Skip to main content
European Commission logo print header

Bioanalytical evaluation, identification, and characterisation of disinfection by-products formed during drinking water treatment

Final Report Summary - AQUABIOTOX (Bioanalytical evaluation, identification, and characterisation of disinfection by-products formed during drinking water treatment)

Chemical drinking water disinfection, to provide drinking water free of pathogens, is associated with the formation of disinfection by-products (DBPs) that have been linked to public health concerns. Drinking water chlorination produces a wide range of toxic DBPs, only a fraction of which has been identified and characterized for their toxic potential. Bioanalytical tools should complement chemical analysis for a comprehensive risk assessment to assure that unidentified DBPs are also captured. In vitro bioassays enable the detection of mixture effects caused by known and unknown DBPs. However, many DBPs are volatile and polar and thus cannot be evaluated with current methods.
This study was organized in 6 work packages (WPs) with the following objectives:
WP1) to evaluate and adapt techniques to enrich volatile and polar DBPs
WP2) to identify and adapt bioassays to assess toxic DBPs
WP3) to evaluate and adapt fractionation methods to isolate toxic DBPs
WP4) to evaluate process and precursor conditions for DBP mitigation and DBP removal options
WP5) to quantify known DBPs for complementing WP1-4
WP6) effect directed analysis to identify and characterize (groups of) toxicologically relevant DBPs

WP1) Enrichment methods for bioanalytical water quality assessment result in the loss of volatile and hydrophilic disinfection by-products (DBPs) and hence underestimate biological effects in disinfected drinking water. We developed methods for extracting non-volatile (solid-phase extraction) and volatile DBPs (purge and cold-trap method) to minimize the loss of analytes. Both methods delivered excellent recoveries of 50–100% for known DBPs as well as for the adsorbable organic halides (AOX) from tap water. Results of WP1 have been published in ENVIRONMENTAL SCIENCE AND TECHNOLOGY (Stalter, D., Peters, L.D. O'Malley, E., Tang, J.Y.M. Revalor, M., Farre, M.J. Watson, K., Von Gunten, U. and Escher, B.I. (2016) Sample enrichment for bioanalytical quality assessment of disinfected drinking water: concentrating the polar, the volatiles, the unknowns. Environmental Science & Technology, 50 (12), 6495–6505). The developed method for the enrichment of volatile compounds provides an effective way to include the volatile fraction without enriching matrix constituents and can also be useful for the bioanalytical assessment of other volatile organic contaminants in water samples. The consecutive combination of the purge-and-trap method with SPE allowed for the assessment of the toxicological relevance of the non-purgeable versus the purgeable fraction.
WP2) A set of nine in vitro cellular bioassays indicative of differing stages of the toxicity pathway was applied to 50 disinfection by-products (DBPs) to obtain a better understanding of the commonalities and differences in the molecular mechanisms of the reactive toxicity of DBPs. 98% activated the NRf2-mediated oxidative stress response and 68% induced an adaptive stress response to genotoxic effects as indicated by the activation of the tumor suppressor protein p53. The energy of the lowest unoccupied molecular orbital ELUMO as reactivity descriptor was linearly correlated with oxidative stress induction for trihalomethanes (r2 = 0.98) and haloacetamides (r2 = 0.58) indicating that potency of theses DBPs is connected to their electrophilicity. For HAAs, we additionally accounted for speciation by including the acidity constant with ELUMO in a two-parameter multiple linear regression model. This increased r2 to > 0.80 indicating that HAAs’ potency is connected to both, electrophilicity and bioavailability. Based on the activation of oxidative stress response and the soft electrophilic character of most tested DBPs we hypothesize that indirect genotoxicity—e.g. through oxidative stress induction and/or enzyme inhibition—is more plausible than direct DNA damage. Results of WP2 have been published in WATER RESEARCH (Stalter, D., O'Malley, E., von Gunten, U. and Escher, B.I. (2016) Fingerprinting the reactive toxicity pathways of 50 drinking water disinfection by-products. Water Research 91, 19-30.). Additionally, we applied a selection of bioassays to assess mixture effects of DBPs to test if DBPs act together in a concentration additive manner. The ratio between experimental and predicted effect concentrations using the concentration addition model were within a factor of 2 demonstrating that the concept of concentration addition is valid for DBPs. The main implication for this result is that the concept of concentration addition could be applied to prioritize toxicologically relevant DBPs for further risk assessments and regulations. In whole drinking water samples the sum of detected DBPs explained <6% of the effect in most cases. Therefore, future research should focus on the identification of toxicologically more relevant DBPs. Results of this study are currently prepared for submission to ENVIRONMENT INTERNATIONAL (Stalter, D., O'Malley, E., Tang, J., von Gunten, U. and Escher, B.I.: Mixture effects of disinfection by-products with diverse reactive modes of action can be predicted by concentration addition).
WP3) To facilitate a subsequent chemical identification of toxicologically relevant DBPs we selected a fractionation procedure to separate toxic compounds. We decided to focus on HPLC as most promising fractionation method. Based on an extensive literature research we selected the HPLC column Kinetex® EVO C18 from Phenomenex. We fractionated a representative tap water sample by use of a methanol gradient from 5–100% in an aqueous phase over 30 minutes and collected fractions by time every 2 minutes. The resulting UV chromatogram (270 nm) of the fractionation experiments indicated that organic compounds are spread across several fractions. Adsorbable organic halides (AOX), quantified in all fractions, indicated that halogenated organic compounds are collected mainly in fractions six to 10. However, we found no clear separation of toxic compounds and no clear peaks with the Microtox assay. This is possibly because a lot of different compounds of variable polarity are contributing to the total effect. With the umuC assay, as a more specific test for genotoxicity, we did not detect any significant effects in the fractions (WP6).
WP4) Point of use filters can effectively remove regulated disinfection by-products from tap water but the removal of unknown DBPs and toxicity has not been assessed before. We evaluated 10 commercial tap water filters for their efficacy to reduce adsorbable organic halides (AOX, as sum parameter of known and unknown halogenated contaminants such as DBPs), and fluoride. Biological effects were quantified in water samples enriched with solid phase extraction by use of the Microtox assay for cytotoxicity, the AREc32 assay for oxidative stress induction, and the umuC assay for genotoxicity. Six out of 11 filters effectively removed chlorinated and brominated organic halogens by >60% with reverse osmosis and one activated carbon based gravity filter being most effective (>94%). Fluoride was removed by >83% only by reverse osmosis and two filters specifically designed for fluoride removal by use of activated alumina. Seven out of 11 filters reduced cytotoxicity, oxidative stress induction, and genotoxicity by >60%. Activated carbon based tap water filters could provide an important public health benefit through removal of halogenated organic compounds, such as disinfection by-products but regular filter cartridge exchange is critical to maintain a good filter efficacy. This study has been published in ENVIRONMENTAL SCIENCE: WATER RESEARCH & TECHNOLOGY (Stalter, D., O’Malley, E., von Gunten, U., Escher, B.I. (2016) Point-of-use water filters can effectively remove disinfection by-products and toxicity from chlorinated and chloraminated tap water. Environmental Science: Water Research & Technology, DOI: 10.1039/c6ew00068a.). Furthermore, another study about DBP mitigation strategies was recently published in WATER RESEARCH to assess the effect of ozone versus hydroxyl radical reactions on the formation of DBPs (De Vera, G.A. Stalter, D., Gernjak, W., Keller, J. and Farré , M. (2015) Towards reducing DBP formation potential of drinking water by favouring direct ozone over hydroxyl radical reactions during ozonation. Water Research 87, 49-58.). A study about the removal of precursors for DBP formation during drinking water production has been published in JOURNAL OF ENVIRONMENTAL ENGINEERING (Farre, M.J. Lyon, B.A. de Vera, G.A. Stalter, D. and Gernjak, W. (2016) Assessing adsorbable organic halogen formation and precursor removal during drinking water production. Journal of Environmental Engineering 142(3), 04015087.). Additionally, a case-study is currently running with the main objective to assess the fraction of unknown toxicity which cannot be explained by the known DBPs and how this is affected by different disinfection methods.
WP5) WP5 is a tool for WP1, WP3, and WP4.
WP6) For WP6 we selected a representative tap water sample from Brisbane, Australia, that has been characterized in previous studies. We extracted and fractionated the extract with the method described in WP3. Next, we tested the fractions with different bioassays and analyzed the AOX concentrations in the fractions as well as in the initial extract. With the bioassays we could not detect any effects in the fractions while the initial extract was quite toxic. A likely explanation is that effects in whole extracts are not caused by a few very potent compounds but by a multitude of different DBPs which act together in a mixture but get distributed across different fractions and thus diluted below the limit of detection of the bioassay. Another possible explanation for the absence of effect is that reactive DBPs degrade over the course of the sample preparation, in particular during fractionation and blow down of the fractions before analysis. Therefore, this effect directed analysis (EDA) methods does not seem to be appropriate for the identification of toxicologically relevant DBPs. In follow-up studies size exclusion chromatography could be applied as alternative fractionation method. Results of this work package are currently prepared for publication.