Integrated Probabilistic Risk Assessment (IPRA) for carcinogens : A first exploration

Abstract

In 2007 the National Institute for Public Health and the Environment (RIVM) and Wageningen University developed the IPRA-method (Integrated Probabilistic Risk Assessment) to estimate which fraction of the population is affected by non-carcinogenic substances in food. Following research commissioned by the Dutch Food and Consumer Product Safety Authority (nVWA) the RIVM shows that the IPRA-method can also be applied to carcinogenic substances. In the IPRA-method the uncertainty in the available data is translated into confidence limits of the results. This gives a more realistic view on the potential health effect. This report describes how the required input data for the IPRAmethod and the results thereof need to be interpreted. As a result of the severe nature of the effect 'cancer' it is desirable that the extra risk of cancer following exposure to substances is very small, for example 1 in a million. Measuring such low cancer incidences would require animal testing at a scale that is too large to be feasible. Therefore, these low risks cannot be measured in animal studies. In practice, measurable cancer incidences from animal experiments are linearly extrapolated to the desired low (non-measurable) cancer incidences. A case study with the carcinogenic mycotoxin aflatoxin B1 illustrates that the uncertainties in risk estimates related to carcinogenic substances are indeed very large. The currently applied linear extrapolation technique results in a single, supposedly conservative, risk estimate, without showing the associated uncertainties. The IPRA-method on the other hand does provide an indication of the uncertainty in the risk estimate. As such it may be a very promising tool for risk managers. The outcome of the method more realistically reflects to what extent quantitative statements on the risk can be made, given the available information. This allows the risk manager to make better informed decisions.