Last month’s Society for Risk Analysis Nano Risk (II) Workshop focused on how to move the field forward in terms of using alternative testing strategies for risk analysis (alternatives to animal testing). My colleagues at Vireo Advisors prepared a case study analysis with nanoscale titanium dioxide. A well studied substance used in many applications, we reviewed 96 papers that involved alternative testing or comparisons to animal testing and developed them into a database. When we tried to compare the studies, we couldn’t, because each was performed a little differently, or the details were not well enough described. So, while there is a pretty large database of studies, many of the data are not all that consistent and do not allow for quantitative risk analysis. This is most disappointing given that 1) people have been calling for standardization in nano-toxicology testing for years, 2) we limited analysis to studies from the last 5-7 years, figuring they would be of higher quality, 3) some of these studies represent significant government investment (and of course government investment comes from tax dollars, including yours and mine!).
So, when I saw the headlines about the paper, “Nanosafety research – are we on the right track?” [Krug 2014Agewandte Chem Ind Ed 53:2-18], I jumped right in to see what perspective this renowned expert would offer. His review of about 1000 of the more than 10,000 publications on environmental health and safety of engineered nanomaterials published since 2000 concludes that, “Most of these studies, however, do not offer any kind of clear statement on the safety of nanomaterials.” Are you shocked by this revelation? This has been my perspective, but I find the amount of evidence overwhelming. It certainly resonates with our recent TiO2 experiment.
Others of Krug’s conclusions also resonate. In our analysis of TiO2, we found only one study that included doses judged to be “environmentally relevant” or at realistic exposure levels. Krug reported many studies were conducted under “overload” conditions, rendering the results meaningless for interpreting toxicity. It’s a commonly accepted fact that the dose makes the poison, so by running experiments at too high doses, substances can appear more toxic than they may actually be. It would be like trying to run a diabetes diet study on Halloween, when subjects are full of candy.
The most significant of Krug’s findings for me is that of all of this work, little contributed to the knowledge of a specific “nano-effect”. That is, from all of this effort and investment, including building the infrastructure for several large nanocenters, we still cannot say definitely whether there are any unique effects from exposure to nanomaterials.
Does this obviate the need for the field of nanotoxicology? No, but it calls on researchers to pay due attention to the real world implications of their work. Many of the studies were reportedly not performed by toxicologists, which is a problem. They didn’t use negative or positive controls, they failed to measure or report physical or chemical properties in a way that allows doses to be confidently characterized, and the assay methods used were not standardized. Our TiO2 analysis corroborates these findings as well. When non-specialists review such studies, they may not have the perspective to interpret the quality of the study. What would help is if researchers and reviewers would pay more attention to the relevance of these studies for understanding the impacts nanomaterials in real world applications, and report what they do and do not tell us about health risks. A novel assay not tested outside of one’s lab is not going to provide conclusive evidence. At a minimum, poorly controlled studies have no place in the peer reviewed literature.
While a major issue for funders and publishers (should these studies be accepted/published?), it’s an even bigger issue for risk assessors and end users. I’ve been calling the results of toxicology studies on nanomaterials equivocal for a decade, and will continue to in the near term. I will also continue to keep the bullet point in my slide deck that “no nano-specific effect has been identified”. Because of poor characterization, the job of determining acceptable dose levels is even more challenging, reinforcing my approach to focus on mitigating potential exposure through the use of nano-life cycle risk assessment to minimize risks from nanomaterials. I predicted in 2005 that it would be 20 years before we answered some of the critical toxicology questions for nanomaterials, and 10 years in, we appear to be on exactly such a course. The bottom line, the state of the science for toxicology of nanomaterials is still in adolescence in terms of environmental relevance.