Validation of a liquid chromatography–high-resolution mass spectrometry method for the analysis of ceftiofur in poultry muscle, kidneys and plasma: A unique accuracy profile for each and every matrix

• A LC–HRMS method was developed for the analysis of ceftiofur in poultry.
• LC coupled to full-scan high-resolution Orbitrap MS analysis at 60,000 FWHM.
• The method was fully validated using total error approach and accuracy profiles.
• Reliable dosage of ceftiofur in plasma, muscle and kidney samples with only one regression model.
• The method was successfully applied during a ceftiofur depletion study.

A robust, selective and specific liquid chromatography–high resolution mass spectrometry (LC–HRMS) method was developed for the quantification of total residues of ceftiofur, an antibiotic belonging to the 3rd generation cephalosporins in plasma, muscle and kidney of poultry. Ceftiofur and conjugates in samples were firstly hydrolyzed with dithioerythritol into desfuroylceftiofur, which was then stabilized by derivatization with iodoacetamide into desfuroylceftiofur acetamide. Sample were then submitted to a solid phase extraction followed the accurate mass analysis of desfuroylceftiofur acetamide by LC–HRMS in full scan mode using a linear trap quadrupole (LTQ)-Orbitrap mass spectrometer with a resolving power 60,000 full width at half maximum (FWHM). The method was fully validated over a dosing range between 100 and 2000 μg kg−1 (or μg L−1) using the total error approach. Accuracy profiles a graphical decision-making tool were built by computing results of validation procedure with acceptance limits set at ±60%, and β-expectation tolerance intervals, i.e. the interval assuming to contain a β % of future measurements (β = 90% in this study). Total measurement error including trueness, repeatability and intermediate precision were evaluated. Relative bias of trueness was never exceeding the threshold of 6% in all matrices at all level of concentration. The mean relative standard deviation for repeatability was lower than 16% at all levels of concentration for all matrices; the mean relative standard deviation for intermediate precision was lower than 25% at all levels of concentration for all matrices. This validation approach proved that the method is reliable for the quantification in each and every matrix (i.e. plasma, kidneys and muscle of chicken) thanks to only one single regression model (i.e. linear) obtained from external calibration standards (without matrix) with deuterated labelled internal standard. The developed method was applied during a depletion study of ceftiofur in chicken tissues and plasma.