S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ loca te / sc i to tenv

An investigation into the acute and chronic toxicity of elevenpharmaceuticals (and their solvents) found in wastewatereffluent on the cnidarian, Hydra attenuata

Brian Quinn⁎, François Gagné, Christian BlaiseSt-Lawrence Centre, Environment Canada, 105 McGill, Montréal, Québec, Canada H2Y 2E7

A R T I C L E I N F O

⁎ Corresponding author. Birches Lane, BlackroE-mail address: quinnbrian@gmail.com (B.

0048-9697/$ – see front matter © 2007 Elsevidoi:10.1016/j.scitotenv.2007.08.038

A B S T R A C T

Article history:Received 10 July 2007Received in revised form16 August 2007Accepted 23 August 2007Available online 10 October 2007

Summary: Pharmaceuticals previously identified in the effluent from the wastewatertreatment plant (WWTP) in Montreal discharging into the St. Lawrence river, were tested foracute and chronic toxicity using the cnidarian Hydra attenuata. Acute toxicity was based onthe established technique looking at morphological changes in the Hydra, while recentlydeveloped endpoints of feeding behaviour, attachment and growth (hydranth number) wereused to measure chronic effects. The compounds under investigation (ibuprofen, naproxen,gemfibrozil, bezafibrate, carbamazepine, sulfamethoxazole, sulfapyridine, oxytetracycline,novobiocin, trimethoprim and caffeine) were tested individually in controlled laboratoryexposures with LC50 and EC50 results calculated. All compounds tested had relatively highLC50 values with gemfibrozil, ibuprofen and naproxen having the lowest at 22.36 mg/L andEC50 values based on morphology of 1.18 to 2.62 mg/L (all concentrations are nominal). TheEC50 values based on feeding were similar to those based onmorphology but with increasedsensitivity for carbamazepine, bezafibrate and novobiocin. A trend of a reduction in feedingwith deterioration in morphology was observed in the Hydra, with the exception ofnovobiocin, where a lower than expected EC50 of 13.53 mg/L was found with no negativeeffect on morphology. Significant reductions in attachment and hydranth number wereseen at concentrations of 1 and 5mg/L for gemfibrozil and ibuprofen respectively. A toxicitythreshold (TT) of 320 μg/L was calculated for ibuprofen, only a factor of 102 or 10 higher thanthe concentration found in the effluent in the present study (1.19 μg/L) and in otherCanadian effluents studied (22 μg/L [Brun GL, Bernier M, Losier R, Doe K, Jackman P, Lee HB,Pharmaceutically active compounds in Atlantic Canadian sewage treatment plant effluentsand receiving waters and potential for environmental effects as measured by acute andchronic aquatic toxicity. Environ Toxicol Chem 2006; 25(8): 2163–2176.] respectively. UsingEU directive 93/67/EEC the pharmaceuticals under investigation can be classified as toxic(gemfibrozil, ibuprofen and naproxen), harmful (carbamazepine, bezafibrate, sulfapyridine,oxytetracycline and novobiocin) and non-toxic (sulfamethoxazole, trimethoprim andcaffeine) and their potential toxicity for the aquatic environment is discussed.

© 2007 Elsevier B.V. All rights reserved.

Keywords:Hydra attenuataPharmaceuticalsSolventsChronic effectsFeeding

ck, Dundalk, Co. Louth. Ireland. Tel.: +353 87 2860874; fax: +353 42 9323063.Quinn).

er B.V. All rights reserved.

307S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

1. Introduction

Municipal effluents and environmental pollutants have be-come more complex and as a result environmental toxicologyis moving away from the obvious acute ‘fish kill’ type ofpollution, to study more subtle effects reducing the health ofan individual or population. This is aided by the developmentof more sensitive analytical techniques, allowing the study ofmore subtle effects from a wide array of emerging pollutants.One such pollutant that has come into focus in recent years ispharmaceutical compounds, everyday drugs ingested bywestern populations that are becoming ubiquitous in theaquatic environment (Sanderson et al., 2004). Many of thepharmaceuticals applied in human medical care are notcompletely eliminated by the body, often excreted onlyslightly transformed or even unchanged (Heberer, 2002).Once excreted, the majority end up passing through awastewater treatment plant (WWTP), from where they areintroduced to surface waters with the treated effluent. Asconventional sewage treatment is generally not effective inremoving many pharmaceuticals and their metabolites, thiseffluent is considered an important and continuous source ofdrug input into the aquatic environment (Brun et al., 2006).

There are numerous original papers (Ternes, 1998; Joneset al., 2002; Metcalfe et al., 2003a, 2003b; Roberts and Thomas,2006; Lishman et al., 2006) and reviews (Heberer, 2002;Kümmerer, 2004; Brun et al., 2006; Fent et al., 2006; Hernandoet al., 2006) on the occurrence and quantities of pharmaceu-ticals found in the aquatic environment. Around 80–100pharmaceuticals and their metabolites have been measuredin both effluent and surface waters in numerous countries(Fent et al., 2006). The most commonly found drugs and theirmetabolites included anti-inflammatory drugs (ibuprofen,naproxen), lipid regulators (bexafibrate, gemfibrozil), anti-convulsion drugs (carbamazepine) and various antibiotics(trimethoprim, oxytetracycline, sulfamethoxazole) (Fentet al., 2006). These compounds are mostly detected in thehigh ng/L, low μg/L range (Heberer, 2002), butmay be sufficientto induce toxic effects (Hernando et al., 2006). Pharmaceuticalshave similar physio-chemical characteristics as harmfulxenobiotics, e.g. they can pass through membranes and arerelatively persistent (Sanderson et al., 2004). Most are specif-ically designed to affect target organs/systems in the humanbody and are designed to be persistent so as to retain theirchemical structure long enough to do their therapeutic work(Ternes and Hirsch, 2000). Generally little is known about theiraffect on non-target species in the environment. The majorityof studies looking at the biological effects of pharmaceuticalshave thus far concentrated on acute studies. As previouslyobserved by Blaise et al. (2007), there is a notable recurrence ofthe use of several well-documented and popular bioassays inthe literature e.g. V. fischeri luminescence inhibition test, P.subcapitata growth inhibition test and D. magna immobiliza-tion and reproduction tests. It has also been observed that dueto their more or less continuous presence at low concentra-tions in the aquatic environment, pharmaceuticals will mostlikely have chronic rather than acute toxic effects (Crane et al.,2006), e.g. changing behaviour that reduces individual fitnessof an organism (Jones et al., 2002). Therefore standardisedacute tests may not be the most appropriate basis for the

ecotoxicological hazard assessment of pharmaceuticals (Fer-rari et al., 2004), but currently studies on chronic effects arelacking for most pharmaceuticals (Carlsson et al., 2006).

In the present study we looked at both the acute andchronic effects of 11 compounds identified and quantified inthe (primary) treated effluent from the Montreal WWTP in aseries of small scale toxicity tests using the freshwatercnidarian Hydra attenuata. Hydra are easily cultured andmaintained in the laboratory and being diploblastic, they aresensitive environmental indicators. For these reasons, Hydrahave been extensively used in toxicity tests to study theeffects of various freshwater pollutants including effluents(Blaise and Kusui, 1997; Pardos et al., 1999), heavy metals(Holdway et al., 2001; Karntanut and Pascoe, 2002), estrogeniccompounds (Pascoe et al., 2002; Pachura-Bouchet et al., 2006)and pharmaceuticals (Pascoe et al., 2003; Blaise et al., 2007).We used the standardised toxicity test based on morphology(Wilby, 1988) and the newly developed endpoints looking atfeeding behaviour (Quinn et al., 2007), attachment and growth(hydranth number). We also discuss the hazard potential ofthese compounds by comparing our results to previouslypublished environmental risk assessments, to classify theirpotential toxicity for the aquatic environment.

2. Materials and methods

2.1. Test organism

Cultures of H. attenuata have been maintained in the Centresince the mid-nineteen nineties and used extensively forvarious toxicity tests by this and other laboratories (Blaise andKusui, 1997; Trottier et al., 1997; Pardos et al., 1999; Quinn et al.,2007). Hydra were maintained and grown in glass bowelscontaining 0.5 L of Hydra medium (110 mg/L TES [N-Tris(hydroxymethyl) methyl 1-2-aminoethanesulfonic acid],147 mg/L CaCl2·2H2O, pH 7) at 20±1 °C with a 16 h light and8 h dark photoperiod, following the procedure adapted fromTrottier et al. (1997).

2.2. Test solutions and exposure

A previous study (Gagné et al., 2006) identified a number ofpharmaceuticals found in the treated effluent from Montrealmunicipal sewage treatment works. These included thecholesterol lowering agents gemfibrozil and bezafibrate (59and 72 ng/L respectively), the analgesic ibuprofen andnaproxen (1191 and 217 ng/L respectively), the anti-epilepticcarbamazepine (33 ng/L), the antibiotics sulfapyridine, oxytet-racycline, novobiocin, sulfamethoxazole and trimethoprim(46, 440, 330, 99 and 63 ng/L respectively) and the stimulantcaffeine (22,187 ng/L). These pharmaceuticals were purchasedfrom Sigma-Aldrich (Canada), dissolved in solution andinvestigated for their lethal and sub-lethal effects at variousconcentrations between 0.1 mg/L and 50/100 mg/L. Nominalconcentrations were used throughout this experiment as aprevious study by Pascoe et al. (2003) reported the measuredconcentrations of all drug solutions (a different suite of drugs,with only ibuprofen being in common) within 10% of thenominal concentration, indicating the relative persistence of

308 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

these compounds during exposure. The solvent DMSO wasused for gemfibrozil, naproxen (0.31% final concentration),sulfapyridine and sulfamethoxazole (0.62% final concentra-tion), while acetone was used for carbamazepine (0.31% finalconcentration). Ibuprofen, bezafibrate and trimethoprim wereall dissolved in ethanol (0.31% final concentration) andoxytetracycline, novobiocin and caffeine were dissolved inHydra media. Medium and solvent (where appropriate) con-trols were used in all exposures. Sub-lethal and lethal toxicitytests for the 3 solvents (0.16 to 1.25%) were also undertaken intriplicate to ensure no solvent effects. 4 mL media containingthe relevant chemical in solution were added to 3 wells in a 12well multi-well plate as previously described (Trottier et al.,1997; Quinn et al., 2007). Healthy Hydra of similar size havingone bud (2 hydranths) spotted with a binocular microscope(×6), were used in each exposure, with 3 Hydra in each of the 3wells per exposure concentration (n=9). The multi-well plateswere covered with parafilm to prevent evaporation andrandomly placed on a work bench at room temperature (20±

Fig. 1 – Effects of various concentrations of the three solvents (DMattachment and hydranth number. Each point represents the mep≤0.01; *** = ≤0.001 for attachment and + = ≤0.05; ++ = ≤0.01 for h

2 °C) for 96 h, after which time they were observed using abinocular microscope (×6 enlargement).

2.3. Lethal and sub-lethal endpoints

Lethality and sub-lethal effects on Hydra morphology, hy-dranth number, attachment and ability to ingest prey after96 h exposure were observed. Hydra were not fed for 24 hbefore exposure. Toxicity in Hydra is measured by drasticchanges in morphology, with the contraction of tentacles andthe body, measured and scored from 10 (normal, elongatedtentacles and body) to 0 (disintegrated) on a scale devised byWilby (1988). As the amount of toxicant increases Hydraprogressively exhibit signs of toxicity, such as clubbedtentacles (score 8), shortened tentacles (score 6), a tulipphase (score 5), loss of osmoregulation (score 2) and disinte-gration (score 0). Scores 10–6 are reversible, sub-lethalindicators while the tulip phase (score 5 and below) isconsidered irreversible and used as the endpoint for lethality

SO, acetone and ethanol) on Hydra morphology (score 10–6),an score (n=3)±standard error. Significance at * = p≤0.05; ** =ydranth number.

Fig. 2 – The effect for each of the 3 solvents (DMSO, acetone and ethanol) on the ability of Hydra to feed on Artemia. Each pointrepresents the mean score (n=3)±standard error. Significance at * = p≤0.05;*** = p≤0.001.

309S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

(Blaise and Kusui, 1997). The feeding test followed thetechnique previously described by Quinn et al. (2007). Brieflyafter 96 h exposure nine Hydra were placed individually intoeach well of a 12 well multi-well plate containing 4 mL ofmedium and their morphology noted. At T0, precisely 5 Arte-mia salinawere added to each well and the number of ingestedprey observed every 30 min for 120 min.

2.4. Statistics

The 96 h LC50 and EC50 values were calculated with their 95%confidence limits using the Trimmed Spearman–Karberprogramme in the ToxTest version 2, US EPA software. Thesub-lethal LOEC (Lowest Observable Effect Concentration) wasreported for ≥2 Hydra with score 8 or below and the NOEC (NoObservable Effect Concentration) was based on b2 Hydra withscore 8, both calculated based on visual observation followingTrottier et al. (1997). A toxicity threshold (TT) was determinedfrom the LOEC and NOEC using the following equation: TT=(NOEC×LOEC)1/2 (US EPA, 1989). Variability in feeding betweenthe exposed and controlHydrawas tested by one-way analysisof variance (ANOVA) using Microsoft Excel. All exposureeffects were compared to solvent controls where appropriate.Significance was set at p≤0.05. Solvent toxicity tests wereundertaken in triplicate with mean scores presented±standard error calculated using Microsoft Excel.

Table 1 – Acute (LC50) and chronic (EC50, LOEC, NOEC) respomorphology for Hydra attenuata exposed for 96 h to the 11 pha

Pharmaceutical 96 h LC50 95% CI 96 h E

Gemfibrozil 22.36 (22.36–22.36) 1.1Ibuprofen 22.36 (22.36–22.36) 1.6Naproxen 22.36 (22.36–22.36) 2.6Carbamazepine 29.4 (32.83–26.32 15.Bezafibrate 70.71 (70.71–70.71 25.Sulfapyridine N100 NC 21.Oxytetracycline N100 NC 40.Novobiocin N100 NC NCSulfamethoxazole N100 NC NCTrimethoprim N100 NC NCCaffeine N100 NC NC

NC = not calculable, results didn't fit the appropriate test. Toxicity thresh

3. Results

3.1. Solvent exposure

Three different types of solvent were used to suspend thepharmaceuticals under study depending on their solubility. Inorder to ensure the solvents did not affect the endpointsmeasured, they were tested for effects on morphology,attachment and hydranth number (Fig. 1) and Hydra feeding(Fig. 2). At 0.31% there was a similar slight shift in morphologyfrom score 10 to 7 for all three solvents (Fig. 1). This effect isslightly increased at 0.62% with the emergence of a very lownumber of score 6 for both DMSO and ethanol indicatingrelatively weak toxicity. This concentration also sees asignificant decrease in attachment for ethanol, but nosignificant effects on hydranth number. At 1.25% there is alarge shift to score 6 for both DMSO and ethanol indicatingtoxicity, and a significant decrease in attachment andhydranth growth for both of these solvents. Acetone howeverremains relatively non-toxic at this concentrationwith a smalleffect on morphology and no significant decrease in attach-ment or hydranth growth. Results for the feeding test show nosignificant decrease in feeding at 0.31% or 0.62% for all threesolvents (Fig. 2). In fact, at 0.62% acetone was found tosignificantly increase feeding.

nses (mg/L) with 95% confidence interval (CI) based onrmaceuticals under investigation

C50 95% CI LOEC NOEC TT

8 (7.15–0.19) 1 0.1 0.325 (2.82–0.96) 1 0.1 0.322 (3.62–1.9) 5 1 2.2452 (29.02–8.3) 5 1 2.2485 NC 1 0.1 0.3261 (34.23–13.64) 5 1 2.2413 (46.95–34.3) 100 50 70.71

NC 100 50 70.71NC 10 5 7.07NC N100 N100 NCNC N100 N100 NC

old TT=(NOEC×LOEC)1/2.

Fig. 3 – Series of graphs showing the number of Hydra at each morphology state (shift from score 10 to 0) and the mean (n=9) feeding (ingestion) rate (with standard error) atvarious concentrations for each pharmaceutical individually.

310SC

IEN

CE

OF

TH

ET

OT

AL

EN

VIR

ON

MEN

T389

(2008)

306–314

311S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

The only significant reduction in feeding was found afterexposure to 1.25% DMSO, with results for both acetone andethanol similar to the control for this concentration.

3.2. Pharmaceutical exposure

3.2.1. Acute toxicity — effect on morphologyThe LC50 and EC50 values with their 95% confidence limitsbased on morphology for all eleven compounds are shown inTable 1. Gemfibrozil, ibuprofen and naproxen all show thesame LC50 value of 22.36mg/L and similar EC50 values between1.18 and 2.62 mg/L. Gemfibrozil and ibuprofen are the mosttoxic pharmaceuticals tested with a LOEC of 1 mg/L and atoxicity threshold (TT) value of 0.32 mg/L. Naproxen, carba-mazepine and bezafibrate are the next most toxic althoughbezafibrate does have a high LC50 value of 70.71mg/L but a lowTT (0.32 mg/L) due to its low LOEC and NOEC. EC50 values werealso calculated for sulfapyridine and oxytetracycline (21.61and 40.13 mg/L respectively) but LC50 could not be calculatedas less than 50% of the exposed population were killed, evenafter 96 h exposure to 100 mg/L. This was the same fornovobiocin, sulfamethoxazole, trimethoprim and caffeine,where no 96 h EC50 could be calculated. Sulfamethoxazoledid however show a relatively low TT of 7.07 mg/L due to itslow LOEC and NOEC values.

3.2.2. Chronic toxicity — effect on feeding, attachment andhydranth growthThe general trend of decreasing morphology from score 10towards 0 with increasing exposure can be seen for eachcompound in Fig. 3. This graph shows that for the five mosttoxic compounds (gemfibrozil, ibuprofen, naproxen, carbama-zepine and bezafibrate) the majority if not all of the Hydrahave a score of 1 or 0 at the highest exposure concentration. Itis also worth noting that novobiocin and caffeine have a verysmall impact on morphology with the majority of exposedHydra being unaffected (score 10) up to 100 mg/L exposure.From this graph a trend in decreasing prey ingestion withincreased toxicity is obvious for the 5 most toxic compounds.Generally as morphology shifts to score 8 or below there is adecrease in feeding. However this is not the case forsulfapyridine or oxytetracycline where despite a small de-

Table 2 – Results for the 96 h chronic toxicity test for Hydra ainvestigation showing the 96 h EC50 (mg/L) based on feeding wiwhere there was a significant (p≤0.05) reduction in feeding res

Pharmaceutical 96 h EC50 (feeding) 95% CI

Gemfibrozil 1.76 (0.9–3.42)Ibuprofen 3.85 (0.0–0.0)Naproxen 2.68 (0–1.7)Carbamazepine 3.76 (0.17–85.4)Bezafibrate 8.59 (3.78–19.51)Sulfapyridine NC NCOxytetracycline NC NCNovobiocin 13.53 (0.0–0.0)Sulfamethoxazole NC NCTrimethoprim NC NCCaffeine NC NC

NC = not calculable, results didn't fit the appropriate test. ↑ indicates inc

crease in morphology, feeding remains high. The concentra-tion of each compound having a significant decrease onfeeding is shown in Table 2. As with morphology it isgemfibrozil, ibuprofen, naproxen, carbamazepine and bezafi-brate that show a significant reduction in feeding at 10,10, 50,50 and 100 mg/L respectively. Novobiocin also shows asignificant reduction in feeding at 100 mg/L despite havingvery little effect on morphology (Fig. 3). Table 2 also shows theEC50 results (with 95% confidence limits) based on feeding.These results show a similar trend and similar sensitivity asthe EC50 based on morphology with the 5 most toxiccompounds having the lowest scores, but with increasedsensitivity for carbamazepine, bezafibrate and novobiocin.Novobiocin has a relatively low EC50 of 13.53 mg/L and adecrease in Hydra attachment at 50 mg/L. A similar trend oftoxicity can also be seen in the attachment results (Fig. 4) witha significant reduction in attachment at 1, 10, 10 and 25 mg/Lfor gemfibrozil, ibuprofen, naproxen and carbamazepinerespectively (Table 2). Interestingly caffeine significantlydecreases attachment at 25 mg/L but significantly increasesit at 50 mg/L. A significant reduction in hydranth number wasonly observed at 5, 10 and 100 mg/L for ibuprofen, naproxenand bezafibrate respectively (Table 2), showing this not to be aparticularly sensitive endpoint (Fig. 4).

4. Discussion

Although there re several routes for pharmaceuticals to enterthe aquatic environment including the application of variousdrugs in farming and aquaculture, leachate from landfill andeffluent from hospitals, WWTP effluent is known to be themost significant of these. Chemical analysis of the primarytreated effluent from Montreal WWTP identified numerouspharmaceutical compounds, previously reported by Gagné etal. (2006) and described in depth by Blaise et al. (2007). Briefly,there were five categories of pharmaceuticals found in theeffluent examined, anti-inflammatory (ibuprofen andnaproxen), lipid regulators (gemfibrozil and bezafibrate),anti-convulsant (carbamazepine), antibiotics (sulfamethoxa-zole, sulfapyridine, oxytetracycline, novobiocin and trimeth-oprim) and a stimulant (caffeine). In general the compounds

ttenuata exposed to each of the 11 pharmaceuticals underth 95% confidence interval (CI) and the concentration (mg/L)ponse, hydranth number and attachment

Feeding response Hydranth Attachment

10 N100 110 5 1050 10 1050 N50 25100 100 N100N100 N100 N100N100 N100 N100100 N100 50N100 N100 N100N100 N100 N100N100 N50 25 ↓; 50 ↑

rease; ↓ indicates decrease.

Fig. 4 – Thenumber ofHydra attached (bars) and thehydranthnumber (lines) after 96 hexposure to each of the 11 pharmaceuticalsunder investigation.

312 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

and concentrations found in this effluentwere similar to thosereported in other studies (for review see Fent et al., 2006).Ibuprofen, naproxen, bezafibrate, gemfibrozil and carbamaze-pine are ubiquitously found in effluents and surface waters inGermany in the ng/L range and are among the mostcommonly found worldwide (Fent et al., 2006). Ibuprofen wasfound up to 85 μg/L in effluents (Farré et al., 2001), and inseawater in Norway up to 0.1–20 μg/L (Weigel et al., 2004).Naproxen had a high concentration in Canadian effluentswith a median level of 12.5 μg/L (Metcalfe et al., 2003a).Concentrations up to 4.6 and 3.1 μg/L of bezafibratewere foundin effluent and surface water respectively (Ternes, 1998).Carbamazepine was found in every Canadian STP effluent atconcentrations up to 2.3 μg/L (Metcalfe et al., 2003b) and inGerman wastewater up to 6.3 μg/L (Ternes, 1998) and has beenconsistently found in effluents throughout the world. In arecent Canadian study bezafibrate, gemfibrozil, ibuprofen andnaproxen were found at concentrations up to 0.8, 1.4, 22 and14 μg/L respectively, in WWTP treated effluent (Brun et al.,2006).

In the present study all of the compounds identified in theeffluent exceeded the 10 ng/L cut off value inWWTP effluents,requiring the implementation of the second-tier assessmentbased on ecotoxicity data (Hernando et al., 2006). Initially, 96 hacute toxicity tests based on morphological changes in theHydra provided LC50 values ranging from 22.36 mg/L toN100 mg/L (Table 1). As previously observed by other authorswith such high and environmentally unrealistic LC50 valuesthe risk of acute toxic effects in the environment are unlikely(Carlsson et al., 2006; Fent et al., 2006). However due to theirbiological activity and continuous release into the environ-ment, environmental classification should be based onchronic rather than acute toxicity (Carlsson et al., 2006;Crane et al., 2006). The EC50 values based on morphologywere generally an order of magnitude lower than the LC50

ranging from 1.18 mg/L up to 40.13 mg/L and were roughlysimilar to those reported by Blaise et al. (2007), withdifferences recorded for ibuprofen, sulfapyridine (the presentstudy beingmore sensitive) and bezafibrate (Blaise et al. (2007)more sensitive). There are several review papers summarizingthe biological effects of pharmaceuticals on numerous species

from various taxa (Carlsson et al., 2006; Crane et al., 2006; Fentet al., 2006, Blaise et al., 2007) where it was reported that fishare less sensitive than algae and invertebrates, with LC50

values at the mg/L level.Little is known about the chronic effects of these com-

pounds on non-target species, particularly invertebrates,where the relevant pathways (e.g. COX enzymes, effects onthe CNS) have yet to be studied. This is further complicatedsince pharmaceuticals form a heterogeneous group consistingof compounds with diverse chemical properties and biologicaleffects, with little known about their chronic effects. For thisreason we developed a toxicity test based on chronic end-points. Animal feeding behaviour has already been identifiedas a potential endpoint for the study of themore subtle effectsof pharmaceuticals (Fent et al., 2006). The results for thefeeding test were similar to those for the EC50 based onmorphology but showed increased sensitivity (lower EC50) forcarbamazepine, bezafibrate and novobiocin (Table 2). Thefeeding results for novobiocin were particularly interesting,showing an EC50 (based on feeding) of 13.53 mg/L with little orno effect on morphology, indicating that exposure to thisantibiotic affects feeding behaviour unrelated to morpholog-ical effects. A previous study by Pascoe et al. (2003) also lookedat the effect of feeding using a different suite of pharmaceu-ticals (except ibuprofen) at lower concentrations (0.01–10 mg/L) for an extended period (7 days) and reported a significantreduction in feeding for ibuprofen and acetylsalicylic acid. Inthe present study the lowest concentration showing asignificant reduction in feeding was 10 mg/L (gemfibroziland ibuprofen). However gemfibrozil was found to significant-ly reduce attachment at 1mg/L and gemfibrozil, ibuprofen andbezafibrate had a NOEC based on morphology of 0.1 mg/L.

In numerous studies, environmental risk assessment hasbeen used to estimate the risk of pharmaceuticals mostfrequently found in wastewater effluents and surface waters,based on the comparison between the predicted environmen-tal concentrations (PEC) and theworst-case predicted no effectconcentrations (PNEC) estimated from toxicity assays (Joneset al., 2002; Sanderson et al., 2004; Hernando et al., 2006). A keypiece of information needed in risk assessment is theconcentration at which a chemical produces adverse effects

Table 3 - Level of toxicity for each of the 11 pharmaceuticals under investigation, based on the chronic EC50 results from Hydraattenuata in the current study using the classification from EU directive 93/67/EECTable 3 - Level of toxicity for each of the 11 pharmaceuticals under investigation, based on the chronic EC50 results fromHydra attenuata in the current study using the classification from EU directive 93/67/EEC

313S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

on the organism. This toxicity data can be interpreted usingEU directive 93/67/EEC, which classifies substances accordingto the measured effective concentrations (EC50 value) (CEC,1996). When this scheme was applied to the data obtained inthe current study gemfibrozil, ibuprofen and naproxen wereclassified as toxic (Table 3), having an EC50 between 1 and10 mg/L. Carbamazepine, bezafibrate, sulfapyridine, oxytetra-cycline and novobiocin were all classified as harmful (EC50

between 10 and 100 mg/L) and sulfamethoxazole, trimetho-prim and caffeine were considered non-toxic (EC50N100 mg/L)(Table 3). These results are similar to other studies thatconsidered ibuprofen, naproxen, gemfibrozil and carbamaze-pine as high risk pollutants in WWTP effluents (Hernando etal., 2006; Blaise et al., 2007). In the present study the toxicitythreshold value for ibuprofen was 320 μg/L, only a factor of 102

or 10 higher than the concentration found in this effluent(1.19 μg/L) and in other Canadian effluents (22 μg/L (Brun et al.,2006)) respectively. In a previous study, exposure of Hydra tocarbamazepine for 48 h induced oxidativemetabolismwith anincrease in lipid peroxidation at a threshold concentration of7.1 μg/L (Quinn et al., 2004), similar to maximum concentra-tion of 6.3 μg/L reported in WWTPs (Ternes, 1998). Theseresults add to a growing volume being published indicatingthat pharmaceuticals may be present in effluents entering theenvironment at concentrations high enough to have chroniceffects. As these compounds enter the environment as amixture or cocktail of drugs it is possible that they have acumulative effect, with each compound behaving in anadditive fashion, contributing to the total effect of the mixtureand further increasing its toxicity. The toxicity of thesepharmaceutical mixtures is currently being investigated. AsHydra occupy one of the lower trophic levels within freshwaterfoodwebs, changes in their population could have an indirectbut significant effect on the rest of the freshwater community.

As mentioned above, the pharmaceuticals assessed in thisstudy comprise of 11 different compounds with differentchemical structures. This array of compounds producedpractical complications as each one had to be suspended insolution up to relatively high concentrations (50/100 mg/L),before exposure to Hydra. To achieve this we used 3 differentsolvents (DMSO, acetone and ethanol) and tested each one onour acute and chronic endpoints. All toxicity results for thepharmaceuticals were compared against the solvent controls

and separate toxicity tests undertaken to ensure no solventeffects. It is surprising to note that numerous publicationsmake no reference to what solvent was used for the dissolvingthe pharmaceuticals. Pascoe et al. (2003) used up to 10%ethanol which resulted in Hydra mortality. Other solventsused include dimethyl sulfoxide for the suspension ofgemfibrozil (Zurita et al., 2007). Little difference in toxicitywas observed between the three solvents tested with nosignificant effects observed at the concentrations used.

5. Conclusion

It is generally thought that the aquatic environment isparticularly at risk to pollution by pharmaceuticals due totheir continued introduction into surface waters fromWWTPs. Although it is unlikely that these pollutants will befound at concentrations high enough to illicit an acute effect,evidence is mounting to suggest that they may be present insufficient concentrations to cause chronic effects. In thepresent study the five most common pharmaceuticals (ibu-profen, naproxen, bezafibrate, gemfibrozil and carbamaze-pine) were found to be the most toxic to H. attenuata with EC50

results in the low mg/L level and toxicity threshold as low as320 μg/L. Based on environmental risk assessment gemfibro-zil, ibuprofen and naproxen may be categorized as toxiccompounds, with their toxicity threshold only one or twoorders of magnitude above concentrations found in theenvironment and may therefore be considered high riskcompounds in WWTP effluents, particularly as they may actcumulatively in a mixture found in effluents. However thecurrent lack of chronic toxicity data for the majority ofpharmaceuticals is a major hindrance to their adequate riskappraisal and needs to be addressed by the use of chronictoxicity tests and the development of new chronic endpoints.

Acknowledgements

The authors are grateful to Jasmine Nahrgang and PavletaPavlota for their help in conducting these experiments. Thisresearch project was undertaken under the Saint-LawrenceCentre's Municipal Effluent Research Program and partly

314 S C I E N C E O F T H E T O T A L E N V I R O N M E N T 3 8 9 ( 2 0 0 8 ) 3 0 6 – 3 1 4

funded by an NSERC visiting Fellowship in Canadian Govern-ment Laboratories.

R E F E R E N C E S

Blaise C, Kusui T. Acute toxicity assessment of industrial effluentswith amicroplate-based Hydra attenuata assay. Environ ToxicolWater Qual 1997;12:53–60.

Blaise C, Gagne F, Eullaffroy P, Ferard JF. Ecotoxicity of selectedpharmaceuticals of urban origin discharged to the Saint-Lawrence river (Quebec, Canada): a review. Brazilian J Aquat SciTech 2007;10(2):29–51.

Brun GL, Bernier M, Losier R, Doe K, Jackman P, Lee HB.Pharmaceutically active compounds in Atlantic Canadiansewage treatment plant effluents and receiving waters andpotential for environmental effects as measured by acute andchronic aquatic toxicity. Environ Toxicol Chem 2006;25(8):2163–76.

Carlsson C, Johansson AK, Alvan G, Bergman K, Kuhler T. Arepharmaceuticals potent environmental pollutants? Part I:environmental risk assessments of selected activepharmaceutical ingredients. Sci Total Environ 2006;364:67–87.

CEC (Commission of the European Communities). Technicalguidance document in support of commission directive 93/67/EEC on risk assessment for new notified substances. Part II,environmental risk assessment. Luxembourg: Office for officialpublication of the European Communities; 1996.

Crane M,Watts C, Boucard T. Chronic aquatic environmental risksfrom exposure to human pharmaceuticals. Sci Total Environ2006;367:23–41.

Fent K, Weston AA, Caminada D. Ecotoxicology of humanpharmaceuticals. Aquat Toxicol 2006;76:122–59.

Ferrari B, Mons R, Vollat B, Fraysse B, Paxeus N, Lo Giudice R, PollioA, Garric J. Environmental risk assessment of six humanpharmaceuticals: are the current environmental riskassessment procedures sufficient for the protection of theaquatic environment? Environ Toxicol Chem 2004;23(5):1344–54.

Ferré ML, Ferrer I, Ginebreda A, Figueras M, Olivella L, Tirapu L,Vilanova M, Barcelo D. Determination of drugs in surface waterand wastewater samples by liquid chromatography–massspectrometry: methods and preliminary results includingtoxicity studies with Vibrio fischeri. J Chromatogr 2001;A 938(1/2):187–97.

Gagné F, Blaise C, Andre C. Occurrence of pharmaceutical products ina municipal effluent and toxicity to rainbow trout (Oncorhynchusmykiss) hepatocytes. Ecotoxicol Environ Saf 2006;64:329–36.

Heberer T. Occurrence, fate, and removal of pharmaceuticalresidues in the aquatic environment: a review of recentresearch data. Toxicol Lett 2002;131:5–17.

Hernando MD, Mezcua M, Fernandez-Alba AR, Barcelo D.Environmental risk assessment of pharmaceutical residues inwastewater effluents, surface waters and sediments. Talanta2006;69:334–42.

Holdway DA, Lok K, Semaan M. The acute and chronic toxicity ofcadmium and zinc of two Hydra species. Environ Toxicol2001;16:557–65.

Jones OAH, Voulvoulis N, Lester JN. Aquatic environmentalassessment of the top 25 English prescription pharmaceuticals.Water Res 2002;36:5013–22.

Karntanut W, Pascoe D. The toxicity of copper, cadmium and zincto four different Hydra (Cnidaria: Hydrozoa). Chemosphere2002;47:1059–64.

Kümmerer K. Pharmaceuticals in the environment, sources, fate,effects and risks. 2nd ed. Berlin: Springer-Verlag; 2004. 528 pp.

Lishman L, Smyth SA, Sarafin K, Kleywegt S, Toito J, Peart T, Lee B,Servos M, Beland M, Seto P. Occurrence and reductions ofpharmaceuticals and person care products and estrogens bymunicipal wastewater treatment plants in Ontario, Canada. SciTotal Environ 2006;367:544–58.

Metcalfe CD, Koenig BG, Bennie DT, Servos M, Ternes TA, Hirsch R.Occurrence of neutral and acidic drugs in the effluents ofCanadian sewage treatment plants. Environ Toxicol Chem2003a;22(12):2872–80.

Metcalfe CD, Miao XS, Koenig BG, Struger J. Distribution of acidicand neutral drugs in surface waters near sewage treatmentplants in the lower Great lakes, Canada. Environ Toxicol Chem2003b;22(12):2881–9.

Pachura-BouchetS,BlaiseC,VasseurP.ToxicityofnonylphenolonthecnidarianHydra attenuata and environmental risk assessment.Environ Toxicol 2006;21:388–94.

Pardos M, Benninghoff C, Gueguen C, Thomas R, Dobrowolski J,Dominik J. Acute toxicity assessment of Polish (waste) waterwith a microplate-based Hydra attenuata assay: a comparisonwith the Microtox test. Sci Total Environ 1999;243/244:141–8.

Pascoe D, Carroll K, Karntanut W, Watts MM. Toxicity of 17α-ethinylestradiol and bisphenol A to the freshwater CnidarianHydra vulgaris. Arch Environ Contam Toxicol 2002;43:56–63.

Pascoe D, Karntanut W, Muller CT. Do pharmaceuticals affectfreshwater invertebrates? A study with the cnidarian Hydravulgaris. Chemosphere 2003;51:521–8.

Quinn B, Gagne F, Blaise C. Oxidative metabolism activity inHydra attenuata exposed to carbamazepine. Fres Environ Bull2004;13(8):783–8.

Quinn B, Gagne F, Blaise C. Validation of a multi-well plate toxicitytest to assess feeding behaviour of the cnidarian, Hydraattenuata. Fres Env Bull 2007;16(9a):1100–7.

RobertsPH,ThomasKV.Theoccurrenceof selectedpharmaceuticalsin wastewater effluent and surface waters of the lower Tynecatchment. Sci Total Environ 2006;356:143–53.

SandersonH, JohnsonDJ, ReitsmaT, Brain RA, SolomonKR. Rankingand prioritization of environmental risks of pharmaceuticals insurface water. Regul Toxicol Pharmacol 2004;39:158–83.

Ternes TA. Occurrence of drugs in German sewage treatmentplants and rivers. Water Res 1998;32(11):3245–60.

Ternes T, Hirsch R. Occurrence and behaviour of X-ray contrastmedia in sewage facilities and the aquatic environment.Environ Sci Technol 2000;34:2741–8.

Trottier S, Blaise C, Kusui T, Johnson EM.Acute toxicity assessmentof aqueous samples using a microplate-based Hydra attenuataassay. Environ Toxicol Water Qual 1997;12:265–71.

U.S. Environmental Protection Agency. Short-term methods forestimating the chronic toxicity of effluents and receivingwastes to freshwater organisms. EPA/600/4-89/001. Cincinnati,OH: Office of Research and Development; 1989. 248pp.

Wiegel S, Aulinger A, Brockmeyer R, Harms H, Loffler J, Reincke H,Schmidt R, Stachel B, Von Tumpling W, Wanke A.Pharmaceuticals in the river Elbe and its tributaries.Chemosphere 2004;57(2):107–26.

Wilby OK. The Hydra regeneration assay. Proceedings of workshoporganisedbyAssociationFrancaisedeTeratologie; 1988.p. 108–24.

Zurita JL, Repetto G, Jos A, Salguero M, Lopez-Artiguez M, CameanAM. Toxicological effects of the lipid regulator gemfibrozil infour aquatic systems. Aquat Toxicol 2007;81:106–15.