Determination of glyphosate and its derivative, aminomethylphosphonic acid, in human urine by gas chromatography coupled to tandem mass spectrometry and isotope pattern deconvolution

Abstract

An analytical method for the simultaneous determination of glyphosate (GLY) and its main derivative, aminomethyl-phosphonic acid (AMPA), in human urine has been developed using gas chromatography coupled to tandem mass spectrometry (MS/MS) operated in multiple reaction monitoring mode (MRM). Sample preparation involved dilution of urine with water and derivatization with a mixture of trifluoroacetic acid anhydride and trifluoroethanol. Derivatization conditions such as reaction time and temperature, derivative stability, injection solvent, MS ionization mode and MS-MS transitions, among others, were studied to obtain the highest method sensitivity. The target compounds were initially quantified by the isotope dilution method using isotopically labelled analogs of GLY and AMPA as internal standards. However, due to spectral overlap between GLY and labelled GLY in the selected quantitative transition, a quantification method based on isotope pattern deconvolution (IPD) has been developed. The instrumental limits of detection were 0.05 ng mL−1 for both compounds, while the method detection limits were 0.39 and 0.25 ng mL−1, for AMPA and GLY, respectively. The mean recoveries from urine and water spiked at different concentrations were 77 and 69% for AMPA and 90 and 102% for GLY, respectively, with mean relative standard deviations of 8–10% (urine samples, n = 12) and 3.6–4% (water samples, n = 6). Once validated, the feasibility of the method was tested by determination of AMPA and GLY in human urine samples from people living close to agricultural areas. The developed method affords the determination of these compounds at trace concentrations in complex matrices such as urine, avoiding elaborate handling and cleanup steps. Isotope pattern deconvolution has proven to be a successful alternative to calibration curve for GLY overriding the method uncertainties associated to spectral overlapping.