DISCUSSION AND CONCLUSION

The landscaping exercise demonstrated that tools and data exist to estimate exposures from consumer products but the amount of data and tools available varies per applicability domain and tier of exposure assessment. The tabulated information provides a useful resource for individuals seeking to perform consumer exposure estimation. An evergreen version of this or a similar reference resource would be useful for the exposure assessment community in general.

In practice, the purpose of the exposure assessment and the required level of detail define the framework and algorithm of exposure modelling and, in particular, the strategy of aggregation of exposure (e.g. sum of worst-cases, or consideration of more refined data like products co-use and chemical occurrence of the chemical). This information should be borne in mind at all points in an exposure assessment, in turn informing what tools and data are best used for a given purpose. Different models and data are best suited for specific purposes, and knowing what tier given tools can to be used for is a vital consideration when performing exposure estimations. Said otherwise, the exposure assessment should be fit-for-purpose.

At lower tiers, for chemicals with a large margin of exposure, it may be sufficient to simply sum worse case consumer exposures, bearing in mind that the output will not be realistic. Lower tier models are purposefully designed estimates of exposure considered to be conservative, that is a high estimate of exposure potential so that they can quickly prioritise where more detailed exposure estimation may be useful. Thus, adding individual low tier estimates will provide a quantitative value that is not likely to reflect an actual exposure, but rather an inflated exposure estimate. At higher tiers, aggregation ideally should adhere to a person-oriented approach (Delmaar et al, 2006) to maintain consistency and to avoid unreasonable overestimation of exposure. If exposure potentially occurs via different sources/pathways, the combination of the pathways considered in the assessment should represent a realistic situation for the considered individual. Sources/pathways that in reality would never co-occur should not be combined (for example, the occupational exposure of an industrial worker should not be combined with the hand-mouth contact exposure of an infant).

Another important aspect for consideration when aggregating exposure is the toxicity of the chemical under study. The timescale on which the exposure is assessed should be consistent with the exposure durations for which health effects are observed. If acute toxicity is a critical endpoint, the assessment should estimate exposures on acute timescales (e.g. one day). Here, details on the temporal and spatial correlations of single exposure events become important, since e.g. two or more exposures occurring simultaneously along different pathways may in combination lead to a peak exposure exceeding some tolerable level, although each exposure event individually may remain below this level. If longer (e.g. one year) timescales are considered, adding the average exposures from different pathways without explicit reference to the temporal correlations between the exposure events can be acceptable. However, in the case of a highly variable profile, the time-averaged value may not only depend on the length of the averaging interval but also on the commencement and termination of this interval (e.g. a weekly average from Monday to Monday or from Sunday to Sunday).

Besides the timescale, the aggregation strategy is very much determined by exposure route(s), which in turn is also governed by the chemical toxicity profile. Generally, when the health effects differ among exposure routes, aggregation should be performed for each individual route separately (ECHA, 2016), followed by the integration of the commonly expressed route-specific aggregate exposures into a ‘collective aggregate exposure’. To accurately aggregate the route-specific doses e.g. to derive the total systemic dose, one needs to calculate the uptake, i.e. the amount of substance that can penetrate the outer barrier of the body (such as skin, lung or gut). Therefore, it is essential to distinguish between the exposure that describes the situation when the human body gets in contact with a chemical and the (internal) dose that actually describes the amount of chemical taken up into the human body as a result of exposure event. The route-specific uptake rates, which are usually measured by means of in vitro or in vivo animal studies, may not necessarily reflect the true absorption or penetration, since realistic exposure scenarios typically differ from the experimental conditions of these studies. The differences may arise due to studying of the pure substance instead of product mixtures, application of high/infinite doses, translating in vitro results to in vivo situations (Blaauboer, 2010; Yoon et al, 2012).

In conclusion, doing aggregate exposure in a meaningful way, using higher tier approaches, requires a high level of detail, in both exposure and hazard aspects, to support the development of realistic summed estimates of exposure via multiple sources. This level of information can be resource intensive to collect. The literature examples provide cases where this has been done for select product categories. Few studies provide aggregation on a total level from all sources, which would require an even greater level of information on co-exposure patterns and likelihood of those patterns, covering the entire range of possible exposure sources for a substance.

This evaluation of the state of science suggests that a most useful first effort might be further development of guidelines for understanding when the additional information provided by an aggregate estimate is most warranted. For example, if population level biomonitoring data are available for a substance and that substance is shown to have low risk potential, that substance may be a lower priority for directing resources to obtain detailed information needed to support aggregate modelling. If, however that substance is considered to have a risk potential that suggests exposure reductions may be appropriate and a specific exposure source is not recognised to be dominant, then it may be appropriate to perform an aggregate assessment to identify key source contributions that could be acted upon. Other types of criteria, such as those mentioned earlier, could be considered for inclusion in a systematic approach to identifying when the level of resources needed to perform aggregate exposure estimation will provide information needed to adequately characterise risk or safety potential, depending upon the purposes of the assessment being performed.