Pharmaceutical contamination in residential, industrial, and agricultural waste
streams: Risk to aqueous environments in Taiwan
Angela Yu-Chen Lin
*, Tsung-Hsien Yu, Cheng-Fang Lin
National Taiwan University, Graduate Institute of Environmental Engineering, No. 1, Section 4, Roosevelt Road, Taipei 10617, Taiwan
a r t i c l e
i n f o
Article history: Received 16 June 2008
Received in revised form 11 August 2008 Accepted 14 August 2008
Available online 30 September 2008
Keywords: Pharmaceuticals Antibiotics Wastewaters Hospitals Aquacultures Ecotoxicity
a b s t r a c t
This is a comprehensive study of the occurrence of antibiotics, hormones and other pharmaceuticals in water sites that have major potential for downstream environmental contamination. These include res-idential (hospitals, sewage treatment plants, and regional discharges), industrial (pharmaceutical produc-tion facilities), and agricultural (animal husbandries and aquacultures) waste streams. We assayed 23 Taiwanese water sites for 97 targeted compounds, of which a significant number were detected and quantified. The most frequently detected compounds were sulfamethoxazole, caffeine, acetaminophen, and ibuprofen, followed closely by cephalexin, ofloxacin, and diclofenac, which were detected in >91% of samples and found to have median (maximum) concentrations of 0.2 (5.8), 0.39 (24.0), 0.02 (100.4), 0.41 (14.5), 0.15 (31.4), 0.14 (13.6) and 0.083 (29.8)lg/L, respectively. Lincomycin and acetaminophen had high measured concentrations (>100lg/L), and 35 other pharmaceuticals occurred at thelg/L level. These incidence and concentration results correlate well with published data for other worldwide loca-tions, as well as with Taiwanese medication usage data, suggesting a human contamination source. Many pharmaceuticals also occurred at levels exceeding predicted no-effect concentrations (PNEC), warranting further investigation of their occurrence and fate in receiving waters, as well as the overall risks they pose for local ecosystems and human residents. The information provided here will also be useful for develop-ment of strategies for regulation and remediation.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
In recent years, a significant body of work has identified trace pharmaceuticals in natural aquatic environments. Some degraded
naturally through hydrolysis (Kuehne et al., 2000), biodegradation
(Morrall et al., 2004; Lin et al., 2006), or direct and indirect
photol-ysis (Tixier et al., 2002; Quanrud et al., 2004; Lin and Reinhard,
2005; Lin et al., 2006), while others had the potential to be sorbed
and eliminated from aqueous environments (Morrall et al., 2004;
Sassman and Lee, 2005), but many compounds persisted in aque-ous systems. Ubiquitaque-ous occurrences of these compounds (antibi-otics, hormones, non-steroidal anti-inflammatory drugs (NSAIDs), lipid regulators, b-blockers, psychiatric drugs, etc.) have been
re-ported for rivers, lakes, and reservoirs around the world (Golet
et al., 2002; Kolpin et al., 2002; Gross et al., 2004; Brown et al., 2006; Lin et al., 2006; Kim et al., 2007). For example, Heberer iden-tified ng/L levels of clofibric acid in tap water from Berlin, Germany (Heberer, 2002), while other studies detected significant and mea-surable concentrations of ibuprofen, diclofenac, and
carbamaze-pine in finished drinking waters (Ternes et al., 2002; Loraine and
Pettigrove, 2006; Kim et al., 2007).
Waste streams from hospitals, wastewater treatment plants (WWTPS), and sewage treatment plants (STPs) have been identi-fied as major contributors to environmental contamination with
human-derived medications (Kummerer and Henninger, 2003;
Gross et al., 2004; Miao et al., 2004; Brown et al., 2006; Gomez et al., 2006; Kim et al., 2007; Xu et al., 2007). Few studies have attempted to correlate the use of veterinary medicines in areas supporting dairy and agricultural industries with occurrence of
these compounds in local waters.Brown et al. (2006)identified
l
g/L levels of lincomycin in two dairies studied in New Mexico.Managaki et al. (2007)found macrolide and sulfonamide antibiot-ics at ng/L levels in Vietnam’s Mekong Delta and compared them with levels in the Tamagawa River in Japan. However, the existing information on contamination sources is spotty and limited, and normally several methods are needed to identify multiple pharma-ceutical classes which are present only at trace levels (ng/L).
Taiwan is a densely populated island, with a population of 22.77
million people residing in a mere 36 188 km2. The island has 19 433
hospitals, whose total drug consumption for the year 2005 is
esti-mated at 13.2 billion doses (Directorate-General of Budget, 2008;
National Research Health Research Database, 2008). Hospital efflu-ents and regional discharges can be important sources of pharma-ceutical contaminations, especially in urban regions. In rural areas, however, agriculture, animal husbandry, and aquaculture are an
0045-6535/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2008.08.027
* Corresponding author. Tel.: +886 2 3366 3729; fax: +886 2 2392 8830. E-mail address:yuchenlin@ntu.edu.tw(A.Y.-C. Lin).
Chemosphere 74 (2008) 131–141
Contents lists available atScienceDirect
Chemosphere
important segment of the economy, and waste from these sectors represents an additional potential source of contamination with veterinary medicines such as antibiotics and estrogens.
Our target analytes include human and veterinary medicines from ten antibiotic classes (sulfonamides, tetracyclines, lincosa-mides, macrolides, penicillins, cephalosporins, imidazoles,
quino-lones, chloramphenicols and others), and from estrogens,
b-blockers, psychostimulants, vasodilators, psychiatric agents,
NSAIDs, and lipid-regulator/cholesterol-lowering medication
groups. These compounds were chosen for their high detection fre-quencies, as reported by investigators worldwide, and for their sig-nificant usage by the global population and by Taiwanese especially. Samples were taken at important water sites that are susceptible to pharmaceutical contaminations and that have the potential to become contamination sources in themselves. Because it would be impractical to attempt individual analyses for each of our 97 target drugs on samples from >20 sites, we used an analyt-ical method to measure a sweep of all 97 pharmaceutanalyt-icals simulta-neously, even if present at only trace levels (low ng/L). Aquatic toxicity results were integrated with local medicine usage data and compared for potential environmental and human risk.
2. Experiment
2.1. Site selection and sampling
Limited data are available on the environmental occurrence in Taiwan of most of our targeted pharmaceuticals. As a consequence, we selected sampling sites from areas/places that were likely to be contaminated with human (hospitals, STPs, regional discharges), industrial (pharmaceutical production facilities), or agricultural (animal husbandry and aquaculture) wastewaters, and that were likely to pass their pharmaceutical load to downstream environ-mental waters. Regional discharges were the contents of sizeable sewage pipes collecting mainly household wastewaters and rain-water before being released to rivers and oceans. Triplicate grab samples were collected from 23 selected sites (waste streams and effluents from three pharmaceutical production facilities, six hospitals, five STPs, five regional discharges, two animal husband-ries, and three aquacultures) in one-liter amber glass bottles and stored in ice-packed coolers. Detailed descriptions of each
sam-pling site are provided in the Supplementary material section.
Briefly, waste stream samples were collected at the exit points (fi-nal effluent) of each facility with or without being treated. Waste-water effluents from hospitals and pharmaceutical production facilities were only treated to meet minimal effluent standards: BOD 30 mg/L, COD 100 mg/L, S.S-30 mg/L, true color 550, E. coli 200 000 CFU/100 mL. STPs were equipped with secondary treat-ment units, while agricultural wastewaters were not treated before releasing to downstream rivers and oceans. Eight mL of 0.125 M EDTA–2Na were added to sampling bottles prior to sample collec-tion. All samples were collected from Jan. 15–31st, 2008, and anal-ysis was completed within two weeks of sample collection.
2.2. Analytical methods
The following method was developed to identify and quantify up to 97 targeted pharmaceuticals simultaneously.
2.2.1. Sample preparation
All water and wastewater samples were vacuum-filtered
through a 0.45
l
m and 0.22l
m cellulose acetate membrane filter,acidified to pH 4.0 using sulfuric acid (2 N), and stored at 4 °C until analysis. Oasis HLB cartridges (500 mg, 6 mL, Waters, Milford, MA, USA) used for solid phase extraction (SPE) were preconditioned with 6 mL of methanol and 6 mL of deionized (DI) water. Aliquots
of 400 mL water samples were spiked with13C
6-sulfamethazine
(employed as a surrogate) and loaded to the cartridges with a flow rate of 3–6 mL/min. After sample passage, cartridges were rinsed with 6 mL DI water to remove excess EDTA–2Na and dried with a flow of nitrogen gas. After drying, analytes were eluted with 4 mL of methanol and 4 mL of methanol–diethylether (50:50, v/ v). The eluates were collected, evaporated to dryness with nitrogen stream, and reconstituted to 0.4 mL with 25% aqueous methanol.
Final solutions were filtered through a 0.45
l
m PVDF membranefilter before liquid chromatography/tandem mass spectrometry (HPLC–MS/MS) analysis.
2.2.2. LC-ESI-tandem MS analysis
Chromatographic separation of analytes was performed using an Agilent 1200 module (Agilent Technologies, Palo Alto, CA,
USA) equipped with a Phenomenex Luna C18 column
(150 4.6 mm, 5
l
m). A binary gradient with a flow rate of1.0 mL/min was used. Mobile phase A contained 0.1% formic acid (v/v) in water. Mobile phase B contained 0.1% formic acid (v/v) in methanol. The gradient started with 0% of mobile phase B for 0.5 min, increased to 40% from 0.5–3 min, to 70% from 3.0– 7.5 min, to 95% from 7.5–9.0 min, remained at 95% until 11 min, decreased to 0% from 11–12 min, and remained at 0%. All target compounds were eluted out of the column within 15 min. The
sample injection volume was 50
l
L. The autosampler was operatedat room temperature.
The mass spectrometric measurements were carried out on a Sciex API 4000 (Applied Biosystems, Foster City, CA, USA) equipped with a turbo ionspray source. Analyses were performed in negative mode for 18 compounds (all NSAIDs except for acetaminophen; mepirizole and famotidine; all chloramphenicol antibiotics; clofi-bric acid, gemfibrozil, benzafibrate, pravastatin; oxacillin and clox-acillin) and positive mode for all others. Ions were acquired in multiple reaction monitoring (MRM) mode with a dwell time of 10 (50) ms. The mass spectrometer conditions were as follows: ion spray voltage: 5.5 ( 4.5) kV, curtain gas: 10 L/h, nebulizer gas: 60 (50) L/h, turbo gas: 50 (60) L/h, heated capillary tempera-ture: 500 °C, interface heater: ON, and collisionally activated disso-ciation: 5. *note: when different from positive mode, values for negative modes are reported in parentheses.
After selecting the precursor ions, product ions were obtained and optimized with these four key parameters: declustering poten-tial, entrance potenpoten-tial, collision energy, and collision cell exit po-tential by direct infusion of the pure analytes to the MS–MS compartment.
2.2.3. Detection, quantification, and quality control
Identification of pharmaceuticals was performed with HPLC– MS/MS with multiple reaction monitoring (MRM), using the two highest characteristic precursor ion/product ion transition pairs. Compounds were identified using the LC retention time ±30% of retention time of a standard. The MRM pair used for quantification
is listed inTable 1.
Recovery experiments were performed on both DI and river water spiked with the 50 ng/L target analytes to estimate this method’s percent recovery, which was determined by comparing the concentrations of the spiked DI and river water before and after
SPE extraction.13C
6-sulfamethazine was used as an internal
stan-dard for the quantification. The stanstan-dard calibration curve was constructed by spiking waters with pharmaceutical standard
solu-tions in the 0.5–2000
l
g/L range, and the linearity of calibrationcurves was estimated by fitting a linear mode, least-squares regression analysis (y = a + bx). The method detection limits (MDLs) were determined with the minimum concentration of analyte in the linear range with a signal-to noise ratio of P3:1 in a river water matrix.
Table 1
Occurrence, usage, and toxicity data for 97 pharmaceuticals in six potential contamination sources
MRM Rec (%) MDL (ng/L) Frequency (%) Max (ng/L)
Med Lowest PNEC
reportedc
(ng/L)
NHI 2005 medicine usage Drug production facilities (n = 3) (ng/L) Hospitals (n = 6) (ng/L) STPs (n = 4) (ng/L) Regional discharges (n = 5) (ng/L) Animal husbandries (n = 2) (ng/L) Aquacultures (n = 3) (ng/L) Dosee (103 ) Contentf Sulfonamide antibiotics Sulfanilamide 173.0/ 108.0 67.9 10 78 207 50 19 26 25 48 ND 140 5 gm Sulfaguanidine 215.0/ 156.0 57.4 1 43 145 4 ND 2 ND 26 ND Sulfadiazine 251.0/ 156.0 79.4 1 78 353 19 50 6 10 21 ND 4,369 10 mg/gm Sulfamethoxazoleb 254.0/ 156.0 69.1 1 96 5823 22 647 226 282 10 229 20000 16023 80, 400 mg; 20, 40 mg/mL Sulfathiazole 256.0/ 156.0 79.2 1 48 9637 ND ND 2 ND 4844 7 Sulfamerazine 265.0/ 172.0 80 1 22 1660 ND 1 ND ND 831 ND 116000 Sulfisoxazole 268.0/ 156.0 83 1 0 ND ND ND ND ND ND ND Sulfamethiazole 271.0/ 156.0 72.7 1 0 ND ND ND ND ND ND ND Sulfisomidine 279.0/ 186.0 80.8 1 13 2 ND ND ND ND ND ND Sulfamethazine 279.0/ 186.1 81 1 35 370 1 ND 2 ND 77 ND 1:2 106d Sulfamethoxypyridazine 281.0/ 156.0 76.7 1 13 2 ND ND ND ND ND ND Sulfamonomethoxineb 281.0/ 156.1 82.9 1 22 20 ND ND ND ND ND 2 Sulfaquinoxaline 301.0/ 156.0 60.2 1 4 1.1 ND ND ND ND ND ND Sulfadimethoxine 311.0/ 156.0 78.7 2.5 35 354 ND ND 2 ND 17 21 248000d 21 250, 500 mg Sulfinpyrazone 405.0/ 279.0 46 10 0 ND ND ND ND ND ND ND 106 2619 100 mg Tetracycline antibiotics Tetracyclinea,b 445.0/ 410.0 72.4 5 87 1570 5 59 21 38 1129 ND 90 7284 250 mg, 125 mg/gm Oxytetracyclinea,b 461.0/ 426.0 83 5 39 15133 ND ND ND ND 8318 12 200 95 50, 250 mg; 50 mg/mL Chlortetracyclinea,b 479.0/ 444.0 73.3 5 22 5637 ND ND ND ND 2821 11 219000d 30 250 mg; 1, 10 mg/gm Lincosamide antibiotics Lincomycina 407.0/ 359.0 70.7 5 70 111667 ND 24 13 25 56760 70 >106d 180 300 mg/mL Clindamycin 425.0/ 377.0 81.6 2.5 70 1150 10 341 51 199 ND ND 500 34516 150 mg, 150 mg/mL Macrolide antibiotics Spiramycina,b 422.5/ 145.0 36.7 10 0 ND ND ND ND ND ND ND Azithromycin 375.5/ 591.5 34.4 5 48 1067 ND 227 15 ND ND ND 150 296 250 mg Oleandomycina 398.2/ 158.0 31.9 10 0 ND ND ND ND ND ND ND
(continued on next page)
A.Y.-C. Lin et al. /Chemosphere 74 (2008) 131–141 133
Table 1 (continued) MRM Rec (%) MDL (ng/L) Frequency (%) Max (ng/L)
Med Lowest PNEC
reportedc
(ng/L)
NHI 2005 medicine usage Drug production facilities (n = 3) (ng/L) Hospitals (n = 6) (ng/L) STPs (n = 4) (ng/L) Regional discharges (n = 5) (ng/L) Animal husbandries (n = 2) (ng/L) Aquacultures (n = 3) (ng/L) Dosee (103) Contentf Erythromycin-H2Oa 716.5/ 558.0 108.9 1 83 5177 107 676 695 705 20 ND 40 15358 250 mg, 10 mg/gm, 75 mg/mL Clarithromycin 748.5/ 590.5 84.5 0.5 83 2450 87 721 172 515 6 ND 40 3013 250, 500 mg; 25 mg/mL Josamycin 828.5/ 174.0 116 2.5 13 861 4 ND ND ND 432 ND 0.2 100 mg/gm Roxithromycin 837.5/ 679.6 84.2 2.5 17 215 ND ND 3 ND ND ND 150 64 150 mg Tylosina 916.7/ 772.5 101.1 5 22 1997 ND ND 7 ND 1001 ND >106d Penicillin antibiotics Penicilline Ga 335.0/ 176.0 68.2 5 9 105 ND ND ND ND 55 ND 53 1, 3, 10 mu Ampicillina,b 350.0/ 160.0 63.6 5 48 5800 ND 53 7 42 ND ND 75 10052 250, 500 mg Oxacillina 432.0/ 354.0 105.7 2.5 9 7 ND ND ND ND ND ND 162 125, 250,500 mg Cloxacillina 434.0/ 293.0 65 5 17 56 ND ND 7 ND ND ND 1698 125, 250,500 mg Cephalosporin antibiotics Cephalexin 348.0/ 174.0 58.3 1 91 31433 27 2457 283 610 ND 12 2500 75252 250, 500 mg Cephradine 350.0/ 176.0 54.7 1 70 2170 1 113 12 128 ND 2 10821 250, 500, 1000 mg Cephapirina 424.0/ 292.0 73.8 2.5 17 10 ND 5 ND ND ND ND 1957 500 mg; 1 gm Cefazolin 455.0/ 323.0 96.8 5 74 42933 ND 6221 27 5892 53 15 1250 361 500 mg; 1, 2 gm Cefotaxime 456.0/ 396.0 76.9 5 30 41900 ND 413 7 ND ND ND 40 3 500 mg; 1 gm Imidazole antibiotics Dimetridazolea 142.0/ 96.0 75.8 1 65 589 22 19 9 10 6 ND Metronidazole 172.0/ 128.0 94.7 1 74 7850 ND 1591 100 314 2 ND 1300 9369 250, 500, 5 mg/mL, 7.5 mg/gm Fenbendazolea 300.0/ 268.0 40.6 1 4 1.0 ND ND ND ND ND ND Oxfendazolea 316.0/ 191.0 80.7 1 4 1.2 ND ND ND ND ND ND Quinolone antibiotics Nalidixic acid 233.0/ 215.0 93 5 83 67567 7 186 178 178 71 ND 3841 50, 250, 500 mg; 50 mg/ mL Flumequinea,b 262.0/ 244.0 87 2.5 57 1041 3 3 15 ND 534 ND Oxolinic acidb 262.0/ 244.1 114.7 2.5 39 19 4 5 ND ND ND ND Pipemidic acid 304.0/ 286.0 54.9 5 78 5527 10 178 70 159 ND ND 2980 200, 250, 400 mg Norfloxacin 320.0/ 302.0 47 1 65 1877 9 131 14 163 219 ND 150 1364 100, 200, 400 mg; 3 mg/ mL Ciprofloxacin 332.0/ 314.0 19.2 1 74 9640 396 751 42 ND ND 2 20 1621 2, 10 mg/mL; 250, 500, 750 mg 134 A.Y.-C. Lin et al. /Chemosphere 74 (2008) 131–141
Pefloxacin 334.0/ 316.0 75.3 5 30 950 12 62 ND ND ND ND 0.1 80 mg/mL Enrofloxacin 360.0/ 342.0 52.7 5 48 342 8 40 ND ND 173 ND Ofloxacin 362.0/ 318.0 58.7 1 91 13633 853 1088 123 1065 ND 4 40 7025 100, 200, 400 mg;3 mg/ mL Marbofloxacin 363.0/ 320.0 49.8 2.5 30 7 3 3 ND ND ND ND Sarafloxacina 386.0/ 368.0 54.6 1 17 17 4 ND ND ND ND ND Difloxacin 400.0/ 382.0 64.6 5 9 10 6 ND ND ND ND ND Chloramphenicol antibiotics Chloramphenicol 323.0/ 152.0 120 0.5 52 114 ND 1 50 1 ND ND 1600 6156 10, 250 mg/gm; 2.5, 5, 125, 250 mg/mL Thiamphenicola,b 354.0/ 290.0 94.2 0.5 70 192 1 4 138 7 ND ND 5048 250, 500 mg; 100 mg/ gm Florfenicola,b 358.0/ 338.0 108.7 0.5 9 36 ND ND ND ND ND 8 Other antibiotics Morantela 221.0/ 123.0 115.3 2.5 0 ND ND ND ND ND ND ND Carbadoxa 263.0/ 231.0 92.5 5 4 6 ND ND ND ND ND ND Trimethoprima 291.0/ 261.0 82.7 2.5 87 2993 ND 1040 321 803 32 85 1000 13042 80, 400 mg Colchicines 400.0/ 358.0 134.3 5 17 37 ND 9 ND ND ND ND b-agonists Terbutaline 226.0/ 152.0 15.5 5 35 53 ND 38 9 ND ND ND 240000 16564 2.5, 5 mg; 0.2, 2.5 mg/ mL Tulobuterol 230.0/ 156.0 86.8 1 13 137 1 ND ND ND ND ND Salbutamol 240.0/ 222.0 11.7 0.5 61 173 1 22 5 9 ND ND 240000 22894 2, 4 mg; 0.2, 0.4, 0.5, 1 mg/mL Clenbuterola 277.0/ 168.0 85 2.5 0 ND ND ND ND ND ND ND Brombuterol 367.0/ 293.0 76.7 2.5 0 ND ND ND ND ND ND ND Ractopamine 302.0/ 284.0 78.6 5 4 48 ND ND ND ND ND ND Estrogens P-Estradiola 255.0/ 159.0 65.8 10 0 ND ND ND ND ND ND ND 0.020 19106 0.2, 3 mg; 2, 5 Estriola 271.0/ 253.0 96 5 39 22633 ND 4651 ND 4648 7169 ND 0.8 149 0.1,0.5, 10 mg; 1 mg/gm Estrone 271.0/ 253.1 107.3 10 4 345 ND ND ND ND ND ND 18d 15 1.25 mg 17a-ethynylestradiol 279.0/ 159.0 54.5 10 4 487 ND ND ND ND ND ND 0.02 801 0.035 mg Progesteronea 315.0/ 109.0 59.9 5 22 53 ND 9 ND ND ND ND 2160 25,50, 125 mg/mL; 10, 20, 50 mg b-blockers Propranolol 260.0/ 183.0 69 2.5 43 171 ND 42 50 ND ND ND 500 156637 10, 40 mg Atenolol 267.0/ 190.0 35.2 5 78 2260 16 1607 411 1025 52 ND 10 106
(continued on next page)
A.Y.-C. Lin et al. /Chemosphere 74 (2008) 131–141 135
Table 1 (continued) MRM Rec (%) MDL (ng/L) Frequency (%) Max (ng/L)
Med Lowest PNEC
reportedc
(ng/L)
NHI 2005 medicine usage Drug production facilities (n = 3) (ng/L) Hospitals (n = 6) (ng/L) STPs (n = 4) (ng/L) Regional discharges (n = 5) (ng/L) Animal husbandries (n = 2) (ng/L) Aquacultures (n = 3) (ng/L) Dosee (103 ) Contentf Metoprolol 268.0/ 191.0 70 5 52 631 ND 145 161 ND ND ND 8800 1561 50, 100 mg Sotalol 273.0/ 255.0 68.9 10 17 619 ND ND 11 ND ND ND 221 80, 160 mg Acebutolol 337.0/ 319.0 105.1 2.5 70 1323 6 185 223 223 ND ND 8017 100, 200, 400 mg Psychostimulants Caffeine 195.0/ 138.0 115.3 0.5 96 23967 5 15600 243 7560 45 36 >10 106 396 10, 20, 25, 50, 100 mg Vasodilators Pentoxifylline 279.0/ 181.0 84.6 1 70 9767 1370 302 90 188 ND ND 33613 100, 400 mg; 20 mg/mL Psychiatric drugs Carbamazepine 237.0/ 194.0 93.4 2.5 87 10933 7810 163 152 138 3 ND 25000 30019 100, 200, 300, 400 mg; 20 mg/mL Fluoxetin 310.0/ 148.0 37.1 5 17 371 154 ND ND ND 13 ND 14221 10, 20 mg Paroxetine 330.0/ 192.0 31.5 2.5 22 35 3 ND ND ND ND ND 6571 12.5, 20, 30 mg NSAIDs Acetaminophen 152.0/ 110.0 84.9 5 96 100433 9 36950 16 8060 12 21 9200 568839 500 mg; 24 mg/mL Mepirizole 235.0/ 220.0 104.1 5 0 ND ND ND ND ND ND ND 2402 Famotidine 338.0/ 259.0 26.9 5 61 252 25 94 8 14 ND ND Ibuprofen 205.0/ 161.0 62.3 5 96 14500 101 282 1758 747 836 50 5000 131359 200, 400, 600 mg Naproxen 229.0/ 185.0 61.9 2.5 70 8463 ND 470 548 278 1766 ND 37000 22655 100, 250 mg Fenoprofen 241.0/ 197.0 77 7.5 43 214 ND ND 24 0 8 ND 1 200 mg Ketoprofen 253.0/ 209.0 78.3 10 35 319 ND ND ND 0 164 ND 15:6 106d Fenbufen 253.0/ 209.1 67.3 10 22 43 31 15 ND ND ND ND 2923 200, 300 mg Fenoprop 267.0/ 195.0 108.5 2.5 0 ND ND ND ND ND ND ND Diclofenac 296.0/ 252.0 67.4 2.5 91 29767 20733 286 61 184 4 4 10000 214672 25, 50, 75, 100 mg; 25 mg/mL; 10 mg/gm Piroxicam 330.0/ 266.0 76.9 1 0 ND ND ND ND ND ND ND 14920 10, 20 mg; 20 mg/mL; 5, 10 mg/gm Indomethacin 356.0/ 312.0 68.3 5 22 231 ND ND ND ND ND ND 14218 25, 50, 75, 100 mg; 10 mg/gm Lipid-regulator/cholesterol-lowering drug Clofibic acid 213.0/ 127.0 93.3 1 65 2400 ND 9 154 35 ND ND 12000 Gemfibrozil 249.0/ 127.0 74.6 1 52 10393 443 134 ND 36 ND ND 100000 25015 300, 600 mg 136 A.Y.-C. Lin et al. /Chemosphere 74 (2008) 131–141
2.3. Taiwan’s medication usage data
Data concerning Taiwan’s medication usage in 2005 (the most recent data available) were recorded by the National Health
Re-search Institutes (NHRI) and underwent statistical analysis (Table
1). Data from years prior to 2005 were also checked, and very
sim-ilar trends were observed. In Taiwan, approximately 1578 distinct medicines are used annually in the National Health Insurance (NHI) program, which provides comprehensive health care for all Taiwanese citizens. The NHI was implemented in March 1995, and by the end of 2005, more than 97.5% of the population had en-rolled in the program, while 91% of health care providers had
con-tracted with the Bureau of NHI (BNHI) (Department of Health,
2008). The NHRI maintains a comprehensive, current database on
quantities and types of drugs used in hospital and clinic partici-pants in the NHI program, thus providing an excellent representa-tion of medicarepresenta-tion use patterns for the entire Taiwanese population.
3. Results and discussion
The 97 target compounds are listed inTable 1with their generic
classes, MRM pair (for quantification), percent recoveries, and method detection limits (MDLs) in environmental aqueous sam-ples. The majority (86%) of the target analytes were >50% recovered with this novel method, while more than half had recoveries rang-ing from 75–120%. MDLs ranged from 0.5 to 10 ng/L with linearity >0.9911. Recoveries and MDLs for DI water were also performed (data not shown), with similar results but slightly better recoveries and lower MDLs for DI water in general.
Table 1summarizes detection frequencies and median values for all compounds found in each potential pharmaceutical contam-ination source, as well as each compound’s maximum
concentra-tion out of all 23 sites studied. A large number of
pharmaceuticals (84 out of 97) were detected at least once during this study; only six antibiotics, two b-agonists, one estrogen, three NSAIDs, and one lipid-regulator/cholesterol-lowering drug were not seen at all. For detected compounds, measured maximum con-centrations ranged from a few ng/L to approximately 0.1 mg/L. Compounds that were frequently detected also had relatively high-er maximum concentrations. The two exceptions whigh-ere estriol and gemfibrozil, which were detected only half of the time but had concentrations up to 22 633 and 10 393 ng/L (seen in hospital efflu-ents). This likely represents the usage of these two specific medi-cines in their source hospitals.
The 97 target compounds comprised a total of 18 distinct pharmaceutical classes (ten classes of antibiotics and eight other
classes of pharmaceuticals). For each class, Fig. 1a shows
fre-quency of detection and percent of total measured concentration (normalized to number of sampling locations in each of the six potential sources). It should be noted that the total mass contri-bution of pharmaceuticals from each source/compound group was not quantified because volumetric flows were not measured during sampling. Eighteen classes were detected in samples from at least ten of 23 sampling sites. Estrogens (43%) were the least frequently detected class, followed by penicillin antibi-otics (57%). Sulfonamides, quinolones, and NSAIDs were the three most ubiquitous classes and were present in all samples analyzed. NSAIDs (26.7%), lincosamides (18.5%), cephalosporins (13.8%), and quinolones (12.4%) dominated the other drug clas-ses in concentration, together comprising >70% of the total
nor-malized concentrations measured (Fig. 1a). Penicillins,
chloramphenicol antibiotics, and b-agonists contributed the least mass, together comprising <0.5% of the total normalized mea-sured concentrations. Bezafibrate 360.0/ 274.0 82.2 1 35 17 ND 1 ND 1 ND ND 100000 3665 200, 400 mg Pravastatin 423.0/ 321.0 70.5 1 26 187 ND ND 1 ND ND ND 6390 5, 10, 20, 40 mg Mevastatin 391.0/ 271.0 64 10 0 ND ND ND ND ND ND ND MRM – MRM pair selected for quantification; Rec – recovery in river water; MDL – method detection limit; Max – maximum concentration; Med -median concen tration in each potential source category; n – number of samples; PNEC – predicted no effect concentration; ND – not detected; mu – million interna tional unit. a Maximum Residues Limits for Veterinary Drug ( Department of Health, 2008 ). b Aquaculture Drug Use Guidance ( Council of Agriculture, 2008 ). c The lowest PNEC refer to(We bb, 1998, 2000; Golet et al., 2002; Jone s et al., 2002 ; Ferrari et al., 2003; Kummerer and Henninger, 2003; Carlsson et al., 2006; Quinn et al., 2008 ). d LC 50 /EC 50 (Brain et al., 2004; Drewes et al., 2005; Carlsson et al., 2006 ). e Dose are number of times a medication was prescribed. fReported as tablets (mg and mg/g m) and liquid (mg/mL) forms.
Fig. 1b depicts each class’s distribution among six potential con-tamination sources. Lincosamides were almost entirely (98%) de-rived from animal husbandry effluents. The majority of the cephalosporins and quinolones were from hospitals (73% and 12%, respectively) and drug production facilities (24% and 83%). Concentrations of NSAIDs were more evenly distributed among effluents of hospitals (53%), drug production facilities (21%), and regional discharges (16.5%). NSAIDs are widely used for humans, and many are available without a prescription, which may explain their prevalence in regional discharges. According to the National
Health Insurance Research database (NHIR) (National Research
Health Research Database, 2008), many NSAIDs are among the top-used medicines in hospitals belonging to the NHI (i.e. acetami-nophen, ibuprofen, diclofenac, and naproxen), with frequencies of use reflecting the detection frequencies reported here.
Penicillin-family medications were detected at trace amounts only. However a Taiwanese government investigation of
antibiot-ics usage in 2006 (National Antibiotics, 2006), reported that in
2004–2005, penicillins and cephalosporins were manufactured or imported in greater amounts than all other human-use antibiotics, and there is no evidence to believe that medication usage patterns have dramatically shifted away from these since that time. Cepha-losporins are also used in veterinary medicine, which should fur-ther increase their concentrations in hospital effluents. However the apparent contradiction between the significant quantities of penicillins used, manufactured, or imported, and the trace or en-tirely absent amounts found in the aqueous environment can be
explained by their easily hydrolyzed nature, as reported byHirsch
et al. (1999).
Fig. 2lists the pharmaceuticals found at each source and de-scribes the distribution of the various drug classes for each source. The ten antibiotic classes were grouped into one, while classes with fewer members were combined into an ‘‘other pharmaceuti-cals” category, which included psychiatric drugs, vasodilators,
0 5 10 15 20 25 30 1 5 3 2 8 4 5 4 1 2 3 4 5 6 5 1 1 3 1 2 5 Drug production facilities
Hospitals STPs
Regional discharge Animal husbandries Aquacultures
Sulfonamide antibiotics (100%)Tetracycline antibiotics (96%)Lincosamide antibiotics (91%) Macrolide antibiotics (91%) Penicillin antibiotics (57%) Cephalosporin antibiotics (97%) Imidazole antibiotics (87%) Quinolone antibiotics (100%) Chloramphenicol antibiotics (83%) Other antibiotics (87%)
β-agonists (70%)Estrogens (43%)β-blockers (83%)
Psychostimulants (96%) Vasodilators (70%)
Psychiatric drugs (91%)
NSAIDs (100%)Lipid-regulator/
Relative Distribution in Six Contaminantion Sources (%)
0 20 40 60 80 100 Cholesterol lowering drug (78%)
Percent of Total Measured
Concentration (%)
a
b
Fig. 1. Percent of total measured concentration (normalized to number of sampling locations) of 18 classes of pharmaceuticals (1a), and their relative distributions in six sites with potential for downstream transmission of contaminants (1b). The detection frequency for each class is shown next to the class name, and the number of compounds in each class is noted above the bar.
b-blockers, lipid-regulator/cholesterol-lowering drugs, and b-ago-nists. The highest total normalized drug concentrations were found in waste streams from hospitals (37.5%), followed by animal husbandries (27.9%) and drug production facilities (23.4%). Aqua-culture effluents were the most free of pharmaceutical contamina-tion, contributing <1.0% of the total normalized concentrations. In hospital effluents, antibiotics (39.3%), NSAIDs (37.7%),
psychostim-ulants (11.2%), and estrogens (7.2%) were the four dominant groups. In animal husbandry effluents, antibiotics comprised 88.5% of the water composition, while estrogens accounted for an additional 8.2%. In drug production facility effluents, significant amounts of vasodilators (5.0%) and psychiatric drugs (8.8%) were found, in addition to the ubiquitous antibiotics and NSAIDs. Com-pounds in the ‘‘other pharmaceuticals” group distributed similarly in waste effluents, with the exception of b-agonists, which contrib-uted only 0.05% to the total concentration. The medications in the ‘‘other pharmaceuticals” group are used exclusively in humans, accounting for their occurrence in hospitals, STPs, and regional dis-charges, and their lower concentrations or complete absence from animal husbandry and aquaculture effluents.
Fig. 3 shows the 27 most frequently detected (>70%) com-pounds. Multiple pharmaceutical classes were represented, of which 37.0% had medians >100 ng/L and 77.8% had 75th-percentile concentrations >100 ng/L. Except for three, all of the 27 com-pounds had maximum values >1000 ng/L, the exceptions being sulfanilamide, sulfadiazine, and thiamphenicol, which had the low-est maximum concentration (192 ng/L). The high concentrations observed indicate that significant amounts of pharmaceuticals were present in these six potential contamination sources.
Sulfamethoxazole, caffeine, acetaminophen, and ibuprofen were the most frequently detected compounds, present in all effluents but one, at concentrations up to 5823, 23 957, 100 433 and, 14 500 ng/L, respectively. Cephalexin, ofloxacin and diclofenac were the next most often detected compounds, with concentrations up to 31 433, 13 633, and 29 767 ng/L. The prevalence of these seven med-ications is consistent with findings of previous investigators. For
in-stance,Kolpin et al. (2002)targeted 95 compounds in a survey of
139 streams in the United States and detected caffeine, acetamino-phen and sulfamethoxazole at frequencies of 70.6%, 23.8% and
19.0%.Managaki et al. (2007)reported that in Vietnamese urban
drainage samples, sulfamethoxazole (0.190–0.369
l
g/L) wasDrug production facilities Hospitals
STPs
Regional dischargesAnimal husbandries Aquacultures
Percent of total measured concentration ( %)
0 10 20 30 40 Antibiotics Estrogens NSAIDs Psychostimulants Other Pharmaceuticals
Fig. 2. The proportion of pharmaceuticals found at six potential contamination sources. The percent of total measured concentration (normalized to number of sampling locations) and the distribution of each pharmaceutical class are indicated for each source.
0.1 1 10 100 1000 10000 100000 1000000 2450 5177 67567 207 353 5527 2260 42933 7850 9640 111667 1150 2170 192 1323 9767 8463 10933 2993 1570 29767 13633 31433 14500 100433 23967 5823 Naproxen (70%) Pentoxifylline (70%) Acebutolol (70%) Thiamphenicol (70%) Cephradine (70%) Clindamycin (70%) Lincomycin (70%) Ciprofloxacin (74%) Metronidazole (74%) Cefazolin (74%) Atenolol (78%) Pipemidic acid (78%) Sulfadiazine (78%) Sulfanilamide (78%) Nalidixic acid (83%) Clarithromycin (83%) Erythromycin H O (83%) 2 Carbamazepine (87%) Trimethoprim (87%) Tetracycline (87%) Diclofenac (91%) Ofloxacin (91%) Cephalexin (91%) Ibuprofen (96%) Acetaminophen (96%) Caffeine (96%) Sulfamethoxazole (96%) Concentration (ng/L) maximum 75% median 25% MDL
Target pharmaceutical name (frequency of detection)
Fig. 3. Measured concentrations for the 27 most frequently detected pharmaceuticals. Box plots show the distribution of concentrations; values for the maximum, 75th percentile, median, 25th percentile, and method detection limits (MDLs) are indicated. The frequency of detection of each compound is also indicated next to the compound.
present at concentrations higher than all other pharmaceuticals as-sayed, indicating a primary derivation from human use. Moreover, Sacher et al. (2001)showed that of 60 pharmaceuticals targeted, sulfamethoxazole was one of the most frequently detected
antibi-otics, with concentrations up to 0.410
l
g/L in groundswell inBa-den-Württemberg, Germany.Brown et al. (2006)studied eleven
antibiotics and frequently found sulfamethoxazole and ofloxacin in hospital effluents, at concentrations ranging from 400 to 2100
and 4900 to 35 500 ng/L. Another study (Gomez et al., 2006) of
hos-pital effluents also revealed high concentrations of acetaminophen,
ibuprofen, and diclofenac (up to 29, 151 and 1.9
l
g/L).Variations in MDLs and recoveries can influence detection fre-quencies and detected concentrations. For example, ciprofloxacin (a quinolones) was only 19.2% recovered with our method, with concentrations up to 9640 ng/L. If an extraction method more suit-able for ciprofloxacin were applied, we would expect to see
in-creased detection frequency and aqueous concentrations.
However, although results are not perfect for all compounds as-sayed, the novel pharmaceutical screen described here obviates the need to perform individually tailored analyses for each of the many target drugs.
Detection frequencies and maximum effluent concentrations in many cases correlate well with Taiwan’s medication usage
pat-terns, as recorded by the NHRI (Table 1). A more precise correlation
between these contamination metrics and actual amounts of med-ication prescribed would have been even more illuminating, but this comparison presents several insurmountable hurdles. First, the NHRI data records only the number of times a medication was prescribed and does not provide doses. Second, although indi-vidual hospitals or pharmacies may have collected similar data for internal use, and although drug companies themselves certainly kept dosage data for marketing purposes, such data are likely to be proprietary, in multiple different formats, or otherwise unavail-able for study. Therefore only the raw medication usage data are
presented (Table 1). Despite the bias introduced by using
prescrip-tion data as a surrogate for actual dosages, we demonstrated that in general, medicines that were often prescribed correlated with higher detection frequencies and observed concentrations in many human-derived waste effluents. Analysis of usage data from the year 2005 for 60 of our target compounds revealed that acetamino-phen, an NSAID, was the most-used medicine of 2005, with a total of 568.8 million doses prescribed (500 mg, 24 mg/mL). Diclofenac, atenolol, ibuprofen, cephalexin, clindamycin, pentoxifylline, and carbamazepine were also frequently prescribed, with usages in the top 100 of 1578 distinct drugs. These eight medicines were
among the most frequently detected groups (Fig. 3) in this study.
All were found at relatively high concentrations in human-impacted contamination sources (drug production facilities, hospi-tals, STPs and regional discharges), with maximum concentrations ranging from 1150 to 100 433 ng/L.
Taiwan has long regulated the use of pharmaceuticals in agri-culture. Twenty-five antibiotics, one b-agonist, and two estrogens are listed in the government’s Maximum Residues Limits for
Veter-inary Drugs (Department of Health, 2008) and Aquaculture
Phar-maceutical Use Guidance (Council of Agriculture, 2008)
regulations and are noted with superscripts ‘a’ and ‘b’ inTable 1.
When regulated veterinary medicines were queried, only 12 anti-biotics were detected in agricultural effluents: erythromycin–
H2O, tylosin, florfenicol, flumequine, lincomycin, chlortetracycline,
oxytetracycline, tetracycline, sulfadimethoxine, sulfamonome-thoxine, dimetridazole, and trimethoprim. Each compound fell into one of the ten antibiotics classes studied, and all ten classes were represented; for veterinary medications, these 12 compounds were taken as representatives for each class. Other investigators have also reported the use and occurrence of these 12 drugs in
agricul-ture (Brown et al., 2006; Managaki et al., 2007). In our samples,
lincomycin and the tetracyclines were the most prevalent antibiot-ics in the two animal husbandry effluents, with concentrations up to 111,667 (lincomycin) and 15,133 ng/L (oxytetracycline). The remaining antibiotics occurred in much lower concentrations, especially in aquaculture waste streams. These findings are consis-tent with a previous study of similar antibiotic classes in eight dairy effluents in New Mexico, in which lincomycin was the only
antibiotic detected (concentration 700–6600 ng/L) (Brown et al.,
2006).
The actual number of distinct pharmaceuticals detected ranged from 36–42, 41–54, 39–49, 30–49, 14–30, and 13–28, for samples from drug production facilities, hospitals, STPs, regional discharges, animal husbandries, and aquacultures, respectively. These counts, in combination with the fact that the majority of the target com-pounds were human-use medicines, demonstrate that our targeted contaminants were derived mainly from human activities. In con-trast, agricultural industries use smaller amounts of pharmaceuti-cals, and agriculture-specific medications were detected at correspondingly low levels or not at all.
The many types and high concentrations of pharmaceuticals identified in the waste streams sampled indicate that our target compounds were not fully removed in wastewater treatment pro-cesses. Further studies on treatment efficiencies and natural atten-uation processes will be needed to determine the fate of these pharmaceuticals in receiving environmental waters.
3.1. Risk assessment
The risk associated with residual pharmaceuticals transmitted from effluents to receiving water bodies is usually characterized as the ratio of environmental concentrations (predicted or mea-sured) to predicted no-effect concentrations (PNECs). A PNEC is normally estimated by dividing the lowest no-observed-effect con-centration (NOEC) for the most sensitive species by a safety factor (Carlsson et al., 2006). However, if NOEC data is not available, val-ues such as minimal inhibitory concentrations (MICs), lowest ob-servable effect concentrations (LOEC), or toxicity thresholds (TT)
can be used to estimate PNECs (Jones et al., 2002; Kummerer and
Henninger, 2003; Quinn et al., 2008). PNEC data for target
pharma-ceuticals are compiled and compared inTable 1, which lists the
lowest PNEC values reported in the literature; when PNEC values
were not available, the lowest acute and chronic LC50and EC50data
were listed as references.
PNEC values for a third of our target analytes are reported, and significant numbers of compounds had effluent concentrations in excess of or very close to PNEC. Three estrogens (estriol, estrone,
and 17
a
-ethynylestradiol), three NSAIDs (acetaminophen,ibupro-fen, and diclofenac), and 16 antibiotics occurred at levels higher than the PNEC reported; estrogens (estriol in particular) exceeded
the EC50by more than a factor of 104. Maximum concentrations of
many b-blockers, psychiatric drugs, and lipid regulators were less than a factor of 10 below reported PNECs. The ubiquity and high environmental concentrations which we report here are causes for concern and further investigation, as many of these medica-tions are frequently prescribed. In addition, effluents from drug production facilities, hospitals, STPs, regional discharges, animal husbandries and aquacultures are the six most important potential contamination sources of pharmaceuticals in Taiwanese surface water systems, implying a high likelihood of similarly ubiquitous occurrence and high concentrations of pharmaceuticals in down-stream surface water systems. Countries with similar economic activities and ways of life are likely to display similar contamina-tion profiles. Key pharmaceuticals with high ubiquity and concen-tration may be investigated for their potential application as early indicators of total pharmaceutical burden; acetaminophen, the most frequently used, identified, and concentrated contaminant
here (median concentrations >PNEC), is one such compound. With many other pharmaceuticals exceeding PNEC values at the source of contamination (and many more without any aquatic toxicity data at all), further investigations should include studies of their occurrence in receiving waters and the possible ecological and hu-man health effects which may result.
Acknowledgments
Financial support was generously provided by the Environmen-tal Protection Administration, Executive Yuan, Taiwan, ROC, under Grant No. EPA-96-G106-02-237, and by the National Science Coun-cil, Taiwan, ROC, under Grant No. NSC96-2221-E-002-042-MY3.
Appendix A. Supplementary material
Supplementary material associated with this article can be
found, in the online version, at doi:10.1016/j.chemosphere.2008.
08.027.
References
Brown, K.D., Kulis, J., Thomson, B., Chapman, T.H., Mawhinney, D.B., 2006. Occurrence of antibiotics in hospital, residential, and dairy, effluent, municipal wastewater, and the Rio Grande in New Mexico. Science of the Total Environment 366, 772–783.
Carlsson, C., Johansson, A.K., Alvan, G., Bergman, K., Kuhler, T., 2006. Are pharmaceuticals potent environmental pollutants? Part I: Environmental risk assessments of selected active pharmaceutical ingredients. Science of the Total Environment 364, 67–87.
Council of Agriculture, 2008. http://www.coa.gov.tw/show_lawcommond.php? serial=9_webuser1_20050512115944&code=A11&type=A&print=1 (accessed April 2008).
Department of Health, 2008.http://dohlaw.doh.gov.tw/Chi/FLAW/FLAWDAT0202. asp(accessed April 2008).
Directorate-General of Budget, Accounting and Statistics, 2008. http:// www.dgbas.gov.tw/ct.asp?xItem=13213&CtNode=3504(accessed April 2008). Golet, E.M., Alder, A.C., Giger, W., 2002. Environmental exposure and risk
assessment of fluoroquinolone antibacterial agents in wastewater and river water of the Glatt Valley Watershed, Switzerland. Environmental Science & Technology 36, 3645–3651.
Gomez, M.J., Petrovic, M., Fernandez-Alba, A.R., Barcelo, D., 2006. Determination of pharmaceuticals of various therapeutic classes by solid-phase extraction and liquid chromatography-tandem mass spectrometry analysis in hospital effluent wastewaters. Journal of Chromatography A 1114, 224–233.
Gross, B., Montgomery-Brown, J., Naumann, A., Reinhard, M., 2004. Occurrence and fate of pharmaceuticals and alkylphenol ethoxylate metabolites in an effluent-dominated river and wetland. Environmental Toxicology and Chemistry 23, 2074–2083.
Heberer, T., 2002. Occurrence, fate, and removal of pharmaceutical residues in the aquatic environment: a review of recent research data. Toxicology Letters 131, 5–17. Hirsch, R., Ternes, T., Haberer, K., Kratz, K.L., 1999. Occurrence of antibiotics in the
aquatic environment. Science of the Total Environment 225, 109–118. Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2002. Aquatic environmental assessment of
the top 25 English prescription pharmaceuticals. Water Research 36, 5013– 5022.
Kim, S.D., Cho, J., Kim, I.S., Vanderford, B.J., Snyder, S.A., 2007. Occurrence and removal of pharmaceuticals and endocrine disruptors in South Korean surface, drinking, and waste waters. Water Research 41, 1013–1021.
Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999–2000: a national reconnaissance. Environmental Science & Technology 36, 1202–1211.
Kuehne, M., Ihnen, D., Moeller, G., Agthe, O., 2000. Stability of tetracycline in water and liquid manure. Journal of Veterinary Medicine Series A 47, 379–384. Kummerer, K., Henninger, A., 2003. Promoting resistance by the emission of
antibiotics from hospitals and households into effluent. Clinical Microbiology and Infection 9, 1203–1214.
Lin, A.Y.C., Reinhard, M., 2005. Photodegradation of common environmental pharmaceuticals and estrogens in river water. Environmental Toxicology and Chemistry 24, 1303–1309.
Lin, A.Y.C., Plumlee, M.H., Reinhard, M., 2006. Natural attenuation of pharmaceuticals and alkylphenol polyethoxylate metabolites during river transport: Photochemical and biological transformation. Environmental Toxicology and Chemistry 25, 1458–1464.
Loraine, G.A., Pettigrove, M.E., 2006. Seasonal variations in concentrations of pharmaceuticals and personal care products in drinking water and reclaimed wastewater in Southern California. Environmental Science & Technology 40, 687–695.
Managaki, S., Murata, A., Takada, H., Tuyen, B.C., Chiem, N.H., 2007. Distribution of macrolides, sulfonamides, and trimethoprim in tropical waters: ubiquitous occurrence of veterinary antibiotics in the Mekong Delta. Environmental Science & Technology 41, 8004–8010.
Miao, X.S., Bishay, F., Chen, M., Metcalfe, C.D., 2004. Occurrence of antimicrobials in the final effluents of wastewater treatment plants in Canada. Environmental Science & Technology 38, 3533–3541.
Morrall, D., McAvoy, D., Schatowitz, B., Inauen, J., Jacob, M., Hauk, A., Eckhoff, W., 2004. A field study of triclosan loss rates in river water (Cibolo Creek, TX). Chemosphere 54, 653–660.
National Antibiotics Usage Survey Project Department of Health, 2006. Executive Yuan, Taiwan.
National Research Health Research Database, 2008d.http://www.nhri.org.tw/nhird/ index.htm(accessed April 2008).
Quanrud, D.M., Quast, K., Conroy, O., Karpiscak, M.M., Gerba, C.P., Lansey, K.E., Ela, W.P., Arnold, R.G., 2004. Estrogenic activity and volume fraction of waste water origin in monitoring wells along the Santa Cruz River, Arizona. Ground Water Monitoring and Remediation 24, 86–93.
Quinn, B., Gagne, F., Blaise, C., 2008. An investigation into the acute and chronic toxicity of eleven pharmaceuticals (and their solvents) found in wastewater effluent on the cnidarian, Hydra attenuata. Science of the Total Environment 389, 306–314.
Sacher, F., Lang, F.T., Brauch, H.J., Blankenhorn, I., 2001. Pharmaceuticals in groundwaters – analytical methods and results of a monitoring program in Baden-Wurttemberg, Germany. Journal of Chromatography A 938, 199–210. Sassman, S.A., Lee, L.S., 2005. Sorption of three tetracyclines by several soils:
Assessing the role of pH and cation exchange. Environmental Science & Technology 39, 7452–7459.
Ternes, T.A., Meisenheimer, M., McDowell, D., Sacher, F., Brauch, H.J., Gulde, B.H., Preuss, G., Wilme, U., Seibert, N.Z., 2002. Removal of pharmaceuticals during drinking water treatment. Environmental Science & Technology 36, 3855– 3863.
Tixier, C., Singer, H.P., Canonica, S., Muller, S.R., 2002. Phototransformation of triclosan in surface waters: a relevant elimination process for this widely used biocide – laboratory studies, field measurements, and modeling. Environmental Science & Technology 36, 3482–3489.
Xu, W.H., Zhang, G., Li, X.D., Zou, S.C., Li, P., Hu, Z.H., Li, J., 2007. Occurrence and elimination of antibiotics at four sewage treatment plants in the Pearl River Delta (PRD), South China. Water Research 41, 4526–4534.
