OBJECTIVETo develop and verify a model diet for consumers with high exposure to food chemicals.

METHODSBased on the assumption that a person might consume average amounts of several different foods but only two or three at high levels, a model diet were put forward to calculate the high exposures to certain food chemical.Important parameters of this model were selected by comparing the outputs of this model with those of a reference model-semi-probabilistic model that is based on individual data. The concentration data of lead, cadmium and mercury (45 832, 43 862 and 25 243 samples respectively) from the national risk surveillance for chemical contaminations and harmful factors (2011), and the consumption data from the national survey on nutrition and health status in Chinese population (2002) were used in this model optimization process. The final model was verified using concentration data of eight chemical contaminations from national surveillance database of 2010.

RESULTSWhen 90, 95 and 97.5 percentile exposure being calculated under the conditions that 7, 12, 20 and 30 food categories was used respectively, the model containing two high consumption foods (2+x model) got results lower than the reference values (the relative deviation < 0) , the model containing three high consumption foods (3+x model) got results higher than the reference values (the relative deviation > 0 ),but the relative deviation of the two models were both within 20%. The verification results showed that the dietary exposure results of 8 food contaminations got from 3+x model were all higher than and(or) close to those got from semi-probabilistic model, and the relative deviations were between -5% and 25%.

CONCLUSIONThe 3+x model diet can fulfill the demands of "conservative" and "accurate" on exposure assessment model, and can give reliable estimations of high exposure to food chemicals occurred in various food categories.