Problems with Modes of Action in Cumulative Risk Assessment. Expert Commentary.

August 18, 2011 at 11:04 am | Posted in Comment | 1 Comment

Dr Kristian SybergKristian Syberg, M.Sc. Ph.D. Assistant Professor, Department of Environmental, Social and Spatial Change, Roskilde University, Denmark. This commentary is a response to our post “Assessing Risk Posed by Chemicals in Mixtures“.

Mixture toxicity assessment is an important topic that is currently receiving some attention in EU. The European Commission recently published a “State of the art” report on the topic (Kortenkamp et al. 2009) and the Scientific Committees Preliminary Opinion on assessing risk posed by chemicals acting in mixtures (DG Health 2011).

The following commentary focuses on one specific aspect of the recommendations from the Scientific Committees – the use of Mode of Action (MoA) as a foundation for assessing mixture toxicity. [Editor’s note: See the US EPA website for a detailed definition of MoA.]

As I read it, the Scientific Committees recommend that chemicals are grouped as either similar or dissimilar acting, based on knowledge of their individual mode of action.

In cases where such data is available these two groups can then be assessed using two different models: dose/concentration addition (DA) for chemicals with the same MoA and independent action (IA) for different MoAs.

Furthermore, the Committees distinguish between human health and environmental risk. Regarding human health mixture toxicity assessment the Scientific Committees recommend that DA is applied for similar mixtures and that individual thresholds are adequate for mixtures of independently acting chemicals. Here, no specific mixture assessment should thus be required.

For environmental mixture toxicity assessment it is recommended that DA is applied for chemical mixtures which act similarly and IA for mixtures which do not. In situations where sufficient knowledge of mode of action is lacking the Committees recommend that DA is applied as default, since it is more protective than IA.

Even though I think many of the considerations are valid there are problems with these recommendations, which I would like to address. Requiring that MoA be known in order to group chemicals can drastically slow down the process for assessing risk posed by mixtures, since it is in reality impossible to obtain this kind of knowledge for chemical mixtures in the environment.

It does not make scientific sense to determine a single mode of action for many industrial chemicals that end up in the Environment. Two pesticides might both inhibit photosysthesis in plants – and thus be similar acting – but that does not imply that they have similar MoAs for other non-target organisms like fish.

In regard to human health, it might be easier to determine specific MoAs e.g. for endocrine disruptors. But even in such scenarios the number of MoAs might differ. For example, one endocrine disruptor might also be carcinogenic where another is not. In such a situation these two chemicals might have the same MoA in regard to their endocrine disrupting properties, but be dissimilar in regard to carcinogenic potential. Are they then similar or dissimilar?

What furthermore complicates this topic is the discussion about the level at which MoA should be defined: do chemicals have to have the same target site to be similar, or can two chemicals with similar effect but different target site be grouped as similar?

Another aspect in the proposal is to disregard human health mixture effects of dissimilarly acting chemicals below threshold concentrations. The argument for this is basically that, since human health risk assessments aim at protecting individuals, the thresholds set by the process are so low they ensure no effect at all. And in situations where the effect equals zero, there can be no mixture effect.

In principle this holds true, but only if (and this is a big if) the effect really is zero. The discussion as to whether or not thresholds in human health risk assessments ensure zero effect is beyond the scope of this text, but it is important to keep in mind that if chemicals have effects that do not equal zero, then by disregarding the mixture effect we will underestimate their toxicity – even if the chemicals in a mixture do not share the same MoA.

I believe it would be feasible to focus on testing chemicals in regard to similar endpoint instead of grouping them based on MoA. The chemicals that produce effects on similar endpoints should then be grouped based on DA. DA has been shown to be a scientifically solid approach with a margin of error that is low compared to other uncertainties inherent in the risk assessment system.

When applying models we often do it after the principles of Ockham’s Razor “Pluralitas non est ponenda sine neccesitate” (entities should not be multiplied unnecessarily). I think this principle can be adopted reasonably in this context as well. Since the empirical evidence indicates that there is no significant difference between applying DA and IA in most cases, we should choose the simple solution and apply DA by default.

In this regard it is important to keep in mind that we have moved drastically away from the original theoretical foundation for the two models, and that the evidence for applying them is more empirical than theoretical now.

Finally there is the important issue of synergistic effects. It seems feasible to apply prediction models to estimate mixture effects in most cases, even though they do not account for synergistic interactions. In scenarios with more than ten chemicals deviations from effect predicted by DA, whether synergistic or antagonistic, is seldom seen.

However, there might be some scenarios where a few chemicals can interact resulting in (synergistic) deviations from additivity. It is therefore important to evaluate this, in scenarios where few chemicals dominate the overall mixture toxicity.

I agree with the Scientific Committees proposal that this must be done case-by-case. But in general using DA for calculating mixture effects is a good and conservative approach – especially if the alternative is to neglect the mixture effect. In that regard it is encouraging to see the Scientific Committees are proposing DA be used when there is lack of knowledge of MoAs.

Edited for grammar, consistency with taxonomy already used in H&E, relevant references added.

1 Comment »

RSS feed for comments on this post. TrackBack URI

  1. It is photosynthesis, please correct,

    Another question is different target sites but same effect, this cannot be used, many chemicals can be involved in causing obesity, =same effect, but different target sites. For instance BPA as estrogenic agent and other chemicals like dioxins, that are involved in the developing process of adipose tissue, using a different pathway. Another example, both dioxinlike and non-dioxinlike PCBs with proven different pathways have both a negative effect on thyroid hormone metabolism. So the same effect is questionable.
    Last but not least the threshold question, for endocrine disruption a threshold isn’t there unless proven otherwise, this is the more so, because these endocrine disruptors have effects via the epigenoom. That means that you have to know what epigenomic changes are made on what genes. There is a paper of National research Council 2008 telling that thresholds cannot be used, Janna

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at
Entries and comments feeds.

%d bloggers like this: