《美国FDA农残分析手册》（第一卷英文版）APPENDIX II

Pesticide Analytical Manual Vol. I APPENDⅨ APPENDIX II: PROTOCOLS AND REPORTING FORMS FOR TESTING CHEMICALS THROUGH PAMI MULTIRESIDUE METHODS INTRODUCTION: MULTIRESIDUE METHOD TESTING Use of any multiresidue method(MRM)is supported by available information about how potential residues behave through the steps of the method To provide that support for PAM I MRMS, additional chemicals are continually tested through the method steps and the resulting data compiled in a single database. All PAM I tables in Chapters 3 and 4 and Appendix I are produced from that database This Appendix provides directions for performing such tests and forms for report- ing results The effort spent on the testing of MRMs and compilation of results is justified by the advantages such compilations offer the analytical chemist. When analytical behavior data for numerous chemicals through the method in use are known, the analyst is better equipped to identify residues that may be present in a sample of unknown treatment history. In situations where the likelihood of some particular residue is known, the data lists for several methods can be consulted to help choose which method should be used Regulatory agencies often must assess the incidence of residue occurrence. This effort is also assisted by compilations of method behavior data. The absence of many chemicals from the sample can be ascertained when it is known that those chemicals could have been detected had they been present It has been found advisable to define protocols for developing data on MRM behavior. In order to compile data into usable formats, it is imperative that all contributing laboratories perform the tests uniformly. The goal of this method- testing is not to find the optimum conditions for the one chemical currently being tested, but to be able to describe how the chemical will behave when determined by the precisely defined method This Appendix includes one protocol for determining GlC characteristics of chemi- cals and six protocols for testing their behavior through individual MRMs. Forms for reporting the results of testing by each protocol are also included. Each pro- tocol references the PAM I method(s) involved, the types of chemicals to which applies, and the PAM I table(s) in which previously collected data are published Some PAM I MRMs are applicable to a wide variety of residues, while others are targeted to those with specific chemical structures. A Decision Tree is included in this Appendix to direct the user to the most appropriate protocol(s) for each chemical being tested. Follow the Decision Tree in deciding which protocol(s)to Follow the steps of these protocols, in the order written Report data on a copy of the appropriate form and send it to: PAM I Editors, HFS-337, Food and Drug Administration, 200 C Street SW, Washington, DC 20204 ppendⅸ‖1

Pesticide Analytical Manual Vol. I APPENDIX II Transmittal No. 94-1 (1/94) Form FDA 2905a (6/92) Appendix II–1 APPENDIX II: PROTOCOLS AND REPORTING FORMS FOR TESTING CHEMICALS THROUGH PAM I MULTIRESIDUE METHODS INTRODUCTION: MULTIRESIDUE METHOD TESTING Use of any multiresidue method (MRM) is supported by available information about how potential residues behave through the steps of the method. To provide that support for PAM I MRMs, additional chemicals are continually tested through the method steps and the resulting data compiled in a single database. All PAM I tables in Chapters 3 and 4 and Appendix I are produced from that database. This Appendix provides directions for performing such tests and forms for report￾ing results. The effort spent on the testing of MRMs and compilation of results is justified by the advantages such compilations offer the analytical chemist. When analytical behavior data for numerous chemicals through the method in use are known, the analyst is better equipped to identify residues that may be present in a sample of unknown treatment history. In situations where the likelihood of some particular residue is known, the data lists for several methods can be consulted to help choose which method should be used. Regulatory agencies often must assess the incidence of residue occurrence. This effort is also assisted by compilations of method behavior data. The absence of many chemicals from the sample can be ascertained when it is known that those chemicals could have been detected had they been present. It has been found advisable to define protocols for developing data on MRM behavior. In order to compile data into usable formats, it is imperative that all contributing laboratories perform the tests uniformly. The goal of this method￾testing is not to find the optimum conditions for the one chemical currently being tested, but to be able to describe how the chemical will behave when determined by the precisely defined method. This Appendix includes one protocol for determining GLC characteristics of chemi￾cals and six protocols for testing their behavior through individual MRMs. Forms for reporting the results of testing by each protocol are also included. Each pro￾tocol references the PAM I method(s) involved, the types of chemicals to which it applies, and the PAM I table(s) in which previously collected data are published. Some PAM I MRMs are applicable to a wide variety of residues, while others are targeted to those with specific chemical structures. A Decision Tree is included in this Appendix to direct the user to the most appropriate protocol(s) for each chemical being tested. Follow the Decision Tree in deciding which protocol(s) to use. Follow the steps of these protocols, in the order written. Report data on a copy of the appropriate form and send it to: PAM I Editors, HFS-337, Food and Drug Administration, 200 C Street SW, Washington, DC 20204

APPENDIX‖l Pesticide Analytical Manual Vol I Decision Tree for MRM Testing Do data already exist (Index to Methods, Appendix I]? evelop only data nd t Does the compound have this structure or characteristic Yes N-methylcarbamate? Test per Protocol A Yes naturally fluorescent? Test per Protocol A. acid or phenol? Yes Test per Protocol B Yes substituted urea? Test per Protocol G Does it chromatograph on GLC systems? vel I Determine GLC characteristics per Protocol C, Level I Yes Discontinue tests Does it chromatograph on at least one Level system in a reasonable time [o3<rrt<5.0)? No Report rrts for Level I DG modules tested, then retest using appropriate Level ll DG module[s] Is the product a nonfatty (27 fat] food? Yesh Test per Protocol F for fatty food Test per Protocol D. Is the chemical recovered through Section 302 extraction? No Yesh Discontinue tests Test per protocol e for nonfatty food ppendⅸ|2 vo.941(1 2905a(6

Transmittal No. 94-1 (1/94) Appendix II–2 Form FDA 2905a (6/92) APPENDIX II Pesticide Analytical Manual Vol. I Decision Tree for MRM Testing Do data already exist (Index to Methods, Appendix I)? Develop only data not yet available. Yes No Does the compound have this structure or characteristic: Yes Yes Yes Yes Determine GLC characteristics per Protocol C, Level I. Does it chromatograph on GLC systems? No Discontinue tests. No Yes Does it chromatograph on at least one Level I system in a reasonable time (0.3<rrt<5.0)? No Yes Report rrts for Level I DG modules tested, then retest using appropriate Level II DG module(s). Is the product a nonfatty (2% fat) food? No Yes Test per Protocol F for fatty food. Test per Protocol D. Is the chemical recovered through Section 302 extraction? No Yes Discontinue tests. Test per Protocol E for nonfatty food. N-methylcarbamate? naturally fluorescent? acid or phenol? substituted urea? Test per Protocol A. Test per Protocol A. Test per Protocol B. Test per Protocol G

Pesticide Analytical Manual Vol. I APPENDⅨ SUGGESTONS FOR PRODUCING QUALITY DATA The following suggestions were developed in response to data received and to pecific questions that have been raised Decisions on what Protocols to follow As directed in the first step of the Decision Tree, review existing data on the chemical before performing experiments Examine PAM I Index to Methods for entries, then review the details in appropriate Chapter 8 or 4 table(s); review Appendix I for available GLC data. If these sources reveal gaps in the data, perform the tests necessary to provide missing data, but do not repeat experiments unless specifically asked to do so or unless current data reflect variability Develop glC data(Protocol C) with system(s) likely to detect the chemi- cal ie. of the dg modules listed. choose at least one whose detector selective to elements in the molecule. The electron capture detector may not be suitable to examine uncleaned extracts from Section 302 Protocol D), so DGl, 19, and 18 are often insufficient. If an electron capture detector is the only one that responds to the chemical, apply caution when injecting extracts from Section 302 Choose the GlC system that provides the best chromatography and sensitivity for examining solutions from recovery studies; it is not nec. essary to re-examine the extracts by multiple glc systems The Decision Tree provides basic criteria for making decisions about which methods to test Where situations exist that make testing or con- tinuation of testing illogical, suspend testing and report the reasons Examples If previous studies with radiolabelled chemical clearly show that the residue will partition into the(discarded) water layer of the method, nethod recovery tests need not be run; however, collection of GLC data should still be attempted If the only commodity of interest is fatty, do not attempt to perform tests on the product with methods designed for nonfatty foods. If the lodity is ly Section 304 El(Protocol F)or, if the hemical is an acid or phenol, Section 402 El(Protocol B) Suspend testing if the GlC tests(Protocol C) indicate that even the most sensitive GLC system is insensitive to the compound As a gen- eral rule, suspend testing if the minimum weight of compound that causes 10% full scale deflection(FSD) is equivalent to 210 times the tolerance for jection of the normal mg sample equivalent de- scribed in the method If the method being tested proves not amenable to analysis of a particular commodity, as evidenced by severe emulsions, failure to form distinct phases, etc, suspend testing ppendix l-3

Pesticide Analytical Manual Vol. I APPENDIX II Transmittal No. 94-1 (1/94) Form FDA 2905a (6/92) Appendix II–3 SUGGESTIONS FOR PRODUCING QUALITY DATA The following suggestions were developed in response to data received and to specific questions that have been raised. Decisions on What Protocols to Follow • As directed in the first step of the Decision Tree, review existing data on the chemical before performing experiments. Examine PAM I Index to Methods for entries, then review the details in appropriate Chapter 3 or 4 table(s); review Appendix I for available GLC data. If these sources reveal gaps in the data, perform the tests necessary to provide missing data, but do not repeat experiments unless specifically asked to do so or unless current data reflect variability. • Develop GLC data (Protocol C) with system(s) likely to detect the chemi￾cal; i.e., of the DG modules listed, choose at least one whose detector is selective to elements in the molecule. The electron capture detector may not be suitable to examine uncleaned extracts from Section 302 (Protocol D), so DG1, 13, and 18 are often insufficient. If an electron capture detector is the only one that responds to the chemical, apply caution when injecting extracts from Section 302. • Choose the GLC system that provides the best chromatography and sensitivity for examining solutions from recovery studies; it is not nec￾essary to re-examine the extracts by multiple GLC systems. • The Decision Tree provides basic criteria for making decisions about which methods to test. Where situations exist that make testing or con￾tinuation of testing illogical, suspend testing and report the reasons. Examples: – If previous studies with radiolabelled chemical clearly show that the residue will partition into the (discarded) water layer of the method, method recovery tests need not be run; however, collection of GLC data should still be attempted. – If the only commodity of interest is fatty, do not attempt to perform tests on the product with methods designed for nonfatty foods. If the commodity is meat, use only Section 304 E1 (Protocol F) or, if the chemical is an acid or phenol, Section 402 E1 (Protocol B). – Suspend testing if the GLC tests (Protocol C) indicate that even the most sensitive GLC system is insensitive to the compound. As a gen￾eral rule, suspend testing if the minimum weight of compound that causes 10% full scale deflection (FSD) is equivalent to ≥10 times the tolerance for an injection of the normal mg sample equivalent de￾scribed in the method. – If the method being tested proves not amenable to analysis of a particular commodity, as evidenced by severe emulsions, failure to form distinct phases, etc., suspend testing

APPENDIX‖l Pesticide Analytical Manual Vol I Proper Application of Methods Do not combine a GlC column that contains cyano groups with a nitro- gen-selective detector Adjust GlC column temperature to establish the correct relative retention time (rrt)of the marker chemical(s). Absolute retention times(min)are provided in DG modules to provide an indication of normal behavior, but the temperature should not be adjusted to achieve this Measure retention times on glc columns from the solvent front wher ever possible for both the compound being tested and the marker com- pound. If the instrumentation in use precludes measurement from the solvent front, state this fact in the report. When submitting chromatograms with reporting forms, label them clearly indicate how many nanograms(ng) or picograms(pg)are represented by the peak. Do not label a chromatogram of a standard in ppm For chro- matograms that result from recovery studies, indicate on the chromato- gram label how much sample weight equivalent(mg) was injected For accurate quantitation, detector response to the reference standard should be within +25% of the response to the analyte in the sample, based on observable(on-scale) peak heights. Usual GLC linearity does not sup- port quantitation that compares responses differing in size by more than that amount Injection to elute before the next injection is made. peaks from one Make sure that chromatograms do not overlap; allow when the chromatogram contains multiple peaks representing different components in the standard, report the rrt of each peak >5%0 FSD It is not necessary to evaporate the solvent from the fortification solution once it has been added to the sample, unless there is some reason to expect it to interfere with the analysis. It is preferable for the fortifying solution to be made from the same solvent used to extract the sample Use only the 60/100 mesh PR grade Florisil specified by Sections 303 and 304(Protocols E and F). Other grades of commercially available Florisil are likely to result in different elution patterns Florisil column in a polar solvent, which will affect elution patern. ch During tests of elution from Florisil, do not add reference material to Test Florisil elution by both ethyl ether/petroleum ether (Sections 303 Cl, 304 Cl and C3)and methylene chloride($0% C2, 304 C2 and C4) elution systems when following the steps of Protocols E and F. Even when florisil elution tests with reference standards indicate no elu tion with later eluants amine by GLC all the Florisil eluates of the recovery test from the fortified sample. Sometimes the presence of extract changes the Florisil elution pattern ppendⅸ|4 vo.941(1 2905a(6

Transmittal No. 94-1 (1/94) Appendix II–4 Form FDA 2905a (6/92) APPENDIX II Pesticide Analytical Manual Vol. I Proper Application of Methods • Do not combine a GLC column that contains cyano groups with a nitro￾gen-selective detector. • Adjust GLC column temperature to establish the correct relative retention time (rrt) of the marker chemical(s). Absolute retention times (min) are provided in DG modules to provide an indication of normal behavior, but the temperature should not be adjusted to achieve this. • Measure retention times on GLC columns from the solvent front wher￾ever possible for both the compound being tested and the marker com￾pound. If the instrumentation in use precludes measurement from the solvent front, state this fact in the report. • When submitting chromatograms with reporting forms, label them clearly: indicate how many nanograms (ng) or picograms (pg) are represented by the peak. Do not label a chromatogram of a standard in ppm. For chro￾matograms that result from recovery studies, indicate on the chromato￾gram label how much sample weight equivalent (mg) was injected. • For accurate quantitation, detector response to the reference standard should be within ±25% of the response to the analyte in the sample, based on observable (on-scale) peak heights. Usual GLC linearity does not sup￾port quantitation that compares responses differing in size by more than that amount. • Make sure that chromatograms do not overlap; allow peaks from one injection to elute before the next injection is made. • When the chromatogram contains multiple peaks representing different components in the standard, report the rrt of each peak >5% FSD. • It is not necessary to evaporate the solvent from the fortification solution once it has been added to the sample, unless there is some reason to expect it to interfere with the analysis. It is preferable for the fortifying solution to be made from the same solvent used to extract the sample. • Use only the 60/100 mesh PR grade Florisil specified by Sections 303 and 304 (Protocols E and F). Other grades of commercially available Florisil are likely to result in different elution patterns. • During tests of elution from Florisil, do not add reference material to the Florisil column in a polar solvent, which will affect elution pattern. • Test Florisil elution by both ethyl ether/petroleum ether (Sections 303 C1, 304 C1 and C3) and methylene chloride (303 C2, 304 C2 and C4) elution systems when following the steps of Protocols E and F. • Even when Florisil elution tests with reference standards indicate no elu￾tion with later eluants, examine by GLC all the Florisil eluates of the recovery test from the fortified sample. Sometimes the presence of extract changes the Florisil elution pattern

Pesticide Analytical Manual Vol. I APPENDIX I/PROTOCOL A PROTOCOL A: PROCEDURE FOR TESTING CHEMICALS THROUGH SECTION 401 BACKGROUND Methods: Section 401 El+cl+ dll or dl2 Chemical Type: Applicable to chemicals with N-methylcarbamate structure (DLl) and to some chemicals that are naturally fluorescent(DL2) soybeans and pe: Applicable to nonfatty foods and to certain fatty foods such as PAM I Tables: Tables 401-a. 401-b DATA DEVELOPMENT HPLC Analytical Behavior N-Methylcarbamates: Set up HPLC with post-column fluorescence labeling and fluorescence detector, as described in Section 401 DLl; check for proper opera- tion using system suitability test described Fluorescent Pesticides: Set up HPLC and fluorescence detector, as described in Section 401 Dl2 Develop information on test chemical(s)as follows Dilute with methanol for HPLC working standards e stock solutions Determine amount (ng) of chemical that causes detector response of 50% full scale deflection(FSD) on recorder or printer/plotter. Note peak shape of response to determine adequacy of chromatography Determine linear response range of detector to chemical Determine stability of chemical in methanol Short term. Prepare l ug/mL methanol solution of chemical. Use actinic glassware. Inject 5-10 uL injections of this solution into HPLC ystem over 8 hr to measure short term stability of chemical in solu- tion Report results as peak height(mm)response Long term. Using same solution as above, inject 10 uL into system once a day for 1-2 wk to measure long term stability of solution. Store solution on laboratory bench during day and in refrigerator over night. Report results as peak height(mm)response to test chemical, normalized to peak height for carbofuran standard injected on same APPENDIX‖-5

Pesticide Analytical Manual Vol. I APPENDIX II/PROTOCOL A APPENDIX II–5 Transmittal No. 94-1 (1/94) Form FDA 2905a (6/92) PROTOCOL A: PROCEDURE FOR TESTING CHEMICALS THROUGH SECTION 401 BACKGROUND Methods: Section 401 E1 + C1 + DL1 or DL2 Chemical Type: Applicable to chemicals with N-methylcarbamate structure (DL1) and to some chemicals that are naturally fluorescent (DL2). Commodity Type: Applicable to nonfatty foods and to certain fatty foods such as soybeans and nuts. PAM I Tables: Tables 401-a, 401-b DATA DEVELOPMENT HPLC Analytical Behavior N-Methylcarbamates: Set up HPLC with post-column fluorescence labeling and fluorescence detector, as described in Section 401 DL1; check for proper opera￾tion using system suitability test described. Fluorescent Pesticides: Set up HPLC and fluorescence detector, as described in Section 401 DL2. Develop information on test chemical(s) as follows: • Dissolve reference standard in methanol to prepare stock solutions. Dilute with methanol for HPLC working standards. • Determine amount (ng) of chemical that causes detector response of 50% full scale deflection (FSD) on recorder or printer/plotter. Note peak shape of response to determine adequacy of chromatography. • Determine linear response range of detector to chemical. • Determine stability of chemical in methanol: – Short term. Prepare 1 µg/mL methanol solution of chemical. Use actinic glassware. Inject 5-10 µL injections of this solution into HPLC system over 8 hr to measure short term stability of chemical in solu￾tion. Report results as peak height (mm) response. – Long term. Using same solution as above, inject 10 µL into system once a day for 1-2 wk to measure long term stability of solution. Store solution on laboratory bench during day and in refrigerator over￾night. Report results as peak height (mm) response to test chemical, normalized to peak height for carbofuran standard injected on same day