Comparison of extraction and quantification methods of perfluorinated compounds in human plasma, serum, and whole blood.

ABSTRACT

Perfluorinated compounds are ubiquitous in the environment and have been reported to occur in human blood. Accurate risk assessments require accurate measurements of exposures, but identification and quantification of PFCs in biological matrices can be affected by both ion suppression and enhancement in liquid chromatography-tandem mass spectrometry techniques (LC/MS-MS). A study was conducted to quantify potential biases in LC/MS-MS quantification methods. Using isotopically labeled perfluorooctanoic acid ([(13)C(2)]-PFOA), perfluorononanoic acid ([(13)C(2)]-PFNA), and ammonium perfluorooctanesulfonate ([(18)O(2)]-PFOS) spiked tissues, ion-pairing extraction, solid-phase extraction, and protein precipitation sample preparation techniques were compared. Analytical accuracy was assessed using both solvent calibration and matrix-matched calibration for quantification. Data accuracy and precision of 100+/-15% was demonstrated in both human sera and plasma for all three sample preparation techniques when matrix-matched calibration was used in quantification. In contrast, quantification of ion-pairing extraction data using solvent calibration in combination with a surrogate internal standard resulted in significant analytical biases for all target analytes. The accuracy of results, based on solvent calibration was highly variable and dependent on the serum and plasma matrices, the specific target analyte [(13)C(2)]-PFOA, [(13)C(2)]-PFNA, or [(18)O(2)]-PFOS, the target analyte concentration, the LC/MS-MS instrumentation used in data generation, and the specific surrogate internal standard used in quantification. These results suggest that concentrations of PFCs reported for human blood using surrogate internal standards in combination with external solvent calibration can be inaccurate unless biases are accounted for in data quantification.