Abteilung Umweltchemie

A model to predict the mass flows and concentrations of pharmaceuticals predominantly used in hospitals across a large number of sewage treatment plant (STP) effluents and river waters was developed at high spatial resolution. It comprised 427 geo-referenced hospitals and 742 STPs serving 98% of the general population in Switzerland. In the modeled base scenario, domestic, pharmaceutical use was geographically distributed according to the population size served by the respective STPs. Distinct hospital scenarios were set up to evaluate how the predicted results were modified when pharmaceutical use in hospitals was allocated differently; for example, in proportion to number of beds or number of treatments in hospitals. The hospital scenarios predicted the mass flows and concentrations up to 3.9 times greater than in the domestic scenario for iodinated X-ray contrast media (ICM) used in computed tomography (CT), and up to 6.7 times greater for gadolinium, a contrast medium used in magnetic resonance imaging (MRI). Field measurements showed that ICM and gadolinium were predicted best by the scenarios using number of beds or treatments in hospitals with the specific facilities (i.e., CT and/or MRI). Pharmaceuticals used both in hospitals and by the general population (e.g., cyclophosphamide, sulfamethoxazole, carbamazepine, diclofenac) were predicted best by the scenario using the number of beds in all hospitals, but the deviation from the domestic scenario values was only small. Our study demonstrated that the bed number-based hospital scenarios were effective in predicting the geographical distribution of a diverse range of pharmaceuticals in STP effluents and rivers, while the domestic scenario was similarly effective on the scale of large river-catchments.

Determining optimal ozone doses for organic micropollutant elimination during wastewater ozonation is challenged by the presence of a large number of structurally diverse micropollutants for varying wastewater matrice compositions. A chemical kinetics approach based on ozone and hydroxyl radical (·OH) rate constant and measurements of ozone and ·OH exposures is proposed to predict the micropollutant elimination efficiency. To further test and validate the chemical kinetics approach, the elimination efficiency of 25 micropollutants present in a hospital wastewater effluent from a pilot-scale membrane bioreactor (MBR) were determined at pH 7.0 and 8.5 in bench-scale experiments with ozone alone and ozone combined with H₂O₂ as a function of DOC-normalized specific ozone doses (gO₃/gDOC). Furthermore, ozone and ·OH exposures, ·OH yields, and ·OH consumption rates were determined. Consistent eliminations as a function of gO₃/gDOC were observed for micropollutants with similar ozone and OH rate constants. They could be classified into five groups having characteristic elimination patterns. By increasing the pH from 7.0 to 8.5, the elimination levels increased for the amine-containing micropollutants due to the increased apparent second-order ozone rate constants while decreased for most micropollutants due to the diminished ozone or ·OH exposures. Increased ·OH quenching by effluent organic matter and carbonate with increasing pH was responsible for the lower ·OH exposures. Upon H₂O₂ addition, the elimination levels of the micropollutants slightly increased at pH 7 (<8%) while decreased considerably at pH 8.5 (up to 31%). The elimination efficiencies of the selected micropollutants could be predicted based on their ozone and ·OH rate constants (predicted or taken from literature) and the determined ozone and ·OH exposures. Reasonable agreements between the measured and predicted elimination levels were found, demonstrating that the proposed chemical kinetics method can be used for a generalized prediction of micropollutant elimination during wastewater ozonation. Out of 67 analyzed micropollutants, 56 were present in the tested hospital wastewater effluent. Two-thirds of the present micropollutants were found to be ozone-reactive and efficiently eliminated at low ozone doses (e.g., >80% for gO₃/gDOC = 0.5).

A pilot-scale hospital wastewater treatment plant consisting of a primary clarifier, membrane bioreactor, and five post-treatment technologies including ozone (O₃), O₃/H₂O₂, powdered activated carbon (PAC), and low pressure UV light with and without TiO₂ was operated to test the elimination efficiencies for 56 micropollutants. The extent of the elimination of the selected micropollutants (pharmaceuticals, metabolites and industrial chemicals) was successfully correlated to physical-chemical properties or molecular structure. By mass loading, 95% of all measured micropollutants in the biologically treated hospital wastewater feeding the post-treatments consisted of iodinated contrast media (ICM). The elimination of ICM by the tested post-treatment technologies was 50–65% when using 1.08 g O₃/gDOC, 23 mg/L PAC, or a UV dose of 2400 J/m² (254 nm). For the total load of analyzed pharmaceuticals and metabolites excluding ICM the elimination by ozonation, PAC, and UV at the same conditions was 90%, 86%, and 33%, respectively. Thus, the majority of analyzed substances can be efficiently eliminated by ozonation (which also provides disinfection) or PAC (which provides micropollutants removal, not only transformation). Some micropollutants recalcitrant to those two post-treatments, such as the ICM diatrizoate, can be substantially removed only by high doses of UV (96% at 7200 J/m²). The tested combined treatments (O₃/H₂O₂ and UV/TiO₂) did not improve the elimination compared to the single treatments (O₃ and UV).

A pilot-scale membrane bioreactor (MBR) was installed and operated for one year at a Swiss hospital. It was fed an influent directly from the hospital’s sanitary collection system. To study the efficiency of micropollutant elimination in raw hospital wastewater that comprises a complex matrix with micropollutant concentrations ranging from low ng/L to low mg/L, an automated online SPE-HPLC-MS/MS analytical method was developed. Among the 68 target analytes were the following: 56 pharmaceuticals (antibiotics, antimycotics, antivirals, iodinated X-ray contrast media, antiinflamatory, cytostatics, diuretics, beta blockers, anesthetics, analgesics, antiepileptics, antidepressants, and others), 10 metabolites, and 2 corrosion inhibitors. The MBR influent contained the majority of those target analytes. The micropollutant elimination efficiency was assessed through continuous flow-proportional sampling of the MBR influent and continuous time-proportional sampling of the MBR effluent. An overall load elimination of all pharmaceuticals and metabolites in the MBR was 22%, as over 80% of the load was due to persistent iodinated contrast media. No inhibition by antibacterial agents or disinfectants from the hospital was observed in the MBR. The hospital wastewater was found to be a dynamic system in which conjugates of pharmaceuticals deconjugate and biological transformation products are formed, which in some cases are pharmaceuticals themselves.

Point-source measures have been suggested to decrease pharmaceuticals in water bodies. We analyzed 68 and 50 alternatives, respectively, for a typical Swiss general and psychiatric hospital to decrease pharmaceutical discharge. Technical alternatives included reverse osmosis, ozonation, and activated carbon; organizational alternatives included urine separation. To handle this complex decision, we used Multiple-Criteria Decision Analysis (MCDA) and combined expert predictions (e.g., costs, pharmaceutical mass flows, ecotoxicological risk, pathogen removal) with subjective preference-valuations from 26 stakeholders (authorities, hospital-internal actors, experts). The general hospital contributed ca. 38% to the total pharmaceutical load at the wastewater treatment plant, the psychiatry contributed 5%. For the general hospital, alternatives removing all pharmaceuticals (especially reverse osmosis, or vacuum-toilets and incineration), performed systematically better than the status quo or urine separation, despite higher costs. They now require closer scrutiny. To remove X-ray contrast agents, introducing roadbags is promising. For the psychiatry with a lower pharmaceutical load, costs were more critical. Stakeholder feedback concerning MCDA was very positive, especially because the results were robust across different stakeholder-types. Our MCDA results provide insight into an important water protection issue: implementing measures to decrease pharmaceuticals will likely meet acceptance. Hospital point-sources merit consideration if the trade-off between costs and pharmaceutical removal is reasonable.

In this paper, we evaluated the ecotoxicological potential of the 100 pharmaceuticals expected to occur in highest quantities in the wastewater of a general hospital and a psychiatric center in Switzerland. We related the toxicity data to predicted concentrations in different wastewater streams to assess the overall risk potential for different scenarios, including conventional biological pretreatment in the hospital and urine source separation. The concentrations in wastewater were estimated with pharmaceutical usage information provided by the hospitals and literature data on human excretion into feces and urine. Environmental concentrations in the effluents of the exposure scenarios were predicted by estimating dilution in sewers and with literature data on elimination during wastewater treatment. Effect assessment was performed using quantitative structure-activity relationships because experimental ecotoxicity data were only available for less than 20% of the 100 pharmaceuticals with expected highest loads. As many pharmaceuticals are acids or bases, a correction for the speciation was implemented in the toxicity prediction model.
The lists of Top-100 pharmaceuticals were distinctly different between the two hospital types with only 37 pharmaceuticals overlapping in both datasets. 31 Pharmaceuticals in the general hospital and 42 pharmaceuticals in the psychiatric center had a risk quotient above 0.01 and thus contributed to the mixture risk quotient. However, together they constituted only 14% (hospital) and 30% (psychiatry) of the load of pharmaceuticals. Hence, medical consumption data alone are insufficient predictors of environmental risk. The risk quotients were dominated by amiodarone, ritonavir, clotrimazole, and diclofenac. Only diclofenac is well researched in ecotoxicology, while amiodarone, ritonavir, and clotrimazole have no or very limited experimental fate or toxicity data available. The presented computational analysis thus helps setting priorities for further testing.
Separate treatment of hospital wastewater would reduce the pharmaceutical load of wastewater treatment plants, and the risk from the newly identified priority pharmaceuticals. However, because high-risk pharmaceuticals are excreted mainly with feces, urine source separation is not a viable option for reducing the risk potential from hospital wastewater, while a sorption step could be beneficial.

Little is known about the significance of hospitals as point sources for emission of organic micropollutants into the aquatic environment. A mass flow analysis of pharmaceuticals and diagnostics used in hospitals was performed on the site of a representative Swiss cantonal hospital. Specifically, we analyzed the consumption of iodinated X-ray contrast media (ICM) and cytostatics in their corresponding medical applications of radiology and oncology, respectively, and their discharge into hospital wastewater and eventually into the wastewater of the municipal wastewater treatment plant. Emission levels within one day and over several days were found to correlate with the pharmacokinetic excretion pattern and the consumed amounts in the hospital during these days. ICM total emissions vary substantially from day to day from 255 to 1259 g/d, with a maximum on the day when the highest radiology treatment occurred. Parent cytostatic compounds reach maximal emissions of 8−10 mg/d. A total of 1.1%, 1.4%, and 3.7% of the excreted amounts of the cytostatics 5-fluorouracil, gemcitabine, and 2′,2′-difluorodeoxyuridine (main metabolite of gemcitabine), respectively, were found in the hospital wastewater, whereas 49% of the total ICM was detected, showing a high variability among the compounds. These recoveries can essentially be explained by the high amount administered to out-patients (70% for cytostatics and 50% for ICM); therefore, only part of this dose is expected to be excreted on-site. In addition, this study emphasizes critical issues to consider when sampling in hospital sewer systems. Flow proportional sampling over a longer period is crucial to compute robust hospital mass flows.

A method for solid phase extraction and HPLC–MS/MS of the cytostatics 5-fluorouracil, cytarabine, and gemcitabine and human metabolites uracil 1-β-D-arabinofuranoside and 2′,2′-difluorodeoxyuridine in wastewater was established. Wastewater samples from a Swiss hospital were analyzed for 5-fluorouracil, gemcitabine and 2′,2′-difluorodeoxyuridine. The limits of quantification were 5.0, 0.9, and 9.0 ng/L and the maximum concentrations detected were 27, 38, and 840 ng/L, respectively. Along with the method development, retention mechanisms on the hydrophilic interaction chromatography (HILIC) stationary phase were studied. Both partitioning and adsorption play a role in the retention on the tested sulfoalkylbetaine modified silica HILIC column material. The contribution of these two processes is changing over the 1.6–40% range water in the mobile phase. Although the specific break point is difficult to determine, adsorption becomes more significant as the fraction of water in the mobile phase decreases below approximately 16%.

ln the framework of elimination of micropollutants, resiclua of pharmaceuticals in waste water require special attention. ln Switzerland, 18 % of all sold drugs are given out in hospitals. lt is selfeviclent that hospitals are regarded as point source of residua of pharmaceuticals in waste water. However, there are only a few drugs significantly more often used in hospitals: contrast media and cytostatics. ln order to eliminate the residua, there are two strategies: upgrading the communal waste water treatment plants and, as a special case, the separate treatment of hospital waste water.