A collaborative study was conducted to evaluate the effectiveness of an immunoaffinity column cleanup liquid chromatography (LC) method for determination of aflatoxin B1 in a milk powder based infant formula at a possible future European regulatory limit (0.1 ng/g). The test portion was extracted with methanol–water (8 + 2 [v + v]), filtered, diluted with water, and applied to an immunoaffinity column. The column was washed with water to remove interfering compounds, and the purified aflatoxin B1 was eluted with methanol. The separation and determination of the aflatoxin B1 was performed by reversed-phase LC and detected by fluorescence after postcolumn derivatization (PCD) involving bromination. PCD was achieved with either pyridinum hydrobromide perbromide (PBPB) or an electrochemical (Kobra) cell by addition of bromide to the mobile phase. The baby food (infant formula) test samples, both spiked and naturally contaminated with aflatoxin B1, were sent to 14 laboratories in 13 different European countries. Test portions were spiked at levels of 0.1 and 0.2 ng/g for aflatoxin B1. Recoveries ranged from 101 to 92%. Based on results for spiked test samples (blind pairs at 2 levels) and naturally contaminated test samples (blind pairs at 3 levels), the relative standard deviation for repeatability (RSDr) ranged from 3.5 to 14%. The relative standard deviation for reproducibility (RSDR) ranged from 9 to 23%. Nine participants used PBPB derivatization, and 5 particpants used the Kobra cell. There was no evidence of method performance depending on the derivatization method used. The method showed acceptable within- and between-laboratory precision for baby food matrix, as evidenced by HORRAT values, at the target levels of determination for aflatoxin B1.

Nine laboratories from Denmark, Finland, Norway, and Sweden participated in a collaborative study to evaluate an immunoaffinity column cleanup/liquid chromatographic determination of aflatoxins B and G. Samples of peanut, fig, maize gluten, soya expel-ler, and copra spiked with total aflatoxin concentrations ranging from 1.37 to 28.60 ng/g were distributed. Altogether, 13 samples were analyzed. Of these, 6 were pairs of blind duplicates, 1 was an unknown negative sample containing ¼0.1 ng/g, and 2 were practice samples. A statistical presentation of the results is given. Although the results obtained were individually corrected for recovery, they were lower than expected in most cases. In some cases, the recoveries were unacceptably low, particularly for aflatoxin G2. Relative standard deviations for method repeatability (RSDr) and reproducibility (RSDR) were good but generally better for the peanut and fig samples than for the others. Results were better for aflatoxins B1 and G1 than for aflatoxins B2 and G2. RSDR values varied from 9 to 35%, and RSDR values varied from 17 to 81%. Horwitz ratio values were >2.00 for 1 RSDR value and <1.00 for 23 of the 35 calculated RSDR values.