Seventeen quinolone antibiotics were determined in cows' milk. A method of high sensitivity, selectivity and accuracy was developed. Accuracy (trueness and precision), linearity, sensitivity, selectivity, decision limit and detection capability were established following the recommendations of the Commission Decision 2002/657/EC and the Food and Drug Administration (FDA) guideline. The use of polar stir-bar sorptive extraction (SBSE) prior to UHPLC-MS/MS analysis is proposed. The variables that affect SBSE were optimised using multivariate optimisation strategies. The ionic strength, the extraction time and the sample volume were studied. pH and stir-bar coating (polydimethylsiloxane, PDMS, and polyethyleneglycol modified silicone, PEG) were studied. PEG showed the best extraction yield at pH 6. For validation, a matrix-matched calibration and a recovery assay were carried out. Limits of quantification from 0.5 μg kg-1 for nalidixic acid, flumequine and piromidic acid, to 4.0 μg kg-1 for sarafloxacin were calculated. The precision (%, RSD) was lower than 15% for all antibiotics. Recoveries in fortified samples were between 88 and 114%.
Keywords: Food analysis; UHPLC–MS/MS; polar stir-bar sorptive extraction (SBSE); quinolone antibiotics; raw cows’ milk analysis.