Département Chimie de l’environnement

Wastewater as a medium contains information on both circulating pathogens and drug consumption at the population level. This study combines tracking of respiratory viruses and quantification of pharmaceuticals as untargeted indicators of symptoms related to acute respiratory infections and influenza-like illnesses such as coughing, fever and pain. From January 2021 to June 2024, raw wastewater samples from ten locations covering 23% of the Swiss population were analysed. This encompassed 15 pharmaceuticals and four priority respiratory viruses including severe acute respiratory syndrome coronavirus virus-2 (SARS-CoV-2), respiratory syncytial virus (RSV), influenza A and influenza B viruses. The pharmaceutical compounds dextromethorphan, pheniramine, clarithromycin, acetaminophen and codeine showed a strong correlation with respiratory virus loads in wastewater. This enabled the estimation of pathogen-specific and cumulative symptom treatment in the population. In 2021 and 2024, notable increases in pharmaceutical loads without corresponding increases in viral loads signalled high community symptoms linked to unsurveilled pathogens. This study demonstrates that pharmaceutical surveillance can inform respiratory disease burden and highlights the value of integrated surveillance for assessing emerging public health threats beyond those routinely monitored.

Antibiotics are of environmental concern. Their concentrations in the aquatic environment are frequently studied, while their occurrence in human excreta-derived fertilizers is less investigated. Therefore, levels of antibiotics, preservatives with antimicrobial properties, and various other micropollutants were determined in sewage sludge and in human fecal compost. Digested sludge of 29 Swiss wastewater treatment plants was analyzed, representing about 2.6 Mio people (30% of the Swiss population). This was compared with residues found in compost with dry toilet content after thermophilic composting, representing about 10 000 people. Fluoroquinolones and preservatives dominate in Swiss sewage sludge with weighted mean concentrations of 6500 μg kg⁻¹ and 2300 μg kg⁻¹. Levels of macrolides (240 μg kg⁻¹), β-lactam transformation products (35 μg kg⁻¹) and sulfonamides (15 μg kg⁻¹) were lower. Pollution patterns in digested sewage sludge were relatively constant throughout Switzerland. Levels of contamination in fecal compost were approximately 30 times lower than in sewage sludge. Pollution patterns differed between compost and sludge. Chemicals used in down-the-drain-applications (e.g., preservatives from personal care products or corrosion inhibitors) are less relevant in compost. Based on the Swiss consumption and excretion data, a mass flow analysis was carried out for antibiotics and pharmaceuticals in sludge and compost. The mass flow analysis in sludge showed a good agreement of predicted and measured concentrations for compounds that tend to sorb to organic matter (e.g., fluoroquinolones). Currently, there is no specific legislation that regulates the use of fecal compost from dry toilets as fertilizer. However, the one to two order of magnitude lower levels of contaminants in fecal compost compared to sludge and manure indicate a lower environmental risk when applying it as fertilizer.

Oxygen isotope ratios of O₂ are important tracers for assessing biological activity in biogeochemical processes in aquatic environments. In fact, changes in the ¹⁸O/¹⁶O and ¹⁷O/¹⁶O ratios of O₂ have been successfully implemented as measures for quantifying photosynthetic O₂ production and biological O₂ respiration. Despite evidence for light-dependent O₂ consumption in sunlit surface waters, however, photochemical O₂ loss processes have so far been neglected in the stable isotope-based evaluation of oxygen cycling. Here, we established the magnitude of the O isotope fractionation for abiotic photochemical O₂ elimination through formation of singlet O_2, ¹O₂, and the ensuing oxygenation and oxidation reactions with organic compounds through experiments with rose bengal as the ¹O₂ sensitizer and three different amino acids and furfuryl alcohol as chemical quenchers. Based on the kinetic analysis of light-dependent O₂ removal in the presence of different quenchers, we rationalize the observable O isotope fractionation of O₂ and the corresponding, apparent ¹⁸O kinetic isotope effects (¹⁸O-AKIE) with a pre-equilibrium model for the reversible formation of ¹O₂ and its irreversible oxygenation reactions with organic compounds. While ¹⁸O-AKIEs of oxygenation reactions amount to 1.03, the O isotope fractionation of O₂ decreased to unity with increasing ratio of the rates of oxygenation reaction of ¹O₂ vs ¹O₂ decay to ground state oxygen, ³O₂. Our findings imply that O isotope fractionation through photochemical O₂ consumption with isotope enrichment factors, ¹⁸O-ϵ, of up to −30‰ can match contributions from biological respiration at typical dissolved organic matter concentrations of lakes, rivers, and oceans and should, therefore, be included in future evaluations of biogeochemical O₂ cycling.

MLinvitroTox is an automated Python pipeline developed for high-throughput hazard-driven prioritization of toxicologically relevant signals detected in complex environmental samples through high-resolution tandem mass spectrometry (HRMS/MS). MLinvitroTox is a machine learning (ML) framework comprising 490 independent XGBoost classifiers trained on molecular fingerprints from chemical structures and target-specific endpoints from the ToxCast/Tox21 invitroDBv4.1 database. For each analyzed HRMS feature, MLinvitroTox generates a 490-bit bioactivity fingerprint used as a basis for prioritization, focusing the time-consuming molecular identification efforts on features most likely to cause adverse effects. The practical advantages of MLinvitroTox are demonstrated for groundwater HRMS data. Among the 874 features for which molecular fingerprints were derived from spectra, including 630 nontargets, 185 spectral matches, and 59 targets, around 4% of the feature/endpoint relationship pairs were predicted to be active. Cross-checking the predictions for targets and spectral matches with invitroDB data confirmed the bioactivity of 120 active and 6791 nonactive pairs while mislabeling 88 active and 56 non-active relationships. By filtering according to bioactivity probability, endpoint scores, and similarity to the training data, the number of potentially toxic features was reduced by at least one order of magnitude. This refinement makes the analytical confirmation of the toxicologically most relevant features feasible, offering significant benefits for cost-efficient chemical risk assessment.