Document v1J0drk11dn2Kpjz3vKaVXQLm

fi 5 3i OOOOG M U G - 3H-0Z TRIP REPORT: National Medical Services Laboratory (NMS), Willow Grove, Pennsylvania Tuesday, January 29, 2002 Mary A. Kaiser, Ph.D. DuPont Corporate Center for Analytical Sciences Central Research & Development E302/117B Meeting Participants from NMS: Laura Labay (supervisor), Harry Garcia-Flores (analyst), Bob Klinger (QA manager), and Paul Delany (former analyst) Summary: The probable reason for some of the differences in the levels of perfluorooctanoic acid (PFOA) in blood from 2000 to 2001 is the use of a new and im proved analytical method for the 2001 samples. Dr. Daryl Cobranchi, DuPont CR&D CCAS, and I visited NMS to determine why some blood PFOA levels from 2000 and 2001 did not agree, although the same laboratory ran them. The samples were taken from individuals at the DuPont Washington Works facility in West Virginia. The date below show that the results from 2000 for some o f the same individuals are generally significantly lower than the results from 2000, Job Description 2001 2000 Granular Polykettle Operator 3.7 776395 3.7 776394 0.76 776413 2.9 776419 1.4 776388 0,55 776412 0.91 776415 8.5 495972 9.0 495964 1.7 495935 6.2 495955 0.80 495917 Fine Powder/Dispersion Autoclave Operator 2.9 776408 1,2 776406 1.9 776405 2.6 776397 1.9 776398 0.79 776399 1.9 776407 1.1 776411 0.49 776409 4.7 495925 4.3 495946 4.9 495916 2.4 495947 Other DuPont Employees 1.1 776403 0.22 776396 0.39 776410 0.14 776401 0.33 776404 0.07 776414 0.45 m m 0.76 776413 0.60 776423 0.12 776416 0.23 776417 1.5 776421 1.3 776420 2.8 495941 0.33 495906 0.22 495950 000064 1.6 776422 ONYX Employees (* = PSA) 2.2 776392 1.3 776393 0.48 776386 2.5 776391 0.66 776387 0.26 776389 0.71 776385 1.2* 776402 The most probable reason for the difference in the reported levels in the two sample sets is that NMS introduced a new method of sample preparation for our 2001 samples. The two methods (old and new) are described below. OldMethod Perfluorooctanoic Acid (Freeze-drying) Method#3425 Effective date: 12/92 New Method Perfluorooctanoic Acid (w/out Freeze drying) Method#3425 Effective date: November 1,2001 The method was changed to reduce the interferences from the acetyl chloride peak impurity from the tat used in the old analysis. The interfering peak came immediately before the PFOA (methyl ester) peak, sometimes making the peak integration so difficult that the analyst had to manually integrate the PFOA peak. Because o f the interference, the laboratory often could not reproducibly detect PFOA at the 10 ppb (10 ng/mL) reporting lim it In foe new method the analyst makes up methylation reagent from highperformance liquid chromatographic grade methanol and concentrated (37%) hydrochloric acid. The data show that there is no interfering peak. The new method does a betterjob homogenizing the blood so that one can more easily extract a representative sample for gas chromatographic analysis with an electroncapture detector. The new method also uses less extraction solvent, improving the dilution factor fourfold. A comparison o f the two methods is given below: New Method Old Method whole blood GCCD PFDA* internal standard 5890 GC HPLC-grade methanol/37% HC1 0.5 mL hexane no freeze drying noNaOH no sonication no water added 65-75% completeness o f methylation (estimate) whole blood GC/ECD PFDA internal standard 6890 GC Alltech instant mefoanolic kit 2.0 mL hexane freeze drying NaOH sonicate for 30 minutes I mL of water added 100% completeness (estimate) *perflurodecanoic acid (PFDA) Overall, the new method seems to be more precise with better recoveries of spikes, especially at the lower levels. 000065 A characteristic o f the new method is lower % methylation reaction completion, although tins is somewhat compensated for with the PFDA internal standard. Both the old and the new method use PFDA, which has measurable amounts o f PFOA as an impurity. Retention o f the PFOA methyl ester peak could be improved in both methods, perhaps by using a pefluorinated stationary phase. Both methods use the electron capture detector, which is not specific for PFQA (other compounds could co-elute causing an enhanced peak and higher reporting levels from what is actually present in the sample). This is especially important since the retention time is so short. Analytes with short retention times have increased probability o f coelution with another material. Conclusion: The NMS new method is an improvement over the old method. The 2001 results are more reliable than the 2000 results for our two sample sets. The determination could be improved by increasing both the sensitivity (10 ng/mL reporting level) and specificity. ) /? 000066