DETERMINATION OF MELAMINE IN MILK INTER-LABORATORY COMPARISON PROGRAMME HKGL 0803
FINAL REPORT
6 November 2008
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
CONTENTS
Page
1.
Summary of Results
.……………………………………………………. 1
2.
Introduction
3.
Objectives
4.
Test Materials
…………………………………………………....………… 3
5.
Test Required
……..…………………………………….………………. 4
6.
Statistical Evaluation
7.
Results and Discussions ……………………………………….………… 5
8.
Disclaimer
9.
Acknowledgement
10.
References
11.
Tables I to VI
12.
Charts I to II
13.
Appendices I to VII
….……………………………………………………………. 2 …………………………………………………….……….……. 2
…………………………….………………………. 4
…………………………….………………………...…………. 7 ……………………………………………………… 7
……………………………….……………….………………… 8 ……………………………………………………………
9-16
……………………………………………………… 17-19 ……………………………………………………
20-28
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
Summary of Results 1.
The inter-laboratory comparison programme (HKGL 0803) aimed at evaluating the testing capability of participants on the quantitative analysis of melamine in milk sample. Fourteen Hong Kong testing laboratories registered to participate in the programme and thirteen of them returned the results to the organizer.
2.
The z-score is used as the numerical indicator to assess individual participant’s competence relative to the others in the programme. The assigned values of melamine were derived from the quantities of melamine gravimetrically spiked into the samples. The target standard deviations for the assessment were calculated using the Horwitz Equation9.1.
3.
The z-scores achieved by participants are summarized as follows: Performance Achieved |z| ≤ 2 2 < |z| < 3 |z| ≥ 3
Laboratory Code Sample 1
Sample 2
Sample 3
Sample 4
1, 4, 7
1, 2, 3, 4, 5, 7, 9, 10, 11, 13
*
1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13
---
12
*
---
2, 6, 8, 12
6, 8
*
6
“*” not evaluated for a qualitative analysis. Assessment results for Sample 3 are listed in clause 6.3.3.
1
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
1. Introduction 1.1.
The common industrial use of melamine is the production of melamine resin which is a very durable thermosetting plastic through its polymerisation with formaldehyde. Melamine is also used in the production of melamine foam, a polymeric cleaning product. The end products include countertops, dry erase boards, fabrics, glues, housewares and flame retardants. Melamine is also one of the major components in Pigment Yellow 150, a colorant in inks and plastics. It is unethically added to food products in order to increase the apparent protein content. Standard methods such as the Kjeldahl test estimate protein levels by measuring the nitrogen content only. As such, the measurements of nitrogen content can be misled by adding nitrogen-rich compounds such as melamine.
1.2.
Concerning the analyses for melamine in food matrices, the determinations are commonly carried out using chromatographic methods such as LC-UV, LC-MS/MS or GC with derivatization.
1.3.
In view of the absence of commercially available inter-laboratory programme offered for melamine in food, the Hong Kong Government Laboratory (HKGL) organized an inter-laboratory comparison programme (HKGL 0803) on the quantitative analysis of melamine in milk sample in late September 2008.
1.4.
A total of 14 local testing laboratories enrolled in the programme and 13 of them (Table I) returned their results on or before the deadline. Each participant was confidentially assigned with a unique laboratory code (1 to 13) and the codes were used throughout the programme.
2. Objectives 2.1.
The main objective of this programme is to assess the performance of participants in the analysis of melamine in milk through an inter-laboratory comparison.
2.2.
Test results from participants are evaluated according to the statistical techniques outlined in ISO13528:20059.2 “Statistical methods for use in proficiency testing by inter-laboratory comparisons” and in the international harmonized protocol for the proficiency testing of analytical chemistry laboratories9.3. The z-score is used as a numerical indicator to assess individual participant’s competence relative to the gravimetric nominal values.
2
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
3. Test Materials 3.1.
Two litres of milk were purchased from the local market and confirmed to be free of melamine by the HKGL. The sample was divided into four portions and labelled as Sample 1, 2, 3 and 4.
3.2.
Three standard solutions of melamine (i), (ii) and (iii) were gravimetrically prepared by dissolving appropriate quantities of neat melamine (claimed purity of 99%) in 10% v/v methanol as follows:
3.3.
Standard Solution
Neat Melamine (g)
10% Methanol (g)
Conc. of Melamine (mg/kg)
(i)
0.0037
100.0189
36.99
(ii)
0.0378
100.0033
377.84
(iii)
0.0993
100.0075
991.94
The four separate bulk milk samples as described in clause 3.1. were gravimetrically spiked with the melamine standard solutions (clause 3.2.) to prepare the nominal concentration of melamine at 0, 0.05, 1.0 and 5.0 mg/kg. The test samples were stored at 4 ºC throughout the programme. Sample
Weight of milk (g)
Weight of melamine spiked (g)
Total weight (g)
Nominal conc. of melamine (mg/kg)
1
499.614
0.6767 of std (i)
500.2907
0.050
2
498.528
2.2683 of std (iii)
500.1160
4.497
3
480.223
nil
480.223
0
4
498.080
1.5880 of std (ii)
500.3483
1.200
3.4.
Homogeneity (on 6 October 2008) and stability test (on 27 October 2008) were determined in accordance with the procedures stipulated in APPENDICES I and II. Results indicate the samples are suitable to be used in the programme.
3.5.
Samples together with the “Receipt Form”, “Instructions to Participating Laboratories” and “Result Proforma” were sent to participants who were instructed to collect the samples, kept at 4 ºC, from the HKGL on 8 October 2008. Specimen copies of these documents are shown in APPENDICES III to V.
3.6.
Each participant was provided with four different labelled bottles (Sample 1, Sample 2, Sample 3 and Sample 4 respectively) each containing about 25 ml of the respective milk sample. 3
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
4. Test Required 4.1.
Participants were required to determine, preferably in triplicate, the concentration of melamine in the test sample according to their own validated methods.
4.2.
Participants should report results in mg/kg to three significant figures or three decimal places, whichever was appropriate. Expanded measurement uncertainty and technical information of the method used were also required.
4.3.
All the analytical results and information should be reported in the Result Proforma provided.
5. Statistical Evaluation 5.1.
Participants’ test results 5.1.1. Test results received from participants are presented in TABLES II to V. Their methodologies and techniques employed are tabulated in TABLES VI to VII.
5.2.
Performance evaluation 5.2.1. Performance of each participant was evaluated by means of a performance index, z-score, which is calculated as: z= where
xi m S
= = =
xi - m S
reported mean of individual participant nominal values (as in clause 3.3.) target standard deviation (SD) estimated from Horwitz equation
5.2.2. z-Score is interpreted as follows: (a)
|z| ≤ 2
Satisfactory
(b)
2 < |z| < 3
Questionable
(c)
|z| ≥ 3
Unsatisfactory
5.2.3. Participants having z-score(s) ≥ 3 or ≤ -3 should thoroughly investigate their results and find out the causes of discrepancies. Participants having z-score(s) in the range 2 < |z| < 3 are highly encouraged to review their results. 4
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
6. Results and Discussions 6.1.
With the prediction of occurrence of significant variation of results amongst small number of participants, consensus mean and SD from pooled data are considered inappropriate to be used for z-score assessment in this programme. In accordance with the recommendation in the ISO 13528:20059.2, the assigned values of melamine in the four samples are the nominal values of the gravimetrical preparation as described in clause 3.3.:
Assigned values (mg/kg)
6.2.
Sample 1
Sample 2
Sample 3
Sample 4
0.050
4.497
0
1.200
Participants either used LC-MS/MS (with or without isotopically labelled internal standard) or GC-MS with derivatization as their analytical methods. The median and robust mean of participants’ are tabulated as follows: Sample 1
Sample 2
Sample 3
Sample 4
7
13
*
13
Median (mg/kg)
0.058
4.250
*
1.180
Robust mean (mg/kg)
0.215
4.170
*
1.163
Number of valid data
“*” not evaluated for a qualitative analysis
6.3.
With reference to the assigned values, participants’ results are interpreted as: 6.3.1. Sample 1: Three laboratories (#1, 4 and 7) reported the melamine concentration ranging from 0.043 to 0.058 mg/kg. Four laboratories (#5, 9, 10 and 11) reported the concentration was below their reporting limits at 0.1 mg/kg and one laboratory (#13) reported below a reporting limit at 0.5 mg/kg. However, one laboratory (#3) reported below a reporting limit at 0.01 mg/kg and the other four (#2, 6, 8 and 12) were found to be unable to perform the measurement of melamine at this level. 6.3.2.
Sample 2: Ten laboratories (#1, 2, 3, 4, 5, 7, 9, 10, 11 and 13) showed their capabilities to detect melamine at this relatively high concentration. One laboratory (#12) reported a marginal result and two laboratories (#6 and 8) gave significantly underestimated values. The robust mean was estimated to be 4.170 mg/kg which had a deviation of 7.1% from the assigned value. The spread of data, in terms of relative standard deviation, was relatively high at 25%.
5
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
6.4.
Hong Kong Government Laboratory
6.3.3.
Sample 3: Nine laboratories (#1, 2, 3, 4, 7, 9, 10, 11 and 13) reported below their respective reporting limits. Four laboratories (#5, 6, 8 and 12) reported melamine was present in the concentration range from 0.1 to 0.82 mg/kg. Since the sample was confirmed to be free of melamine, results from lab. #1, 2, 3, 4, 7, 9, 10, 11 and 13 are satisfactory, while results from lab #5, 6, 8 and 12 are considered as false positive or unsatisfactory.
6.3.4.
Sample 4: Twelve laboratories (#1 to 5 and 7 to 13) reported melamine levels ranging from 0.88 to 1.34 mg/kg. Lab. #6 reported the level at 0.23 mg/kg. The robust mean was estimated to be at 1.163 mg/kg which was very close to the assigned value and with a spread of data at 9.6%.
Values of Horwitz’s SD used for z-score assessment:
Sample 1
Sample 2
Sample 3
Sample 4
24.9
12.7
*
15.5
0.0125
0.570
*
0.186
Horwitz’s SD (%) Horwitz’s SD (mg/kg)
“*” not evaluated for a qualitative analysis
6.5.
The z-scores of individual participant are presented in TABLE VIII and illustrated graphically in CHARTS I to III. Performance of participants is summarized as follows: Performance Achieved |z| ≤ 2 2 < |z| < 3 |z| ≥ 3 “*”
6.6.
Laboratory Code Sample 1
Sample 2
Sample 3
Sample 4
1, 4, 7
1, 2, 3, 4, 5, 7, 9, 10, 11, 13
*
1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 13
---
12
*
---
2, 6, 8, 12
6, 8
*
6
not evaluated for a qualitative analysis. Assessment results for Sample 3 are commented in clause 6.3.3.
The z-score results indicate that majority of the laboratories in this programme are able to detect melamine at the legal limit, ie. 1 mg/kg, and at 4.5 mg/kg. However, only three laboratories (#1, 4 and 7) have the capabilities to accurately quantify low level of melamine (0.05 mg/kg).
6
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
6.7.
An MCX cleanup is expected to give a lower limit of quantification (LOQ) for melamine analysis, but the results are not distinct amongst the laboratories in the programme. For example, Lab. #6 did not carry out a MCX cleanup claimed an LOQ of 0.02 mg/kg where Lab #13 claimed an LOQ of 0.5 mg/kg with the MCX cleanup. On the other hand, Lab. #5, 9, 10 and 11 all claimed an LOQ of 0.1 mg/kg but Lab# 5 and 10 (using LC-MS/MS) did not use MCX where #9 and 11 did (using GC-MS with derivatization). It might probably be explained that LC-MS/MS method gives a better sensitivity than that of GC-MS in the analysis of melamine in milk.
6.8.
Recovery is an important parameter in the method validation process. Five laboratories (#1, 5, 7, 8 and 13) did not provide the data. For the other laboratories, the recovery is in the range of 69 to 120% and all reported no correction was made.
6.9.
ISO/IEC 17025 requires laboratories to estimate measurement uncertainty (MU) and report them if necessary. Although MU data are not used in the performance assessment in the programme, participants should estimate MU correctly. Lab. #3 seems to have difficulties to give a reliable MU because the reported value (5.9937 mg/kg) is the same for all the four sample and such value is too large with respect to the concentration of melamine in the samples. Furthermore, three laboratories (#5, 6 and 12) did not attempt to report the MU for their analyses. This indicates that that these laboratories are having hesitation in estimating MU and further training in this area is recommended.
7. Disclaimer 7.1.
The HKGL played the roles as the organizer and participant in this programme. It is solemnly stated that the PT sample preparations and the analysis of the testing materials were undertaken by two independent groups of members. Both groups had no direct contact concerning the details of the PT samples throughout the programme.
8. Acknowledgement 8.1.
Contributions from all participants to this programme are gratefully noted. We are also indebted to Dr. T.L. Ting, Government Chemist of the HKGL, for his support and encouragement to this programme.
8.2.
Participants are welcome to forward any comments on this programme to “
[email protected]” or the following address: Dr. Yiu-chung WONG The Proficiency Test Advisory Board 7
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
Government Laboratory, 7/F. Homantin Government Offices, 88 Chung Hau Street, KLN, Hong Kong.
9. References 9.1.
W. Horwitz. Evaluation of analytical methods used for regulations of food and drugs, Anal. Chem., 1982, 54:67A-76A.
9.2.
ISO 13528 “Statistical methods for use in proficiency testing by interlaboratory comparisons”, 2005, Geneva, Switzerland.
9.3.
M. Thompson, S.L. Ellison, R. Wood. The International harmonized protocol for the proficiency testing of analytical chemistry laboratories (IUPAC technical report), Pure Appl. Chem., 2006, 78:146-196.
8
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
TABLE I: List of Participating Laboratories (in Alphabetical Order)
Name of Laboratories (a)
ALS Technichem (HK) Pty Limited
(b)
Bureau Veritas Hong Kong Limited
(c)
CMA Industrial Development Foundation Limited
(d)
Fugro Technical Services Limited
(e)
Hong Kong Government Laboratory
(f)
Hong Kong Productivity Council
(g)
Hong Kong Standards and Testing Centre
(h)
Intertek Testing Services HK Limited
(i)
Man Kam To Food Laboratory, FEHD
(j)
SGS Hong Kong Limited
(k)
Specialized Technology Resources (HK) Limited
(l)
TUV Rheinland Hong Kong Limited
(m)
Wellab Limited
9
HKGL 0803 Melamine in Milk Inter-laboratory Comparison Programme
Hong Kong Government Laboratory
TABLE II: Participants’ Results for Sample 1 Concentration of Melamine (mg/kg)
Measurement Uncertainty
Lab. Code Data #1
Data #2
Data #3
Mean
Expanded Uncertainty (mg/kg)
Coverage Factor
1
0.038
0.040
0.050
0.043
0.009
2
2
0.0062
0.0061
0.0064
0.0062
0.00065
2
3
