Aflatoxin M1 in milk:: Reliability of the immunoenzymatic assay

To evaluate the reliability of ELISA in aflatoxin M-1 (AFM(1)) determination in milk, more than 15,000 samples were tested and results from 600 of them were compared with those of the HPLC reference method. Extraction efficiency by immunoaffinity columns was checked under various conditions. ELISA and HPLC assays of spiked samples gave same precision (coefficient of variation), recovery, and regression coefficient R 2 values (0.9-8%, 96.8-108%, 0.993, respectively) for samples containing less than 70 ng L-1 of AFM(1). At higher concentrations up to 100 ng L-1, ELISA gave a slight overestimation, with CV 7-21% and R-2 = 0.876. This overestimation was confirmed (R-2 = 0.788) for the 600 comparison samples. Data from this study support ELISA as a reliable routine quality control assay for AFM(1) in milk, quantifying exactly its content for values above, but close to, the legal limit. Regulatory rules require HPLC confirmation of positive samples, but by using ELISA, no cases of false-negative determination should occur. (c) 2006 Elsevier Ltd. All rights reserved.