Evaluation and validation of two fluorometric HPLC methods for the determination of aflatoxin B1 in olive oil

Two methods for the determination of aflatoxin B-1(AFB(1)) in olive oil were tested and compared. In method A the oil sample was mixed with methanol + water (60 + 40), extracted with hexane and then with chloroform. Chloroform was evaporated and the residue was dissolved with dichloromethane which was then transferred for clean-up onto a silica "Sep-Pak' cartridge. The cartridge was pre-washed with hexane, ethyl ether and dichloromethane. AFB(1) was eluted with chloroform + acetone (9+1), and evaporated to dryness. In method B, the oil sample was mixed with methanol + water (80+20), shaken and centrifuged. The supernatant was diluted 1:10 with water and 10 ml of the diluted mixture transferred to an "Aflaprep' immunoaffinity column for the clean-up step. AFB(1) was eluted with acetonitrile and evaporated to dryness. AFB(1) from both methods was derivatized to its hemiacetal (AFB(2a)) and then quantitated by HPLC using a C-18 (60 Angstrom 4.6 x 250 mm) column with fluorescence detection. Both methods are simple, reliable nd efficient, but method A showed a lower detection limit (2.8 ng/kg) than method B (56 ng/kg). With a 95% confidence level there was no significant difference in recovery between the two methods, which was 87.2% for method A and 84.8% for method B. In addition, application of a two-tailed F-test to the variances within spiked samples at concentrations 1, 2, 5 and 10 mu g/kg separately showed that there was no significant difference in the precisions of the two methods. Fifty samples of olive oil of Greek origin produced between 1995 and 1998 were examined with both methods for the presence of AFB(1). When analysing the samples wit method B, the presence of AFB(1) was not detected. The use of method A revealed the presence of AFB(1) in 72% of the samples. The range of contamination was generally found to ber very low (2.8-15.7 ng/kg), however one sample was contaminated with 46.3 ng/kg.