ECHA

Assessment factors play an important role in assessing the health hazards of a chemical substance. They are essential in bridging the gap between toxicological information, usually generated through studies with experimental animals, and the human situation.

This gap is largely determined by the fact that tests are not done on humans and that the real-life exposure of humans often differs from the exposure of the experimental animals. The results of a study with experimental animals can therefore not be directly used to predict what will happen in humans. This holds true in particular for predicting exposure levels and exposure conditions that may or may not lead to any adverse effects observed in an experiment.

Still, we have to get insight into safe human exposure conditions and exposure levels, which is, after all, one of the reasons for performing toxicological experiments. We can use other information on the properties of the substance, for instance information on what the body does with it (uptake from the gut, metabolism in the liver, distribution over the various organs and tissues, excretion by the kidneys, etc.).

General toxicological principles may also help us to bridge the gap, but often we are still not confident when these additional considerations are used. As a result, non-science-based margins are built in to prevent underestimation of the human health hazard.

Additional information on the substance, the general toxicological principles and any residual uncertainty are reflected in the assessment factors. For instance, the average sensitivity of mammals to the toxic effects of substances is related to their size. Large mammals are more sensitive than small ones, at least when the dose of the substance per kilogram of body weight is the same (total dose divided by body weight). Rats, for example, are on average four times less sensitive than humans on a mg/kg basis, and this factor of four is therefore used as an assessment factor when the toxicological experiment is performed with rats. However, differences in sensitivity may be larger (or smaller) than the average difference for reasons specifically related to the nature of the substance. To prevent an underestimation of the toxicological hazard an extra assessment factor is used.

Safe exposure levels 

A measure of human exposure is needed to mark the border between safe or acceptable exposure levels and hazardous or unacceptable exposure levels. In REACH this measure is called the derived no-effect level (DNEL) or the derived minimum effect level (DMEL).

Generally, we only have the results of toxicological studies with experimental animals as a starting point for the derivation of this measure because adequate human data is rarely available. Under the chemical safety assessment defined by REACH (Annex I), the DNEL or DMEL is then compared with the estimated exposure levels and, if the exposure exceeds DNELs or DMELs, further risk management measures are needed to bring the exposure levels down. Therefore DNELs or DMELs play a crucial role in ensuring safe use of the substance and in fulfilling the REACH obligations.

The derivation of a DNEL begins with the selection of an experimental dose level that serves as a starting point. In most cases, this is the no-observed adverse effect level (NOAEL) of a relevant endpoint, meaning the highest dose applied in an experiment that does not lead to an adverse effect. This dose is subsequently ‘corrected' for the differences between the study with experimental animals and the human situation. This correction consists of: 1) scaling; where the NOAEL is converted to a measure of exposure comparable with the outcome of the human exposure assessment and 2) dealing with uncertainty; by applying the assessment factors to the scaled exposure level.

Only use the default assessment factors provided by REACH

ECHA provides extensive guidance on the application of assessment factors to address the differences between an experiment and a human situation, and the uncertainties associated with these differences. A set of default assessment factors is defined in ECHA's guidance.

Default assessment factors should be applied when there is no information on the substance that would allow the definition of substance-specific or case-specific assessment factors. The defaults can only reflect generic information and remaining uncertainty. When information on the substance is available, it should be used to adapt the default assessment factors. The adaptation is necessary because the substance-specific factors have to provide at least the same level of protection as the default ones and therefore, have to be based on the default factors.

The set of default assessment factors plays an essential role in human hazard assessment under REACH, because:

• They are applied when there is no substance-specific information on the differences between experiment and human situation that have to be addressed for the substance in question, which is often the case;
• They are to be adapted to substance-specific assessment factors based on the substance-specific information.

As with most sets of assessment factors currently applied in regulatory frameworks, the REACH default set is based on the choice of the factor of ‘100‘ in the 1950s. This factor was originally chosen to compensate for the uncertainty associated with setting up an accepted concentration of substances (contaminants, additives) in human food, based on a chronic oral toxicity study. This choice was not based in the first instance on scientific information. The factor of 100 has subsequently been differentiated into the default sets currently in use, whereby some generic information has been used.

In many REACH registration dossiers, companies have used different default assessment factors and not those presented in the REACH guidance. Sometimes these assessment factors are based on an explanation provided by the registrant; in most cases, the registrant only refers to a publication and/or an opinion.

Using these alternative default sets seriously undermines the validity of the DNEL derivation for the purpose of REACH. There can only be one default set of assessment factors. This set applies to all DNEL derivations where there is no additional, substance-specific insight into the uncertainties that have to be bridged. Any deviation from the default can only be based on substance-specific information. ECHA cannot accept alternative defaults because to do so would imply that one registration can give rise to different DNELs, depending on the default the registrant has chosen, which is of course an undesirable situation.

The default assessment factors are therefore crucial for DNEL derivation. They have a strong influence on the accepted human levels of exposure to a chemical substance. The establishment of the REACH set of default factors was based on a careful process in which many stakeholders were involved. For the reasons explained above, no other acceptable defaults can be used under REACH. This is not to say that the REACH set cannot change. Development of science or policy may lead to an adaptation in the future. However, at the moment, deviations are only possible when based on scientifically sound information on the substance that increases insight into the uncertainty associated with the DNEL derivation for that specific substance.

Further information:

• Guidance on Information Requirements and Chemical Safety Assessment, Part B – Hazard Assessment
• REACH Regulation, Annex I

Text by Karel de Raat