocaine hydrochloride,
approximately 73 % base (m/m) supplied by the Australian Federal Police. The following
were added by the study coordinator: glucose to all three test samples, procaine to Sample S2
and levamisole to Sample S3.
Eleven participants correctly identified all the adulterants added to the test samples. Some
laboratories are not routinely testing for and did not report benign adulterants such as sugars.
Laboratory 13 incorrectly reported sucrose in Sample S1.
Twenty-eight participants correctly identified procaine in Sample S2 and levamisole in
Sample S3
AQA 13-12 Cocaine 226.6 Participants’ Analytical Methods
Participants were requested to analyse the samples using their normal test methods and to
report a single result for each sample as they would normally report to a client. Results
reported in this way reflect the true variability of results reported to laboratory clients. The
method descriptions provided by participants are presented in Table 1.
A summary of accreditation status, participants’ methods and reference standards is presented
below.
Accredited Laboratory Code
1, 2, 3, 5*, 6, 7, 8, 10, 11, 12, 16, 18, 19, 20, 21, 23,
Yes to ISO 17025 25, 27, 29, 30, 32
*Not accredited for quantification - only for detection confirmation
Yes to ASCLD/Lab International 15, 24
No 4, 13, 14, 17, 26, 28, 31
Sample Mass Used (mg) Laboratory Code
4-10 4, 11, 18, 21, 25, 26, 31
11-30 2, 5, 6, 7, 8, 10, 13, 14, 17, 18, 20, 21, 23, 28, 32
31-50 1, 15, 17, 19, 24, 27, 29, 30, 32
51-100 12, 16, 24
101-≥150 3
Instrument Used for quantification Laboratory Code
GC/FID 3, 11, 12, 13, 14, 19, 20, 21, 24, 25, 29, 30
HPLC 1, 2, 3, 4, 5, 6, 8, 10, 15, 16, 17, 18, 23, 26, 27, 31
UPLC 7, 28, 32
Sources of Calibration Standard Laboratory Code
NMI Australia 1, 7, 10, 12, 15, 16, 20, 24, 28, 29, 32
Lipomed 2, 4, 6, 11, 13, 17, 23,
Sigma Aldrich 5, 18
Other 14, 19. 21, 26, 27, 30, 31
Plots of measurement extraction solvent vs z-score, instrument used vs z-score and calibration
standard vs z-score are presented in Figures 7, 8 and 9. No trends were identified.
23 AQA 13-12 Cocaine10
8
6
4
2
z-Score
0
-2
-4
-6
-8
-10
Chloroform
Acetonitrile
Methanol
Ethanol
Others
Figure 7 Extraction solvent vs z-score
10
8
6
4
2
z-Score
0
-2
-4
-6
-8
-10
GC-FID
HPLC
UPLC
Figure 8 Measurement instrument vs z-score
AQA 13-12 Cocaine 2410
8
6
4
2
z-Score
0
-2
-4
-6
-8
-10
Sigma-Aldrich
Lipomed
Other
NMIA
Figure 9 Calibration standard vs z-score
6.7 Summary of participation and performance in Cocaine Studies
Overall percentages of satisfactory z-scores and En-scores obtained by laboratories since 2006
are presented in Figure 10. The proportion of satisfactory z-scores over 7 years on average is
73% while for En-scores is 72%.
100% 35
30 30
90% 29
28 28 30
80%
70% 22 25
Number of participants
% Satisfactory scores
20 26
60%
20
50%
86% 83% 15
40%
69% 74%
68% 68% 64% 66%
30% 10
20%
5
10%
89% 63% 53% 69% 81% 86% 71% 70%
0% 0
AQA 06-01 AQA 07-06 AQA 08-07 AQA 09-10 AQA 10-09 AQA 11-09 AQA 12-11 AQA 13-12
Satisfactory z-score Satisfactory En-score
Figure 10 Summary of participants’ performance since 2006
25 AQA 13-12 Cocaine7 REFERENCES
[1] ISO/IEC 17043:2010, Conformity assessment – General requirements for proficiency
testing, ISO, Geneva
[2] NMI Chemical Proficiency Testing Study Protocol
http://www.measurement.gov.au → services → chemical proficiency testing
[3] NMI Chemical Proficiency Testing Statistical Manual
http://www.measurement.gov.au → services → chemical proficiency testing
[4] Thompson, M. Ellison, S. L. R. and Wood, R., The international harmonized protocol
for proficiency testing of (chemical) analytical laboratories, Pure Appl. Chem. 78,
145-196, 2005.
[5] ISO/IEC 17025:2005 General requirements for the competence of testing and
calibration laboratories (2005) ISO, Geneva
[6] Eurachem/CITAC Guide Quantifying uncertainty in analytical measurement third
edition, (2012), http://www.eurachem.org/index.php/publications/guides/quam.
[7] ISO 13528:2005(E), Statistical methods for use in proficiency testing by interlaboratory
comparisons.
[8] Thompson, M., and Lowthian, P.J., (1995), A Horwitz-like function describes precision
in a proficiency test, Analyst, 120, 271-272.
[9] Thompson, M. and Fearn, T., A new test for sufficient homogeneity, Analyst, 126, 1414-
1417, 2001.
[10] AQA 10-09 Pesticides in Fruit and Vegetables. November 2010
http://www.measurement.gov.au → services → chemical proficiency testing
[11] ISO/IEC Guide 98:2008 Guide to Expression of Uncertainty in Measurement, ISO,
Geneva
AQA 13-12 Cocaine 26APPENDIX 1 - PARTICIPANT LABORATORIES
ACT Government Analytical Laboratory, ACT Centre for Forensic Science
Health Sciences Authority, SINGAPORE
CHEMCENTRE, WA Environmental Science and Research Ltd
Mt. Albert Science Centre, NEW ZEALAND
Forensic & Analytical Science Services, NSW Forensic Science, SA
Forensic Institute, Odense I.N.C.C., BELGIUM
Syddansk Universitet, DENMARK
Forensic Science Services Tasmania, TAS Instituto Nacional De Toxicologia Y Ciencias Forensics
Departamento de Sevilla, SPAIN
Instituto Nacional de Toxicologia y Ciencias IRCGN, FRANCE
Forenses Departamento de Madrid, SPAIN
Instituto Nacional de Toxicologia Laboratoire Toxlab s.a.s., FRANCE
Departamento de Barcelona, SPAIN
Laboratoire Toxgen, FRANCE Lancashire Constabulary Headquarters, UK
Laboratory of Environmental Hygiene and Forensic Ministry of Justice Netherlands Forensic Institute,
Toxicology, ITALY NETHERLANDS
Ministero dell'interno Dipartimento National Criminal Investigation Service/Kripos,
Pubblica Sicurezza, ITALY NORWAY
MSA Ltd, UK Queensland Health Forensic and Scientific Services,
QLD
National Measurement Institute, NSW Service Commun de Laboratoires
Laboratoire de Lille, FRANCE
University of Copenhagen, DENMARK Skl - Kriminaltekniska Laboratoriet, SWEDEN
Service Commun des Laboratoires Victoria Police Forensic Services Dept., VIC
Laboratoire de Paris, FRANCE
University of Aarhus, Institut of Forensic Medicine
Department of Toxicology and Drug Analysis,
DENMARK
27 AQA 13-12 CocaineAPPENDIX 2 - HOMOGENEITY TESTING
Homogeneity testing was based on that described by Thompson and Fearn,9 which is also the
procedure described in the International Protocol.4
Seven sample bottles were selected at random from Sample S3 and analysed for the purpose
of assessing homogeneity.
Measurements were made using exact-matching isotope dilution with gas chromatography
mass spectrometry (GCMSD), in single ion monitoring mode. The mass fraction of cocaine
was also determined using IDMS with liquid chromatography tandem mass spectrometry
(LCMSMS) in selected reaction mode. The LCMSMS data were only used to investigate
potential method bias. A detailed description of the measurement technique used is presented
in the report of NMI Proficiency Test AQA 10-09.10
The Reference Standard of cocaine base was obtained from NMI’s Chemical Reference
Materials group. The purity data supplied with the material is summarised in Table 10.
Table 10 NMI Reference Standard
Supplier Purity (95 % confidence)
NMI Cocaine base D826B, Batch No 09-D-29 96.1% ± 2.6%
Results for the homogeneity testing of Sample S3 are summarised in Table 11. The average of
results for each sample was used as the homogeneity value.
Table 11 Homogeneity Testing for Sample S3
Bottle Fill Cocaine (% base m/m)
No.
Replicate 1 Replicate 2
303 17.6 16.8
314 17.1 16.2
315 17.0 16.8
324 17.0 17.3
330 16.9 16.6
332 15.5 16.0
333 15.1 17.1
Mean 16.6
CV 4.2%
Homogeneity value for Sample S3 is 16.6 ± 1.1% cocaine base (m/m)a
a
The uncertainty is an expanded uncertainty at 95% confidence level using a coverage factor of k=2.05 calculated using effective degrees of
freedom from the Welch-Satterthwaite equation11.
Thompson and Fearn Homogeneity Tests10
Critical
Test Test Value Result
Value
Cochran 0.42 0.78 Pass
Sa/σ 0.48 0.5 Pass
s2sam 0.199 0.41 Pass
Results on bottle fill no 333 were not included in the test for homogeneity being identified as an analytical outliers (test Sa/σ) due
to the difference between replicates.10
AQA 13-12 Cocaine 28APPENDIX 3 - MEASUREMENT UNCERTAINTY OF THE ASSIGNED VALUE
When the assigned value is calculated as the robust average using the procedure described in
‘ISO13258:2005(E), Statistical methods for use in proficiency testing by interlaboratory
comparisons – Annex C’7, the uncertainty is estimated as:
urob average = 1.25*Srob average / p Equation 4
where:
urob average robust average standard uncertainty
Srob average robust average standard deviation
p number of results
The expanded uncertainty (Urob average) is the standard uncertainty multiplied by a coverage
factor of 2 at approximately 95% confidence level.
A worked example is set out below in Table 10.
Table 10 Uncertainty of assigned value for Sample S1 as % base (m/m)
No. results (p) 30
Robust average 25.73
Srob average 1.87
urob average 0.43
k 2
Urob average 0.86
The assigned value for Sample S1 is 25.7 ± 0.9% cocaine base (m/m).
29 AQA 13-12 CocaineAPPENDIX 4 - ACRONYMS AND ABBREVIATIONS
CRM Certified Reference Material
CV Coefficient of Variation
DAD Diode Array Detector
|En| Absolute value of an En-score
FID Flame Ionization Detector
GC Gas Chromatography
GC-MS Gas Chromatography Mass Spectrometry
HPLC High Performance Liquid Chromatography
ISO International Standards Organisation
LC Liquid Chromatography
Max Maximum value in a set of results
Md Median
Min Minimum value in a set of results
NATA National Association of Testing Authorities
NMI National Measurement Institute Australia
NR Not Reported
NT Not Tested
PDA Photodiode array
PT Proficiency Test
Robust CV Robust Coefficient of Variation
Robust SD Robust Standard Deviation
SI International System of Units
Target SD (σ) Target standard deviation
UPLC Ultra Performance Liquid Chromatography
UV Ultraviolet
|z| Absolute value of a z-score
AQA 13-12 Cocaine 30You can also read
