EFSAs Margin of Exposure

Margin of Exposure

As already mentioned, it was adviced by EFSA in 2012 to evaluate health risks of genotoxic carcinogenic chemical substances by means of the BMDL10. This is different from various other countries. They still use the concept of the slope factor.

The EFSA approach looks quite simple. It states that the exposure of a consumer to the chemical should be at least 10,000 times lower than the BMDL10 derived from a study with experimental animals. In that sense it looks quiet similar to the use of a HBGV where the exposure should be equal or less than the HBGV. So, at first sight, the BMDL seems to be a kind of HBGV using an interpolated NOAEL with a fixed Safety Factor of 10,000. That factor of 10,000 can be understood as the usual extrapolation factors of 100 for a the HBGVs include for extrapolation from experimental animals to humans, and for the variation of sensivity within the human population. For the BMDL one should also include an extrapolation factor of 10 for the BMDL to the dose without effect, and the type of effect (i.e. cancer). However, the HBGV is derived by a number of experts who select the SFs on the basis of expert judement. That expert judgement is now lacking when using a BMDL10 and a fixed critical level of 10,000.

Points missing

It is not described by EFSA how to evaluate the MoE using a BMDL01 or BMDL05, or BMDLs derived from human studies. One can argue, that the critical value can be lower than 10,000 when dealing with a BMDL01 or BMDL05, as experts will select a lower SF than 10 for low dose extrapolation. As the BMDL is calculated using the lower uncertainty limit, the SF for variation can also be lower than 10. And, using a BMDL based on human studies, the SF for extrapolation from animal to human is not needed. So it can be concluded that the critical value of 10,000 is correct for a BMDL10 based on a study with experimental animals, but that the critical value should be lower when using other BMDLs. There is however no guidance from EFSA how to act in such cases. It is adviced to consult a toxicological expert about the critical level if one is considering to compare a Margin of Exposure with another BMDL than the one mentioned by EFSA.

Chemicals without a HBGV

EFSA has noted to use the Margin of Exposure approach when assessing exposure to genotoxic carcinogens in food. Although not stated by EFSA one could consider to use this approach also for other chemicals without a HBGV. This might be most relevant for the evaluation of certain chemicals for which sufficient data is available but where a HBGV was withdrawn. A known example is lead, with a former PTWI of 25 ug per kg per week. This PTWI was withdrawn at the 73th JECFA meeting in 2010. In the EFSA Scientific Opinion on Lead in Food of 20 April 2010 a BMDL01 of 1.5 ug per kg per dag was reported for a study with human volunteers. Such values ar very helpfull for an exposure assessment. In the latter case is can even be concluded that the critical level is far smaller than 10,000.


Although EFSAs approach for the assessment of genotoxic carcinogenic compounds in food and feed by means of the Margin of Safety is far more easy that using a slope factor, one should be aware of the following points.

- The BMDL is a value calculated by statistical method using curve fitting, using data from experimental studies. The underlying functions might lack biological meaning.
- HBGVs are derived by experts from data of experimental studies and safety (i.e. extrapolation) factors. The selection of the effect and extrapolation to human exposure is made on the basis of expert judgement. The BMDL is not derived by expert judement, and does not include the extrapolation from animals to humans.
- The critical level of 10,000 is to be used for a MoE based on a BMDL10 of a study with experimental animals. There is no guidance when using another BMDL or those based on human studies. In these cases the critical level can be smaller than 10,000.
- Although not mentioned in the EFSA statement, the MoE approach can also be used for chemical substances without a HBGV, and for chemicals which HBGVs are withdrawn.