RESEARCH ARTICLE

Prioritisation of organic contaminants in a river basinusing chemical analyses and bioassays

Tvrtko Smital & Senka Terzić & Jovica Lončar &

Ivan Senta & Roko Žaja & Marta Popović & Iva Mikac &

Knut-Erik Tollefsen & Kevin V. Thomas & Marijan Ahel

Received: 4 April 2012 /Accepted: 25 June 2012# Springer-Verlag 2012

Abstract Region-specific contaminant prioritisation is animportant prerequisite for sustainable and cost-effectivemonitoring due to the high number of different contami-nants that may be present. Surface water and sedimentsamples from the Sava River, Croatia, were collected at fourlocations covering a 150-km-long river section characterisedby well-defined pollution gradients. Analysis of contami-nant profiles along the pollution gradients was performed bycombining toxicity screening using a battery of small-scaleor in vitro bioassays, which covered different modes ofaction, with detailed chemical characterisation based ongas chromatography/mass spectrometry (GC/MS) and liquidchromatography/quadrupole time-of-flight mass spectrome-try (LC-QTOF-MS). A large number of contaminants, be-longing to different toxicant classes, were identified in both

analysed matrices. Analyses of water samples showed thatcontaminants having polar character occurred in the highestconcentrations, while in sediments, contributions from bothnon-polar and amphiphilic contaminants should be takeninto account. Estimated contributions of individual contam-inant classes to the overall toxicity indicated that, besidesthe classical pollutants, a number of emerging contaminants,including surfactants, pharmaceuticals, personal care prod-ucts and plasticizers, should be taken into consideration infuture monitoring activities. This work demonstrates theimportance of the integrated chemical and bioanalyticalapproach for a systematic region-specific pollutant prioriti-sation. Finally, the results presented in this study confirmthat hazard assessment in complex environmental matricesshould be directed towards identification of key pollutants,rather than focusing on a priori selected contaminants alone.

Keywords River basin . Priority contaminants . Integratedassessment . Bioassays . GC/MS . LC/MS

Introduction

The European Union Water Framework Directive (EU WFD2000) requires good ecological status in European riverbasins to be achieved by 2015. One of the important stepstowards a realistic and cost–benefit implementation of theWFD should be the identification of the most hazardousenvironmental contaminants. Currently, the list of prioritysubstances contains only 33 contaminants and/or contaminantgroups, which is only the small fraction of a much largernumber of possible hazardous pollutants (Schwarzenbach etal. 2006) that eventually reach aquatic systems. Since target-analyses of classical priority pollutants often fail to explain theobserved (eco)toxicological effects of complex environmental

Responsible editor: Philippe Garrigues

Electronic supplementary material The online version of this article(doi:10.1007/s11356-012-1059-x) contains supplementary material,which is available to authorized users.

T. Smital : S. Terzić : J. Lončar : I. Senta : R. Žaja :M. Popović :I. Mikac :M. Ahel (*)Division for Marine and Environmental Research,Rudjer Boskovic Institute,Bijenicka 54,10000 Zagreb, Croatiae-mail: ahel@irb.hr

K.-E. Tollefsen :K. V. ThomasNorwegian Institute for Water Research (NIVA),Gaustadalléen 21,0349 Oslo, Norway

K.-E. TollefsenDepartment of Plant and Environmental Sciences (IPM),University of Life Sciences (UMB),Post box 5003, 1432 Ås, Norway

Environ Sci Pollut ResDOI 10.1007/s11356-012-1059-x

samples (Brack 2003), a combination of small-scale biologicalanalyses and comprehensive chemical analytical methodshas been recently proposed as a promising tool in theecological risk assessment. This approach has been success-fully demonstrated with various environmental matrices(Thomas et al. 2001, 2002; Brack et al. 2007; Grung etal. 2007).

Implementing the EU WFD will be a demanding task formany of the transition countries in the region of Central andSoutheast Europe (CSE). The key environmental problem,which is common for all transition countries in the Sava andDanube River basins, is the release of contaminated untreat-ed effluents from municipalities and industrial facilitiesconsiderably dominated by old and environmentally un-friendly technologies (Kastelan-Macan et al. 2007).Hazardous chemical contamination represents not only amajor threat to the good environmental status of naturalwaters as required by EU WFD, but it also puts in jeopardythe drinking water supply for the riparian cities (Ahel 1991).Despite such an unfavourable situation, the monitoring ac-tivities and identification capabilities in most of thecountries in the region are often restricted to a very limitednumber of potentially hazardous contaminants. As a conse-quence, the occurrence of potentially toxic substances otherthan priority pollutants at the sites exposed to chemicalcontamination is frequently overlooked and the potentialhazard posed to the environment and human health widelyunderestimated. Recent studies (Terzic et al. 2005; Terzic atal. 2008; Smital et al. 2011) have demonstrated the wide-spread occurrence of a number of emerging contaminants inmunicipal wastewaters of the region, including some prom-inent classes of pollutants such as pharmaceuticals andpersonal care products, surfactants and their degradationproducts, plasticizers, pesticides, insect repellents, andflame retardants. Furthermore, a multidisciplinary study onhepatic biomarker responses to organic contaminants in theSava River by Krca et al. (2007), which focused on the samearea as the present study, showed that classical contaminantssuch as polyaromatic carbons (PAHs) and polychlorinatedbiphenyls (PCBs) show conspicuous gradients of contami-nant concentrations in feral fish. However, the measuredwater concentrations of these contaminants poorly explainthe observed responses in hepatic biomarkers, suggestingpossible contribution of other contaminant classes. Recenttoxicological characterisation of surface water and sedi-ments in the area, suggest that the total toxic potency arevariable and several hotspots were identified (Kallqvist et al.2008; Radic et al. 2010).

There have been several attempts to propose schemes foran efficient prioritisation of contaminants to be monitored.In their report on the nationwide reconnaissance of waste-water contaminants in the USA, Kolpin et al. (2002) putsignificant emphasis on the frequency of occurrence. A

more advanced approach was developed by Reemtsma etal. (2006) who proposed a water cycle spreading index(WCSI) to predict the concentrations of individual waste-water contaminants in receiving surface waters. Goetz et al.(2010) proposed a simple exposure-based methodologybased on publicly available data, which allowed categorisa-tion of a broad range of contaminants in surface waters.Most recently, a systematic risk assessment approach hasbeen demonstrated for the prioritisation of 500 classical andemerging contaminants in four European river basins (vonder Ohe et al. 2011).

In recent years, there has been a growing awareness thatthe combination of biological and chemical methods shouldbe regarded as an indispensable part of prioritisation proce-dures for contaminants (Brack et al. 2007). Along theselines, the work presented in this manuscript aims not onlyat providing the first comprehensive assessment of the haz-ardous chemical contamination in the Sava River watershed,but it demonstrates the importance of combining advancedchemical and bioanalytical tools for a systematic region-specific pollutant prioritisation.

Materials and methods

Study area and sampling

A 150-km section of the Sava River, starting at theCroatian–Slovenian border, was selected for this study dueto the well-defined gradient of pollution, ranging from low-to-moderately polluted sites upstream of the city of Zagreb(1 million inhabitants, heavily industrialized), to the siteslocated downstream from the Zagreb and Sisak city areas,which are characterized by the enhanced loads of variousinorganic and organic contaminants (Dragun et al. 2008;Radic et al. 2010).

A total of four sampling sites were selected: SamoborskiOtok, situated 10 km upstream of the city of Zagreb, close tothe Slovenian border; the Oborovo site, situated 15 kmdownstream of the main wastewater effluent outlet of thecity of Zagreb, representing the mixing point of that input;the Crnac site 2 km downstream from the city of Sisak(50,000 inhabitants, pesticides production facility, iron-works and oil refinery); and the Kosutarica site, 50 kmdownstream from the city of Sisak, close to the confluenceof the Una River, marks the beginning of the large trans-boundary river section, shared between Croatia and Bosniaand Herzegovina (Fig. 1).

Analysed samples included two different matrices: sur-face water and river sediment, collected at the four samplinglocations, and secondary effluent (SE) sample from thewastewater treatment plant (WWTP) of the city of Zagreb.Samples of river water (20 l) were collected in solvent-

Environ Sci Pollut Res

rinsed stainless steel containers from a 0.5-m depth.Approximately 1 kg of surface sediment (top 5–10 cm)sample per site was collected using a stainless steel spatula.A 24-h composite SE sample was collected using automaticsampling devices. All samples were collected in June andJuly 2008.

Sample preparation and extraction

Water samples

Aqueous samples (20 l river water and 10 l biologicallytreated wastewater) were transported to the laboratory,stored at 4 °C and extracted within 24 h using a large-volume solid-phase extraction (SPE) system similar to thatused by Thomas et al. (2001). Briefly, unfiltered watersamples were passed through a glass wool filter to trapparticulate matter followed by a preconditioned 6 g Oasishydrophilic–lipophilic balance (HLB) cartridge (Waters,Milford, MA, USA) at a flow rate of approximately 60–80 ml/min with aid of compressed high-purity N2. After

percolation, the Oasis cartridges were washed with distilledwater, dried using N2 and eluted with methanol (2×100 ml;HPLC grade, Merck, Germany). The eluates were reducedin volume using TurboVap evaporation (Caliper LifeSciences, Hopkinton, MA, USA) under N2 and divided intotwo identical aliquots for chemical and biological analyses,respectively, and stored at 4 °C until further analysis usingmethanol as a keeper solvent.

Sediments

Sediment samples were transferred to the laboratory inZagreb within 5 h, air-dried at the room temperature andpulverized using a mechanical mill. Ground sediment frac-tions having a particle size less than 63 μm were isolated bydry sieving and stored at 4 °C until extraction. Sub-samplesof about 40 g dry sediment were extracted by acceleratedsolvent extraction (ASE) subsequently with methylene chlo-ride and methanol (2,000 lb/in2 pressure, 100 °C) using aDionex ASE 200 (Dionex Corp., Sunnyvale, CA, USA).The resulting extracts were reduced to a small volume under

Sava

Sava

ZAGREB

Kupa

Sava

Una

ZAGREB

OS

OB

SC

KS

150 km

SLOVENIA

BOSNIAAND

HERCEGOVINA

HUNGARY

AUSTRIA

SERBIA

Sisak

Fig. 1 Sava River basin mapwith the expanded area ofinvestigation and samplinglocations in the study. OS OtokSamoborski, OB Oborovo, SCSisak Crnac, KS Kosutarica

Environ Sci Pollut Res

nitrogen using a TurboVap system set at 40 °C and trans-ferred into 4-ml screw-cap vials.

In all subsequent steps, the water and sediment extractswere treated identically. The concentrated extracts weredivided into two identical aliquots for chemical and biolog-ical analyses and stored separately.

Extract fractionation

The extract aliquots, used for chemical characterisation,were subjected to an additional separation step using silicagel deactivated with 15 % water (Smital et al. 2011). Briefly,the total extract was applied to the top of the silica gelcolumn (5 ml) and subsequently eluted with 25 ml n-hex-ane, 30 ml dichloromethane and 30 ml methanol to yieldnon-polar (A), medium-polar (B) and polar fractions (C),respectively. Each of the fractions was reduced in volume byevaporation under the N2 stream and transferred into 1.8-mlvials fitted with Teflon-lined screw caps.

Chemical analyses

The applied analytical approach was designed to provide adetailed characterisation of the collected samples, covering awide range of possible non-target and target contaminants(Terzic et al. 2009). The hexane and dichloromethane silicagel fractions were screened for identifiable compounds us-ing gas chromatography/mass spectrometry (GC/MS), whilethe dichloromethane and methanol fractions were analysedby liquid chromatography/time-of-flight mass spectrometry(LC/QTOF), thus providing a basis for identification andadditional confirmation of the more polar compounds. Thetarget-analyses were applied for selected characteristic clas-ses of organic contaminants such as polycyclic aromatichydrocarbons, PAH (Krca et al. 2007) and pharmaceuticals(Senta et al. 2008). The two classes were selected taking intoaccount two major specific industrial sources of organiccontaminants in the area: the oil refinery in the area ofSisak and the pharmaceutical industry in the area ofZagreb. Details on the comprehensive screening of environ-mental extracts using GC/MS and LC/QTOF MS techniquesare described elsewhere (Terzic et al. 2009; Smital et al.2011; Terzic and Ahel 2011) and can be found inSupplementary material to this paper.

Bioassays

The total extracts of water and sediment samples were sub-jected to toxicity screening in a series of small scale or invitro bioassays designed to characterise the biological re-sponse of hazardous contaminants with different modes ofaction. The bioassays included determinations of cytotoxic-ity using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-

diphenyltetrazolium bromide) reduction assay, ethoxyresor-ufin O-deethylase (EROD) activity; inhibition of the multi-xenobiotic resistance (MXR), genotoxicicity and estrogenicpotential. Details on the applied bioassay procedures aredescribed elsewhere (Smital et al. 2011) and can be foundin Supplementary material to this paper.

Results and discussion

Bioassays

A battery of ecotoxicological in vitro bioassays that encom-passed various toxicity end-points was applied for the as-sessment of the collected water and sediment samples. Thecytotoxicity of the Sava River water extracts to PLHC-1cells was very low at all locations studied and no significantdifferences between the individual sampling stations wereobserved (Fig. 2). SE collected from the Zagreb WWTPshowed much higher acute toxicity as measured by theMTT assay, although the impact of this major source onthe toxicity of the Sava River water was not pronounced. Incontrast, a significant cytotoxicity was detected in all sedi-ment samples, in particular those collected at Kosutarica andSisak-Crnac (Fig. 2). Such a distribution of cytotoxicity is inagreement with the data from bioassay-assisted monitoringof the Sava Rive r us ing the f re shwa te r a lgaePseudokirchneriella subcapitata (Kallqvist et al. 2008) andindicates that the effects may be related to industrial efflu-ents from the Sisak area, in particular effluents originatingfrom oil refinery activities. It should be stressed that theinformation about cytotoxicity is important for the interpre-tation of the results of other bioassays since some specificend-point responses can be partly masked due to the cyto-toxicity at higher extract doses.

The distribution of EROD induction potential is generallyin agreement with the distribution of cytotoxity. Asexpected, a significantly enhanced EROD activity was de-termined in the SE sample. However, all examined riverwater samples were characterised by low EROD inductionpotential with only slightly increased activity at theOborovo location downstream of the main Zagreb citywastewater outlet (Fig. 3). In contrast, high EROD inductionpotential was determined in the sediment samples, in partic-ular at the locations Sisak-Crnac and Kosutarica (Fig. 3).This probably reflected an additional input of CYP1Ainducers such as multi-ring PAHs from the oil refinery,situated downstream of the city of Sisak.

The distribution of MXR inhibitors in the Sava River(Fig. 4) was significantly different from the distribution ofEROD/CYP1A inducers, indicating location-specific differ-ences in compounds causing the bioassay responses thatinhibit MXR. The results revealed that these contaminants

Environ Sci Pollut Res

were primarily associated with the aqueous phase, whiletheir concentrations in analysed sediments were rather low.The highest MXR inhibitory effect was observed in thesamples collected from Sisak Crnac and Kosutarica wherethe inhibitory effect was comparable to 3–5 μM concentra-tions of the model MXR inhibitor cyclosporine A. It isinteresting to note that the SE from Zagreb WWTP did notshow an enhanced inhibitory effect as compared to the SavaRiver water. However, it should be pointed out, that theMXR inhibitory potential of this sample was probablyunderestimated due to cytotoxic effects at higher extractconcentrations (Fig. 2). This was further verified by testinga dilution series of this extract.

The estrogenic potential of both surface water andsediment samples suggested rather modest presence of(xeno)estrogens in the Sava River (estrogenic equivalent[EEQ] <1 ng/l and <1 ng/g, respectively; Fig. 5) withno pronounced spatial distribution dynamics, despite thefact that SE of Zagreb WWTP represented a distinctinput of estrogenic substances (EEQ, 3 ng/l). This situ-ation is a consequence of a large dilution factor of SEin the river water (>50).

The mutagenic/genotoxic potential of the Sava Riversamples was generally very low (Table S1). Weak but sig-nificant mutagenic potential was only observed for the sur-face water sample collected at Kosutarica and the SE sample

0

10

20

30

40

50

60

70

80

90

100

110

120

OS SE OB SC KS

locations

cell

viab

lity

(%)

SE and surface water samples

sediment samples

Fig. 2 Cytotoxicity of the SavaRiver water and sedimentsamples (OS Otok Samoborski,OB Oborovo, SC Sisak Crnac,KS Kosutarica) vs. cytotoxicityof secondary effluent (SE) ofthe city of Zagreb as determinedby the MTT test in PLHC-1cells. Data represent meanswith corresponding SDs fromtriplicate determinations

0.0

0.5

1.0

1.5

2.0

2.5

OS SE OB SC KS

locations

FEL-

TEQ

(ng/

L or

ng/

g)

SE and surface water samples

sediment samples

Fig. 3 CYP1A inductionpotential of the Sava Riverwater and sediment samples(OS Otok Samoborski, OBOborovo, SC Sisak Crnac, KSKosutarica) vs. cytotoxicity ofsecondary effluent (SE) of thecity of Zagreb as determined bythe measurement ofethoxyresorufin O-deethylase(EROD) activity in PLHC-1cells. Data are fixed-effect-leveltoxic equivalents (FEL-TEQ;ng/l or ng/g) determined intriplicate (means ± SDs arepresented)

Environ Sci Pollut Res

from Zagreb city WWTP. No mutagenicity was observed forany of the Sava River sediment samples (not shown), whichwas due to the same reasons as indicated for the estrogenicpotential, i.e., due to the low mutagenic potential in SE asthe main input as well as a large dilution factor (>50) of thewastewater effluents.

Chemical screening

Water and sediment extracts from the Sava River containedcomplex assemblages of contaminants. The identified com-pounds listed in Table 1 are grouped in several contaminantclasses and sorted according to increasing polarities (adetailed list of identified contaminants can be found in

Table S2). They cover a wide range of chemical struc-tures and physicochemical properties, ranging from non-polar and hydrophobic compounds, such as petroleumhydrocarbons and PAHs, to the polar and amphiphiliccompounds, including pharmaceuticals and surfactants.The semi-quantitative estimates of concentration areshown in Table 1 and clearly illustrate a large variabil-ity of the concentration ranges among the individualcontaminants. Compared to some previous studies(Thomas et al. 2002; Biselli et al. 2005), a significantadvantage of the approach used in this study is inidentification of an extended range of contaminants,which was achieved by the combination of GC/MS andUPLC/Q-TOF analyses. This allowed the identification of a

0

1

2

3

4

5

OS SE OB SC KS

locations

Ca-

AM

acc

um

ula

tio

n (

fold

incr

ease

ove

r co

ntr

ol)

SE and surface water samples

sediment samples

Fig. 4 MXR inhibitorypotential of the Sava Riversurface water samples,sediment samples (OS OtokSamoborski, OB Oborovo, SCSisak Crnac, KS Kosutarica)and the secondary effluentsample (SE) of the city ofZagreb as determined by theincrease in calcein-AMaccumulation in PLHC-1 cellsexposed to correspondingextracts. Data are expressed asfold increase in fluorescenceover the control value (cellsexposed to calcein-AM only)and represent means withcorresponding SDs fromtriplicate determinations

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

OS SE OB SC KSlocations

EE

Q (n

g/L

or n

g/g)

SE and surface water samples

sediment samples

Fig. 5 Estrogenic potential ofthe Sava River surface watersamples, sediment samples(OS Otok Samoborski, OBOborovo, SC Sisak Crnac, KSKosutarica) and the secondaryeffluent sample (SE) of the cityof Zagreb as determined by theYES test. Data are expressed in17β-estradiol (E2) estrogenicequivalents (EEQ, ng/l or ng/g)and represent means withcorresponding SDs fromtriplicate determinations

Environ Sci Pollut Res

number of polar contaminants, some of whichwere among themost abundant compound classes identified.

It is very difficult to assess the possible contributions ofindividual contaminant classes to the measured biological

Table 1 Main groups of specific organic compounds identified in the Sava River water and sediments in relation to their composition in thesecondary effluent of the WWTP of the city of Zagreb (sampling July 2008)

Compound WWTPsecondaryeffluent

Location

OS OB SC KS

Riverwater

Riversediments

Riverwater

Riversediments

Riverwater

Riversediments

Riverwater

Riversediments

Nonpolar compounds — hydrocarbons

UCM (mainly petroleum-derivedhydrocarbons)

+++++ ++++ +++++ ++++ +++++ ++++ +++++ ++++ +++++

Linear alkylbenzenes C10–C14 (LAB) +++ ++ +++ +++ +++ +++ +++ +++ +++

Diisopropylnaphthalenes +++ ++ +++ ++ +++ ++ +++ ++ +++

Polyaromatic hydrocarbons (PAH); 2–6 rings ++ ++ ++++ ++ ++++ ++ +++++ ++ +++++

Polychlorinated biphenyls (PCBs) + + ++ + ++ + ++ + +

Diverse medium polar compounds

Alkylphenols (C1–C9) +++ ++ ++ ++ ++ ++ ++ ++ ++

Phenoxyethanol +++ ++ +++ ++ +++ ++ +++ ++ +++

3-Hydroxy-4-methoxybenzaldehyde ++ ++ +++ ++ ++ ++ +++ ++ +++

Indole ++ + ++ + ++ + ++ + ++

Skatole ++ + ++ + ++ + ++ + ++

Coumarine ++ ++ + ++ + ++ + ++ +

Parabens +++ + + + + + + + +

Benzophenone +++ +++ +++ +++ +++ ++ ++ ++ ++

Benzotriazoles +++ + ++ +++ ++ ++ ++ ++ ++

Sterols ++++ ++++ ++++ ++++ +++++ ++++ +++++ ++++ +++++

Polycyclic musk fragrances ++++ ++ +++ ++ +++ ++ +++ ++ +++

Methyl esters of common fatty acids +++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++

Phthalates +++++ +++ +++++ ++++ +++++ ++++ +++++ ++++ +++++

Diethyltoluamide ++++ ++ + ++ + ++ + ++ +

Pesticides

Triazine herbicides ++++ ++ + ++ + ++ + ++ +

Chloroacetanilide herbicides ++++ ++ + ++ + ++ + ++ +

Pharmaceutical compounds

Propyphenazone ++++ + + + + + + + +

Caffeine +++ ++ + ++ + ++ + ++ +

Nicotine ++ + + + + + + + +

Sulfonamides +++ ++ + ++ + ++ + ++ +

Fluoroquinolones ++++ ++ +++ ++ +++ ++ +++ ++ +++

Macrolides ++++ ++ +++ +++ ++++ +++ ++++ +++ ++++

Surfactants

Linearalkylbenzene sulfonates (LAS) +++++ +++++ ++++ +++++ +++++ +++++ ++++ +++++ ++++

Linear alcohol polyethoxylates (LAEO) +++++ +++++ +++++ +++++ +++++ +++++ +++++ +++++ +++++

Alkylphenol polyethoxylates (APEO) +++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++

Polyethylene glycols (PEG) +++++ +++++ ++++ +++++ ++++ +++++ ++++ +++++ ++++

River water and wastewater samples: +, <0.01 μg/l; ++, 0.01–0.1 μg/l; +++, 0.1–1 μg/l; ++++, 1–10 μg/l; +++++, >10 μg/l; sedimentsamples: +, <1 ng/g; ++, 1–10 ng/g; +++, 10–100 ng/g; ++++, 100–1,000 ng/g; +++++, >1,000 ng/g

nd not determined, OS Otok Samoborski, OB Oborovo, SC Sisak–Crnac, KS Kosutarica

Environ Sci Pollut Res

endpoints in a fully quantitative manner, in particular bear-ing in mind that all bioassays were performed on raw(unfractionated) extracts that contained a wide variety ofchemical compounds having different modes of action(Table 1). However, the results obtained in this studyallowed us to indicate specific classes of compounds beingof higher concern in the Sava River basin.

EROD activity correlated well with the distribution ofPAH in sediments. This is in accordance with literaturereports which showed that, PAHs or related planar hydro-phobic contaminants, such as PCBs or dioxins, are typicalcandidates for the enhanced EROD activity in aquatic sedi-ments (Brack et al. 2000; Thomas et al. 2002). The PAHconcentration in the Sava River water was very low and theobserved EROD activity should primarily be assigned tosome other, more polar compound classes (Smital et al.2011). Possible candidates contributing to EROD activity,detected in the water samples, include chloroacetanilideherbicides, acetochlor and metolachlor (Dierickx 1999),ubiquitous municipal wastewater contaminant nicotine(Wei et al. 2002) and the antimicrobial sulfamethoxazole(Laville et al. 2004). However, these medium polar contam-inant classes were only sparsely present in sediments, andtherefore cannot be regarded as significant contributors tothe overall EROD activity.

The measurements of estrogenicity in the Sava Riversamples were in good agreement with a relatively lowestrogenic potential of the SE of the Zagreb WWTP (EEQ,3 ng/l) (Smital et al. 2011). An earlier study of the sameWWTP (Grung et al. 2007) revealed a predominant contri-bution of natural estrogenic hormones such as 17β-estradioland progesterone, which belong to readily biodegradablecompounds. Moreover, Labadie and Hill (2007) have shownthat it is unlikely that these compounds would significantlyaccumulate in river sediments. Typical xenobiotic candi-dates for the estrogenic activity, detected in all analysedmatrices, were alkylphenols (Tollefsen 2007; Tollefsen etal. 2008), benzophenone (Kunz et al. 2006) and phthalates(Harris et al. 1997). Another specific contribution to estro-genic activity could have been from the natural compoundslike phytosterols (Tremblay and Van der Kraak 1998).Despite the fact that some of these compounds (sterols andphthalates) were relatively abundant in Sava River sedi-ments (Table 1), all investigated sediment samples showeda low estrogenic activity (<1 ng/g EEQ).

As already indicated by the relationship between theMXR inhibitory potential of water and sediment extracts(Fig. 3), this end-point was mainly affected by water-solublecompounds. This observation is in agreement with the dis-tribution of MXR inhibition potential in different polarityfractions of the wastewater effluents of the city of Zagreb(Smital et al. 2011). As to the possible chemical candidatesfor the observed MXR inhibition, various compound

classes, including pharmaceuticals, were shown to interferewith the cellular detoxification by active transport over thecell wall (Seelig et al. 2005). However, recent reports haveshown that the typical concentrations of pharmaceuticals inCroatian municipal wastewaters rarely exceeded low micro-grams per litre levels (Terzic et al. 2008), while the onlypharmaceutical class that occurred in the Sava River atelevated concentrations (100–1,000 ng/l) were macrolideantibiotics (Senta 2009). The testing of macrolides in ourlaboratory showed that they cannot be considered strongMXR inhibitors (unpublished data). Therefore, the observedMXR inhibitory effects are either derived from some othercompounds or, more likely, they actually represented aneffect caused by a complex mixture of various polar con-taminants (Smital et al. 2011).

Prioritisation of contaminants using combinationof chemical analyses and bioassays

Most of the compounds detected in the analysed water andsediment samples from the Sava River, cannot be associatedwith the specific endpoints tested. It is reasonable to assumethat non-specific biotests such as cytotoxicity determina-tions are related to the most abundant compounds classesfound in the samples, including PAHs, phthalates, sterolsand surfactants. Except for PAHs, the other groups of prom-inent chemicals are not highly toxic. Surfactants are, gener-ally, only moderately toxic to aquatic life (Ying 2006), butthey should not be neglected when assessing the overalltoxic potential since their concentrations in the river waterare often 1,000 times higher than the concentrations of theclassical hydrophobic contaminants. The observed ratios ofmeasured environmental concentrations (MECs) and pre-dicted no effect concentrations (PNEC) for moderately toxicchemicals can often be higher than the corresponding ratiosof the classical pollutants, indicating that less toxic contam-inants may well be responsible for the observed adverseeffects. Indeed, our preliminary risk assessment data indi-cate that this scenario might be correct for the Sava River.As shown in Table 2, the risk quotients (RQs) calculated forselected organic contaminants identified in this study clearlyindicate that, besides PAHs, linear alkylbenzenesulfonates(LAS), cationic surfactants and alkylphenol polyethoxylates(APEO) may represent the greatest risks for aquatic organ-isms in the Sava River. The PNEC values in Table 2 werecalculated conservatively, i.e., dividing LC50 values forDaphnia magna with a safety factor of 1,000. Moreover,recent in vitro studies have shown that surfactants canenhance the toxicity of other contaminants (Harris et al.2009), most probably because surfactants may enhance thetransfer of contaminants through the cell membrane(Hellenius and Simons 1975) or enhance their bioavailabil-ity (Kile and Chiou 1989). However the environmental

Environ Sci Pollut Res

Table 2 Preliminary risk assessment of selected organic contaminants determined in the Sava River

Substance Source LC50 (mg/L)

PNEC (µg/L)

MEC-water

(µg/L)RQ

EQS-water (µg/L)

EU WFD compliance

Phenanthrene Pyrolytic 0.04 0.04 0.005 0.13 na na

Anthracene Pyrolytic 0.035 0.04 0.001 0.029 0.4 Yes

Pyrene Pyrolytic 0.2 0.20 0.005 0.025 na na

Chrysene Pyrolytic 1 1.00 0.005 0.005 na na

Benzo(k)fluoranthene Pyrolytic 0.00065 0.0007 0.001 1.5 0.03 Yes

Benzo(a)pyrene Pyrolytic 0.006 0.01 0.001 0.17 0.05 Yes

Linearalkylbenzene sulfonates (LAS) Surfactants 7 7 20 2.8 na na

Benzalkonium cationic surfactants Surfactants 0.45 0.45 2 4.4 na na

Alkylphenol polyethoxylates (APEO) Surfactants 15 15 5 0.33 na na

Polyethylene glycols (PEG) Surfactants 1000 1000 20 0.020 na na

Polypropylene glycols (PPG) Surfactants 1700 1700 5 0.003 na na

Palmitic acid Natural, Soap 150 150 5 0.033 na na

Diethylhexylphthalate Plasticizer 11 11 5 0.46 1.3 No

Phenoxyethanol Ind. Solvent 290 290 0.05 0.000 na na

Metolachlor Pesticide 2 2 0.05 0.025 na na

Terbuthylazine Pesticide 4 4 0.05 0.013 na na

Atrazine Pesticide 4.5 5 0.05 0.011 0.6 Yes

Diethyltoluamide Repellent 75 75 0.05 0.001 na na

Benzophenone UV-filter 0.28 0.28 0.05 0.18 na na

Galaxolide Fragrance 0.14 0.14 0.05 0.36 na na

4-Nonylphenol Biotrans. APEO 0.13 0.13 0.005 0.038 2 Yes

Sulfamethoxazole Pharmaceutical 177 177 0.05 0.0003 na na

Erythromycin Pharmaceutical 22 22 0.05 0.002 na na

LC50 values were obtained from the Tox Net Data base and are related to acute toxicity tests with either Daphnia sp. or fish species. PNEC valueswere calculated from LC50 values using an assessment factor (AF) of 1,000. MEC are best estimates taken from the Table S2. Risk quotients (RQs)were defined as MEC/PNEC

Color coding of RQ values of individual contaminants: red highly important, orange very important, yellow important, green moderate, blue weak,pale blue not important. Compliance with EUWFD was assessed from the EQS/MEC ratio. EQS values are taken from the Annex I of the Directive2000/60/EC. na not applicable (substances not on the EU WFD priority list)

Environ Sci Pollut Res

significance of this mechanism needs to be confirmed.Along these lines, it is also interesting to note that surfac-tants were the most abundant contaminants in the SavaRiver sediments (Table 1). Obviously, their hydrophobicmoieties allow an efficient adsorption onto river sediments,which is in agreement with other reports (Tabor and Barber1996). All of these aspects warrant the careful monitoringof surfactant contaminants in order to assess the overallindices of water quality. Apart from surfactants, compar-atively high RQs were obtained for the personal careproducts benzophenone and galaxolide, indicating thatmunicipal wastewater is a major source for discharge ofpollutants to the Sava River. A high RQ was alsoobtained for the environmentally ubiquitous plasticizerdiethylhexylphthalate, which even exceeded the EUWFD recommended maximum allowable concentrationin the present Sava River water samples.

As clearly demonstrated in numerous recent studies, in-tegrated evaluation of chemical and biological data repre-sents the ultimate approach for a reliable ecotoxicologicalhazard assessment of various classes of emerging contami-nants (e.g., Aerni et al. 2004; Blasco and Pico 2009; Wang

et al. 2012). Table 3 presents a list of the selected substan-ces, representative of different contaminant categories in theSava River, which have been ranked using three differentspecific endpoints: EROD induction potential (EROD),MXR inhibitory potential (MXR) and estrogenic potential(ER). The relative contributions of individual contaminantswere calculated by multiplying their estimated concentra-tions in the Sava River samples by the corresponding rela-tive potencies. Most of the relative potencies applied in theTable 3 were determined in our laboratory as the lowestobserved effect concentrations (LOEC) for selected contam-inants relative to the reference contaminants: TCDD forEROD (LOECTCDD00.01 nM); cyclosporine A for MXR(LOECCA030 nM); 17β-estradiol for ER (LOECE200.03nM). As can be seen, the relative contributions of the se-lected representative contaminants to the overall responsevaried across wide ranges of several orders of magnitude,which have been appropriately color-coded to indicate themost critical contaminant classes for each biological end-point. Some of the results, shown in the Table 3, were highlyexpected and confirmed the importance of some classicalcontaminants. PAHs were the strongly predominant

Table 3 Relative potencies (expressed as induction/inhibition equiva-lence factors [IEFs]) of characteristic categories of contaminants in theSava River water and sediments and their estimated relative

contribution to the selected biological endpoints — EROD inductionpotential (EROD), MXR inhibitory potential (MXR) and estrogenicpotential (ER)

SubstanceRelative potency (IEF) MEC-

water (ng/L)

Relative contribution-Sava water

MEC-sedim. (ng/g)

Relative contribution-Sava sediment

EROD MXR ER EROD MXR ER EROD MXR ER

Benzo(a)pyrene 0.033 nd 0.00003 1 0.033 nd 0.00003 50 1.65 nd 0.002

Linear alkylbenzene sulfonates (LAS) .00000010 .0 012 nd 20000 0.0020 240 nd 500 0.00005 6 nd

Benzalkonium cationic surfactants

0.0000001 0.0003 nd 2000 0.0002 0.60 nd 100 0.00001 0.03 nd

Alkylphenol polyethoxylates (APEO)

0.0000001 0.05 0.0000002 5000 0.0005 250 0.0010 500 0.00005 25 0.0001

Polyethylene glycols (PEG) 0.000002 0.06 nd 20000 0.040 1200 nd 500 0.001 30 nd

Palmitic acid 0.0000001 0.012 nd 5000 0.0005 60 nd 500 0.00005 6 nd

.0 0000001 .0 03 .0 0000002 5000 0.0005 150 0.0010 500 0.00005 15 0.0001

Metolachlor 0.0000002 nd 0.00002 50 0.00001 nd 0.0010 1 0.0000002 nd 0.00002

Atrazine .0 00001 .0 00075 .0 00003 50 0.00050 0.038 0.0015 1 0.00001 0.00075 0.00003

Benzophenone 0.0000001 nd 0.00001 50 0.000005 nd 0.00050 5 0.0000005 nd 0.00005

Galaxolide 0.0000033 0.3 0.000035 50 0.00017 15 0.0018 50 0.000165 15 0.002

4-nonylphenol nd nd 0.0003 5 nd nd 0.0015 1 nd nd 0.0003

Sulfamethoxazole 0.000000125 0.0002 nd 50 0.000006 0.010 nd 1 0.0000001 0.0002 nd

Erythromycin nd 0.00015 nd 50 nd 0.0075 nd 5 nd 0.00075 nd

Diethylhexylphthalate

Presented examples for water and sediment refer to the samples collected at Sisak Crnac and Oborovo, respectively. IEFs were calculated bycomparing the experimental lowest observed effect concentrations (LOEC) for selected contaminants relative to the reference contaminants: TCDDfor EROD (LOECTCDD00.01 nM); cyclosporine A for MXR (LOECCA030 nM); 17β-estradiol for ER (LOECE200.03 nM). Relative contributionsof individual contaminants were calculated by multiplying their measured concentrations (MEC) in the samples by the corresponding relativepotencies (IEF)

Color coding of the relative contaminant contributions: red highly important, orange very important, yellow important, green moderate, blue weak,pale blue not important

Environ Sci Pollut Res

contributors to the EROD activity, in particular in sedi-ments. Other contaminant classes showed either a week ornegligible EROD activity. A very pronounced contributionto the MXR inhibitory potential was found for varioussurfactants and galaxolide. Most of these compounds arecomparatively polar chemicals, but their contribution to theoverall activity was high also in sediments. The estrogenicactivity in both sediments and water samples was rather low.Among the potential xenoestrogens, the highest contribu-tions were estimated to be associated with galaxolide, whilenonylphenol showed relatively low contribution in the ex-amined samples. Interestingly, benzo(a)pyrene (BaP) wasfound to significantly contribute to otherwise modest estro-genicity, based on relative binding affinity for the estrogenreceptor. Although BaP and/or the hydroxylated metaboliteof BaP have been associated with ER-mediated responses invarious bioassays (Hirose et al. 2001; van Lipzig et al.2005), the mode of action of this compound has not beenclearly identified and warrants further studies.

Although it would be premature to use these data for thefully quantitative estimates, the new insights regarding pos-sible importance of the less common classes of contami-nants should be taken into account in further prioritisationactivities and ultimately for revisiting the current monitoringscheme in the Sava River basin.

Conclusion

Apart from being the first comprehensive assessment ofcontaminants in the Sava River watershed, this study clearlyemphasizes the possible importance of certain emergingclasses of organic contaminants, which are not included inthe European and national monitoring strategies. This isparticularly true for the most polar fraction, which repre-sents the least studied fraction in environmental matrices,while their potential for being bioavailable in the aquaticenvironment is rather high compared to the classical hydro-phobic pollutants (Reemtsma et al. 2006). Consequently,there is a need to include typical representatives of this classsuch as surfactants and pharmaceuticals in the future region-specific monitoring activities. The results presented in thisstudy confirm that hazard assessment in complex environ-mental matrices should be directed towards identification ofkey pollutants, rather than focusing on a priori selectedcontaminants alone. Moreover, due to the presence of highlycomplex mixtures of potentially toxic substances in waterand sediment extracts, possible synergistic or antagonisteffects should also be taken into account.

Acknowledgements The work presented here has been supportedthrough the NATO Science for Peace and Security Program, ProjectSfP 982590 – “Assessment of hazardous chemical contamination in the

Sava River basin” (http://www.irb.hr/nato-savariver/). In addition, thiswork was partially supported by the Ministry of Science, Educationand Sports of the Republic of Croatia, Project Nos: 098-0982934-2745and 098-0982934-2712 and with institutional funding from the Nor-wegian Institute for water Research (NIVA). We thank Nenad Muhinfor the technical assistance.

References

Aerni HR, Kobler B, Rutishauser BV, Wettstein FE, Fischer R, GigerW, Hungerbuhler A, Marazuela MD, Peter A, Schonenberger R,Vogeli AC, Suter MJ, Eggen IR (2004) Combined biological andchemical assessment of estrogenic activities in wastewater treat-ment plant effluents. Anal Bioanal Chem 378:688–696

Ahel M (1991) Infiltration of organic pollutants into groundwater: fieldstudies in the alluvial aquifer of the Sava River. Bull EnvironContam Toxicol 47:586–593

Biselli S, Reineke N, Heinzel N, Kammann U, Franke S, HuehnerfussH, Theobald NJ (2005) Bioassay-directed fractionation of organicextracts of marine surface sediments from the North and BalticSea — Part I: Determination and identification of organic pollu-tants. J Soils Sed 5:171–181

Blasco C, Pico Y (2009) Prospects for combining chemical and bio-logical methods for integrated environmental assessment. TrendsAnal Chem 28:745–757

Brack W (2003) Effect–directed analysis: a promising tool for theidentification of organic toxicants in complex mixtures. AnalBioanal Chem 377:397–407

Brack W, Klamer HJC, De Alda ML, Barceló D (2007) Effect-directedanalysis of key toxicants in European river basins: a review.Environ Sci Pollut R 14:30–38

Brack WH, Segner M, Moder M, Schuurmann G (2000) Fixed-effect-level toxicity equivalents - A suitable parameter for assessingethoxyresorufin-O-deethylase induction potency in complexenvironmental samples. Environ Toxicol Chem 19:2491–2501

Dierickx PJ (1999) Glutathione-dependent cytotoxicity of the chloroa-cetanilide herbicides alachlor, metolachlor, and propachlor in ratand human hepatoma-derived cultured cells. Cell Biol Toxicol15:325–332

Dragun Z, Raspor B, Roje V (2008) The labile metal concentrations inSava River water assessed by diffusive gradients in thin films.Chem Spec Bioavail 20:33–46

EU WFD (2000) Directive 2000/60/EC of the European Parliamentand of the Council of 23 October 2000 establishing a frameworkfor Community action in the field of water policy. Available at:h t tp : / / eur- lex .europa .eu /LexUr iServ /LexUr iServ.do?uri0CELEX:32000L0060:EN:NOT

Goetz CW, Stamm C, Fenner K, Singer H, Scharer M, Hollender J(2010) Targeting aquatic microcontaminants for monitoring: ex-posure categorization and application to the Swiss situation.Environ Sci Pollut 17:341–354

Grung M, Lichtenthaler R, Ahel M, Tollefsen KE, Langford K,Thomas KV (2007) Effects-directed analysis of organic toxicantsin wastewater effluent from Zagreb, Croatia. Chemosphere67:108–120

Harris CA, Henttu P, Parker MG, Sumpter JP (1997) The estrogenicactivity of phthalate esters in vitro. Environ Health Perspect105:802–811

Harris CA, Brian JV, Pojana G, Lamoree M, Booy P, Marcomini A,Sumpter JP (2009) The influence of a surfactant, linear alkylben-zene sulfonate, on the estrogenic response to a mixture of(xeno)estrogens in vitro and in vivo. Aquat Toxicol 91:95–98

Hellenius A, Simons K (1975) Solubilization of membranes by deter-gents. Biochim Biophys Acta 415:29–79

Environ Sci Pollut Res

Hirose T, Morito K, Kizu R, Toriba A, Hayakawa K, Ogawa S, InoueS, Muramatsu M, Masamune Y (2001) Estrogenic/antiestrogenicactivities of benzo[a]pyrene monohydroxy derivatives. J HealthSci 47:552–558

Kallqvist T, Milacic R, Smital T, Thomas KV, Vranes S, Tollefsen KE(2008) Chronic toxicity of the Sava River (SE Europe) sedimentsand river water to the algae Pseudokirchneriella subcapitata.Water Res 42:2146–2156

Kastelan-Macan M, Ahel M, Horvat AJM, Jabucar D, Jovancic P(2007) Water resources and waste water management in Bosniaand Herzegovina, Croatia and the State Union of Serbia andMontenegro. Water Policy 9:319–343

Kile DE, Chiou CT (1989) Water solubility enhancement of DDT andtrichlorobenzene by some surfactants below and above the criticalmicelle concentration. Environ Sci Technol 23:832–838

Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, BarberLB, Buxton HT (2002) Pharmaceuticals, hormones, and otherorganic wastewater contaminants in U.S. streams, 1999–2000: anational Rrconnaissance. Environ Sci Technol 36:1202–1211

Krca S, Zaja R, Calic V, Terzic S, Grubesic MS, Ahel M, Smital T(2007) Hepatic biomarker responses to organic contaminants inferal chub (Leuciscus cephalus) – laboratory characterization andfield study in the Sava River, Croatia. Environ Toxicol Chem26:2620–2633

Kunz PY, Galicia HF, Fent K (2006) Comparison of in vitro and invivo estrogenic activity of UV filters in fish. Toxicol Sci 90:349–361

Labadie P, Hill EM (2007) Analysis of estrogens in river sediments byliquid chromatography-electrospray ionisation mass spectrometry— comparison of tandem mass spectrometry and time-of-flightmass spectrometry. J Chromatogr A 1141:174–181

Laville N, Ait-Aissa S, Gomez E, Casellas C, Porcher JM (2004)Effects of human pharmaceuticals on cytotoxicity, EROD activityand ROS production in fish hepatocytes. Toxicology 196:41–55

Radic S, Stipanicev D, Vujcic V, Rajcic MM, Sirac S, Pevalek-KozlinaB (2010) The evaluation of surface and wastewater genotoxicityusing the Allium cepa test. Sci Total Environ 408:1228–1233

Reemtsma T, Weiss S, Mueller J, Petrovic M, Gonzalez S, Barcelo D,Ventura F, Knepper TP (2006) Polar pollutants entry into thewater cycle by municipal wastewater: a European perspective.Environ Sci Technol 40:5451–5458

Schwarzenbach RP, Escher B, Fenner K, Hofstetter TB, Johnson CA,von Gunten U, Wehrli B (2006) The challenge of micropollutantsin aquatic systems. Science 313:1072–1077

Seelig A, Gatlik-Landwojtowicz E (2005) Inhibitors of multidrugefflux transporters: their membrane and protein interactions.Mini Rev Med Chem 5:135–151

Senta I (2009) Occurrence and behaviour of sulfonamides, fluoroqui-nolones, macrolides and trimethoprim in wastewater and naturalwaters. PhD thesis, University of Zagreb, 120pp

Senta I, Terzic S, Ahel M (2008) Simultaneous determination ofsulfonamides, fluoroquinolones, macrolides and trimethoprim inwastewater and river water by LC-tandem-MS. Chromatographia68:747–758

Smital T, Terzic S, Zaja R, Senta I, Pivcevic B, Popovic M, Mikac I,Tollefsen KE, Thomas KV, Ahel M (2011) Assessment of toxi-cological profiles of the municipal wastewater effluents usingchemical methods and bioassays. Ecotoxicol Environ Saf74:844–851

Tabor CF, Barber LB (1996) Fate of linear alkylbenzene sulfonate inthe Mississippi River. Environ Sci Technol 30:161–171

Terzic S, Ahel M (2011) Nontarget analysis of polar contaminants infreshwater sediments influenced by pharmaceutical industry usingultra-high-pressure liquid chromatography – quadrupole time-of-flight mass spectrometry. Environ Pollut 159:557–566

Terzic S, Matosic M, Ahel M, Mijatovic I (2005) Elimination ofaromatic surfactants from municipal wastewaters – comparisonof conventional activated sludge treatment and membrane biolog-ical reactor. Water Sci Technol 51:447–453

Terzic S, Senta I, Ahel M, Gros M, Petrovic M, Barcelo D, Müller J,Knepper TP, Martí I, Ventura F, Jovancic P, Jabucar D (2008)Occurrence and fate of emerging wastewater contaminants inWestern Balkan Region. Sci Total Environ 399:66–77

Terzic S, Mikac I, Senta I, Ahel M (2009) Comprehensive molecularcharacterization of organic contaminants in freshwater sedimentsby gas chromatography/mass spectrometry and liquid chromatog-raphy/time-of-flight mass spectrometry. In: Li S, Wang Y, Cao F,Huang P, Zhang Y (eds) Progress in environmental science andtechnology, vol II, Part A. Science Press, Beijing, pp 397–405

Thomas KV, Hurst MR, Matthiessen P, Waldock MJ (2001)Characterization of estrogenic compounds in water samples col-lected from United Kingdom estuaries. Environ Toxicol Chem20:2165–2170

Thomas KV, Balaam J, Lavender J, Jones C (2002) Characterisation ofpotentially genotoxic compounds in sediments collected fromUnited Kingdom estuaries. Chemosphere 49:247–258

Tollefsen KE (2007) Binding of alkylphenols and alkylated non-phenolics to the rainbow trout (Oncorhynchus mykiss) plasmasex steroid-binding protein. Ecotox Environ Saf 68:40–48

Tollefsen KE, Eikvar S, Finne EF, Fogelberg O, Gregersen IK (2008)Estrogenicity of alkylphenols and alkylated non-phenolics in arainbow trout (Oncorhynchus mykiss) primary hepatocyte culture.Ecotoxicol Environ Saf 71:370–383

Tremblay L, Van der Kraak G (1998) Use of a series of homologous invitro and in vivo assays to evaluate the endocrine modulatingactions of β-sitosterol in rainbow trout. Aquat Toxicol 43:149–162

van Lipzig MMH, Vermeulen NPE, Gusinu R, Legler J, Frank H,Seidel A, Meerman JHN (2005) Formation of estrogenic metab-olites of benzo[a]pyrene and chrysene by cytochrome P450 ac-tivity and their combined and supra-maximal estrogenic activity.Environ Toxicol Pharmacol 19:41–55

von der Ohe PC, Dulio V, Slobodnik J, De Deckere E, Kuhne R, EbertRU, Ginebreda A, De Cooman W, Schuurmann G, Brack W(2011) A new risk assessment approach for the prioritization of500 classical and emerging microcontaminants as potential riverbasin specific pollutants under the European Water FrameworkDirective. Sci Total Environ 409:2064–2077

Wang L, Ying GG, Chen F, Zhang LJ, Zhao JL, Lai HJ, Chen ZF, TaoR (2012) Monitoring of selected estrogenic compounds and es-trogenic activity in surface water and sediment of the YellowRiver in China using combined chemical and biological tools.Environ Pollut 165:241–249

Wei C, Caccavale RJ, Weyand EH, Chen S, Iba MM (2002) Inductionof CYP1A1 and CYP1A2 expressions by prototypic and atypicalinducers in the human lung. Cancer Lett 178:25–36

Ying GG (2006) Fate, behavior and effects of surfactants and theirdegradation products in the environment. Environ Int 32:417–431

Environ Sci Pollut Res