Screening of 2,073 mono-constituent organic substances REACH registered by June 2017 in tonnages at or above 10 tpa / EU manufacturer or importer for persistency, mobility and toxicity properties was performed by matching proposed criteria for being very persistent and very mobile (vPvM) or persistent, mobile and toxic (PMT) by Umwelt Bundesamt (UBA (1)) as well as a range of alternative criteria with algorithms employing QSAR predictions from the Danish (Q)SAR database. The (Q)SAR algorithms for the persistency and mobility part of the criteria sets were quite simple whereas those applied for T were rather complex. In addition to the T-related QSAR prediction algorithms, we also used harmonized EU classifications for CMR and STOT. The analysis of the results indicated that the water solubility part of UBA´s proposed set of criteria could be omitted and that UBA´s proposed set of criteria selects a lower number of substances than those by alternative sets of criteria where the cutoff values for minimum log D values between 4 and 9 are more in line with log Koc cut-off values of 3 and 4 for vM and M, respectively. The number of selected potential vPvM/PMT substances was not particular sensitive to whether the environmentally relevant pH range was set to the ranges of 5-8 or 4-9, but because the latter pH-range is more often used in environmental assessments for regulatory purposes, we propose to use that range. Our analysis included an additional screening of the identified potential vPvM/PMT-substances with respect to their possible endocrine disruptive properties (based on information in the Danish (Q)SAR Database) and other screenings relating to the environmental emission/exposure potential based on aggregated “best case” estimations of annual lowest tonnage and environmental exposure potential score category based on aggregated information from REACH registrations and information from the SPIN database (from the product registries of Norway, Sweden, Denmark and Finland). We also made a limited comparison of our (Q)SAR-based screening with a screening performed by Arp et al. (9) and Berger et al. (8) and conclude that their screenings selected around 25% of the substances selected by one of our vPvM/PMT-screening scenarios where the vP/P cut-off was set in a strict manner (i.e. “catching more”) and the vM/M- criteria matched those employing log Dmin, pH:4-9 proposed by UBA. Substances identified by both us and Arp et al. (9) or Berger et al. (8) may be of particular priority for further in depth vPvM/PMT- assessment. Finally, we have created an overview table given in Appendix 1 as well as a downloadable extended table containing chemical ID information, key P-, M-, T-, ED- and environmental exposure-related data (mainly (Q)SAR predictions used for screening), and screening selection flags for potential environmental exposure potential, whether the substances were also identified by Arp et al. (9) or Berger et al. (8) and vPvM/PMT –scenario selection “flags” for the most important of the many different vPvM/PMT-screening selection scenarios we have employed. We hope this work can contribute to further work e.g. by ECHA´s PBT Expert Group for identification and assessment of vPvM/PMT-substances which may be problematic for our ground- and surface water.
|Place of Publication||Kgs. Lyngby|
|Publisher||Danmarks Tekniske Universitet|
|Number of pages||79|
|Publication status||Published - 2021|