ecotoxicology of human pharmaceuticals

8/9/2019 Ecotoxicology of Human Pharmaceuticals

1/38

Aquatic Toxicology 76 (2006) 122–159

Review

Ecotoxicology of human pharmaceuticals

Karl Fent a, b, ∗, Anna A. Weston a , c, Daniel Caminada a, d

a University of Applied Sciences Basel, Institute of Environmental Technology, St. Jakobs-Strasse 84, CH-4132 Muttenz, Switzerland b Swiss Federal Institute of Technology (ETH), Department of Environmental Sciences, CH-8092 Z¨ urich, Switzerland

c Springborn Smithers Laboratories Europe AG, Seestrasse 21, CH-9326 Horn, Switzerland d University of Z¨ urich, Institute of Plant Biology, Limnology, Seestrasse 187, CH-8802 Kilchberg, Switzerland

Received 21 February 2005; received in revised form 1 September 2005; accepted 1 September 2005

Abstract

Low levels of human medicines (pharmaceuticals) have been detected in many countries in sewage treatment plant (STP)efuents, surface waters, seawaters, groundwater and some drinking waters. For some pharmaceuticals effects on aquatic organ-isms have been investigated in acute toxicity assays. The chronic toxicity and potential subtle effects are only marginally known,however. Here, we critically review the current knowledge about human pharmaceuticals in the environment and address severalkey questions. What kind of pharmaceuticals and what concentrations occur in the aquatic environment? What is the fate insurface water and in STP? What are the modes of action of these compounds in humans and are there similar targets in loweranimals? What acute and chronic ecotoxicological effects may be elicited by pharmaceuticals and by mixtures? What are theeffect concentrations and how do they relate to environmental levels? Our review shows that only very little is known aboutlong-term effects of pharmaceuticals to aquatic organisms, in particular with respect to biological targets. For most humanmedicines analyzed, acute effects to aquatic organisms are unlikely, except for spills. For investigated pharmaceuticals chroniclowest observed effect concentrations (LOEC) in standard laboratory organisms are about two orders of magnitude higher thanmaximal concentrations in STP efuents. For diclofenac, the LOEC for sh toxicity was in the range of wastewater concentra-tions, whereas the LOEC of propranolol and uoxetine for zooplankton and benthic organisms were near to maximal measuredSTP efuent concentrations. In surface water, concentrations are lower and so are the environmental risks. However, targetedecotoxicological studies are lacking almost entirely and such investigations are needed focusing on subtle environmental effects.This will allow better and comprehensive risk assessments of pharmaceuticals in the future.© 2005 Elsevier B.V. All rights reserved.

Keywords: Pharmaceuticals; Ecotoxicological effects; Environmental toxicity; Chronic effects; Environmental risk assessment

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1232. Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

∗ Corresponding author. E-mail address: [email protected] (K. Fent).

0166-445X/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.aquatox.2005.09.009


2/38

K. Fent et al. / Aquatic Toxicology 76 (2006) 122–159 123

3. Fate in the environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1274. Environmental concentrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

4.1. Analgesics and antiinammatory drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1324.2. Beta-blockers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1334.3. Blood lipid lowering agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1334.4. Neuroactive compounds (antiepileptics, antidepressants) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1334.5. Antineoplastics and antitumor agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1334.6. Various other compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1344.7. Steroidal hormones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134

5. Modes of actions in humans and mammals and occurrence of target biomolecules in lowervertebrates and invertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355.1. Analgesics and non-steroidal antiinammatory drugs (NSAID) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355.2. Beta-blockers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1365.3. Blood lipid lowering agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1375.4. Neuroactive compounds (antiepileptics, antidepressants) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1385.5. Cytostatics compounds and cancer therapeutics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138

5.6. Various compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1386. Ecotoxicological effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

6.1. Acute effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1396.1.1. Analgesics and non-steroidal antiinammatory drugs (NSAID) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1406.1.2. Beta-blockers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1406.1.3. Blood lipid lowering agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1416.1.4. Neuroactive compounds (antiepileptics, antidepressants) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1416.1.5. Cytostatic compounds and cancer therapeutics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

6.2. Chronic effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1426.2.1. Analgesics and non-steroidal antiinammatory drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1446.2.2. Beta-blockers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1446.2.3. Blood lipid lowering agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144

6.2.4. Neuroactive compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1446.3. In vitro studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1466.4. Toxicity of pharmaceutical mixtures and community effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

7. Comparison of environmental concentrations and ecotoxicological effects concentrations . . . . . . . . . . . . . . . . . . . . . . 1478. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1489. Conclusions and future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

1. Introduction

It came as a surprise when an unusually high deathrate among three species of vulture in India and Pak-istan was reported in 2004 to be caused by diclofenac,a widely used analgesic and antiinammatory drug(Oaks et al., 2004 ). The oriental white-backed vul-ture (Gyps bengalensis ) is one of the most commonraptors in the Indian subcontinent and a populationdecline of >95% makes this species as being criticallyendangered. Whereas a population decline has startedin the 1990s, recent catastrophic declines also involve

Gyps indicus and Gyps tenuirostris across the Indian

subcontinent ( Prakash et al., 2003; Risebrough, 2004 ).High adult and subadult mortality and resulting popu-lation loss is associated with renal failure and visceralgout, theaccumulation of uric acid throughout thebodycavity following kidney malfunction. A direct correla-tion between residues of diclofenac and renal failurewas reported both by experimental oral exposure andthrough feeding vultures diclofenac-treated livestock.Hence, the residues of diclofenac were made respon-sible for the population decline ( Oaks et al., 2004 ).This drug has recently come into widespread use in


3/38

124 K. Fent et al. / Aquatic Toxicology 76 (2006) 122–159

these countries as a veterinary medicine, but is alsowidely used as in human medicine since the 1970s.Vultures are natural scavengers feeding on carrion of

wildlife and domestic livestock and cattle. The threevulture species continue to decline in Pakistan, India,Bangladesh andsouthern Nepal. Apart from this severecase, never having been anticipated, potential ecotox-icological effects of drug residues in the environmenton wildlife are largely unknown.

Pharmaceuticals are a class of emerging environ-mental contaminants that are extensively and increas-ingly being used in human and veterinary medicine.These chemicals are designed to have a specic modeof action, and many of them for some persistence in thebody. These features among others make pharmaceu-ticals to be evaluated for potential effects on aquaticora and fauna. The current investigations are mainlydriven by advances in environmental residue analy-sis, particularly after the establishment of chemicalanalysis methods able to determine more polar com-pounds such as liquid chromatography–tandem massspectrometry, which allows the identication of tracequantities of polar organic pollutants without derivati-zation (Ternes et al., 1998, 2001; Kolpin et al., 2002;Kümmerer, 2004 ). Accordingly, many environmentalanalyses have been performed in various countries,

which are summarized by various reports (e.g. Halling-Sorensen et al., 1998; Daughton and Ternes, 1999;Kümmerer, 2004 ). These monitoring studies demon-strate that drug residues in treated wastewater and sur-face water are very widespread.

In contrast, only little is known about ecotoxicologi-cal effects of pharmaceuticals on aquatic and terrestrialorganisms and wildlife,anda comprehensive reviewonecotoxicological effects is lacking. Aquatic organismsare particularly important targets, as they are exposedvia wastewater residues over their whole life. Standard

acute ecotoxicity data have been reported for a num-ber of pharmaceuticals, however, such data alone maynot be suitable for specically addressing the questionof environmental effects, and subsequently in the haz-ard and risk assessment ( Fent, 2003 ). The current lack of knowledge holds in particular for chronic effectsthat have only very rarely been investigated. In spite of the sizeable amounts of human drugs released to theenvironment, concise regulations for ecological risk assessment are largely missing. Only in the last fewyears, regulatory agencies have issued detailed guide-

lines on how pharmaceuticals should be assessed forpossible unwanted effects on theenvironment. Therstrequirement for ecotoxicity testing as a prerequisite

for registration of pharmaceuticals was established in1995 according to the European Union (EU) Directive92/18 EEC and the corresponding “Note for Guid-ance” (EMEA, 1998 ) for veterinary pharmaceuticals.The European Commission released a draft guide-line (Directive 2001/83/EC) specifying that an autho-rization for a medicinal product for human use mustbe accompanied by an environmental risk assessment(EMEA, 2005 ). The U.S. Food and Drug Administra-tion (FDA) published a guidance for the assessments of humandrugs;according to this, applicants in theU.S.A.are required to provide an environmental assessmentreport when the expected introduction concentrationof the active ingredient of the pharmaceutical in theaquatic environment is ≥ 1 g/L (FDA-CDER, 1998 ),which corresponds to about 40 t as a trigger level.In contrast, environmental assessments of veterinarypharmaceuticals is required by the U.S. FDA since1980 (Boxall et al., 2003 ).

The objective of our paper is to compile and crit-ically review the present knowledge about the envi-ronmental occurrence and fate of human pharmaceu-ticals in the aquatic environment, to discuss poten-

tial mechanisms of action based on knowledge frommammalian studies, and to describe the acute andchronic ecotoxicological effects on aquatic organisms.We also identify major gaps in the current knowl-edge and future research needs. We concentrate onpharmaceuticals used in human medicine, some of which are also applied in veterinary medicine, therebyfocusing on environmentally important compoundsbelonging to different drug categories, namely non-steroidal antiinammatory drugs, beta-blockers, bloodlipid lowering agents, cancer therapeutics and neu-

roactive compounds. These classes differ for theirmodes of actions and were chosen because of theirconsumption volumes, toxicity and persistence in theenvironment. We will not address the environmentaleffects of antibiotics and biocides ( Halling-Sorensenet al., 1998; Daughton and Ternes, 1999; Hirsch etal., 1999 ), hormones (used in contraceptives and intherapy) ( Damstra et al., 2002 ) and special veteri-nary pharmaceuticals ( Montforts et al., 1999; Boxallet al., 2003) as the cited reports provide detailedinformation.


4/38


5/38

Table 1Annual consumption of different classes of prescribed drugs for different countries

Compounds Germany1999 a

Germany2000a

Germany2001a

Austria1997b

Denmark 1997c

Australia1998d

England2000e

Analgesics, antipyretics and anti-inammatoryAcetylsalicylic acid 902.27 (1) 862.60 (1) 836.26 (1) 78.45 (1) 0.21 (7) 20.4 (9) Salicylic acid 89.70 (12) 76.98 (17) 71.67 (17) 9.57 (11)

Paracetamol 654.42 (2) 641.86 (2) 621.65 (2) 35.08 (2) 0.24 (6) 295.9 (1) 390.9 (1) Naproxen 4.63 (16) 22.8 (7) 35.07 (12)Ibuprofen 259.85 (5) 300.09 (5) 344.89 (5) 6.7 (13) 0.03 (19) 14.2 (13) 162.2 (3) Diclofenac 81.79 (16) 82.20 (14) 85.80 (14) 6.14 (15) 26.12 (16)

-BlockerAtenolol 28.98 (13)Metoprolol 67.66 (18) 79.15 (16) 92.97 (11) 2.44 (20)

AntilipidemicGembrazol 20 (10) Bezabrate 4.47 (17)

NeuroactiveCarbamazepine 86.92 (13) 87.71 (13) 87.60 (12) 6.33 (14) 9.97 (18) 40.35 (8) Diazepam 0.21 (8)

AntiacidicRanitidine 85.41 (15) 89.29 (12) 85.81 (13) 33.7 (5) 36.32 (10) Cimetidine 35.65 (11)

DiureticsFurosemide 3.74 (1)

SympatomimetikaTerbutalin 0.46 (3) Salbutamol 0.17 (9)

VariousMetformin 368.01 (4) 433.46 (4) 516.91 (3) 26.38 (3) 90.9 (2) 205.8 (2) Estradiol 0.12 (13)

Iopromide 64.93 (19) 63.26 (19) 64.06 (19) For every country a top 20 sold-list is taken into account. Data in bracket represent the position in the ranking list within a country. Data are in t/year.

a Huschek et al. (2004) .b Sattelberger (1999) .c Stuer-Lauridsen et al. (2000) .d Khan and Ongerth (2004) .e Jones et al. (2002) .f Calamari et al. (2003) .g ©IMS Health Incorporated or its afliates. All rights reserved. MIDAS–02/03/05.


6/38


3. Fate in the environment

The behavior and fate of pharmaceuticals and their

metabolites in the aquatic environment is not wellknown. The low volatility of pharmaceuticals indi-cates that distribution in the environment will occurprimarily through aqueous transport, but also via foodchain dispersal. In wastewater treatment , two elimi-nation processes are generally important: adsorptionto suspended solids (sewage sludge) and biodegrada-tion. Adsorption is dependent on both hydrophobic andelectrostatic interactions of the pharmaceutical withparticulates and microorganisms. Acidic pharmaceu-tical such as the NSAID acetylsalicylic acid, ibupro-fen, fenoprofen, ketoprofen, naproxen, diclofenac andindomethacin having p K a values ranging from 4.9 to4.1, as well as clobric acid, bezabrate (p K a 3.6) andgembrozil occur as ion at neutral pH, and have littletendency of adsorption to the sludge. But adsorptionincreases with lower pH. At neutral pH, these nega-tively charged pharmaceuticals therefore occur mainlyin thedissolved phase in thewastewater. For these com-pounds and the antitumor agent ifosfamide sorptionby non-specic interactions seems not to be relevant(Kümmerer et al., 1997; Buser et al., 1998b ). In gen-eral, sorption of acidic pharmaceuticals to sludge is

suggested to be not very important for the elimina-tion of pharmaceuticals from wastewater and surfacewater. Therefore, levels of pharmaceuticals in digestedsludge and sediments are suggested to be relativelylow, as was demonstrated in several monitoring studies(Ternes et al., 2004; Urase and Kikuta, 2005 ). How-ever, basic pharmaceuticals and zwitterions can adsorbto sludge to a signicant extent, as has been shownfor uoroquinolone antibiotics ( Golet et al., 2002 ). Forthe hydrophobic EE2 (log K ow 4.0) sorption to sludgeis likely to play a role in the removal from wastew-

ater. Degradation in sludge seems not signicant. Asa consequence, EE2 occurs in digested sludge, whereconcentrations of 17 ng/g were reported ( Temes et al.,2002 ).

In case a pharmaceutical is occurring mainly in thedissolved phase, biodegradation is suggested to be themost important elimination process in wastewater treat-ment. It can occur either in aerobic (and anaerobic)zones in activated sludge treatment, or anaerobically insewage sludge digestion. In general, biological decom-position of micro-pollutants includingpharmaceuticals

increases with increase in hydraulic retention time andwith age of the sludge in the activated sludge treat-ment. For example, diclofenac was shown to be sig-

nicantly biodegraded only when the sludge retentiontime was at least 8 days ( Kreuzinger et al., 2004 ). Incontrast, data from Metcalfe et al. (2003a,b) indicatethat the neutral drug carbamazepine, which is hardlybiodegradable, is onlypoorly eliminated(normallylessthan10%), independent fromhydraulicretentiontimes.Pharmaceuticals are often excreted mainly as non-conjugated and conjugated polar metabolites. Conju-gates can, however, be cleaved in sewage treatmentplants (STP), resulting in the release of active parentcompound as shown for estradiol ( Panter et al., 1999;Ternes et al., 1999 ), and the steroid hormone in the con-traceptive pill, 17 -ethinylestradiol ( D’Ascenzo et al.,2003 ).

Studies on the elimination rates during the STPprocess are mainly based on measurements of inu-ent and efuent concentrations in STPs, and they varyaccording to the construction and treatment technol-ogy, hydraulic retention time, season and performanceof the STP. Some studies ( Ternes, 1998; Stumpf etal., 1999; Carballa et al., 2004 ) indicate eliminationefciencies of pharmaceuticals to span a large range(0–99%). The average elimination for specic pharma-

ceuticals varied from only 7 to 8% for carbamazepine(Ternes, 1998; Heberer, 2002; Clara et al., 2004 ) upto 81% for acetylsalicylic acid, 96% for propranolol,and 99% for salicylic acid ( Ternes, 1998; Ternes etal., 1999; Heberer, 2002 ). Lowest average removalrates were found for diclofenac (26%), the removal of bezabrate was 51%, but varied signicantly betweenSTPs, and high removal rates were found for naproxen(81%) ( Lindqvist et al., 2005 ). Table 2 shows thatremoval rates are variable, even for the same pharma-ceutical between different treatment plants. Very high

total elimination of 94–100% of ibuprofen, naproxen,ketoprofen and diclofenac was found in three STPsin the U.S.A. ( Thomas and Foster, 2004 ). Efcientremoval took place mainly in the secondary treat-ment step (51–99% removal), whereas in the primarytreatment only 0–44% were removed. X-ray contrastmedia(diatrizoate, iopamidol, iopromide, iomeprol), tothe contrary, were not signicantly eliminated ( Ternesand Hirsch, 2000 ). Also, the anticancer drug tamox-ifen (antiestrogen) was not eliminated ( Roberts andThomas, 2005 ). This variation in elimination rates is


7/38


Table 2Inuent and efuent concentrations and removal efciency in sewage treatment plants (different equipment, different countries, sampling indifferent seasons)

Compound Inuentconcentration( g/L)

Efuentconcentration( g/L)

Maximalremoval(%)

Reference

Analgesics and antiinammatory drugsAcetylsalicylic acid 3.2 0.6 81 Ternes et al. (1999)

Salicylic acid 57 0.05 99 Metcalfe et al. (2003a) a

330 3.6 Carballa et al. (2004)

Dextropropoxyphene 0.03 0.06 0 Roberts and Thomas (2005) a

Diclofenac 3.0 2.5 17 Heberer (2002)n.r. n.r. 69 Ternes (1998) b

0.33–0.49 n.r. 75 (10–75) Andreozzi et al. (2003a) c

[5] [1.5] 53–74 Strenn et al. (2004) a

1.3 n.r. Metcalfe et al. (2003a) a0.47–1.9 0.31–0.93 Buser et al. (1998b)2.8 1.9 23 ± 30 Quintana et al. (2005) b

0.4–1.9 0.4–1.9 0 Tauxe-Wuersch et al. (2005) c

0.35 ± 0.1 0.17–0.35 9–60 Lindqvist et al. (2005) c

1.0 0.29 71 Roberts and Thomas (2005) a

Ibuprofen 3 96 Buser et al. (1999)38.7 4 >90 Metcalfe et al. (2003a) a

9.5–14.7 0.01–0.02 99 Thomas and Foster (2004)[0.54] [0.08–0.28] 22–75 99 (52–99) Andreozzi et al. (2003a) c

[1.5] [0.01] 12–86 Strenn et al. (2004) a

2.6–5.7 0.9–2.1 60–70 Carballa et al. (2004) a

5.7 0.18 97 ± 4 Quintana et al. (2005) b

28.0 3.0 98 Roberts and Thomas (2005) a2–3 0.6–0.8 53–79 Tauxe-Wuersch et al. (2005) c

13.1 ± 4 0–3.8 78–100 Lindqvist et al. (2005) c

Ketoprofen 0.41–0.52 0.008–0.023 98 Thomas and Foster (2004)[0.55] [0.18–0.3] 48–69 Stumpf et al. (1999) b

5.7 n.r. Metcalfe et al. (2003a) a

0.47 0.18 62 ± 21 Quintana et al. (2005) b

0.25–0.43 0.15–0.24 8–53 Tauxe-Wuersch et al. (2005) c

2.0 ± 0.6 0–1.25 51–100 Lindqvist et al. (2005) c

Mefenamic acid 1.6–3.2 0.8–2.3 2–50 Tauxe-Wuersch et al. (2005) c

0.20 0.34 0 Roberts and Thomas (2005) a

Naproxen 66 Ternes (1998) b

40.7 12.5 40–100 Metcalfe et al. (2003a)10.3–12.8 n.d.-0.023 100 Thomas and Foster (2004)[0.6] [0.1–0.54] 15–78 Stumpf et al. (1999) b

93(42–93) Andreozzi et al. (2003a) c

1.8–4.6 0.8–2.6 40–55 Carballa et al. (2004) a

0.95 0.27 71 ± 18 Quintana et al. (2005) b

4.9 ± 1.7 0.15–1.9 55–98 Lindqvist et al. (2005) c

Paracetamol 6.9 0 100 Roberts and Thomas (2005) a

-BlockerMetoprolol n.r. n.r. 83 Ternes (1998) b

n.r. n.r. 10 (0–10) Andreozzi et al. (2003a) c


8/38


Table 2 ( Continued )

Compound Inuentconcentration( g/L)

Efuentconcentration( g/L)

Maximalremoval(%)

Reference

Propranolol n.r. n.r. 96 Ternes (1998) b

70 304 0 Roberts and Thomas (2005) a

Atenolol n.r. n.r.


9/38


reactions are probably more important. Whereashydrolysis is generally negligible for environmentallyrelevant human drugs, photodegradation sometimes

plays an important role at the water surface. Photolysishas been shown to be the main removal process fordiclofenac in surface water ( Buser et al., 1998b ).For additional pharmaceuticals (sulfamethoxazole,ooxacin and propranolol) laboratory experimentsindicate direct and indirect photolysis as an importantremoval process ( Andreozzi et al., 2003b ). Carba-mazepine and clobric acid, both compounds thatare marginally processed in STP, have been shown toundergo slow photodegradation in salt- and organic-free water with estimated half-lives in the range of 100 days at latitudes of 50 ◦ N in winter (Andreozziet al., 2003b ). The efciency of photodegradationdepends, besides substance properties, on the strengthof the solar irradiation, and therefore on latitudeand season, and on constituents present in the waterthat may act as photosensitizers generating hydroxylradicals and singlet oxygen (i.e. nitrates, humic acids).Some adsorption to particles may occur. Laboratorybatch studies to characterize the sorption behaviorof carbamazepine, diclofenac and ibuprofen in sandysediments show that sorption coefcients were gen-erally quite low ( Scheytt et al., 2005 ). Diclofenac and

ibuprofen are carboxylic acids with p K a values of 4.16and 4.52 and these weak acids are negatively chargedat pH of ambient water and sediment.

There is no information about the bioaccumulationpotential of pharmaceuticals in biota or food webs withtheexception of diclofenac, accumulating in theprey of vultures ( Oaks et al., 2004 ), uoxetine, sertraline andthe SSRI metabolites noruoxetine and desmethylser-traline detected in sh ( Brookset al., 2005 ). Diclofenacbioconcentration factors were 10–2700 in the liver of sh and 5–1000 in the kidney, depending on exposure

concentrations ( Schwaiger et al., 2004 ).A few cases were reported, where pharmaceuticalswere detected in drinking water (Heberer and Stan,1996 ) and groundwater ( Holm et al., 1995; Terneset al., 2001 ). Ozonation, granulated activated carbon,and advanced oxidation have been shown as efcientremoval processes. In drinking water, this has beenshown for diclofenac, while clobric acid and ibupro-fen were oxidized in laboratory experiments mainly byozone/H 2O2 (Zwiener and Frimmel, 2000 ). The elim-ination of selected compounds (bezabrate, clobric

acid,carbamazepine, diclofenac) during drinkingwatertreatment was investigated in laboratory experimentsand waterworks ( Ternes et al., 2002 ). No signicant

removal was observed in batch experiments with sand,indicating low sorption properties and persistence.Flocculation using iron(III) chloride was ineffective,but ozonation was in some cases very effective in elim-inating thesepolarpharmaceuticals. However, clobricacid was stable and not eliminated, even with ltrationusing granular activated carbon, which was effectivefor the other compounds. The removal of pharmaceu-ticals and other polar micro-pollutants can thereforeonly be assured using more advanced techniques suchas ozonation, activated carbon or membrane ltration(Ternes et al., 2002 ). However, the economic conse-quences have to be evaluated carefully before investinginto these advanced treatment technologies on a largerscale.

4. Environmental concentrations

The occurrence of pharmaceuticals was rstreported in the U.S.A. in treated wastewater, whereclobric acid in the range of 0.8–2 g/L was found(Garrison et al., 1976 ). Subsequently, pharmaceuticals

were detected in the U.K. in 1981 in rivers up to 1 g/L(Richardson and Bowron, 1985 ), and ibuprofen andnaproxen were identied in wastewaters in Canada(Rogers et al., 1986 ). In the last few years, knowledgeabout theenvironmental occurrence of pharmaceuticalshas increased to a large extent due to new analyti-cal techniques able to determine polar compounds attrace quantities. This also holds for the steroid hor-mones contained in contraceptive pills such as 17 -ethinylestradiol (EE2), which is linked to biologicaleffects in sh ( Stumpf et al., 1996; Desbrow et al.,

1998 ). Data on environmental concentrations up to2004 have been compiled and reviewed (e.g. Halling-Sorensen et al., 1998; Daughton and Ternes, 1999;Kümmerer, 2001; Heberer, 2002; K ümmerer, 2004 ).Here, we give a summary on environmental concentra-tions focusing on most recent analytical data with theultimate aim to relate them to ecotoxicological data.First, we give a general overview on the occurrenceof pharmaceuticals in general and in different environ-mental media, and subsequently present data on thedifferent pharmaceutical classes.


10/38


11/38


Fig. 1. Concentration of pharmaceuticals in treated wastewater (a) and surface water (b). References : Halling-Sorensen et al. (1998) , Ternes(1998) , Stuer-Lauridsen et al. (2000) , Jones et al. (2002) , Kolpin et al. (2002) , Andreozzi et al. (2003b) , Calamari et al. (2003) , Metcalfe et al.(2003a,b) , Gross et al. (2004) , Khan and Ongerth (2004) , Kümmerer (2004) , Stackelberg et al. (2004) , Thomas and Hilton (2004) , Weigel et al.(2004) , Lindqvist et al. (2005) , Quintana et al. (2005) , Roberts and Thomas (2005) and Tauxe-Wuersch et al. (2005) .

propiphenazone and clobric acid were found in sam-ples of potable water collected in the vicinity of Berlin,Germany ( Heberer and Stan, 1997; Reddersen et al.,2002 ). Several polar pharmaceuticals such as clob-ric acid, carbamazepine, and X-ray contrast media canoccur in groundwater. In the following, current knowl-edge about major pharmaceuticals of different thera-peutic classes is summarized.

4.1. Analgesics and antiinammatory drugs

The widely used non-steroidal antiinammatorydrugs (NSAID) ibuprofen, naproxen, diclofenac andsome of their metabolites (e.g. hydroxyl-ibuprofen andcarboxy-ibuprofen) are very often detected in sewageand surface water. Ternes (1998) reported levels insewage exceeding 1 g/L, and in efuents of con-ventional STP (mechanical clarication and biologicaltreatment) concentrations often approach or exceed

0.1 g/L in the U.S.A. ( Gross et al., 2004 ). The deacy-lated, more active form of acetylsalicylic acid, salicylicacid, has been found in many municipal wastewatersat levels up to 4.1 g/L (Ternes, 1998 ), 13 g/L (Farr éet al., 2001; Heberer, 2002 ) or even 59.6 g/L withmedian levels of 3.6 g/L (Metcalfe et al., 2003a ).However, salicylic acid may also derive from othersources. Similar to acetylsalicylic acid, acetaminophen

(paracetamol) is well removed from STP. However,up to 10 g/L (median 0.11 g/L) acetaminophen hasbeen found in 24% of samples from U.S. streams(Kolpin et al., 2002 ). The analgesic codeine wasdetected in 7% of samples at median concentrationsof 0.01 g/L.

In many countries diclofenac was frequentlydetected in wastewater in the g/L range, and in sur-face water at lower levels ( Heberer and Stan, 1997;Buser et al., 1998b ; Ternes, 1998; Stumpf et al., 1999;Farr é et al., 2001; Sedlak and Pinkston, 2001; Heberer,


12/38


2002 ). This also holds for ibuprofen ( Heberer andStan,1997; Ternes, 1998; Buser et al., 1999; Stumpf etal., 1999 ). Sometimes, high levels of up to 85 g/L

(Farr é et al., 2001 ), or 24.6 g/L (median 4.0 g/L)were detected in STP efuents ( Metcalfe et al., 2003a ).In Norway, ibuprofen and its metabolites occurred inall sewage samples, and in seawater at concentrationsof 0.1–20 g/L (sum of ibuprofen and metabolites)(Weigel et al., 2004 ). In U.K. estuaries maximal con-centration of 0.93 g/L (median 0.05 g/L) occurred(Thomas and Hilton, 2004 ). Ibuprofen is signicantlyremoved during sewage treatment, and metabolitessuch as hydroxy-ibuprofen occur in STP efuents.Kolpin et al. (2002) f ound ibuprofen in 10% of streamwater samples with maximal concentrations of 1 g/L(median 0.2 g/L). In two stormwater canals levels of ibuprofen were up to 674 ng/L and of naproxen up to145 ng/L ( Boyd et al., 2004 ). Naproxen was also foundat much higher level in Canadian STP efuents withmedian levels of 12.5 g/L and maximal levels of upto 33.9 g/L (Metcalfe et al., 2003a ). Moreover, sev-eral other analgesics have been detected in sewage andsurface water, but also in ground water and drinkingwater samples.

4.2. Beta-blockers

Several beta-blockers were identied in wastewater(Ternes, 1998; Sedlak and Pinkston, 2001 ). Propra-nolol, bisoprolol and metoprolol were found at highestlevels (0.59, 2.9 and 2.2 g/L, respectively, in surfacewater), with lower levels of nadolol (in surface water)and betaxolol (0.028 g/L in surface water) ( Ternes,1998 ). Propranolol, metoprolol and bisoprolol havealso been found in surface water, andsotalol in ground-water (Sacher et al., 2001 ).

4.3. Blood lipid lowering agents

Clobric acid, the active metabolite from a series of widely used blood lipid regulators (clobrate, etofyllinclobrate, etobrate) belongs to the most frequentlyfound and reported pharmaceutical in monitoring stud-ies. It hasbeen found in numerous wastewaters, surfacewaters, in seawater ( Stumpf et al., 1996; Buser et al.,1998a ; Ternes, 1998 ), and at rather high concentra-tions in groundwater (4 g/L) (Heberer andStan,1997 )and drinking water (0.07–0.27 g/L) (Stumpf et al.,

1996; Heberer and Stan, 1997 ). Bezabrate occurredin maximal concentrations of up to 4.6 and 3.1 g/L(median 2.2 and 0.35 g/L, respectively) in wastewa-

ter and surface water, respectively ( Stumpf et al., 1996;Ternes, 1998 ). In addition, gembrozil, clobric acidandfenobricacid(metabolite of fenobrate) have alsobeen detected in sewage up to the g/L level and in sur-face water ( Ternes, 1998; Stumpf et al., 1999; Farr é etal., 2001; Heberer, 2002 ). Gembrozil was detected in4% of streams at maximal levels of 0.79 g/L (Kolpinet al., 2002 ).

4.4. Neuroactive compounds (antiepileptics,antidepressants)

Of this category, the antiepileptic carbamazepinewas detected most frequently and in highest concen-tration in wastewater (up to 6.3 g/L) (Ternes, 1998 ),and at lower levels in other media ( Heberer et al.,2002; Andreozzi et al., 2003b; Metcalfe et al., 2003b;Wiegeletal., 2004 ). Carbamazepine wasfound in everyCanadian STP efuent sample at concentration up to2.3 g/L (Metcalfe et al., 2003b ). This compound wasfound to occur ubiquitously in the river Elbe and itstributaries, Germany ( Wiegel et al., 2004 ), exceed-ing 1 g/L in other German surface waters ( Ternes,

1998; Heberer, 2002 ) and occurred in groundwater(Seiler et al., 1999; Sacher et al., 2001; Ternes et al.,2001 ). In U.S. rivers average levels were 60 ng/L inwater and 4.2ng/mg in the sediment ( Thaker, 2005 ).Carbamazepine was also found at average levels of 20.9ng/mg solids of STP. Diazepam was present in8 of 20 STPs in Germany at relatively low concen-trations of up to 0.04 g/L (Ternes, 1998 ) whereas inBelgium it was found at concentration up to 0.66 g/L(vanderVenetal.,2004 ). Theantidepressantuoxetinewas also detected in STP efuents samples in Canada

(Metcalfe et al., 2003a ), and in U.S. streams, medianconcentrations of 0.012 g/L were estimated ( Kolpinet al., 2002 ). Primidone, an antiepileptic drug, hasalso been detected in sewage up to 0.6 g/L (Heberer,2002 ).

4.5. Antineoplastics and antitumor agents

Pharmaceuticals usedin cancer chemotherapyoccurprimarily in hospital efuents and only at lower con-centrations in municipal wastewater. Ifosfamide and


13/38


cyclophosphamide occur in concentrations of up to4.5 g/L in hospital wastewaters ( Steger-Hartmannet al., 1997 ), and at ng/L in municipal wastewater

(Kümmerer et al., 1997; Steger-Hartmann et al., 1997 ).The occurrence of the antiestrogen tamoxifen used inbreast cancer therapy was reported in U.K. wastewater,where concentrations in STP efuents ranged between146and369ng/L( Roberts andThomas, 2005 ). Tamox-ifen was not reduced in the STP, and even foundin estuarine waters (Tye estuary) at concentrationsranging from 27 to 212 ng/L with a median levelof 53ng/L ( Thomas and Hilton, 2004; Roberts andThomas, 2005 ).

4.6. Various other compounds

Many additional pharmaceuticals have beendetected in sewage and surface water ( Daughton andTernes, 1999; Heberer, 2002; Kolpin et al., 2002 ).Here only a few of them will be mentioned. Thestimulant caffeine and the nicotine metabolite cotininewere generally present in STP efuents and surfacewaters contaminated by drugs ( Metcalfe et al., 2003b ).Caffeine was generally found in U.S. streams atmaximal levels of 6.0 g/L (median 0.1 g/L) (Kolpinet al., 2002 ) and this compound can even serve as

an anthropogenic marker in aquatic systems due toits ubiquity in surface water, seawater ( Weigel et al.,2004 ), and also in groundwater ( Fig. 1). The antiacidcimetidine and ranitidine were estimated to occur inU.S. streams at concentrations of 0.58 and 0.01 g/L,respectively, and they were detected at a frequency of 10 and 1%, respectively ( Kolpin et al., 2002 ). X-raycontrast media are very persistent. Iopamidol has beenfound in municipal wastewater as high as 15 g/L, insurface water (0.49 g/L) and groundwater ( Putschewet al., 2000; Ternes and Hirsch, 2000 ). Iopromide

was detected at 2–4 g/L in surface water, and up to21 g/L in STP (Putschew et al., 2000 ), but showeddegradation in the laboratory ( Steger-Hartmann etal., 2002 ). Hospital wastewater was also a sourceof gadolinium ( Kümmerer and Helmers, 2000 ). Theantidiabetic compound metformin was observed in5% of stream water samples with estimated levelsof 0.11 g/L (Kolpin et al., 2002 ). Bronchodilators( 2-sympathomimetics terbutalin and salbutamol)were also detected in sewage in a few cases notexceeding 0.2 g/L (Ternes, 1998 ).

4.7. Steroidal hormones

Steroidal hormones have been reported on in many

reports, and in our review we only summarize knowl-edge about the synthetic estrogen EE2 and mestranolcontained in contraceptive pills. These steroids havebeen found in numerous studies in many countries inEurope, Canada, the U.S.A., Japan, Brazil, etc. both inwastewater and surface water. A survey in the U.S.A.showed that maximal and median EE2 concentrationswere as high as 831 and 73 ng/L, respectively, and lev-els of mestranol were 407 and 74 ng/L, respectively(Kolpin et al., 2002 ). They were detectable in 16 and10% of the streams sampled. Generally, median con-centrations are much lower being in the range of non-detectable up to 9 ng/L in treated wastewater in severalcountries ( Baronti et al., 2000 ). Typical wastewaterefuent concentrations are 0.5 ng/L and they are evenlower in surface water. However, these concentrationsmust put into the perspective of their high biologicalactivity accounting for potential estrogenic effects insh.

Exposure and fate models are increasingly beingused to estimate environmental concentrations with-out analytical chemical measurements. Some exposuremodels have been developed for drugs (e.g. PhATE),

others have been extended from general chemicals topharmaceuticals (e.g. EPIWIN, GREAT-ER). Thesetools have been developed both for estimation of pre-dicted environmental concentrations (PEC) and thebehavior of pharmaceuticals in the environment. Apharmaceutical assessment and transport evaluationmodel (PhATE) was developed to estimate concen-trations of active pharmaceutical ingredients in U.S.surface waters ( Anderson et al., 2004 ). The PhATEmodel uses some for most hydrologic regions of theU.S. representative watersheds. For European sur-

face waters an exposure simulation was developed forpharmaceuticals with the GREAT-ER (Geo-referencedRegional Exposure Assessment Tool for EuropeanRivers) model, a tool developed for use within eco-logical risk assessment (ERA) schemes and riverbasin management ( Schowanek and Webb, 2002 ). TheGREAT-ER software calculates the distribution of PEC’s of consumer chemicals in surface waters, forindividual stretches, as well as representative aver-age PEC’s for entire catchments. The system usesan ARC/INFO-ArcView (ESRI ®) based Geographical


14/38


Information System (GIS) for data storage and visu-alization, combined with simple mathematical modelsfor prediction of the fate of chemicals.

For some estimates, measured environmental con-centrations (MEC) are in agreement with the estimatedPEC’s, however, often, they are not as large differ-ences occur between the models and the real worldsituation. The main reason is that different assump-tions are made, which not always correspond to therealconditions in the environment. Consumption gures,metabolism in the organism, removal during sewagetreatment plants and fate in the environment containall uncertainties that may result in inappropriate esti-matesof PEC’s. Moreover, detailedsituationsat a givensite is not reected by models integrating large geo-graphicalareas. Poorprediction performanceof currentmodels for many pharmaceuticals is one of the out-standing scientic issues with regard to the questionof pharmaceuticals in the environment. It is hoped thatthe models are improving by further rening the men-tioned uncertainties and may be developing to a usefulandreadily applicable regulatorytool ( Sandersonet al.,2004b ).

5. Modes of actions in humans and mammals

and occurrence of target biomolecules in lowervertebrates and invertebrates

Here, we briey summarize the modes of actionsof pharmaceutical classes and ask, whether or not sim-ilar target receptors and biomolecules exist in lowervertebrates and invertebrates. Knowledge about simi-lar targets exists primarily forsh. In general, very littleis known about possible counterparts of human targetbiomolecules of pharmaceuticals in invertebrates. Inaddition, some of the side effects in humans are dis-

cussed, giving hints to possible adverse effects in loweranimals.

5.1. Analgesics and non-steroidalantiinammatory drugs (NSAID)

Non-steroidal antiinammatory drugs act byinhibitingeither reversibly or irreversiblyoneor bothof thetwoisoforms of thecyclooxygenaseenzyme (COX-1 and COX-2), which catalyze the synthesis of dif-ferent prostaglandins from arachidonic acid ( Vane and

Botting, 1998 ). Classical NSAID inhibit both COX-1and COX-2 at different degrees, whereas new NSAIDact more selectively on COX-2, the inducible form

responsible for the inammatory reactions. Differencesin binding site size are responsible for the selectivityof these drugs ( Kurumbail et al., 1997; Penning et al.,1997; Gierseet al., 1999 ). NSAIDSarecommonly usedto treat inammation and pain and to relieve fever, andsometimes they are also used for long-term treatmentof rheumatic diseases.

Prostaglandins play a variety of physiologicalroles according to their cells source and targetmolecules. They are known to be involved in pro-cess such as inammation and pain, regulation of blood ow in kidney, coagulation processes and syn-thesis of protective gastric mucosa ( Smith, 1971; Vane,1971; Mutschler, 1996 ). Since NSAID inhibit non-specically prostaglandin synthesis, most side effects,at least after long-term treatment, are related to thephysiological function of prostaglandins. In the kid-ney, prostaglandins are involved in maintenance of theequilibrium between vasoconstriction and vasodilata-tion of the blood vessel that supply glomerular ltra-tion. Renal damages and renal failure after chronicNSAID treatment seems to be triggered by the lack of prostaglandins in vasodilatation-induction. Gastric

damages are thought to be caused by inhibition of bothCOX isoforms ( Wallace, 1997;Wallace et al., 2000 ). Incontrast, liver damages are apparently due to buildingof reactive metabolites (e.g. acyl glucuronides) ratherthan inhibition of prostaglandins synthesis ( Bjorkman,1998 ).

The mode of action of paracetamol is not yetfully elucidated. It seems that this drugs acts mainlyby inhibiting the cyclooxygenase of the central ner-vous system and it does not have antiinammatoryeffects, because of the lack of inhibition of periph-

eral cyclooxygenase involved in inammatory pro-cesses. Adverse effects of paracetamol are mainlydue to formation of hepatotoxic metabolites, primar-ily N -acetyl- p-benzoquinone imine, synthesized whenthe availability of glutathione is diminished in livercells. Acetaminophen widely used in many anal-gesic/antipyretic medications induces proliferation of cultured breastcancercells viaestrogen receptorswith-out binding to them, but has no estrogenic activity inrodents ( Harnagea-Theophilus et al., 1999 ). The con-sequences of these observations are not clear.


15/38


In sh an inducible COX-2 homologue has beenfound to be expressed in macrophages in rainbow trout(Oncorhynchus mykiss ) and the translation product of

the COX gene was found to have a high homology of 83–84 and 77% to its human counterpart COX-2 andCOX-1, respectively ( Zou et al., 1999 ). Also in gold-sh, macrophages express a COX enzyme, which is anequivalent to mammalian COX-2 ( Zou et al., 1999 ).A COX-1 and COX-2 homologue was cloned frombrook trout ovary ( Roberts et al., 2000 ), and recently,a shark COX was cloned in dogsh Squalus acan-thias having 68 and 64% homology to mammalianCOX-1 and COX-2, respectively ( Yang and Carlson,2004 ). Prostaglandins are formed in a diverse rangeof vertebrates and invertebrates. However, in lowerinvertebrates such as corals, their synthesis is inde-pendent of COX, involving other enzymes ( Song andBrash, 1991 ). In arthropods and molluscs, COX-likeactivity is apparently responsible for the formationof prostaglandins, but these enzymes have not beenpuried and characterized ( Pedibhotla et al., 1995 ).In birds, prostaglandins play a role in the biosynthe-sis of egg shells and treatment with the COX-inhibitorindometacineresultedin eggshellthinning( Lundholm,1997 ).

5.2. Beta-blockers

Beta-blocker act by competitive inhibiting beta-adrenergic receptors and they are used in the treatmentof high blood pressure (hypertension), and to treatpatients after heart attackto prevent further attacks. Theadrenergic system is involved in many physiologicalfunctions such as regulation of the heart oxygen needand beating, vasodilatation mechanisms of blood ves-sels, and bronchodilation. Furthermore, the adrenergicsystem is also known to interact with carbohydrate and

lipid metabolisms, mainly in response to stress needssuch as starvation ( Jacob et al., 1998 ).-Adrenoceptors are 7-transmembrane receptor

proteins coupled with different G-proteins that ulti-mately enhance the synthesis of the second messengersignalingmoleculescAMP( Rangetal.,2003 ). Accord-ing to medical needs beta-blockers may selectivelyinhibit one or more -receptors types; for example 2-blockersare usedto treathypertensionavoiding cardiaceffects, since this receptor subtype is not found in theheart. Selectivity is based on difference in chemical

groups added to compounds that are able to enhancethe interactions with amino acids of the transmem-brane domains. Some of the beta-blockers (e.g. pro-

pranolol, a beta1-adrencoceptor antagonist) have theability to cause cell membrane stabilization, whereasother (e.g. metoprolol) have no membrane stabilizingactivity ( Doggrell, 1990 ). Side effects of this therapeu-tic class are mainly bronchoconstriction and disturbedperipheral circulations ( Hoffman and Lefkowitz, 1998;Scholze, 1999 ). Due to their lipophilicity they are sup-posed to pass the blood brain barrier and to act in thecentral nervous system ( Soyka, 1984, 1985 ).

-Adrenoceptors were found in sh ( O. mykiss )liver, red and white muscle with a high degree of sequence conservation with other vertebrate homo-logues. They are also supposed to play similar roleas in humans ( Nickerson et al., 2001 ). The presenceof a 2-adrenoceptor subtype was also suggested bybinding studies to occur in liver membranes of othersh and amphibians. 2-Adrenoceptors of rainbowtrout (Nickerson et al., 2001 ) show a high degree of amino-acid sequence conservation with other verte-brate 2-adrenoceptors. Frog- ( Devic et al., 1997 ) andturkey 1-adrenoceptors ( Yardeny et al., 1986 )aresim-ilar to mammalian 1-adrenoceptors. In rainbow trout,the 2-adrenoceptor gene is highly expressed in the

liver, red and white muscle, with lower expression ingills, heart, kidney and spleen ( Nickerson et al., 2001 ).Clenbuterol or ractopamine that function in mammalsas -agonist were found in rainbow trout to showa somewhat different reaction. Clenbuterol displayedonly partial agonist activities and the small effects of ractopamine may be related to low afnity for the trout

2-adrenoceptor. Agonist regulation of the trout hep-atic 2-adrenoceptors may involve down-regulation of the receptors without affecting responsiveness ( Duganet al., 2003 ). Differences in the structure and function

of the receptors may be responsible for differences inthe afnity with -blockers and mechanisms triggeredby these drugs.

Whereas mammals have three 2-adrenoceptors,ve distinct 2-adrenoceptor genes have been foundexpressed in zebrash ( Ruuskanen et al., 2005 ). Local-ization of the -adrenoceptors in zebrash showsmarked conservation when compared with mammals.The 2-adrenergic system is functional in zebrashas demonstrated by marked locomotor inhibition andlightening of skin color induced by the specic


16/38


2-adrenoceptor agonist dexmedetomidine, similar tomammals. Both effects were antagonized by the spe-cic 2-adrenoceptor antagonist atipamezole. The -

adrenoceptor agonists medetomidine and clonidineare being investigated as potential antifouling agentspreventing the settlement of barnacles on ship halls(Dahlstrom et al., 2004 ). Settlement of larvae is inhib-ited at low concentrations of 0.25–2.5 g/L.Additionalpharmacological and biochemical investigations on -and -adrenoceptors of sh and other lower organismsare needed.

5.3. Blood lipid lowering agents

There arebasically two types of antilipidemic drugs,namely statins and brates, the latter have been tar-geted analytically more often in the aquatic environ-ment than the former. Both types are used to decreasethe concentration of cholesterol (statins and brates)and triglycerides (brates) in the blood plasma. Statinsas inhibitors of cholesterol synthesis act by inhibit-ing the 3-hydroxymethylglutaril coenzyme A reduc-tase (HMG-CoA), responsible for the limiting step inthe cholesterol synthesis, namely the conversion of HMG-CoA to mevalonate ( Laufs and Liao, 1998 ). Asa consequence of the intracellular cholesterol deple-

tion, the expression of LDL receptors in hepatocytemembranes is increased and therefore, the resorptionof LDL-cholesterol from blood plasma. Due to interac-tions of statins with mevalonate metabolism, multipleadditional effects occur (antiinammatory, antioxida-tive). There is also evidence that statins affect juvenilehormone synthesis in insects ( Debernard et al., 1994 ),as uvastatin completely suppressed its biosynthesis invitro, and in the mandibular organo of lobsters ( Li etal., 2003 ).

In contrast, effects of brates are mediated, at

least in part, through alterations in transcription of genes encoding for proteins controlling lipoproteinmetabolism. Fibrates act probably by activating thelipoprotein lipase enzyme, which is mainly respon-sible for the conversion of very low density lipopro-tein (VLDL) to high density lipoproteins (HDL),decreasing therefore plasma triglycerides concentra-tion (Staels et al., 1998 ). Binding of brates to peroxi-some proliferator-activated receptors (PPARs), nuclearreceptors known to be activated during different cellu-lar pathways, stimulates the expression of several lipid

regulatory proteins such as, for example, the lipopro-tein lipase ( Staels et al., 1998 ). To date, three subtypesof PPAR have been described; PPAR is involved in

peroxisome proliferation and plays a pivotal role incontrolling hepatic lipid metabolism ( Schoonjans etal., 1996 ), whereas PPAR has diverse roles in basiclipid metabolism, and PPAR plays a key role in thedifferentiation of adipocytes ( Kersten et al., 2000 ). Het-erodimerization of PPARs with the retinoid X receptorand their binding to response elements in the promoterregions of genes leads to their activation.

Fibrates stimulate cellular fatty acid uptake, con-version to acetyl-CoA derivatives, and catabolism bythe beta-oxidation pathways, which, combined witha reduction in fatty acid and triglyceride synthesis,results in a decrease in VLDL production ( Staels et al.,1998 ). Hepatic damages may occur after chronic expo-suretobratesinrat( Quet al.,2001 ) andthis is thoughtto be related to inhibition of mitochondrial oxidativephosphorylation ( Keller et al., 1992 ). Furthermore,brates caused in rodents a massive proliferation of peroxisomes ( Hess et al., 1965 ). Strong correlationbetween brates exposure and hepatocarcinogenicityin rodents were found, while this was not observedin humans ( Cajaraville et al., 2003 ). These ndingsincrease the interest for ecotoxicological impact of this

therapeutic class of drugs.PPAR genes have been found in sh such as plaice

(Leaver et al., 1998 ) and Atlantic salmon ( Ruyter et al.,1997 ) and zebrash ( Ibabe et al., 2002 ). Fish PPARsdisplay an amino acid sequence identity of 43–48%to the human and amphibian PPAR (Andersen et al.,2000 ). All PPAR forms have been found in zebrash,and PPAR was mainly expressed in hepatocyte andtissues that catabolize high amounts of fatty acids(Ibabe et al., 2002 ). Furthermore, PPAR was shownto be induced in response to clobrate and bezabrate

in salmon hepatocytes ( Ruyter et al., 1997 ), althoughtheir PPAR seem to be less responsive than PPARof rodents ( Andersen et al., 2000 ). All three PPARreceptors were found to already been expressed in thelarval stage, with a similar tissue distribution patternto that found in adult zebrash ( Ibabe et al., 2005a ).Activators of PPAR include a variety of endoge-nously present fatty acids, leukotrienes and hydrox-yeicosatetraenoic acids and drugs, such as brates(Cajaraville et al., 2003 ). PPAR activators includefatty acids, prostaglandin A 2 and prostacyclin. PPAR


17/38


is the most selective receptor and prostaglandin J2 hasbeen described to be a specic ligand ( Ibabe et al.,2005b ). In isolated zebrash hepatocytes, mRNA of

both PPAR and PPAR was induced by clobrateat 0.5–2 mM, although to a low extent ( Ibabe et al.,2005b ). The physiological and toxicological roles of PPARs have yet to be investigated, and their involve-ment in potential effects of lipid lowering drugs is notyet known. With regard to invertebrates, no informa-tion is currently available on the existence of PPARs,although extensive searches for nuclear receptors incnidarians and platyhelminthes have been performed(Escriva et al., 1997 ).

5.4. Neuroactive compounds (antiepileptics,antidepressants)

Among the many drugs interacting with the cen-tral nerve system (CNS), only a few will be consid-ered as with respect to its occurrence in the aquaticenvironment. Antiepileptic drugs act on the CNS bydecreasing the overall neuronal activity. This can beachieved either by blocking voltage-dependent sodiumchannels of excitatory neurons (e.g. carbamazepine),or by enhancing of inhibitory effects of the GABAneurotransmitter by binding on a specic site in the

gamma subunit of the corresponding receptor (e.g.diazepam, member of benzodiazepine family) ( StudyandBarker, 1981; MacDonaldandOlsen,1994;Rogerset al., 1994 ). Evidence of the occurrence of the GABAsystem in sh ( O. mykiss , Cole et al., 1984; Meissl andEkstrom, 1991 ) was found, whereas no studies havebeen found indicating the occurrence of sodium volt-age dependent channels in sh or lower invertebrate.

Fluoxetine is a widely used antidepressant, whichacts by inhibiting the re-uptake of serotonin. This neu-rotransmitter is involved in many mechanisms, namely

hormonal and neuronal, and it is also important infunctions such as food intake and sexual behavior. Apump directs serotonin from the synapse space back to the presynapse, and selective serotonin re-uptakeinhibitors (SSRI) inhibit this pump, thus increasing theserotonin level in the synapse space. Serotonin as aneurotransmitter occurs in lower vertebrates and inver-tebrates ( Fong, 1998 ), however, the effects associatedwith this transmitter are different, and so are possiblythe effects of SSRI. Serotonin mediates, among oth-ers, endocrine functions in aquatic organisms such as

ngernail claims ( Sphaerium striatinum , Fong et al.,1998 ) and Japanese medaka ( Oryzias latipes , Fonget al., 1998; Foran et al., 2004 ). Fluoxetine and ser-

traline and the SSRI metabolites noruoxetine anddesmethylsertraline have been detected in sh sampledfrom wild in the U.S., and therefore reect a bioaccu-mulation potential ( Brooks et al., 2005 ). Whether theaccumulated levels of 1.6 ng/g uoxetine and 4.3 ng/gsertraline found in brain have effects on the nervoussystem of sh has yet to be investigated.

5.5. Cytostatics compounds and cancer therapeutics

Another potential interesting class of compoundis represented by cytostatic pharmaceuticals interact-ing with cell proliferation. There are different modesof actions of the different compounds. For examplemethotrexate acts as a potent inhibitor of the folatedehydroreductase enzyme, which is responsible forthe purine and pyrimidine synthesis ( Schalhorn, 1995;Rang et al., 2003 ). Doxorubicin is an intercalating sub-stance inducing DNA-strand brakes (in humans, heartarrhythmiamaybe a side effect).Tamoxifen asanantie-strogenic drug is used for breast cancer treatment andacts by competitive inhibiting the estrogenic receptor

at least in mammary gland ( Rang et al., 2003 ).

5.6. Various compounds

Cimetidine and ranitidine are compounds, whichactby inhibiting the histamine receptors type 2 in the gas-tric system, thus inhibiting the acid secretion (antacid).These drugs are used to treat gastric ulceration. SinceH2-histamine receptorsarefound also in thebrain,bothdrugs may elicit central nervous system reactions andside effects ( Cannon et al., 2004 ). Peitsaro et al. (2000)

demonstrated the presence of H 3-histamin receptors incentral nervous system of zebrash ( Danio rerio ), butthe lack of histamine in the periphery of this sh wasalso reported. However, interspecies differences mayoccur; cod and carp seem to have histamine and H 2-receptors in the periphery ( Peitsaro et al., 2000 ).

Metformin is an antidiabetic agent, which mecha-nisms of actions are not well understood. It seems thatthis drugs acts by increasing the cellular use of glucoseand inhibiting the gluconeogenesis. Metformin seemsto act on insulin receptor by direct stimulation of the


18/38


insulin receptor or indirectly through inhibition of tyro-sine phosphatase ( Holland et al., 2004 ).

6. Ecotoxicological effects

Pharmaceuticals are designed to target specicmetabolic and molecular pathways in humans and ani-mals, but they often have important side effects too.When introduced into the environment they may affectthe same pathways in animals having identical or sim-ilar target organs, tissues, cells or biomolecules. Asshown above, certain receptors in loweranimals resem-ble those in humans, others however, are different orlacking, which means that dissimilar modes of actionsmay occur in lower animals. It is important in thisrespect to recognize that for many drugs, their specicmodes of actions are not well known and often not onlyone, but many differentmodes of actions occur. Amongother reasons, this makes specic toxicity analysis inlower animals difcult to perform. Despite this, toxi-city experiments should be targeted and designed forspecic targets of thepharmaceuticaleven in lowerver-tebrates and invertebrates, based on the hypothesis of similarity of modes of actions. However, current toxi-city testing is not designed in this way, rather general

and established test systems and traditional organismsaccording to guidelines are being used and traditionalend points such as mortality are assessed.

Thus far, ecotoxicity testing merely providedindications of acute effects in vivo in organisms of different trophic levels after short-term exposure,and only rarely after long-term (chronic) exposures.These data are ultimately used for ecological risk assessments. Because of animal welfare and screen-ing purposes, in vitro analyses are becoming moreimportant, but they are not sufcient for assessing the

toxicological proles of a compound, particularly as abasis for risk analysis ( Fent, 2001 ). Beyond laboratoryinvestigations, some mathematical models were devel-oped to estimate or predict ecotoxicological effects.The most often applied quantitative structure–activityrelationship (QSAR) program is ECOSAR (onlinehttp://www.epa.gov/oppt/newchems/sarman.pdf )(Sanderson et al., 2004b ). Despite serious drawbackssuch as an inadequate structure coverage for pharma-ceuticals, the program has been repeatedly applied toestimate pharmaceutical baseline toxicities ( Jones et

al., 2002; Sanderson et al., 2004b; Cleuvers, 2005 ).Both methods are helpful in estimating potential toxic-ity or the behavior of a compound in the environment,

but they cannot replace in vivo or in vitro assays.The current literature about ecotoxicological effectsof human pharmaceutical deals mainly with the acutetoxicity in standardized tests and it is generally focusedon aquatic organisms. The inuence of environmen-tal parameters such as pH on toxicity has only rarely,or not yet been investigated. Such studies would beof importance for instance for acidic pharmaceuticalsthat may induce different toxicities depending on spe-ciation at different ambient pH. Moreover, effects of drug metabolites have rarely been investigated. Pho-totransformation products of naproxen, for instance,showed higher toxicities than the parent compound,while genotoxicity was not found ( Isidori et al., 2005 ).At contaminated sites, aquatic life is exposed over theentire life cycle to these compounds. Chronic effectsare less investigated and often even related to relativeshort-term exposures. However, long-term exposuresare needed for an accurate environmental risk assess-ment. Here we summarize the current ecotoxicologicaldata, focusing on specic modes of action of differ-ent therapeutic classes of pharmaceuticals, and cover-ing many differences in methods, species and time of

exposure. These data are then related to environmentallevels in order to assess the potential hazard for the dif-ferent classes of pharmaceuticals and identify currentresearch and knowledge gaps.

6.1. Acute effects

Pharmaceuticals areassessed for their acute toxicityby traditional standard tests according to establishedguidelines (e.g. OECD, U.S. EPA, ISO) using estab-lished laboratory organisms such as algae, zooplankton

and other invertebrates and sh. Acute toxicity data of pharmaceuticals were compiled by Halling-Sorensenet al. (1998) and Webb (2001) , whereby in the latter, alist of about 100 human pharmaceuticals from differentsources is given. By comparingdifferent trophic levels,Webb (2001) suggested that algae were more sensitiveto the listed pharmaceuticals than Daphnia magna , andsh were even less sensitive. However, such general-izations do not focus enough on the different modesof actions of a given pharmaceutical, and hence, dif-ferences in toxicity in different phyla. In the attempt

http://www.epa.gov/oppt/newchems/sarman.pdfhttp://www.epa.gov/oppt/newchems/sarman.pdf


19/38


to compare the different classes of pharmaceuticals interms of acute toxicity, Webb (2001) noted that themost toxic classes were antidepressants, antibacterials

and antipsychotics, but the range of responses withineach of these categories was large, typically severalorders of magnitude. In our present review, we provideand summarize additional and new data and discuss itsecotoxicological relevance covering different classesof human pharmaceuticals. The data originate fromdifferent sources, and studies were performed underdifferent quality criteria (i.e. nominal versus mea-sured exposure concentrations), making comparisonsdifcult.

6.1.1. Analgesics and non-steroidalantiinammatory drugs (NSAID)

In general, toxicity data vary for each pharma-ceutical, however, diclofenac seems to be the com-pound having highest acute toxicity within the classof NSAID, since for all the tests performed the effect

concentrations were below 100 mg/L ( Fig. 2). Short-term acute toxicity was analyzed in algae and inver-tebrates ( Webb, 2001; Cleuvers, 2003 ), phytoplank-

ton was found to react more sensitive [lowest EC 50(96h)=14.5mg/L( Ferrari et al., 2004 )] thanzooplank-ton [lowest EC 50 (96 h)= 22.43 mg/L ( Ferrari et al.,2004 )]. There is no correlation between the acute toxi-city in Daphnia and the lipophilicity as represented bylog K ow (Fig. 3). In general, not much is known aboutthe acute toxicity to sh.

6.1.2. Beta-blockersAs shown in Fig. 2, the acute toxicity of beta-

blockers is not extensively studied, with the excep-tion of propranolol. This compound shows the high-est acute toxicity and highest log K ow as compared toother beta-blockers ( Fig. 3). This and the fact that itis a strong membrane stabilizer, whereas other inves-tigated beta-blockers are not, may in part explain itshigher toxicity ( Doggrell, 1990; Huggett et al., 2002 ).

Fig.2. Acutetoxicity of 24 differentpharmaceuticals,belongingto different therapeutic classes to aquatic organisms. EC 50 andLC50 for differentorganisms and different endpoint and exposure time are summarized. See text for details. References : Calleja et al. (1993, 1994) , Lilius et al.(1994) , Henschel et al. (1997) , Fong (1998) , Halling-Sorensen et al. (1998) , Webb (2001) , Huggett et al. (2002) , Brooks et al. (2003) , Cleuvers(2003, 2004) , Villegas-Navarro et al. (2003) , Ferrari et al. (2004) , Henry et al. (2004) , Hernando et al. (2004) , Kümmerer (2004) , Marques et al.(2004a,b) , Nunes et al. (2004) and Isidori et al. (2005) .


20/38


Fig.3. Relation betweenacute toxicity(LC 50 ) of analgesics andanti-inammatory drugs( y= 0.0082 x − 0.0034; R2 = 0.1202; ANOVA notsignicant) (a) and -blockers ( y= 0.1386 x − 0.1709; R2 = 0.4301;ANOVA signicant; p < 0.02) (b) and octanol–water partition coef-

cients of the compounds (log K ow ); calculated and measured valuesare given in different symbols. Acute toxicity of Daphnia magnarefers to immobilization after 48 h (LC 50 value). References —acutetoxicity: Calleja et al. (1993) , Lilius et al. (1994) , Henschel etal. (1997) , Halling-Sorensen et al. (1998) , Huggett et al. (2002) ,Brooks et al. (2003) , Cleuvers (2003) , Villegas-Navarro et al. (2003) ,Cleuvers (2004) , Ferrari et al. (2004) , Hernando et al. (2004) ,Marques et al. (2004a,b) . log K ow , in between parentheses: acetyl-salicylic acid (1.13) ( Sanderson et al., 2003 ); salicylic acid (2.26)(Hansch et al., 1995 ); diclofenac (4.51), ibuprofen (3.97) ( Avdeef et al., 1998 ); naproxen (3.18) ( Cleuvers, 2004 ); paracetamol (0.49)(Henschel et al., 1997 ); atenolol (0.5) ( Grifn et al., 1999 ); betaxolol(2.98)( Sandersonet al., 2003 ); metoprolol (2.15), propranolol (3.56)(Hardman et al., 1996 ); sotalol (0.24) ( Hansch et al., 1995 ).

Comparison of toxicity is difcult in this case, sinceother beta-blockers, except metoprolol, were only ana-lyzed in D. magna (Hernando et al., 2004 ). Meto-prolol and verapamil caused the acceleration of theheart beat rate at low concentration, but lowered itat high concentrations in D. magna (Villegas-Navarroet al., 2003 ). For propranolol it seems that phyto-and zooplankton are more sensitive than sh. Ceri-odaphnia dubia [EC50 (48h)=0.8mg/L; Ferrari et

al., 2004 ] displayed higher sensitivity than D. magna[EC50 (48h) = 1.6 mg/L; Huggett et al., 2002 ] orother zooplankton organisms. Within phytoplankton,

the microorganism Synechococcus leopolensis reactedmost sensitive [EC 50 (96 h)= 0.668 mg/L; Ferrari et al.,2004 ].

6.1.3. Blood lipid lowering agentsSimilar to beta-blockers, acute toxicity of lipid low-

ering agents is not extensively reported. Clobrateshowed LC 50 values in the range of 7.7–39.7mg/Land can be classied as harmful to aquatic organisms.The sh Gambusia holbrooki [LC50 (96h) = 7.7 mg/L;Nunes et al., 2004 ] seems the most sensitive organ-ism to acute clobrate concentrations studied so far.The known rodent peroxisome proliferator gembrozilinjected to rainbow trout led to signicant increasesin fatty acyl-CoA oxidase (FOA) activity at dosesof 46–152mg/kg/day ( Scarano et al., 1994 ). Signi-cant dose-related increases in peroxisomal FOA wereobserved after exposure of rainbow trout primary hep-atocytes to clobric acid, and ciprobrate, but not withgembrozil ( Donohue et al., 1993 ). Thein vitro activityin these shes is weak.

6.1.4. Neuroactive compounds (antiepileptics,

antidepressants)The serotonin re-uptake inhibitor uoxetine is

apparently the most acute toxic human pharmaceu-tical reported so far with acute toxicity rangingfrom EC 50 (48 h, alga) = 0.024 mg/L ( Brooks et al.,2003 ) to LC50 (48h)=2mg/L ( Kümmerer, 2004 ).For benthic organisms, acute toxicity is in the rangeof 15–43mg/kg sediment [ Chironomus tentans LC50(10 days) = 15.2mg/kg, Hyalella azteca LC50 (10days) = 43 mg/kg; Brooks et al., 2003 ]. Fluoxetineseems to stronger affect phytoplankton than other

aquatic organisms.Diazepam and carbamazepine, both antiepileptics,can be classied as potentially harmful to aquaticorganisms, because most of the acute toxicity data arebelow 100mg/L. For both compounds it seems that D.magna is affected more than other species, but the rea-sons for the higher susceptibility is not known. Acutetoxicity of carbamazepine was found at 17.2mg/L in Daphnia and at 34.4mg/L in midges, but growth wasinhibited at 12.7 mg/L in Daphnia and at 9.2mg/L inmidges (Thaker, 2005 ).


21/38


6.1.5. Cytostatic compounds and cancer therapeutics

Acute toxicity of methotrexate on highly prolif-

erative species, namely the ciliate Tetrahymena pyri- formis , indicated acute effects [EC 50 (48 h)= 45mg/L;Henschel et al., 1997 ]. Teratogenicity in sh embryoswas observed at even higher concentrations [EC 50(48h)=85mg/L; Henschel et al., 1997 ].

The acute toxicity data summarized in Fig. 2 showsthat 17% of the pharmaceuticals displayed an acutetoxicity below 100mg/L, and for uoxetine, all toxic-ity values were below 10mg/L. On the other hand, 38%of the pharmaceuticals such as acetylsalicylic acid,betaxolol, sotalol,bezabrate, gembrozil, bezabrate,cimetidine and ranitidine displayedLC 50 values higherthan 100mg/L, which, according to EU-Directive93/67/EEC (Commission of the European Commu-nities, 1996), are classied as not being harmful foraquatic organisms. The other pharmaceuticals (45%)displayed a considerable variability of acute toxicityvalues, spreading over a wide range, thus making aclassication difcult.

Variability of data both withinthesame andbetweendifferent species is obvious. Different actual exposureconcentrations (only nominal concentrations wereusedin the determination of the endpoints), different sen-

sitivities of used clones, different laboratory perfor-mances are among the reasons for variability withinthe same species (for example, clobric acid toxicityin D. magna varies between 72 and 200 mg/L; the LC 50(48 h) of acetylsalicylic acid varies between168 mg/L(Calleja et al., 1994 ) and 1468mg/L ( Lilius et al.,1994 ); the LC50 (24 h) of diazepam varies between9.6 mg/L ( Calleja et al., 1993 )and10000mg/L( Callejaet al., 1994 )). Depending on the quantity and qualityof data, ranges of acute toxicity values span one totwo orders of magnitude, in some cases such as pro-

pranolol or diazepam, the species differences are quitelarge,spanning three tofour ordersof magnitude. Whendifferent categories are compared, a tendency of lowerLC50 (EC50) values is found for beta-blockers andneu-roactive drugs as compared to antiinammatory drugsor various other compounds.

Often, acute toxicity is related to non-specic modeof actions, and not to mechanisms involving spe-cic target molecules. The compounds are thought tointeract with cellular membranes leading to unspecicmembrane toxicity. This general mechanism may be

only one, additional ones (e.g. oxidative stress) comeintoplaywithparticular pharmaceuticals. We evaluatedwhether the acute toxicity data of the different classes

of pharmaceuticals correlate with the log K ow of thecompound, as the lipophilicity determined by log K owis an important parameter formembrane toxicity. How-ever, no correlation was found between the log K ow of pharmaceuticals of a certain category or of all phar-maceuticals, and the acute toxicity either of a certainspecies, a group of organisms, or all of them. Thebest relation between measured and estimated log K owof one class of pharmaceuticals and acute toxicity inone species, D. magna , is depicted in Fig. 3. Rea-sons for the variability of the data are probably basedon laboratory differences, nominal concentration dif-ferences, clone susceptibility differences, but also onthe fact that log K ow may not be the best model forlipophilicity. Thisholds in particular for ionizablecom-pounds, where the pH-dependent speciation is of sig-nicant inuence ( Fent andLooser, 1995; Looser et al.,1998 ).

In conclusion, acute toxicity to aquatic organisms isunlikely to occur at measured environmental concen-trations, as acute effects concentrations are 100–1000times higher than residues found in the aquatic envi-ronment. For example, the lowest acute effect concen-

tration of uoxetine was 20 g/L, whereas the highestestimated environmental concentration was 0.01 g/L;the lowest acute effect of salicylic acid was 37 mg/L,whereas the highest environmental concentration was∼60 g/L. Therefore, acute toxicity is only relevant incase of spills.

6.2. Chronic effects

Many aquatic species are continuouslyexposedoverlong periods of time or even over their entire life

cycle. Evaluation of the chronic potential of micro-pollutants, e.g. pharmaceuticals, is therefore impor-tant. However, there is a lack of chronic data, andwhere available, chronic toxicity is marginally known.The available chronic data do often not investigatethe important key targets, nor do they address thequestion in different organisms. Toxicity experimentsare usually performed according to established guide-lines. More specic investigations including analysisof possible targets of the pharmaceutical, or over dif-ferent life stages, are lacking, or have only rarely


22/38


been performed. Moreover, life-cycle analyses are notreported, except for EE2 ( Länge et al., 2001; ParrottandBlunt, 2005 ), andtoxicityto benthic andsoil organ-

isms have very rarely been evaluated. In this chapter,we review the current literature according to the differ-ent pharmaceutical classes and summarize the data inFig. 4.

The best knowledge exists for the synthetic steroid EE2 contained in contraceptive pills, showing estro-genic effects at extremely low and environmentallyrelevant concentrations. This steroid has been shown

Fig. 4. Chronic toxicity of 10 different pharmaceuticals, belongingto different therapeutic classes.Given are lowestobserved effect con-centrations (LOEC) and no observed effect concentrations (NOEC)for different aquatic organism, different endpoints and exposuretimes. See text for details. References : Webb (2001) , Huggett et al.(2002) , Brooks et al. (2003) , Ferrari et al. (2003, 2004) , Cleuvers(2004) , Henry et al. (2004) , Kümmerer (2004) , Marques et al.(2004a,b) , Schwaiger et al. (2004) and Triebskorn et al. (2004) .

in many sh to induce estrogenic effects at extremelylow concentrations. Exposure of fathead minnowsover their life cycle indicates reproductive effects at

low concentrations of EE2 ( Länge et al., 2001 ). TheNOEC values of the F 0 generation F 1 embryo hatch-ing success and larval survival were ≥ 1 ng/L. Malesh exposed to EE2 at 4 ng/L failed to develop nor-mal secondary sexual characteristics and the sex ratiowas altered. No testicular tissue was observed in anysh exposed to EE2 at 4 ng/L. A recent study showsvitellogenin (VTG) induction in fathead minnows withan EC50 value as low as 1 ng/L; EE2 was 25–30 timesmore potent than estradiol ( Brian et al., 2005 ), con-rming previous reports on VTG induction at con-centrations between 0.1 and 1 ng/L ( Pawlowski etal., 2004 ). Decreased egg fertilization and sex ratio(skewed toward females), both of which were sig-nicantly affected at extremely low concentrations of 0.32 ng/L EE2 ( Parrott andBlunt,2005 ). Thenext mostsensitive parameter was demasculinization (decreasedmale secondary sex characteristic index) of malesexposed to an EE2 concentration of 0.96 ng/L. Fulllife-cycle exposure of zebrash to 3 ng/L EE2 leadto elevation of VTG and caused gonadal feminiza-tion in all exposed sh and thus inhibited reproduction(Fenske et al., 2005 ). Life-long exposure of zebrash

to 5ng/L in the F 1 generation caused a 56% reduc-tion in fecundity and complete population failure withno fertilization. Infertility in the F 1 generation wasdue to disturbed sexual differentiation with males hav-ing no functional testes and intersex gonads ( Nashet al., 2004 ).

In hazard and risk assessment, the ratio betweenacute to chronic toxicity is often taken for evalua-tion of chemicals. For pharmaceuticals, this is difcult,because only very rarely, a systematic analysis of agiven drug in both acute and chronic toxicity in a sin-

gle species is performed. Apart from EE2, there areonly a few NSAID, from which acute to chronic ratioscan be deduced. Table 3 shows that even for similardrugs, these ratios in D. magna vary by two orders of magnitude. For all other drugs, only partial informa-tion is available on a given species. Ratios derived onthe basis of a number of different species are not accu-rate, giving questionable information. The examples inTable 3 conrm again that chronic toxicity cannot bederivedfromacutetoxicity by simple calculations. Thisis often neglected in risk assessment.


23/38


Table 3Ratio between acute and chronic toxicity in Daphnia magna andCeriodaphnia dubia (48 h/21days)

Drug Acute (mg/L) Chronic (mg/L) Ratio

Acetylsalicylic acid 1293.1 1.4 924Salicylic acid 1031.7 13.3 77Clobrate 28.2 0.01 2820Naproxen 66.4 0.33 201Naproxen Na 43.6 0.68 64

Data after Marques et al.(2004a,b) (acetylsalicylic acid and salicylicacid, D. magna ), Webb (2001) (clobrate, D. magna ) and Isidori etal. (2005) (naproxen and naproxen Na, Ceriodaphnia dubia ).

6.2.1. Analgesics and non-steroidalantiinammatory drugs

NSAID inhibit the synthesis and release of prostaglandins via COX inhibition and these com-pounds are the most consumed category of drugs.About NSAID commonly found in the aquatic envi-ronment, most chronic data are reported. Acetylsali-cylic acid affected reproduction in D. magna and D.longispina at concentrations of 1.8 mg/L ( Marques etal., 2004a ). Diclofenac is commonly found in wastew-ater at median concentration of 0.81 g/L (Ternes,1998 ) whereas the maximal concentration in wastewa-ter and surface water is up to 2 g/L (Stumpf et al.,1996; Ternes, 1998; Schwaiger et al., 2004 ). Tradi-

tional chronic toxicity studies with diclofenac werereported in invertebrates ( Ferrari et al., 2003, 2004 ).A recent study demonstrated chronic histopathologi-cal effects in rainbow trout after 28 days of exposure.At the LOEC of 5 g/L renal lesions (degeneration of tubular epithelia, interstitial nephritis) and alterationsof the gills occurred in rainbow trout ( Schwaiger etal., 2004 ), and subtle subcellular effects even at 1 g/L(Triebskorn et al., 2004 ). Impairment of renal and gillfunction is likely to occur after long-term exposure.The kidney was also found to be a target of diclofenac

in vultures, acute renal failure was probably the rea-son for the visceral gout ( Oaks et al., 2004 ) and theoccurrence of extensive deposits of uric acid on andwithin internalorgans( Gilbert et al., 2002 ). In zebrashembryos, no effect of diclofenac on embryonic devel-opment wasobserved,except delayed hatching at 1 and2 mg/L (Hallare et al., 2004 ). Additional side effects of diclofenac have been observed in humans in the liverwith degenerative and inammatory alterations ( Bankset al., 1995 ), in lower gastrointestinal tract and in theesophagus ( Bjorkman, 1998 ), but not in sh.

6.2.2. Beta-blockersAs sh contain 2-receptors in heart and liver

(Gamprel et al., 1994 ) and probably in reproductive

tissues (Haider and Baqri, 2000 ), unspecic antago-nists such as propranolol may be active in sh. Infact, propranolol indicated chronic toxicity not onlyon the cardiovascular system, but also on reproduc-tion. The no-observed-effect-concentration (NOEC)and lowest-observed-effect-concentration (LOEC) of propranolol affecting reproduction in C. dubia were125 and 250 g/L, and reproduction was affected after27 days of exposure in H. azteca at 100 g/L (Huggettet al., 2002 ). In sh O. latipes , signicant changes inplasma steroid levels occurred after 14 days of expo-sure. The number of eggs released by sh was reducedat 0.5 g/L after a 4-week exposure to 0.5 and 1 g/L,but not at 50 and 100 g/L (Huggett et al., 2002 ).No alteration in vitellogenin levels was observed. Itwas suggested that alteration in sex steroids let todecreased oxytocin excretion, which could decreasethe number of eggs released. Propranolol was also ana-lyzed in invertebrates. LOEC and NOEC for differentorganisms span several orders of magnitude ( Fig. 4),partly due to differences between laboratories, but alsospecies differences. These data should be compared tothe environmental concentrations; propranolol, meto-

prolol and nadolol were identied in U.S. wastewa-ter samples up to 1.9, 1.2 and 0.36 g/L, respectively(Huggett et al., 2002 ).

6.2.3. Blood lipid lowering agentsData on this class of compounds are rare. Fibrates

have been evaluated by traditional toxicity tests. Thefollowing NOEC were found for clobric acid in C.dubia [NOEC (7 days) = 640 g/L], the rotifer B. caly-ciorus [NOEC (2 days) = 246 g/L], and in earlylife stages of zebrash [NOEC (10 days) = 70 mg/L]

(Ferrari et al., 2003 ). Gembrozil occurred in bloodplasma of goldsh after exposure over 14 days at 113-times higher levels than in water (bioconcentrationfactor of 113). Plasma testosterone was reduced byover 50% after exposure to 1.5 and 10 mg/L, as wellas levels of steroid acute regulatory protein transcriptin goldsh testes ( Mimeault et al., 2005 ).

6.2.4. Neuroactive compoundsMost data were reported for the antiepileptic carba-

mazepine and selective serotonin re-uptake inhibitors


24/38


(SSRI), other neuroactive compounds were very rarelyor not evaluated ( Fig. 4). Traditional toxicity testsshowed chronic toxicity of carbamazepine in C. dubia

[NOEC (7 days)=25 g/L], in the rotifer B. caly-ciorus [NOEC (2 days) = 377 g/L], and in earlylife stages of zebrash [NOEC (10 days) = 25 mg/L](Ferrari et al., 2003 ). Carbamazepine is considered car-cinogenic in rats but is not mutagenic in mammaliancells. Sublethal effects occurred in Daphnia at 92 g/Land the lethal concentration in zebra sh was 43 g/L(Thaker, 2005 ). In a study with the cnidarian Hydravulgaris , diazepam was shown to inhibit polyp regen-eration at 10 g/L (Pascoe et al., 2003 ).

Most chronic studies focused on SSRI. Serotonin isa neurotransmitter found in lower vertebrates and inver-tebrates, and SSRI may adversely inuence the func-tion of the nervous and associated hormonal systemsof these organisms as well. Besides having importantfunctions as a neurotransmitter, serotonin may directlyact on the immune system, alters appetite, inuencesbehavior and modulates sexual function. The roleof serotonin in reproduction varies between differentphyla and effects of SSRI as well. Fong (1998) f oundthat SSRI (uvoxamine, paroxe

ecotoxicology of human pharmaceuticals

Documents