The types of information available differ for existing and new chemicals. For established chemicals, exposure and adverse effect data may be readily available, e.g., from scientific literature, or poison control centres. Such information is often unavailable for unapproved new substances for which no practical experience exists.
Toxicological information may be gained prospectively or retrospectively. Prospective information may be gained from in vivo (including human), in vitro, or in silico studies. Data may be obtained using standardized test protocols or novel mechanistic techniques, e.g., toxicogenomics. Retrospective information includes epidemiological, post-marketing surveillance, and occupational health data, but also information from non-human sentinel species studies, e.g., fish, wildlife, or livestock.
Before collecting and/or evaluating toxicological information, it is important to assess its source. The quality of the data and the quality control measures used for generating the data should be assessed. An assay should be validated for its reliability and (predictive) relevance, and its mechanistic basis (relevance) should be understood. Finally, the appropriateness of the statistical methods used should be evaluated.
Sources of information should also be evaluated for their relevance to the question being addressed, e.g, taking into account interspecies variability. For example, weaknesses of in vitro systems include the inability to predict pharmacokinetic parameters, whereas in silico models have this capability, but do not on the other hand provide empirical information. Epidemiological studies should be assessed for population variability, and uncertainties over dose and exposure.
Other factors to consider include whether the original data are available for examination and audit. All information should be evaluated for variability and inconsistencies, and their potential causes. Laboratory testing (e.g., toxicogenomics) should adhere to validated, explicit test protocols. It is important also to consider the demonstrated applicability domain limitations of test methods based on physical/chemical properties. Lastly, publication bias should be considered.
Recommendations for evidence-based toxicological assessments include increasing the availability of standardized databases of toxicological information. Integrated testing strategies, including infomatic methods to integrate and analyse data from different methodologies, are also needed. Finally, the scientific community should develop and disseminate test methods and strategies for assessing the safety of novel materials, e.g., nanomaterials.
GRIESINGER Claudius;
HOFFMANN Sebastian;
KINSNER-OVASKAINEN Agnieszka;
COECKE Sandra;
HARTUNG Thomas;
2013-03-19
SAGE PUBLICATIONS LTD
JRC56783
0960-3271,
http://het.sagepub.com/content/28/2-3/149.citation,
https://publications.jrc.ec.europa.eu/repository/handle/JRC56783,
10.1177/0960327109105776,
