Integration of rotating disk sorptive extraction and dispersive-solid phase extraction for the determination of estrogens and their metabolites in urine by liquid chromatography/mass spectrometry

Abstract

The matrix of certain complex samples represents a great challenge in microextraction technology due to its ability to interfere with some step of the sample preparation procedure or in the detection itself. Normally, a suppression of the analytical signal occurs by the presence of certain components of the matrix, which tend to be coextracted with the analytes under study. Urine represents one of these complex matrices. The yellow coloration of urine comes from a main pigment known as urobilin, which can persist in all final extracts after sample treatment and has been correlated with the impossibility of correctly measuring the analytes. In the present study, we propose the determination of estrogens and their hydroxylated metabolites in urine by integration of a dispersive solid phase extraction (d-SPE) to eliminate the matrix effect with the rotating disk sorptive extraction (RDSE) of the analytes, prior to their determination by liquid chromatography/mass spectrometry.The proposed dispersive phase was primary-secondary amine (PSA), which, due to its physicochemical characteristics, allows the extraction of polar compounds (such as urobilin). Estrogens and their hydroxylated metabolites, being molecules of lower polarity, were demonstrated not to be removed by PSA and were extracted only by the styrene–divinylbenzene phase supported in the rotating disk. The amount of dispersive phase was optimized, finding that 300 mg is the best analytical response for 2 mL of urine sample. The equilibrium time for the extraction of the analytes was 60 min.Matrix-matched calibration was used for quantification, obtaining correlation coefficients greater than 0.99, limits of detection and quantification between 0.10 and 0.13 and 0.3–0.5 µg/L, respectively, and recoveries between 72 and 130 % with precision, expressed as relative standard deviation, between 8 and 17 %. The matrix effect was reduced significantly between −24 and −48 % (without the use of d-SPE, most of the analytes were not detected). In addition, the proposed method was applied to the determination of the analytes in real urine samples of different persons. Finally, through enzymatic hydrolysis, it was possible to quantify estrogens and their hydroxylated metabolites in urine samples, both in their total form and in their naturally conjugated form.