Comparison of SSD models for environmental risk assessment of nanomaterials – a case study on two reference materials

Sara Nørgaard Sørensen, Henning Wigger, Alex Zabeo, Danail Hristozov, Bernd Nowack, Dave Spurgeon, Anders Baun

Research output: Contribution to conferenceConference abstract for conferenceResearchpeer-review

35 Downloads (Pure)

Abstract

Environmental risk assessment of engineered nanomaterials (ENMs) requires predicted-no-effect-levels derived through the assessment factor (AF) or species sensitivity distribution (SSD) approach. The reliability of ecotoxicity data applied for such estimations is crucial, especially with ENMs, for which ecotoxicity testing has proven challenging. A reason is that the measurable characteristics uniquely defining ENMs are different than for conventional chemicals, and includes size, shape, surface coating, agglomeration state etc. Also, ENMs behave differently than conventional chemicals in environmental media, often being instable during testing. Bearing these difficulties in mind, the present study investigates how the outcome of SSD modeling is influenced by; 1) Weighing data according to the level of nano-specific characterization as described by Semenzin in al. 2016, and 2) Selecting input data based on nano-specific study quality criteria outlined by Hartmann et al. 2017.

In practice, freshwater ecotoxicity studies were collected for the silver and titanium dioxide reference nanomaterials NM-300K and NM-105, and evaluated according to their level of nano-specific characterization and reliability as referenced above. The resulting database covers 40 entries, 31 studies, 16 species and 6 taxonomic groups. Three datasets were formed: 1) All data, 2) All data categorized reliable, and 3) All NOECs only, and each run in three SSD models: The nano-specific weighed SSD (nSSWD by Semenzin et al., 2016), the probabilistic SSD (pSSD by Wigger et al., submitted) and the “standard” SSD generator (by US EPA available online).

For NM-300K, the majority of studies were evaluated as reliable, however the degree of nano-specific characterization conducted varied greatly. Generally, the standard SSD yielded most conservative HC5 values, but showed poor precision with 95% CI ranges app. 100 times wider than the other models. For the pSSD and nSSWD (without nano-specific weighing) HC5 values were largely similar. Employing nano-specific weighing in nSSWD significantly affected the HC5 outcome compared to both non-weighing and the other models. Moreover, it provided different outcome for the three datasets, illustrating that exclusion of unreliable studies increased HC5, whereas exclusion of data transformation (using NOECs only) decreased HC5. Seemingly, actual long-term NOECs are lower than NOECs estimated from short-term and/or effect data by use of assessment factors.
Original languageEnglish
Publication date2019
Number of pages1
Publication statusPublished - 2019
Event14th International Conference on Environmental Effects of Nanoparticles and Nanomaterials - ICEENN 2019 - Vienna, Austria
Duration: 1 Sep 20194 Sep 2019
Conference number: 14

Conference

Conference14th International Conference on Environmental Effects of Nanoparticles and Nanomaterials - ICEENN 2019
Number14
Country/TerritoryAustria
CityVienna
Period01/09/201904/09/2019

Fingerprint

Dive into the research topics of 'Comparison of SSD models for environmental risk assessment of nanomaterials – a case study on two reference materials'. Together they form a unique fingerprint.

Cite this