Validation of an ion-pair liquid chromatography-electrospray-tandem mass spectrometry method for the determination of heterocyclic aromatic amines in meat-based infant foods

A method based on ion-pair liquid chromatography-electrospray tandem mass spectrometry (LC-MS/MS) is reported for determining heterocyclic aromatic amines (HAAs) in meat-based infant foods. The HAAs encompassed quinoline (IQ, MeIQ), quinoxaline (MelQx), pyridine (PhIP), and carboline derivatives (A alpha C, Harman, Norharman) with d(3)-IQ, C-13(2)-MeIQx, and d(3)-PhIP used as labelled internal standards. The method used extraction into acetone followed by a clean-up on an SCX solid-phase extraction column. LC separation was performed on a TSKgel ODS-80TS column (250 x 2.0 mm, 5 pm), the mobile phase being an ammonium formate-formic acid buffer (3.03 mM ammonium formate, pH = 2.8) aqueous solution-acetonitrile gradient at a flow rate of 0.2 ml min(-1). For unequivocal identification of each analyte, three ions were detected and chosen for selected reaction monitoring (SRM). Validation was carried out on lyophilized meat samples. Mean recoveries ranged between 78 +/- 4% and 98 +/- 2% for different analytes. Limits of quantification generally lower than 8 ng g(-1) were demonstrated in meat samples for the analytes investigated. The method exhibited a good linearity and repeatability. Robustness testing identified those factors which were statistically significant in influencing chromatographic separation and response, and indicated which parameters have to be strictly controlled for a reliable analysis of HAAs. In particular, the mobile-phase flow rate was found to be statistically significant (alpha = 0.05) for the capacity factor (k') of all analytes except for A alpha C peak, whereas the mobile-phase pH resulted to be a critical parameter for the k' values of IQ, MeIQ, and Norharman. The method was proved to be robust vs. resolution between IQ and MeIQ peaks. Among mass-spectrometric parameters, collision energy was found to significantly affect quantitative response of all analytes except that of IQ. The applicability of the method to the analysis of meat-based infant food samples was demonstrated.