Risk calculation

Cancer risk

To calculate an "acceptable" dose level of genotoxic carcinogens, two parameters must be known. First the dose-response line. This is known as the "slope factor"; they are mostly based of the results of toxicological studies which can be found in the scientific publications and collected in various databases (like IRIS). Regarding the IRIS database a background document is presented where the extrapolation is described in more detail.

The second factor is the level of additional cancer risk to be accepted in the population due to exposure of chemicals in food. Different levels exist, leading to confusion by laypeople. Most experts use a risk level of 1 case in a population of one million (1 in 1,000,000) exposed persons per year, being equal to 1 case in ten thousand persons (1 in 10,000) after lifetime exposure. For comparison: the global background cancer risk for all cancers was estimated to be about 20 per 10,000 in 2018.

The animal study

The method to evaluate cancer risk starts with a toxicological study where experimental animals are divided in groups. Each group consist usually of 10 to 20 animals. All but one group are exposed to the chemical substance to be studied, in increasing dose levels. The one group that is not exposed is the "control" group. The animals are sacrificed after a lifetime exposure, and each animal is searched for tumors. Based on the number of animals with tumors the cancer incidence is calculated for each group. The chemical substance is considered carcinogenic when the increase in number of animals with tumors is dose related.

Linear extrapolation

As the study is intended to find the additional cancer risk above the common background incidence, the cancer frequencies of the exposed groups have to be corrected for the background (this is the frequency of animals with tumors in the control group). The group with the lowest increase in tumor incidence above the control is then used in the extrapolation.The figure shows a substance that produces additional tumors above background in 3.1% of the animals (31 per 1000) when administered 55 mg/kg/day. "Datum" points to that result. The extrapolation is based on the assumption that the data extend linearly to the 0, 0 origins. In this case is the additional cancer risk 31 cases per 1,000, which is equal to 31,000 cases per million individuals or 310 cases per ten thousand. In order to have one additional cancer incidence in a million individuals, the dose would be 0.002 mg/kg/day (55 [mg/kg/d] ÷ 31,000 = 0.0018 mg/kg/day). When dealing with a lifetime acceptable cancer risk of 1 in 10,000 (like in the EU) the dose equals to 0.18 mg/kg/day (55 ÷ 310). For this carcinogenic compound dose levels below 0.18 mg per kg per day are therefore to be considered "safe" in the EU.

The slope factor is cancer incidence ÷ dose. In this example the slope factor equals 0.00056 [mg per kg per day]^-1 (31 cases per 1000 = 0.031 ÷ 55 [mg per kg per day]) .

Risk calculation

The risk is intake × slope factor. The cancer risk for the above example at an intake of 0.6 mg per kg per day is 0.00034 (0.6 [mg/kg/day ] × 0.00056 [mg per kg per day]^-1). This situation corresponds with 3 to 4 cases per 10,000, which is above the acceptable risk level of 1 per 10,000 for lifetime exposure in the EU. This conclusion is in line with the cancer limit value of 0.18 mg/kg/day in the EU, as it is exceeded about three times in this example.