Document 0v79mRgx8kkjrdgLVrRLxXZR

Final Draft Guidance Document Telomer Research Program (TRP) AR226-1888 Guidance Document Best Analytical Practices for TRP Studies Author Marilyn Stadalius Critical Path Services 300 Foulk Road, Suite 1C Wilmington, Delaware 19803 Completion Date March 2004 RECEIVED OPPT NCIC 2004 DEC 22 AM 11:20 Sponsor Telomer Research Program Asahi Glass Co., Ltd. Clariant GmbH Daikin Industries, Ltd. DuPont Chemical Solutions Enterprise Page 1 of 8 Final Draft Guidance Document Telomer Research Program (TRP) B est A nalytical P ractices for TRP Studies Including Guidance on Method Validation General Guidance Special care m u st be used in sam pling and analysis to avoid problem s with accuracy and precision of analyses, especially w hen determ ining non-volatile fluorinated com pounds at ppb levels. For more accurate and reproducible analytical determ inations the use of a d u al-lab el, a t a m in im u m , 13C -in tern al s ta n d a rd is highly recom m ended to alert the analyst of interferences from co-eluting peaks. To verify th at unacceptable levels of carry-over contam ination do not occur from run-to-run and th at contam ination is not introduced during sam ple preparation, m ultiple sam ple solvent blanks and at least one reagent blank is recom m ended for each sam ple set. Individuals should take the following precautions w hen handling sam ples: 1) Only previously w ash e d clothing sh o u ld be w orn w hen prep arin g , shipping or analyzing these sam ples. 2) Always w a sh h a n d s th o ro u g h ly w ith w arm w a ter a n d soap p rior to working in sam ple preparation and designated lab areas. 3) Use of nitrile gloves w hen p rep arin g , sh ip p in g a n d analyzing sam p les is highly recom m ended. 4) Avoid ex p o su re to p re-p ack ag ed foods w hen in c o n ta c t w ith d esig n ated working areas or when working with sam ples. A lum inum foil sh o u ld be avoided a t th is tim e. New boxes of zip-lock bags have been found to be free of contam ination. In addition, Post Its should not be used in sam pling area or to label sam ples, and blue ice should not be used with sam ples. Analytical Guidance I. S am pling Before sam pling, define work area for sam pling and clean area with a high purity organic solvent; m ethanol, acetonitrile or isopropanol are suitable. W ork area can be swipe tested before a n d /o r after cleaning to test for contam ination. Before using, evaluate all labw are, including containers and their covers, m easu rem en ts devices or im plem ents th a t will come in contact w ith sam ples for PFOA a n d /o r analytes of interest. See separate discussion on containers. Page 2 of 8 Final Draft Guidance Document Telomer Research Program (TRP) W hen sam pling, focus full attention on sam pling until work is complete. Avoid moving into o ther a re as of the w orkplace during sam pling. Field blanks m ay be required and evaluated for the study. II. Solvents Check all solvent lots used during sam ple preparation and analysis for background to avoid false positives. This includes w ater w hen used as a reagent for low level, i.e. ppt, testing a s well a s a mobile p h ase modifier in HPLC m ethods. C ertain lots of solvents have been found to be unacceptable for use. Solvents should have background significantly lower th a n the LOQ for a stu d y e.g. < 0.1 LOQ. III. C o n tain ers To avoid contam ination from sam pling equipm ent and containers, fluoropolymers should be avoided. Carefully cleaned polypropylene and polystyrene sam ple containers and centrifuge tubes have been satisfactorily used for the analysis of fluorinated com pounds. However, all containers th a t come in contact with treated and untreated sam ples m u st be confirm ed to be free of PFOA. C ontainers th a t are not free from PFOA should be triple w ashed in high purity organic solvent, i.e., m ethanol, acetonitrile or isopropanol, and reconfirm ed to be free of PFOA before use. The use of disposable labw are is recom m ended, i.e., pipets. IV. R eagent B lan k s Run reagent blanks early in the evaluation process of m ethod developm ent as well with each sam ple set analyzed. T hat is, conduct com plete sam ple w ork-up and analysis on sam ple solvent, following current m ethod protocol. Reagent blanks should have background significantly lower th a n the LOQ for a stu d y e.g. <0.1 LOQ. V. Sam ple P re p ara tio n for A nalysis Sam ple preparation area(s) should be w ashed extensively with high purity organic solvents, e.g. m ethanol, isopropanol or acetonitrile, before and after sam ple preparation. C onduct all sam ple preparation work on disposable towels or pads. Weigh and prepare sam ples in a separate location from sta n d ard s - more im portant for ppb determ inations. Optimally, the weigh station and balance should be swipe tested periodically and the swipes extracts analyzed for the test substance to ensure th at contam ination does not exist. Page 3 of 8 Final Draft Guidance Document Telomer Research Program (TRP) W ear nitrile gloves w hen handling sam ples and change gloves often. The use of disposable labw are is recom m ended, i.e., pipets. E nsure th at containers used to store sam ples and solvents used in sam ple preparation are free of contam ination -see discussion on containers, solvents and reagent blanks. VI. S ta n d a rd s a n d S ta n d a rd P rep aratio n Each standard needs a certificate of analysis with its purity and analytical m ethods used during its determ ination. Preferably, the purity of the stan d ard should be evaluated using the sam e m ethodology used to analyze sam ples. At least five s ta n d a rd s sh o u ld be u se d to generate th e calibration curve, bracketing expected levels in sam ples. The lowest stan d ard should be at the LOQ for the m ethod. S tandards m ust be analyzed with each set of sam ples and are generally injected a t sta rt of a sam ple set - from low to high concentration - and standard analyses are generally repeated throughout a n d /o r at the end of the sam ple set. Analysis of instrum ent and reagents blanks should precede stan d ard injections, see exam ple sam ple sequence on page 8. If background levels in u n trea te d sam ples are unacceptable, > 20% LOQ, consider raising the LOQ or s ta n d a rd s m ay need to be m ade u p in m atrix and not neat solvent. S tandards are a comm on source of contam ination for sam ples analyzed at a ppb levels. Some practices to consider w hen weighing stan d ard s and sam ples at the sam e balance: -cover weigh station area with disposable cloths before weighing sam ples or standards -bring stan d ard s and sam ples to weighing station in separate tray -clean weigh station and balance with m ethanol after weighing -wipe test balance for contam ination routinely Optimally, stan d ard s should be weighed at a different weighing station/balance than sam ples. Solvent blanks should be ru n throughout a sam ple sequence to ensure th a t carry-over is not a concern, e.g. inject solvent blank after high standard injection. See suggested sam ple sequence included in section on M ethod Validation. Page 4 of 8 Final Draft Guidance Document Telomer Research Program (TRP) VII. S o u rces of Mobile P h ase a n d In s tru m e n t C o n tam in atio n If th e LC m obile p h a se show s m ea su ra b le q u a n titie s of PFOA th a t c a n n o t be reduced with the use of another solvent lot or change in solvent m anufacturer, consider protecting the analytical colum n with a Hypercarb pre-colum n placed after the HPLC pum p but before the injector to collect PFOA. At low levels of detection, PTFE or PTFE-lined containers or equipm ent m u st not be used, including PTFE-lined vials for the HPLC autosam pler. VIII. R eagent B lan k s A reagent b lan k should be preferably considerably less th a n the LOQ, i.e. < 0.1 LOQ, and should be included with each chrom atographic run. IX. F ilters Filtration of the sam ple m ay be necessary to remove undissolved solids. This should be done only if necessary, since analyte m ay be lost due to absorption onto the filter and should be experim entally verified. Polypropylene filter m edia are preferred, since absorption is generally less th an for other m aterials. Volatile Fluorinated Compounds Some fluorinated com pounds having low w ater solubility an d relatively high vapor pressure, can escape aqueous solutions in the course of hours. For example, after forty eight hours, less th an 10% 8-2 telom er B alcohol w as recovered from an aqueous solution having an initial concentration of 130 n g /m L ; w ith a 50% loss d eterm in ed a t a b o u t 13 h o u rs. As a lm o st no sorptio n occurred to glass wall, loss to head space a n d /o r parafilm a n d /o r diffusion through parafilm w as suspected. - A eration and flow -through test system s will accelerate loss of these analytes. -W hen dissolved in w ater, these analytes will m ost likely be lost to headspace in closed test and assay vessels. -Also, if an assay for these analytes from aq ueous system s is required, back-extract sam ple into solvent like MTBE at tim e of sam pling or as soon as possible to better reflect the concentration at th at sam pling time. -M ix/shake/vortex sam ples thoroughly imm ediately before analysis. -If the aqueous system contains soil or soil-like particles, recoveries can be higher th an expected, as these analytes can preferentially sorb to these particles and be subsequently extracted into organic solvent for quantification. Page 5 of 8 Final Draft Guidance Document Telomer Research Program (TRP) G uidelines for M ethod Validation for A nalysis of A nalytes in T est System s These guidelines assu m es th at m ethods to assay levels of an analyte is a test system can be less stringent th an validation criteria used to assay the active ingredient in a drug form ulation or to set a U.S. EPA tolerance for a pesticide in a crop. However, in all cases, in order for a m ethod to be valid, the following m ust be dem onstrated: Linearity Accuracy Precision (reproducibility) Recovery LOQ a n d LOD For a m ethod to successfully m eet the above criteria, the chrom atography for the analyte should be well-behaved with: -good peak shape with m inim um band broadening -adequate retention with good retention reproducibility -signal to noise ratio of 5:1 a t LOQ -m inim al interference from other su b sta n c es (specificity) -controls, sludge w ithout sample, having area counts of no more than 20% of the LOQ. A nalysis of reagent b lan k s an d solvent b lan k s should be considerably lower th an the LOQ, i.e., < 0.1. The GLP m ethod validation protocol should include the following: I. L inearity a n d A ccuracy -A c alib ratio n curve sh o u ld be g e n erate d for each a n aly te in th e sam ple. U sing five or m ore sta n d a rd s p rep ared in th e injection solvent, b rack et anticipated concentration range of analyte concentration in sam ple to create calibration curve. L ow est sta n d a rd concentration should correspond to the LOQ. Calibration curve can include a blank solvent sam ple with or w ithout internal standard. Sam ples outside the calibration limits should be re-run by diluting a high sam ple, concentrating a low sam ple, or expanding the calibration range. -Use standard curve fitting, by applying the sim plest model th at adequately describes the concentration-response relationship, e.g. linear regression with r2 > 0.992. -Accuracy: Standard concentrations defined by curve should be within 20% a t the LOQ an d w ithin 15% for all o th er p o ints w hen com pared w ith actual standard concentration. Duplicate standard injections are preferred see exam ple of sam ple sequence below. Page 6 of 8 Final Draft Guidance Document Telomer Research Program (TRP) II. Recovery a n d Precision Recovery pertains to the extraction efficiency of an analytical m ethod w ithin the lim its of variability. Recovery experim ents should be perform ed by com paring the analytical resu lts/an aly sis for extracted sam ples at three concentrations (LOQ, m ed iu m a n d high) w ith th e a m o u n t applied. At a m inim um , triplicate recovery analyses should be conducted at each concentration. For exam ple, for an an aly sis w ith an LOQ of 5 ppb, set the dosing of control m atrix at 5 ppb, 50 ppb and 500 ppb, each analyzed in triplicate. A. 3 control sam p les (sludge m a trix w /o sam ple) do sed a t LOQ or 5 p pb B. 3 control sam p les dosed a t 50 ppb C. 3 control sa m p le s dosed a t 500 p pb Actual recoveries should be w ithin 70-120% with RSD (precision / reproducibility) of less th an 20%. III. LOD Use e x p an d ed scale of "solvent b la n k " to determ in e th ree tim es b aselin e noise an d convert to co n centration u n its u sin g calibration curve. LOD can also be calculated; include reference. IV. E xam ple Sam ple S equence To m aintain quality control, the following sequence of sam ples is proposed for each analytic run: Exam ple Sam ple A nalysis Sequence Instrum ent blank Reagent blank Series of standards Solvent blank T hree "A" sam p les Solvent blank* T hree "B" sam p les Solvent blank* T hree "C" sam p les Solvent blank* -Solvent injection to ensure instrum ent is clear of interfering substances -Blank generated from conducting sam ple workup w ithout sample -Injected from lowest to highest concentration Check stan d ard s could be injected throughout the ru n or freshly spiked QC sam ples a t LOQ an d 10x LOQ. If u sed, in tern al sta n d a rd should be added after extraction to account for LC/M S/M S m atrix effects. *With experience, n u m b e r of solvent b la n k s m ay be reduced. Page 7 of 8 Final Draft Guidance Document Telomer Research Program (TRP) Report including validation d ata should be prepared and raw d ata retained in the study records, according to GLP guidelines. V. Sam ple A nalysis According to GLP guidelines, a protocol for the analysis of sam ples should be w ritten and approved before sam ple analysis. Recovery sam ples should be analyzed with each analysis set at LOQ, m edium and high levels to validate d a ta w ith each ru n . Also calibrations sta n d ard s, instrum ent blanks, reagent blanks and solvent blanks sam ples should be analyzed as detailed above with each analysis set. VI. Frozen Sam ples: S hort-T erm Stability If sam ples are to be frozen before analysis, three aliquots of each of the low and high concentrations should be thaw ed at room tem perature and kept at this tem perature from 4 to 24 hours (based upon expected duration th at sam ples will be m aintained a t room tem p eratu re in the intended study) and analyzed. VII. Frozen Sam ples: Long-Term Stability Storage time in a long-term stability evaluation should exceed the time between the d ata of first sam ple collection and the date of last sam ple analysis. Long term stability should be determ ined by storing at least three aliquots of each of the low an d high co n cen tratio n s u n d e r the sam e conditions a s the study sam ples. The volume of sam ple should be sufficient for analysis on three separate occasions. Page 8 of 8