You are on page 1of 7

Environmental Toxicology and Chemistry, Vol. 25, No. 1, pp.

265–271, 2006
q 2006 SETAC
Printed in the USA
0730-7268/06 $12.00 1 .00




Department of Environmental Science and Engineering, International Environmental Research Center, Gwangju Institute of Science and
Technology, 1 Oryong-dong, Buk-gu, Gwangju 500-712, Korea

( Received 17 March 2005; Accepted 16 June 2005)

Abstract—The overall ecotoxicological effect of pharmaceutically active compounds (PhACs) detected in the effluents of Korean
wastewater treatment plants (WWTPs) to Daphnia magna was investigated using biological and chemical analyses. The bioassay
results showed median lethal concentrations and no-observed-effect concentrations ranging from a few to tens of ppm levels for
nine PhACs in 48-h acute and 21-d chronic tests. The mixture effects of pharmaceuticals also were examined by other acute and
chronic tests, which showed no significant toxicity despite a slightly increased combined effect of approximately twofold. The
residual concentrations of nine PhACs were analyzed at concentrations ranging from 10 ng/L to 89 mg/L in the influents and from
10 ng/L to 11 mg/L in the effluents from four metropolitan cities in South Korea between January and November of 2004. Through
repeated investigations of the influents and the effluents from different WWTPs, relatively higher removal efficiencies (23.9–91.3%)
compared with those of previous surveys performed in other countries were observed for most pharmaceuticals, with the exception
of acetaminophen (8.7%). The present study showed no significant risk effects of the effluents from WWTPs containing pharma-
ceuticals (i.e., hazard quotient , 1), even at the 95th percentile contamination range, although a risk assessment factor of 1,000
was applied. Therefore, it can be concluded that the potential risk of pharmaceuticals should be monitored carefully with more
bioassay data, because many uncertainties still exist in the determination and toxicity of metabolites in water environments. No
significant risk was observed, however, from the selected PhACs in the effluents from WWTPs discharged into surface waters.

Keywords—Pharmaceuticals Wastewater treatment plants Toxicity Risk assessment

INTRODUCTION aquatic environments [7–13]. Many metabolites, such as hy-

droxy- and carboxy-ibuprofen and carboxy-hydrotrophic acid
Although large quantities of pharmaceuticals are produced
for ibuprofen [11,12] and 2-(4-chlorophenoxy)-2-methyl pro-
and used as animal and human medicines, little attention has
pionic acid for clofibric acid [3], have been found in raw
been paid to the potential risks of pharmaceuticals as toxic
sewage water and rivers. Ibuprofen and its metabolites are
contaminants in aquatic environments. Pharmaceuticals have degraded efficiently by biological treatment in WWTPs,
similar physicochemical behaviors and cause biological effects whereas diclofenac and clofibric acid are degraded only min-
toward nontarget organisms in the environment (i.e., human imally by biological reaction [3,5,7]. More than 90% of the
and animals) even at low concentrations [1,2]. Recent studies diclofenac in Lake Greifensee, Switzerland, was removed by
have demonstrated that some metabolites are more lipophilic photolytic degradation [7]. Some of the other substances de-
and more persistent than the original drugs from which they tected in rivers include antipyretics (i.e., acetaminophen [5,6]),
were derived [3]. These substances are excreted or discarded b-sympathomimetics [14], antibiotics (e.g., erythromycin
from both humans and domestic animals as a mixture of me- [15,16]), as well as antiepileptics (e.g., carbamazepine
tabolites and unchanged forms via the municipal sewage sys- [12,17,18]) and hypolipidemics (i.e., gemfibrozil [5,6,19]). De-
tem to rivers, bank filtrates, and groundwaters. This implies spite many studies, it remains unclear whether the exposure
that their parent compounds and metabolites may exist in of aquatic biota to these pharmaceuticals have direct or indirect
drinking water through several drug-fate pathways. The bal- effects on the environments. Furthermore, no data are available
ances of the influents and effluents of drug residues detected from Korea, which uses large quantities of drugs because of
in wastewater treatment plants (WWTPs) reveal that many the high population densities and expanding sizes of its cities.
pharmaceuticals are not eliminated completely by traditional The objectives of the present study were to monitor the
treatment processes (i.e., coagulation, flocculation, and sedi- extent of drug residues in Korean WWTPs and to investigate
mentation) [4]. the ecological risks of target compounds. First, the acute and
Recently, major efforts to clarify the ecological influence chronic toxicity of pharmaceuticals, both singularly and as
and occurrence of pharmaceuticals in the aquatic environment mixtures, were evaluated for Daphnia magna. Using ecotox-
have been attempted in North American and European coun- icological results, environmental risk assessments were per-
tries [1,3,5,6]. Diclofenac, ibuprofen, and clofibric acid are formed for drug residues in Korean rivers receiving waters
medical agents used widely in both human and domestic an- from WWTP effluents to predict their potential adverse effects.
imal practices. These pharmaceuticals have been detected fre- MATERIALS AND METHODS
quently at concentrations up to several ng/L or mg/L levels in Sample collection
Samples were collected between January and November of
* To whom correspondence may be addressed ( 2004 using precleaned, amber-glass containers during periods
266 Environ. Toxicol. Chem. 25, 2006 G.H. Han et al.

Table 1. Overview of pharmaceutically active compounds

Compounds CASa no. Molecular weight Molecular structure Therapeutic use

Diclofenac 15307-79-6 318.1 Nonsteroidal anti-inflammatory

Ibuprofen 15687-27-1 206.3 Nonsteroidal anti-inflammatory

Clofibric acid 882-09-7 214.7 Lipid regulator (active metabolite)

Carbamazepine 298-46-4 236.3 Antiepileptic

Salicylic acid 69-72-7 138.1 Analgesic, antipyretic

Gemfibrozil 25812-30-0 250.3 Lipid regulator

Acetaminophen 103-90-2 151.2 Analgesic antipyretic

Bezafibrate 41859-67-0 361.8 Antihyperlipidemic

Tolfenamic acid 13710-19-5 261.7 Nonsteroidal anti-inflammatory

a CAS 5 Chemical Abstracts Service.

of normal WWTP operation. Big cities, including Seoul, Bus- spectrometry (GC-MS; Shimadzu, Kyoto, Japan) in full-scan
an, Gwangju, and Daegu, were selected as study areas because or selected ion-monitoring mode. Separations were carried out
of their high population densities, which were assumed to have using a DB5-type capillary column (length, 30 m; inner di-
exposure to large amounts of pharmaceuticals to their aquatic ameter, 0.25 mm; film thickness, 0.25 mm) with helium
environments because of the discharge of effluents. Treatments (99.999%) as the carrier gas (flow rate, 1.1 ml/min) and the
at all WWTPs consisted of three main steps: Preliminary clar- flow velocity set to 38.1 cm/s in the constant-flow mode. In-
ification, final clarification, and an aeration tank. Gwangju2 jections were performed in the split mode (purge time, 2 min)
had additional treatment steps for phosphate removal and de- with an injection volume of 1 ml. The GC oven was pro-
nitrification treatment. The WWTP in Busan also was equipped grammed as recommended by Ternes et al. [21]. The methyl
with sand filtration, followed by an ultraviolet disinfection esters of the analytes were quantified by monitoring the ions
step, after the final clarification. Water samples were stored in at m/z (mass to charge ratio) 214 and 242 for diclofenac;
an icebox for their return to the laboratory, adjusted to pH less m/z 161 for ibuprofen; m/z 128 for clofibric acid; m/z 80, 109,
than 2 by the addition of concentrated HCl, and preserved at and 151 for acetaminophen; m/z 95, 165, and 193 for carba-
a temperature less than 48C before analysis. Samples were mazepine; m/z 92, 120, and 138 for salicylic acid; and m/z
filtered through glass-fiber filters (0.45 mm) and sucked 122 and 107 for gemfibrozil. All samples were analyzed by
through solid-phase extraction cartridges containing 500 mg five-point calibration over the whole procedure. Using this
of Sep-Pakt Vac 6cc C18 (Waters, Milford, MA, USA) at a method, limits of detection of 10 and 250 ng/L were obtained
rate of approximately 8 ml/min. Cartridges were dried for 2 for the sewage influents and effluents, respectively. The target
h under a moderate stream of nitrogen and then extracted by pharmaceuticals used in the bioassay experiments were se-
elution with three successive 1-ml aliquots of high-perfor- lected with respect to their occurrence, potential ecological
mance liquid chromatography (HPLC)–grade methanol [20]. effect, and broad application (Table 1). Individual stock so-
After elution, the samples were evaporated to a final 1-ml lutions of each target pharmaceutical were prepared by dis-
volume. solving in ethanol for 1 d before use for each experiment, with
the concentration of the solvent in the medium, including the
Analytical procedure control groups, being less than 1%.
All solutions of pharmaceutically active compounds
(PhACs) were prepared in HPLC-grade methanol, purchased Test organisms and culture
from Sigma-Aldrich Chemical (St. Louis, MO, USA), for anal- The test organism used in the present study, D. magna,
ysis. Chemicals were analyzed by gas chromatography–mass was obtained from the Korea Institute of Toxicology (Daejeon,
Occurrence and toxicity of pharmaceuticals from Korean WWTPs Environ. Toxicol. Chem. 25, 2006 267

South Korea). Daphnia magna was cultured with Selenastrum ber, were assigned for each treatment under the same test con-
capricornutum and a mixture of yeast, CEROPHYLL t ditions as used for a single compound.
(Ward’s, Rochester, NY, USA), and trout chow (YCT) supplied
as food. The organisms were cultured and handled according Statistical analysis
to the procedures outlined in the U.S. Environmental Protec- The endpoints of the toxicity tests using D. magna are based
tion Agency (U.S. EPA) manual [22]. Neonates of D. magna on adverse effects on survival and reproduction. The LC50s
less than 24 h old were used for acute and chronic tests, which for the acute tests were calculated using regression analysis
were illuminated with a 16:8-h light:dark photoperiod. Or- following probit analysis. The Dunnett method of pair-wise
ganisms were not fed during the test periods. Reconstituted separation was used to determine the NOEC in the 21-d chronic
hard water (MgSO4, 9.98 3 1024 M; CaSO4, 6.98 3 1024 M; test. The experimental response used in the statistical analysis
NaHCO3, 2.28 3 1023 M; and KCl, 1.07 3 1024 M), with a was the total number of neonates produced until the end of
hardness of 170 6 5 mg/L (mean 6 SD), an alkalinity of 110 the experiment or at the time of death. An animal that died
6 5 mg/L as CaCO3 mg/L, and pH 7.8 6 0.2 at 258C through- before reproducing neonates was included in the analysis, with
out the study. zero entered as the number of young produced [23].

Single-compound bioassay tests Risk assessment

For 48-h static nonrenewal acute toxicity tests, D. magna Ecological risk assessments were performed to characterize
was exposed to a control and eight different concentrations of the degree of contamination and to evaluate the adverse effects
PhACs. Four replicates, containing five organisms per 30-ml of these chemicals in real aquatic environments [24]. In the
test chamber, were assigned to the treatments and control. Test present study, results from the D. magna toxicity tests were
organisms were fed YCT and S. capricornutum for 2 h before compared to those from the GC-MS analyses of selected com-
the test [22]. Mortality was recorded 48 h later. pounds in the influents and effluents of the WWTPs. The haz-
The 21-d chronic effect of each pharmaceutical was as- ard quotient (HQ) was calculated to evaluate the degree of
sessed on the reproduction of D. magna in a static renewal risks for each target PhAC, which was mathematically for-
test according to the U.S. EPA guidelines [23]. The exposure mulated as follows:
concentrations to D. magna in 21-d chronic tests ranged from EC (Exposure concentration)
0.1 to 80 mg/L based on the results of the acute toxicity tests HQ 5
for nine different concentrations and one control. Five repli-
cates, containing one organism per test chamber, were assigned where TRV represents the toxicity reference value (i.e., LC50
to each concentration. Organisms were fed every 2 d with or NOEC). A HQ value of less than 1 when the assessment
YCT and S. capricornutum, with all test solutions renewed factor was applied indicates an insignificant risk. In the present
with fresh culture medium every 2 d. The number of neonates study, we employed a quite moderate assessment factor of
reproduced and the survival of D. magna were counted and 1,000, as recommended by the European Union [25] and the
recorded on each of the 21 d. Organization for Economic Co-operation and Development
Mixture bioassay tests
To elucidate the combined effects of the pharmaceutical
Bioassay tests
mixtures to which the aquatic ecosystems were exposed, the
toxic unit (TU) concept was used as follows: Single-compound toxicity. The mortality (LC50) and repro-
duction inhibition (NOEC) were measured in D. magna ex-
concentration (mg/L) posed to nine pharmaceuticals. The results indicated that most
TU 5 (acute test)
LC50 pharmaceuticals tested were toxic and harmful to aquatic or-
concentration (mg/L) ganisms (Table 2). Generally, toxicity levels of chemicals are
TU 5 (chronic test) classified as follows: very toxic to aquatic organisms for less
than 1 mg/L, toxic to aquatic organisms for 1 to 10 mg/L, and
where LC50 is the single-substance concentration causing 50% harmful to aquatic organisms for 10 to 100 mg/L [25]. Com-
mortality when applied alone and NOEC is the no-observed- paring the measured and predicted LC50s using the Ecological
effect concentration. The total combined effect can be cal- Structure–Activity Relationships (ECOSAR) model from the
culated by summation of each TU. website of the U.S. EPA [1], the toxicity values showed rel-
For mixture acute toxicity tests, the methodology of equi- atively significant differences for high-lipophilic compounds,
toxic addition of chemicals tested was employed. Daphnia including diclofenac, ibuprofen, and tolfenamic acid, than for
magna was exposed to the control and five TUs, ranging from the hydrophilic compounds. For diclofenac, Sanderson et al.
0.18 to 1.05, in 48-h static nonrenewal tests. For example, to [1] used an extremely low octanol–water partition coefficient
prepare 0.6 TU of a three-chemical mixture, 0.2 TU for each (e.g., log Kow of 0.7; in the present study, log Kow of 4.51 was
chemical (diclofenac, ibuprofen, and clofibric acid) was com- used), which caused a large difference in the toxicity value.
bined. Four replicates, containing five organisms per test cham- To investigate the relationship of toxicity with respect to hy-
ber, were assigned to the treatments and the control. Other drophobicity, the LC50 versus log Kow values were plotted, as
experimental procedures, such as illumination, feeding regime, shown in Figure 1. This figure indicated a linear relationship
and endpoint measurement, were the same as those for a single between the toxicity values and hydrophobicities (r2 5 0.79).
compound. That is, the LC50s decreased (i.e., toxicity increases) with
For the mixture chronic toxicity test, D. magna was ex- increasing log Kow of the pharmaceuticals. However, the result
posed to less than 1 TU for the five treatments and one control for acetaminophen was exceptional, showing high toxicity for
for 21 d. Ten replicates, containing one organism per test cham- an extremely low log Kow of 0.27.
268 Environ. Toxicol. Chem. 25, 2006 G.H. Han et al.

Table 2. Toxicity values of pharmaceuticals to Daphnia magna in 48-h acute and 21-d chronic tests

Literature review
Chemicals Log Kowa (mg/L) Hazard classc (mg/L) LC50e NOEC

Diclofenac 4.51 80.1 Harmful 10 5057.0 —

Ibuprofen 3.97 132.6 Nontoxic 20 38.0 3f
Clofibric acid 2.57 141.2 Nontoxic 40 293.0 —
Carbamazepine 2.45 111 Nontoxic — 111.0 —
Salicylic acid 2.26 111.7 Nontoxic — 59.0 —
Gemfibrozil 4.77 10.4 Harmful — 6.0 —
Acetaminophen 0.27 20.1 Harmful — 42.0 —
Bezafibrate 4.25 30.3 Harmful — 25.0 —
Tolfenamic acid 5.17 7.4 Toxic — 1.7 —
a Values from Howard and Meylan [33] and Sanderson et al. [1].
b LC50 5 median effective concentration.
c Classified in European Union Directive 93/67/EEC (European Economic Communities) [25] in the
basis of LC50 values.
d NOEC 5 no-observable-effect concentration.
e Values from Sanderson et al. [1] calculated by the structure–activity relationship model using Ecological

Structure–activity relationships (ECOSAR).

f Value from Halling-Sorensen et al. [28].

In the 21-d chronic toxicity tests, the concentration–re- toxic but harmful to aquatic organisms). These values were
sponse curves for the single substances were sigmoid but had 5,000- to 100,000-fold higher than the median concentrations
different sharpness and shapes. When D. magna was exposed detected in aquatic environments [5]. Accordingly, considering
to 80 mg/L of each PhAC, the mortality rates were 100%. All the presence of low concentrations of these pharmaceutical
compounds showed good reproduction rates at 10 mg/L. The residues in real environments, such as rivers and wastewater
observed mortality rate with diclofenac was 100% at the rel- effluents containing various PhACs within the range 0.1 to 10
atively low concentration of 40 mg/L. The single chronic tox- mg/L, we can conclude that the acute effects, or even chronic
icity results indicated that diclofenac was the most toxic phar- effects (NOECs), of single compounds are negligible.
maceutical in the present study. The mortality by the end of Mixture toxicity. To evaluate the combined effect of mix-
the experiment never exceeded 10% for the controls. The tures, toxicity tests were carried out on D. magna with a com-
NOECs were 10, 20, and 40 mg/L for diclofenac, ibuprofen, bination of the three drugs (diclofenac, ibuprofen, and clofibric
and clofibric acid, respectively. Compared with the acute tox- acid). The observed and calculated toxicities, based on the TU
icities, the NOECs were not low, as shown by the acute to concept, were compared (Fig. 2). At greater than 0.5 TU, the
chronic ratio (LC50:NOEC), which ranged from 3.5 to 8. This TU observed in bioassay experiment showed approximately
implies that the tested chemicals (i.e., diclofenac, ibuprofen, twofold higher toxicity values than the TU predicted by cal-
and clofibric acid) show no serious long-term effects toward culation. For example, addition of the mixture concentration
aquatic ecosystems [27,28]. corresponding to 1.0 TU (i.e., 50% mortality in calculation)
The LC50s, even the chronic values (NOECs), were quite for three pharmaceuticals resulted in 2.0 TU (i.e., an observed
high for single compounds, in the range 10 to 100 mg/L (not 100% mortality) to D. magna. The measured toxicity values

Fig. 2. Comparison of the observed and calculated percentage mor-

Fig. 1. The relationship between the toxicity (median lethal concen- tality based on the toxic unit (TU) in 48-h acute mixture toxicity tests
tration [LC50]) and hydrophobicity (log Kow) for nine pharmaceuticals to Daphnia magna. The dashed line shows a 1:1 slope. The observed
in the Daphnia magna 48-h acute toxicity tests. The solid line rep- TU was determined from the percentage mortality results, with the
resents the linear regression of the data (r2 5 0.79). The closed circle calculated TU indicating the sum of the TUs of individual pharma-
indicates the data for acetaminophen. ceuticals based on their median lethal concentrations.
Occurrence and toxicity of pharmaceuticals from Korean WWTPs Environ. Toxicol. Chem. 25, 2006 269

Occurrence of pharmaceuticals from WWTP effluents

Grab samples of the influents and effluents at five WWTPs
in metropolitan cities in South Korea were collected between
January and November of 2004 (Table 3). Recovery of the
pharmaceuticals ranged from 72 to 89%, with standard devi-
ations (1s, n 5 3) between 5 and 20%, following spiking of
the water with 1 mg/L of each compound. The present results
showed that residual concentrations of seven PhACs were de-
tected at concentrations ranging from 0.01 to 88.99 mg/L in
the influents and from 0.01 to 10.96 mg/L in the effluents.
Salicylic acid, ranging between 0.12 and 88.99 mg/L, had the
highest detected concentrations in the present study. Gemfi-
brozil appeared at nonquantifiable levels in the multiple sam-
ples from all sampling sites because of the low limit of de-
tection. The results in Table 3 also show highly fluctuating
removal rates, between 8.7 to 91.3%, indicating that the phar-
Fig. 3. Comparison of the observed and calculated percentage inhi- maceuticals in WWTPs may not be eliminated effectively by
bitions, based on the toxic unit (TU), in 21-d chronic mixture toxicity the use of traditional treatment processes (i.e., activated sludge,
tests to Daphnia magna. The dashed line shows a 1:1 slope. The coagulation, flocculation, and sedimentation) at WWTPs [4].
observed TU was determined from the relative inhibition value cor-
responding to the no-observed-effect concentration (NOEC) from the Low removal rates were found for several PhACs in repeated
mixture tests, with the calculated TU indicating the sum of the TUs observations. Very low removal rates of 8.7 and 23.9% were
of individual pharmaceuticals based on their NOECs. observed for acetaminophen and diclofenac, respectively, in
all sampling areas, whereas that for carbamazepine in the con-
ventional treatment system, which discharged into surface wa-
ters at a concentration as low as 0.16 mg/L, was 91.3%. This
for the mixture were somewhat underestimated compared with was a notable result compared to that in another study [5],
those predicted [27]. which showed a very low removal rate (7%) for carbamazepine
In the 21-d chronic mixture tests, the inhibited reproduction in WWTPs. However, through analytical investigations on the
rate was used as the endpoint, which can be determined by influents and effluents from different WWTPs in South Korea,
the ratio of the number of neonates in various mixture series relatively high removal efficiencies were observed for most
of the three pharmaceuticals to those in the control. Figure 3 pharmaceuticals, with the exception of acetaminophen, al-
shows the expected and observed toxicities toward D. magna though no notable differences were found between WWTPs
of the diclofenac, ibuprofen, and clofibric acid mixtures. The compared to results of previous surveys in other countries
present results illustrate that the measured chronic toxicities [3,8]. This may have been caused by differences in the elim-
for the mixtures were greater than we predicted. In statistical ination rates of the pharmaceutical residues because of sorption
analysis, the observed TU had high standard deviations over and biodegradation. Meanwhile, treatment technologies, such
all the ranges tested. Overall, the measured toxicities of the as oxidation with chlorine and ozone, activated carbon, and
mixtures of the three pharmaceuticals in the present study were membrane filtration, have been demonstrated as effective
not significantly different from the predicted toxicities for ei- methods for the elimination of antibiotics [29] and some phar-
ther the acute or chronic toxicities, although a twofold increase maceuticals [21,30].
in the combined effect of PhACs on D. magna was observed. Figure 4 shows the effluent concentration profiles for seven
Note, however, that the result of the mixture toxicity could be pharmaceuticals following treatment at the WWTPs in Korean
serious if the combined toxic effects of unknown metabolized metropolitan cities (Gwangju1, Gwangju2, Busan, and Daegu).
chemicals and large number of PhACs in real water environ- Most residual PhAC concentrations were in the range 0.01 to
ments are considered. 0.4 mg/L, including approximately 40% not-detectable points.

Table 3. Concentrations (mg/L) of selected pharmaceuticals in wastewater treatment plant influents and effluents from South Korea during the
2004 survey

Influent (mg/L) Effluent (mg/L)

REC 6 1sb Removal rate
Pharmaceuticals LOD a (%) Samples (n) Mean (max, min) n . LOD c Mean (max, min) n . LODc (%)

Diclofenac 0.01 85 6 7 17 2.59 (9.87, ND)d 14 1.97 (10.96, ND) 12 23.9

Ibuprofen 0.01 79 6 10 17 0.30 (0.58, ND) 14 0.07 (0.31, ND) 12 77.8
Clofibric acid 0.01 89 6 5 17 1.51 (4.38, 0.03) 17 0.31 (0.74, ND) 15 79.5
Carbamazepine 0.05 75 6 15 16 1.84 (9.42, ND) 9 0.16 (0.97, ND) 8 91.3
Salicylic acid 0.1 73 6 20 16 23.92 (88.99, 0.12) 15 2.43 (6.73, ND) 13 89.8
Gemfibrozil 0.5 83 6 9 16 ND (ND, ND) 0 ND (ND, ND) 0 ND
Acetaminophen 0.01 72 6 18 16 0.07 (0.26, ND) 14 0.06 (0.16, ND) 12 8.7
a LOD 5 limit of detection.
b Recoveries (REC) and standard deviation (1s, n 5 3) of pharmaceuticals, following spiking of the water with 1 mg/L of each compound.
c Number of samples having values greater than the LOD.
d ND 5 not detectable.
270 Environ. Toxicol. Chem. 25, 2006 G.H. Han et al.

Fig. 4. Comparative mean concentration (mg/L) profiles for the target Fig. 5. Graphical illustration of the calculated hazard quotients (HQs)
pharmaceuticals in the effluents from metropolitan cities (Gwangju, for the selected pharmaceuticals in the wastewater treatment plants
Busan, and Daegu) in South Korea during 2004. An asterisk indicates effluents. A HQ of 1 indicates the critical value to determine a sig-
no detectable concentration. The data were less than the detection nificant risk of pharmaceuticals. A 5 acetaminophen; Carb 5 car-
limits in Seoul (South Korea). bamazepine; Clof 5 clofibric acid; D 5 diclofenac; G 5 gemfibrozil;
I 5 ibuprofen; S 5 salicylic acid.

However, several hundredfold higher concentrations were ob-

served for salicylic acid, up to 2.2 and 6.7 mg/L at Gwangju1 ranges of 1021 to 1024 when an application factor of 1,000 was
and Busan, respectively. These high concentrations in the ef- applied. From our results, it may be concluded that no risks
fluents might affect ecosystems because of the large volumes to D. magna were caused by the pharmaceuticals in Korean
of water discharged daily into the nearby river. Interestingly, WWTP effluents. Risk assessments also were conducted by
none of the PhACs mentioned above were detected in the applying a factor of 100 for the chronic effect on the basis of
effluents at the Seoul site (not shown in graph). Although only NOECs for the three pharmaceuticals (diclofenac, ibuprofen,
one sample was taken from Seoul, the results were unexpected, and clofibric acid). The present results indicated that the three
considering Seoul is the largest and most industrialized city pharmaceuticals had no significant ecotoxicological impact on
in Korea. long-term chronic exposure (not shown in graph). Moreover,
Many pharmaceuticals containing polar functional groups pharmaceuticals may be of less risk if the dilution, because of
are thermally stable and do not readily lend themselves to GC the mixing of effluents and the receiving stream, is considered.
analysis. Therefore, most GC analysis of polar pharmaceuticals Many uncertainties, however, exist in risk assessments, es-
must incorporate a derivatization step to overcome this limi- pecially those attributed to the difficulties in identifying the
tation [30], which makes the sample preparation laborious and presence of pharmaceuticals and their quantification in water
time-consuming, increases the probability of contamination discharged from WWTPs containing a mixture of chemicals.
and experimental errors, and leads to degradation of labile Currently, a method (i.e., extraction and derivatization) used
compounds. In addition, pharmaceuticals are easily changed for the detection of pharmaceuticals in WWTP effluents with
to their metabolites by glucuronidation [31] or photodegra- lower detection limits is being developed. Metabolites can be
dation [32], although considerable amounts of their unchanged one of the concerns in the ecological impact of pharmaceu-
forms are excreted via urine or feces. Therefore, further mea- ticals; therefore, an analytical method for the detection of their
surements and identification of pharmaceuticals and their me- forms should be developed for more accurate risk assessments.
tabolites in WWTP effluents and rivers will be required to In addition, more toxicological data for various pharmaceu-
evaluate the contamination and occurrence of drug residues in ticals must be obtained from bioassay experiments using aquat-
surface waters. ic species, such as algae, daphnids, and fish. Sanderson et al.
[1] demonstrated that the susceptibility of aquatic species to
Risk assessment in the aquatic environment pharmaceuticals were in the following order: Algae, daphnids,
The toxicity index (i.e., HQ) was calculated by comparing and fish. According to their results, approximately 10% of
the maximum measured concentrations of field data collected pharmaceuticals in the environment had a HQ greater than 1
from the WWTPs with the acute toxicity values to D. magna. when multiplied by an application factor of 1,000. Conse-
Figure 5 shows the effect of WWTP effluents that contained quently, the potential risk of pharmaceuticals should be mon-
pharmaceuticals was not a significant risk, even at the 95th itored carefully, with more bioassay data, although no risk was
percentile contamination range, although a risk assessment fac- found in the WWTP effluents from large Korean cities.
tor of 1,000, as recommended by the European Union and the
Organization for Economic Co-operation and Development, CONCLUSION
was applied. For salicylic acid and diclofenac, a difference of The overall ecotoxicological effects of PhACs detected in
less than one order of magnitude between maximum environ- WWTP effluents were investigated by biological and chemical
mental and effect concentrations was represented by the HQ analyses. The bioassay results showed LC50s and NOECs
values. None of the HQ values of the drugs tested in the present ranging from levels of several ppm to several tens of ppm in
study exceeded 1. On average, the HQ values of the phar- the acute and chronic tests, respectively, for the selected
maceuticals measured in the WWTP effluents were within the PhACs. The combined toxicity of pharmaceuticals showed
Occurrence and toxicity of pharmaceuticals from Korean WWTPs Environ. Toxicol. Chem. 25, 2006 271

slight synergistic effects in both the acute and chronic tests. mental behavior of the pharmaceutical drug ibuprofen in surface
waters and in wastewater. Environ Sci Technol 33:2529–2535.
These bioassay results would be valuable in risk assessments
13. Kosjek T, Heath E, Krbavcic A. 2005. Determination of nonste-
with respect to pharmaceutical residues in surface waters, tak- roidal anti-inflammatory drug (NSAIDs) residues in water sam-
ing into account the insufficient toxicological data regarding ples. Environ Int 31:679–685.
pharmaceuticals. The concentrations of PhACs in discharged 14. Sedlak DL, Pinkston KE. 2001. Factors affecting the concentra-
effluents indicated highly variable distributions depending on tions of pharmaceuticals released to the aquatic environment. Wa-
ter Resources Update 120:56–64.
the site, although relatively high removal efficiencies for most 15. Hirsch R, Ternes TA, Haberer K, Kratz KL. 1999. Occurrence of
pharmaceuticals (except for acetaminophen) were observed in antibiotics in the aquatic environment. Sci Total Environ 225:
Korean WWTPs compared to those from other countries. The 109–118.
PhAC residues in the effluents might affect aquatic ecosystems 16. Ashton D, Hilton M, Thomas KV. 2004. Investigating the envi-
ronmental transport of human pharmaceuticals to streams in the
because of the large volumes of water discharged daily into United Kingdom. Sci Total Environ 333:167–184.
the nearby river. Finally, the potential risk from the pharma- 17. Miao X-S, Metcalfe CD. 2003. Determination of carbamazepine
ceuticals detected in the effluents from Korean WWTPs in- and its metabolites in aqueous samples using liquid chromatog-
dicated a low biological risk (HQ , 1) to aquatic ecosystems raphy–electrospray tandem mass spectrometry. Anal Chem 75:
when a risk assessment factor of 1,000 was applied to the tests 18. Pedersen JA, Soliman M, Suffet IH. 2005. Human pharmaceu-
on D. magna. The monitored effluent concentrations did not ticals, hormones, and personal care product ingredients in runoff
reach critically harmful levels. It was concluded that no sig- from agricultural fields irrigated with treated wastewater. J Agric
nificant risk from the selected pharmaceuticals existed in Ko- Food Chem 53:1625–1632.
19. Gross B, Montgomery-Brown J, Naumann A, Reinhard M. 2004.
rean WWTPs, despite the presence of uncertainty that should
Occurrence and fate of pharmaceuticals and alkylphenol ethox-
be characterized in the determination and toxicity of metab- ylate metabolites in an effluent-dominated river and wetland. En-
olites. viron Toxicol Chem 23:2074–2083.
20. Metcalfe CD, Koenig BG, Bennie DT, Servos M, Ternes TA,
Hirsch R. 2003. Occurrence of neutral and acidic drugs in three
Acknowledgement—We acknowledge the financial support of the effluents of Canadian sewage treatment plants. Environ Toxicol
Ministry of Environment as The Eco-technopia 21 project through
Chem 22:2872–2880.
21. Ternes TA, Meisenheimer M, Mcdowell D, Brauch HJ, Brigitte
the Korea Institute of Environmental Science and Technology
HG, Preuss G, William U, Zulei-Seibert N. 2002. Removal of
(KIEST) and a grant (code 4-1-2) from the Sustainable Water Re-
pharmaceuticals during drinking water treatment. Environ Sci
sources Research Center of 21st Century Frontier Research Program.
Technol 36:3855–3863.
We also thank Dong-il Shimadzu Corporation for technical and fi-
22. U.S. Environmental Protection Agency. 1993. Method for mea-
nancial support.
suring the acute toxicity of effluents and receiving waters to fresh-
water and marine organism. EPA/600/4-90/027F. Cincinnati, OH.
REFERENCES 23. U.S. Environmental Protection Agency. 1994. Short-term meth-
1. Sanderson H, Johnson DJ, Wilson CJ, Brain RA, Solomon KR. ods for estimating the chronic toxicity of effluents and receiving
2003. Probabilistic hazard assessment of environmentally occur- waters to freshwater organisms. EPA/600/4-91/002. Cincinnati,
ring pharmaceuticals toxicity to fish, daphnids and algae by ECO- OH.
SAR screening. Toxicol Lett 144:383–395. 24. U.S. Environmental Protection Agency. 1992. Framework for
2. Ferrari B, Paxéus N, Giudice RL, Pollio A, Garric J. 2003. Eco- ecological risk assessment. EPA/630/R-92/001. Risk Assessment
toxicological impact of pharmaceuticals found in treated waste- Forum, Washington, DC.
waters: Study of carbamazepine, clofbric acid, and diclofenac. 25. Commission of the European Communities. 1996. Technical guid-
Ecotoxicol Environ Saf 55:359–370. ance document in support of commission directive 93/67/EEC on
3. Heberer T. 2002. Tracking persistent pharmaceutical residues risk assessment for new notified substances and commission reg-
from municipal sewage to drinking water. J Hydrol 266:175–189. ulation (EC) No 1488/94 on risk assessment for existing sub-
4. Doll TE, Frimmel FH. 2003. Fate of pharmaceuticals photodeg- stances. Office for Official Publications of the European Com-
radation by simulated solar UV-light. Chemosphere 52:1757– munities, Luxembourg.
1769. 26. Organization for Economic Co-operation and Development. 1992.
5. Ternes TA. 1998. Occurrence of drugs in German sewage treat- Report of the OECD workshop on the extrapolation of laboratory
aquatic toxicity data to the real environment. Monograph 58.
ment plants and rivers. Water Res 32:3245–3260.
Paris, France.
6. Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD,
27. Cleuvers M. 2003. Aquatic ecotoxicity of pharmaceuticals in-
Barber LB, Buxton HT. 2002. Pharmaceuticals, hormones, and
cluding the assessment of combination effects. Toxicol Lett 142:
other organic wastewater contaminants in U.S. streams, 1999– 185–195.
2000: A national reconnaissance. Environ Sci Technol 36:1202– 28. Halling-Sorensen B, Nielsen SN, Lanzky PF, Ingerslev F, Lutzhoft
1211. HCH, Jorgensen SE. 1998. Occurrence, fate and effect of phar-
7. Buser HR, Poiger T, Muller MD. 1998. Occurrence and fate of maceutical substances in the environment—A review. Chemo-
the pharmaceutical drug diclofenac in surface waters: Rapid pho- sphere 36:357–393.
todegradation in a lake. Environ Sci Technol 32:3449–3456. 29. Adams C, Wang Y, Loftin K, Meyer M. 2002. Removal of an-
8. Stumpf M, Ternes TA, Wilken RD, Rodrigues SV, Baumann W. tibiotics from surface and distilled water in conventional water
1999. Polar drug residues in sewage and natural waters in the treatment processes. J Environ Eng 128:253–260.
state of Rio de Janeiro, Brazil. Sci Total Environ 225:135–141. 30. Boyd GR, Reemtsma H, Grimm DA, Mitra S. 2003. Pharmaceu-
9. Ahrer W, Scherwenk E, Buchberger W. 2001. Determination of ticals and personal care products (PPCPs) in surface and treated
drug residues in water by the combination of liquid chromatog- waters of Louisiana, USA, and Ontario, Canada. Sci Total En-
raphy or capillary electrophoresis with electrospray mass spec- viron 311:135–149.
trometry. J Chromatogr A 919:69–78. 31. Richardson ML, Bowron JM. 1985. The fate of pharmaceutical
10. Öllers S, Singer HP, Fässler P, Müller SR. 2001. Simultaneous chemicals in the aquatic environment. J Pharm Pharmacol 37:
quantification of neutral and acidic pharmaceuticals and pesticides 1–12.
at the low-ng/L level in surface and waste water. J Chromatogr 32. Poiger TH, Buser HR, Muller MD. 2001. Photodegradation of
A 911:225–234. the pharmaceutical drug diclofenac in a lake: Pathway, field mea-
11. Stumpf M, Ternes TA, Haberer K, Baumann W. 1998. Isolierung surement, and mathematical modeling. Environ Toxicol Chem 20:
von ibuprofen-metaboliten and deren bedeutung als kontaminaten 256–263.
in der aquatischen Umwelt. Vom Wasser 91:291–303. 33. Howard PH, Meylan WM. 1997. Handbook of Physical Prop-
12. Buser HR, Poiger T, Muller MD. 1999. Occurrence and environ- erties of Organic Chemicals. Lewis, Boca Raton, FL, USA.