Detection of banned antibacterial growth promoters in animal feed by liquid chromatography-tandem mass spectrometry: optimisation of the extraction solvent by experimental design

Because of the risk that residues of antibacterial growth promoters in edible tissues could lead to allergic reactions and development of resistant strains of bacteria in humans, the European Commission decided to ban bacitracin, olaquindox, spiramycin, tylosin and virginiamycin as feed additives. To control the compliance of this ban, a single multi-analyte liquid chromatography-tandem mass spectrometry method was developed in 2003. Starting from this method, we further tried to optimise the response with the aid of experimental design. Using a central composite design, we searched for the optimal extraction solvent for each of the five components separately. As expected, this optimal composition was different for each of these antibacterial growth promoters. Two groups of compounds could be distinguished. The first group contained those compounds (spiramycin, tylosin and virginiamycin) that had the highest response (peak area) when the extraction solvent consisted of pure methanol (with or without the addition of formic acid). The second group contained the components (bacitracin and olaquindox) that preferred between 50 and 70% (v/v) methanol as extraction solvent. To find a compromise between these two groups, we created weighed utility functions and analysed these data with linear regression (α < 0.05). The extraction solvent that gave the overall best response for the type of feed used during this experiment was very similar to the extraction solvent used during the validation of the multi-analyte method (70% (v/v) methanol and 2% (v/v) formic acid) [C. Van Poucke, K. De Keyser, A. Baltusnikiene, J.D.G. McEvoy, C. Van Peteghem, Anal. Chim. Acta 483 (2003) 99]. Finally, a recovery experiment was set up using different types of feed to identify if the outcome of this experiment was valid for all types of feedstuffs.