Consumers' ability to discriminate aflatoxin-contaminated Brazil nuts.

The objectives of the study were to investigate the extent to which consumers can separate nuts with a high content of aflatoxin from sound nuts, and whether sorting results can be improved by information or whether they are affected by certain factors. A test panel consisting of 100 subjects was asked to crack 300 g Brazil nuts and to sort the nuts into those they considered edible and inedible. The test showed that consumers can, on current behaviour, discriminate aflatoxin-contaminated Brazil nuts to a significant extent. The median and the 95th percentile of the total concentrations of aflatoxins (B1, B2, G1, G2) in the samples before sorting were 1.4 and 557 microg kg(-1), respectively, and in the edible fractions after sorting 0.4 and 56 microg kg(-1), respectively. Given that levels of aflatoxins before sorting exceed either 2 microg aflatoxin B1 kg(-1) or totally 4 microg aflatoxins kg(-1), there was no effect of aflatoxin concentrations before sorting on the probability of exceeding these thresholds in the edible fraction. This means that similar sorting results were obtained for samples with aflatoxin levels exceeding either of the two thresholds, irrespective of if the thresholds were exceeded with a few microg kg(-1) or up to more than 1000 microg kg(-1). None of the tested factors (such as sex, age, level of education, ethnic background or knowledge of mycotoxins) had any effects on the probability of exceeding either of the two aflatoxin thresholds.