Extracting fumonisins from maize: efficiency of different extraction solvents in multi-mycotoxin analytics

The complete extraction of analytes is of utmost importance when analyzing matrix samples for mycotoxins. Mycotoxins consist of substances with widely different physicochemical properties; therefore, the loss of toxins that occurs in multi-mycotoxin methods due to compromises in the extraction solvent is currently a topic under discussion. With regard to fumonisins, several extractants from recently published multi-mycotoxin methods were investigated when analyzing unprocessed and processed maize matrices. All extractants were tested in a validated on-site method and the extraction yields were compared to those of an HPLC-FLD reference method (EN 14352). Most of the compared multi-mycotoxin methods that have been published were only for analyzing fumonisins in maize or maize-meal; we have applied the extractants of these methods to processed, complex maize matrices for the first time. Our results show that, for extractions with aqueous acetonitrile mixtures with the addition of acid, e.g. MeCN/H₂O/acetic acid (79/20/1, v/v/v), higher extraction yields are obtained than with MeCN/H₂O (80/20, v/v), in both spiked and naturally contaminated maize matrices. But compared to the results of the reference method EN 14352, the two extractants did not show a similar extraction efficiency. Overall, the extractant MeCN/MeOH/H₂O (1/1/2, v/v/v) turned out to be the most appropriate extractant applied in all experiments, obtaining the best and most comparable extraction yields and recoveries. Furthermore, our investigations showed that, with some of the tested extraction solvents, e.g. MeCN/H₂O (75/25) containing 50 mmol/l formic acid, stark differences occur when analyzing spiked and naturally contaminated matrices. With spiked matrices, recoveries of approximately 80–110 % were obtained, but with naturally contaminated matrices no results comparable to the EN method have been achieved. In contrast, a double extraction with MeCN/H₂O/formic acid (80/19,9/0,1, v/v/v), followed by a second polar extraction step with MeCN/H₂O/formic acid (20/79,9/0,1, v/v/v), led, for most naturally contaminated samples, to comparable results with the EN method. However, for spiked samples, the same extractant led to raised recoveries of between 120 and 140 %. For some processed matrices, like taco-chips, all tested extractants showed a poor extraction efficiency for fumonisins. By extending the extraction time from 1 to 15 min, a result comparable to that of the reference method could also