What is in the fish? Collaborative trial in suspect and non-target screening of organic micropollutants using LC- and GC-HRMS.

A collaborative trial involving 16 participants from nine European countries was conducted within the NORMAN network in efforts to harmonise suspect and non-target screening of environmental contaminants in whole fish samples of bream (Abramis brama). Participants were provided with freeze-dried, homogenised fish samples from a contaminated and a reference site, extracts (spiked and non-spiked) and reference sample preparation protocols for liquid chromatography (LC) and gas chromatography (GC) coupled to high resolution mass spectrometry (HRMS). Participants extracted fish samples using their in-house sample preparation method and/or the protocol provided. Participants correctly identified 9-69 % of spiked compounds using LC-HRMS and 20-60 % of spiked compounds using GC-HRMS. From the contaminated site, suspect screening with participants' own suspect lists led to putative identification of on average ∼145 and ∼20 unique features per participant using LC-HRMS and GC-HRMS, respectively, while non-target screening identified on average ∼42 and ∼56 unique features per participant using LC-HRMS and GC-HRMS, respectively. Within the same sub-group of sample preparation method, only a few features were identified by at least two participants in suspect screening (16 features using LC-HRMS, 0 features using GC-HRMS) and non-target screening (0 features using LC-HRMS, 2 features using GC-HRMS). The compounds identified had log octanol/water partition coefficient (KOW) values from -9.9 to 16 and mass-to-charge ratios (m/z) of 68 to 761 (LC-HRMS and GC-HRMS). A significant linear trend was found between log KOW and m/z for the GC-HRMS data. Overall, these findings indicate that differences in screening results are mainly due to the data analysis workflows used by different participants. Further work is needed to harmonise the results obtained when applying suspect and non-target screening approaches to environmental biota samples.

Application of a novel prioritisation strategy using non-target screening for evaluation of temporal trends (1969-2017) of contaminants of emerging concern (CECs) in archived lynx muscle tissue samples.

Most environmental monitoring studies of contaminants of emerging concern (CECs) focus on aquatic species and target specific classes of CECs. Even with wide-scope target screening methods, relevant CECs may be missed. In this study, non-target screening (NTS) was used for tentative identification of potential CECs in muscle tissue of the terrestrial top predator Eurasian lynx (Lynx lynx). Temporal trend analysis was applied as a prioritisation tool for archived samples, using univariate statistical tests (Mann-Kendall and Spearman rank). Pooled lynx muscle tissue collected from 1969 to 2017 was analysed with an eight-point time series using a previously validated screening workflow. Following peak detection, peak alignment, and blank subtraction, 12,941 features were considered for statistical analysis. Prioritisation by time-trend analysis detected 104 and 61 features with statistically significant increasing and decreasing trends, respectively. Following probable molecular formula assignment and elucidation with MetFrag, two compounds with increasing trends, and one with a decreasing trend, were tentatively identified. These results show that, despite low expected concentration levels and high matrix effects in terrestrial species, it is possible to prioritise CECs in archived lynx samples using NTS and univariate statistical approaches.

Novel prioritisation strategies for evaluation of temporal trends in archived white-tailed sea eagle muscle tissue in non-target screening.

Environmental monitoring studies based on target analysis capture only a small fraction of contaminants of emerging concern (CECs) and miss pollutants potentially harmful to wildlife. Environmental specimen banks, with their archived samples, provide opportunities to identify new CECs by temporal trend analysis and non-target screening. In this study, archived white-tailed sea eagle (Haliaeetus albicilla) muscle tissue was analysed by non-targeted high-resolution mass spectrometry. Univariate statistical tests (Mann-Kendall and Spearman rank) for temporal trend analysis were applied as prioritisation methods. A workflow for non-target data was developed and validated using an artificial time series spiked at five levels with gradient concentrations of selected CECs (n = 243). Pooled eagle muscle tissues collected 1965-2017 were then investigated with an eight-point time series using the validated screening workflow. Following peak detection, peak alignment, and blank subtraction, 14 409 features were considered for statistical analysis. Prioritisation by time-trend analysis detected 207 features with increasing trends. Following unequivocal molecular formula assignment to prioritised features and further elucidation with MetFrag and EU Massbank, 13 compounds were tentatively identified, of which four were of anthropogenic origin. These results show that it is possible to prioritise new CECs in archived biological samples using univariate statistical approaches.

New extraction method prior to screening of organic micropollutants in various biota matrices using liquid chromatography coupled to high-resolution time-of-flight mass spectrometry.

A new extraction method with limited clean-up requirements prior to screening various matrices for organic micropollutants using liquid chromatography-high resolution mass spectrometry (LC-HRMS) for analysis was developed. First, the performance of three extraction methods (QuEChERS with SPE clean-up, ultrasonication with SPE clean-up, extraction without SPE clean-up) was tested, optimized, and compared using >200 contaminants of emerging concern (CECs) together covering a wide range of physicochemical properties applicable for suspect and non-target screening in biota. White-tailed sea eagle (Haliaeetus albicilla) muscle tissue was used in method development and optimization. The method without SPE clean-up was then applied to European perch (Perca fluviatilis) muscle, heart, and liver tissues. The optimization and application of the method demonstrated a wide applicable domain of the novel extraction method regarding species, tissues, and chemicals. For future applications, the suitability of the method for suspect and non-target screening was tested. Overall, our extraction method appears to be sufficiently simple and broad (relatively non-discriminant) for use prior to analysis of CECs in various biota.

Removal of pharmaceuticals, perfluoroalkyl substances and other micropollutants from wastewater using lignite, Xylit, sand, granular activated carbon (GAC) and GAC+Polonite® in column tests - Role of physicochemical properties.

This study evaluated the performance of five different sorbents (granular activated carbon (GAC), GAC + Polonite® (GAC + P), Xylit, lignite and sand) for a set of 83 micropollutants (MPs) (pharmaceuticals, perfluoroalkyl substances (PFASs), personal care products, artificial sweeteners, parabens, pesticide, stimulants), together representing a wide range of physicochemical properties. Treatment with GAC and GAC + P provided the highest removal efficiencies, with average values above 97%. Removal rates were generally lower for Xylit (on average 74%) and lignite (on average 68%), although they proved to be highly efficient for a few individual MPs. The average removal efficiency for sand was only 47%. It was observed that the MPs behaved differently depending on their physicochemical properties. The physicochemical properties of PFASs (i.e. molecular weight, topological molecular surface area, log octanol water partition coefficient (Kow) and distribution coefficient between octanol and water (log D)) were positively correlated to observed removal efficiency for the sorbents Xylit, lignite and sand (p < 0.05), indicating a strong influence of perfluorocarbon chain length and associated hydrophobic characteristics. In contrast, for the other MPs the ratio between apolar and polar surface area (SA/SP) was positively correlated with the removal efficiency, indicating that hydrophobic adsorption may be a key feature of their sorption mechanisms. GAC showed to be the most promising filter medium to improve the removal of MPs in on-site sewage treatment facilities. However, more studies are needed to evaluate the removal of MPs in field trials.

Per- and Polyfluoroalkyl Substances in Swedish Groundwater and Surface Water: Implications for Environmental Quality Standards and Drinking Water Guidelines.

The aim of this study was to assess per- and polyfluoroalkyl substances (PFASs) in the Swedish aquatic environment, identify emission sources, and compare measured concentrations with environmental quality standards (EQS) and (drinking) water guideline values. In total, 493 samples were analyzed in 2015 for 26 PFASs (∑26PFASs) in surface water, groundwater, landfill leachate, sewage treatment plant effluents and reference lakes, focusing on hot spots and drinking water sources. Highest ∑26PFAS concentrations were detected in surface water (13 000 ng L-1) and groundwater (6400 ng L-1). The dominating fraction of PFASs in surface water were perfluoroalkyl carboxylates (PFCAs; 64% of ∑26PFASs), with high contributions from C4-C8 PFCAs (94% of ∑PFCAs), indicating high mobility of shorter chain PFCAs. In inland surface water, the annual average (AA)-EQS of the EU Water Framework Directive of 0.65 ng L-1 for ∑PFOS (linear and branched isomers) was exceeded in 46% of the samples. The drinking water guideline value of 90 ng L-1 for ∑11PFASs recommended by the Swedish EPA was exceeded in 3% of the water samples from drinking water sources ( n = 169). The branched isomers had a noticeable fraction in surface- and groundwater for perfluorooctanesulfonamide, perfluorohexanesulfonate, and perfluorooctanesulfonate, highlighting the need to include branched isomers in future guidelines.