Pharmaceutically active compounds in the Xiangjiang River, China: Distribution pattern, source apportionment, and risk assessment

Science of the Total Environment 636 (2018) 975–984

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Pharmaceutically active compounds in the Xiangjiang River, China: Distribution pattern, source apportionment, and risk assessment Huiju Lin, Leilei Chen, Haipu Li ⁎, Zhoufei Luo, Jing Lu, Zhaoguang Yang ⁎ Center for Environment and Water Resources, College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, PR China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• First investigation of categories of PhACs on the Xiangjiang River • Urban areas were heavier impacted by pharmaceuticals than rural areas. • Distribution of PhACs was closely related to the temperature and NH3-N. • Domestic sewage was the major contribution of PhACs on the surface water. • Both single and mixture risk approaches were used to evaluate the aquatic risks.

a r t i c l e

i n f o

Article history: Received 31 January 2018 Received in revised form 29 March 2018 Accepted 20 April 2018 Available online xxxx Keywords: Pharmaceutically active compounds Spatiotemporal distribution Occurrence Redundancy analysis Mixture risks

a b s t r a c t The occurrence of 36 pharmaceutically active compounds in surface water of the Xiangjiang River was investigated in two seasons (n = 38). Twenty-five of these compounds were detected, with cefotaxime (maximum concentration 830 ng L−1) the most abundant compound followed by amoxicillin (maximum concentration 710 ng L−1). The spatiotemporal distribution was observed; indicating that pollution hotspots were mostly located in economically developed and densely populated regions such as Changsha City. Lower concentrations were found in summer than winter, which may be attributed to the dilution effect of a flood event and higher water temperatures. The distribution of pharmaceuticals was significantly correlated with temperature and ammonia nitrogen content. A principal component analysis-multiple linear regression model estimated that domestic sewage was the main source of pharmaceuticals, although the source composition varied among different sampling sites. Risk assessment was conducted using both individual and mixture models for preliminary identification of potential hazards. Sulfamethoxazole, clarithromycin, and azithromycin posed a high risk to algae, while sulfamethoxazole, trimethoprim, and erythromycin-H2O showed a medium risk to invertebrates. Moreover, the mixture risk quotients calculated using a concentration addition model ranged from 0.31 to 9.60 in winter, and from 0.06 to 0.61 in summer, indicating a potential risk to the aquatic environment. This study provides scientific support to policy-makers to establish contaminant management priorities and enriches the global data on emerging contaminants. © 2018 Elsevier B.V. All rights reserved.

Abbreviations: ANOVA, analysis of variance; BOD5, five-day biochemical oxygen demand; CODMn, chemical oxygen demand; DCA, detrended correspondence analysis; GDP, gross domestic product; LC-MS/MS, liquid chromatography-tandem mass spectrometry; MDLs, method detection limits; MLR, multiple linear regression; MQLs, method quantitation limits; MRM, multiple reaction monitoring; MRQ, mixture risk quotient; PCA, principal component analysis; RDA, redundancy analysis; STU, sum of the toxic unit; TOC, total organic carbon; TP, total phosphorus; WWTPs, wastewater treatment plants; NH3-N, ammonium nitrogen; NSAIDs, non-steroidal anti-inflammatory drugs. ⁎ Corresponding authors. E-mail addresses: [email protected], (H. Li), [email protected] (Z. Yang).

https://doi.org/10.1016/j.scitotenv.2018.04.267 0048-9697/© 2018 Elsevier B.V. All rights reserved.

976

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

1. Introduction

2. Materials and methods

River systems often serve as important drinking water sources; simultaneously, they are among the most vulnerable water bodies due to natural and anthropogenic activities (Zhang et al., 2010). As reported in a worldwide study, over 65% of the rivers in the world are polluted (Vorosmarty et al., 2010). Emerging contaminants, especially pharmaceutically active compounds (PhACs), which were widely used in agricultural practices, as veterinary additives, and in human health care (Barceló and Petrovic, 2007), heavily impact water quality. Even at trace levels, PhACs exhibit undesired biologically active effects on non-target organisms (Zhou et al., 2016). For example, ciprofloxacin may interfere with the photosynthesis pathway of higher plants, leading to morphological abnormalities or growth inhibition (Aristilde et al., 2010). Diclofenac also has high antiovulatory effects on aquatic vertebrates (Yokota et al., 2015). In addition, previous studies have provided evidence that the prevalence of antibiotics in water bodies was strongly correlated with the population of antibioticresistant bacteria and they can also promote the generation of antibiotic resistance genes (Jiang et al., 2013; Varela et al., 2014; Zhang et al., 2014). These genes could be shared between microorganisms, animals, and even to humans, through horizontal gene transfer (Liu and Wong, 2013). There is a knowledge gap regarding the environmental implications of chemical mixtures since most risk assessments have been based on individual compound (Ding et al., 2017). It should be noted that PhACs are unlikely to exist as independent constituents, a broad range are applied in combination at any real sites (Lopez-Serna et al., 2012; Paiga et al., 2016), engendering a serious mixture effect referred to as the “cocktail effect” (Du et al., 2017). Backhaus and Faust (2012) introduced a tiered assessment approach, which introduces first tier screening for chemical mixtures to determine whether more elaborate mixture risk assessment is needed. Since then, certain studies have paid close attention to the environmental implications of chemical mixtures (Backhaus and Karlsson, 2014; Liu et al., 2015; Yao et al., 2017). To date, several studies have focused on the occurrence of PhACs in the estuary (Yan et al., 2015; Zhao et al., 2017), central and lower areas of the Yangtze River (Wu et al., 2014). These studies give rise to concerns over the potential effects of pharmaceuticals in drinking water sources. However, data that characterize the sources, exposure, and effects of pharmaceuticals in this area are still very limited. In addition, tributaries can be important contributions to pollution in the main stream. To our knowledge, no study has ever systematically addressed the pharmaceutical contaminants in the Xiangjiang River, which is a main tributary of the Yangtze River. The Xiangjiang River is 856 km in length with a catchment area of 94,660 km2, of which 90.2% is located in the Hunan Province. N40 million people live along the river bank. It is an important water resource for drinking, irrigation, industry, fisheries, and transportation. It is also the most important economic belt in the Hunan Province, and is responsible for 70% of the gross domestic product (GDP) of the province. However, water pollution in the Xiangjiang River is markedly exacerbated by increasing population, booming economy, and accelerated urbanization during recent years (Xie, 2016). In this context, this study focused on the spatiotemporal distribution of multi-residue PhACs, which are frequently used for human and animal purposes in proximity to the Xiangjiang River. The linkages between PhAC concentrations and water quality parameters were explored. Potential pollution sources were also interpreted based on the concentration profile. In addition, both the single compound ecological risks and the mixture risks were evaluated for the aquatic environment. This work will serve to enrich the inventories of pharmaceutical pollution on a global scale, elucidate whether aquatic organisms are at risk, and help identify potential PhAC sources in the study region.

2.1. Chemicals and reagents Thirty-six PhACs with high purity grade (N98%) and obtained from Dr. Ehrenstorfer GmbH (Augsburg, Germany), were selected as the targets in this study. Detailed information on the physico-chemical properties is provided in Table S1. Surrogate standards (purity N 99%), including sulfamethoxazole-D4, sulfamethazine-D4, ciprofloxacin-D8, ibuprofen-D3, roxithromycin-D7, and thiabendazole-D4, were purchased from Toronto Research Chemicals (Oakville, Canada), and meclocycline was obtained from Sigma-Aldrich (Steinheim, Germany). Methanol (N99.9%) and acetonitrile (LC-MS Grade) were supplied by Sigma-Aldrich. Formic acid (LC-MS Grade) was acquired from Fisher Scientific (Loughborough, UK). Analytical grade Na2EDTA was purchased from Sinopharm Group Co. Ltd. (Beijing, China). Ultrapure water was produced using an Ultrapure Water Purification System from Ulupure Corporation (Sichuan, China). 2.2. Study area and sample collection The Xiangjiang River basin covers urban and rural areas, husbandry areas, industrial districts, cage culture areas, hospitals, and pharmaceutical factories. It flows through five administrative regions in the Hunan Province, which from south to north are: Yongzhou, Hengyang, Xiangtan, Zhuzhou, and Changsha City, finally enters into the Dongting Lake in Yueyang City. The Chang-Zhu-Tan region is the most urbanized area and is one of the pioneers of urban agglomeration in China (Long et al., 2013). The basic information of the six study areas is presented in Table S2. There are N80 wastewater treatment plants (WWTPs) around the Xiangjiang River basin, and their effluents will flow through the tributaries or the sewer systems into the main stream. Therefore, pharmaceutical pollution in this area could potentially be widespread. The region is heavily influenced by the subtropical monsoon climate and is hot and wet in summer, cold and dry in winter. The average annual rainfall is about 1500 mm, of which 68% occurs from April to September (Xu et al., 2013). The average temperature is 6–9 °C in winter (December to February) and 24–29 °C in summer (June to August). Sampling campaigns were conducted from upstream to downstream on January 6th, 2017 (winter) and August 8th, 2017 (summer, about three weeks after a flood). The sampling sites (n = 38) selected in this study were based on the state or province-controlled transects in the Xiangjiang River basin, Hunan Province. Samples were collected from three parallel sites perpendicular to riverbank (at least 5 m away from the right and left of riverbank and in the middle of the river) to objectively represent water quality within each transects. Sampling was conducted either from a boat or on a bridge. At each site, 2.5 L of water was collected at a depth of 0.5 m below the water surface using stainless steel samplers. The sampling locations are shown in Fig. 1. All samples were transported to the laboratory in coolers. On arrival, the samples were immediately filtered through 0.7 μm GF/F glass fiber filters (Whatman, England), then the pH was adjusted to 3 using 3 M H2SO4. The water samples were stored in a refrigerator at 4 °C before analysis. 2.3. Sample preparation and instrumental analysis The water sample treatment procedure was based on the reported method of Zhou et al. (2012) with some adjustments (Text S1). Briefly, a mixture of isotope-labeled internal standards was added to 1 L water samples, followed by filtration and adjustment of the pH to 3, and the addition of 0.2 g Na2EDTA. Subsequently, samples were extracted and enriched using Oasis HLB 3 cc SPE cartridges (60 mg, Waters, USA). The elutes were concentrated under gentle nitrogen blow-down and reconstituted in 1 mL of a mixture of methanol/0.1% formic acid in ultrapure water (20:80 v/v). Finally, to avoid losses ascribed to the filter

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

977

Fig. 1. Map of sampling locations and the spatial distribution of PhACs in the Xiangjiang River during two sampling campaigns. The letters YZ, HY, XT, ZZ, CS, and YY represent sampling sites located in the Yongzhou, Hengyang, Xiangtan, Zhuzhou, Changsha, and Yueyang sections. The color scheme represents the accumulative concentration ranges (ng L−1).

membrane, the extracts were centrifuged at 18000 rpm for 20 min, and then 0.5 mL of supernatant was extracted for liquid chromatographytandem mass spectrometry (LC-MS/MS) analysis. The separation of target compounds was performed using an Agilent 1260 series HPLC system (Agilent, USA) with an Infinity Lab Poroshell 120 EC-C18 column (3.0 × 50 mm, 2.7 μm particle size) and a Poroshell 120 EC-C18 guard column (4.6 × 5 mm, 2.7 μm particle size). Two mobile phase conditions were used for different ionization modes. For the positive mode, mobile phase A was 0.1% (v/v) formic acid in ultrapure water, mobile phase B was methanol and the flow rate was 0.2 mL min–1. For the negative mode, mobile phase A was 0.001% (v/ v) ammonia in ultrapure water, mobile phase B was methanol and the flow rate was 0.3 mL min–1. The injection volume was 5 μL for each sample, and the column temperature was kept at 40 °C during the analysis. The optimized gradient condition is presented in Table S4. For mass spectrometric analysis, the optimum mass conditions were set as follows: the source temperature was 350 °C, the source gas flow was 8 mL min–1, and the nebulizer gas pressure was 40 psi. Two ionization modes were used; chloramphenicol, thiamphenicol, florfenicol, ibuprofen, diclofenac, and mefenamic acid were analyzed in negative mode, and the other compounds were analyzed in positive mode. The precursor ions, product ions, fragmentor voltages, and collision energies can be found in Table S5. Quantification of the target compounds was conducted using multiple reaction monitoring (MRM) as well as isotope labeled surrogates (except for trimethoprim and flumequine). 2.4. Measurement of water quality parameters Temperature (T, °C), pH, conductivity (Cond, μS m−1) and dissolved oxygen (DO, mg L−1) were measured in situ. The other environmental parameters, including total organic carbon (TOC, mg L−1), ammonium nitrogen (NH3-N, mg L−1), total phosphorus (TP, mg L−1), chemical oxygen demand (CODMn, mg L−1) and five-day biochemical oxygen

demand (BOD5, mg L−1) were measured according to the Methods of Monitoring and Analysis for Water and Wastewater (Bureau, S. E. P., 2002). 2.5. Quality assurance and quality control Strict quality control procedures were performed to identify target PhACs. With each batch of 15 samples, solvent blank, procedure blank, blank spiked, sample spiked duplicate, and independent check standards were analyzed to identify potential background contamination and test the system performance. Analysis of blanks demonstrated that the whole analytical system was free of contamination. The recovery, method detection limits (MDLs), and method quantitation limits (MQLs) are shown in Table S6. 2.6. Ecological risk assessment The potential ecological risks from individual compound were evaluated based on a risk quotient (RQ) approach (Eqs. (1) and (2)): RQ ¼ MEC=PNEC

ð1Þ

PNEC ¼ EC50 =AF

ð2Þ

where MEC is the measured concentration of PhACs in a water sample, PNEC is the predicted no effect concentration in water and EC50 is the median effective concentration sourced from the literature or the Ecological Structure Activity Relationships predictive model (ECOSAR v1.11, US EPA). AF is the assessment factor, and a value of 1000 was used. To calculate the worst-case scenario, the maximum concentration and the lowest EC50 were used. Three different trophic levels (algae, invertebrates, and fish) were selected as representatives of aquatic organisms. Detailed eco-toxicological data are listed in Table S7.

978

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

In addition, a classical concentration addition model was used to calculate the mixture toxicity. Two approaches were used to assess the mixture risk quotients (MRQs) as outlined by Eqs. (3) and (4): n

MRQ MEC=PNEC ¼ ∑i¼1 n

¼ ∑i¼1

MECi PNECi MECi
 min EC50aglae ; EC50invertebrate ; EC50fish i  ð1=AFÞ

ð3Þ


 MRQ STU ¼ max STUa lgae ; STUinvertebrates ; STUfish  AF  n ¼ max ∑i¼1

MECi MECi MECi n n ; ∑i¼1 ; ∑i¼1 EC50i ;algae EC50i ;invertebrates EC50i ;fish

  AF

ð4Þ where STU is the sum of toxicity unit (MEC/EC50). MEC, EC50, and AF are defined as per the individual risk assessment approach. The detailed algorithm can be found in the previous work of Backhaus and Faust (2012). To further elucidate the risk levels, ecological risks were classified into three levels: low risk (b0.1), medium risk (0.1–1), and high risk (N1) (Hernando et al., 2006). Since the data used to calculate MRQs are derived from the data for RQs, they share the same risk ranking criteria. 2.7. Statistical analysis Target compounds with concentrations below the MQL were included in the detection frequency calculation, whereas no detection (ND) values were set to zero and not counted. One-way ANOVA was used to determine the differences between the two sampling seasons. Detrended correspondence analysis (DCA) was used to discriminate whether linear or unimodal numerical methods should be used. If the calculated length of the first ordination gradient is less than three, redundancy analysis (RDA) was chosen to assess the association between the detected concentrations of PhACs and environmental variables. Forward selection with Monte Carlo tests based on 499 unrestricted permutations was run following RDA to evaluate the significance of each potential predictor (Blanchet et al., 2008). Varimax-rotated principal component analysis (PCA) was implemented to identify the potential sources and was followed by multiple linear regression (MLR) to predict the contribution of each source to the total PhACs load. Statistical analysis was performed using IBM SPSS Statistics 19.0 and Canoco for Windows 4.5. The spatial distribution of PhACs during the two sampling campaigns was displayed using ArcGIS software version 10.2 (Redland, CA, USA).

winter, with a maximum concentration of 60 ng L−1 for sulfamethazine, 100 ng L−1 for sulfamethoxazole, and 68 ng L−1 for sulfadiazine. Sulfamethoxazole was the only compound with 100% detection frequency in both sampling campaigns. The prevalence of sulfonamides can be explained by their low degradability and hydrophilic characteristics (Xu et al., 2007; Zou et al., 2011). Trimethoprim, which is a synergist of sulfonamides, was detected at a maximum concentration of 93 ng L−1. Tetracyclines are widely used in aquaculture and livestock activities as antibacterial agents and growth promoters (Lalumera et al., 2004), but they have been rarely found in surface water. This is because they can be strongly adsorbed by the soils, preventing water pollution. In the present study, only oxytetracycline was found with a maximum concentration of 13 ng L−1and a relatively low detection frequency of 5% in the water samples. Measured concentrations of fluoroquinolones were generally below MQLs in this study, as were found in previous studies (Jiang et al., 2011; Li et al., 2014), which may because they are easily photo-degraded in surface water (Kolpin et al., 2002). In addition, the Chinese government regulated the use of certain fluoroquinolones (e.g., norfloxacin) in food animals in 2015 (MAPRC, 2015). Ofloxacin showed a relatively high residue level (ranging from not detected to 23 ng L−1) and a detection frequency higher than 50%. Flumequine was detected at a maximum concentration of 7.5 ng L−1 with a detection frequency of 95% in winter and 5.0% in summer. Florfenicol is a third-generation derivative of chloramphenicol and has been intensively applied in aquaculture since it was first promoted in Japan in 1990 (Li et al., 2014). It was found at a maximum concentration of 23 ng L−1, indicating a potential source from fish or shrimp farming in the study area. All macrolides were detected with high frequencies in winter (100%) with maximum concentrations ranging from b MQL to 190 ng L−1 (roxithromycin). High detection frequencies and relatively high residue levels revealed the wide use of macrolides in this region. Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used in the treatment of symptoms like pain, colds, and inflammation (Singh et al., 2014). Ibuprofen, which is the third most popular drug in the world (Xie, 2016), dominate the NSAID load in this study, with detected concentrations ranging from 2.4 to 320 ng L−1. Other compounds were found with mean concentrations lower than 3 ng L−1 but with relatively high detection frequencies. Sample concentrations in this study were also compared with those reported in the literature from around the world (Table S8). The detected concentrations were lower than those found in the Pearl River Delta and the Yangtze River Estuary (Yang et al., 2017; Zhao et al., 2017), but comparable to those detected in Italy, Portugal, and Spain (da Silva et al., 2011; Osorio et al., 2016; Paiga et al., 2016; Zuccato et al., 2010). The differences between data detected from the study area and other regions could be due to general differences in prescription and pharmaceutical sale patterns.

3. Results and discussion 3.2. Seasonal variation and spatial distribution 3.1. Occurrence and concentrations of PhACs Among the 36 investigated compounds, 25 of them were detected during the two sampling campaigns, indicating their ubiquitous presence in the Xiangjiang River. The compounds which were not detected or concentrations below the MDL in all samples are not discussed in this study. These include sulfathiazole, sulfamerazine, sulfaquinoxaline, difloxacin, pefloxacin, tetracycline, chlortetracycline, doxycycline, tylosin, naproxen, and ketoprofen. As shown in Table 1, the mean concentrations of PhACs in the Xiangjiang River range from 0.72 ng L−1 (mefenamic acid) to 69 ng L−1 (ibuprofen and cefotaxime). Only two β-lactams were included in the target PhACs, but they were the most abundant compounds observed. Cefotaxime had the highest concentration (830 ng L−1), followed by amoxicillin (710 ng L−1), which agrees with their high consumption for human health in China (Zhang et al., 2014). Three sulfonamides were detected in all samples collected in

The seasonal variation of target PhACs is presented in Table 1. Ofloxacin, florfenicol, and diclofenac displayed non-significant concentration fluctuation between different seasons, because these compounds were present at relatively low concentration levels (p N 0.05), while most of the PhACs showed significant seasonal variation (p b 0.05). Considerably higher concentrations were found in the winter compared to summer (except for mefenamic acid). Many factors contribute to the lower residues in the summer. The higher temperature in summer may accelerate biodegradation of pharmaceuticals due to higher microbial activity (Luo et al., 2011). In addition, hydrological conditions (e.g., stream flow) may result in greater dilution of PhACs concentrations. In this study, the summer sampling campaign was undertaken after a flood event. Due to a continuous heavy rainstorm, the water level in many sections of the Xiangjiang River reached a new record. For instance, Changsha reached its highest ever recorded height (39.49 m), which was 3.5 m higher than the security water level

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

979

Table 1 Range, median, mean values and detection frequencies of PhACs in the Xiangjiang River during two sampling seasons (ng L−1). Compounds

Sulfamethazine Sulfamethoxazole Sulfadiazine Trimethoprim Ciprofloxacin Enrofloxacin Ofloxacin Norfloxacin Flumequine Oxytetracycline Erythromycin-H2O Roxithromycin Clarithromycin Azithromycin Amoxicillin Cefotaxime Chloramphenicol Thiamphenicol Florfenicol Ibuprofen Diclofenac Mefenamic acid Paracetamol Indomethacin Antipyrine ∑Chemicals

MQL

0.78 8.49 0.63 0.13 6.55 9.93 7.04 5.08 0.58 11.16 1.70 1.00 0.51 4.46 1.48 2.74 0.20 0.42 0.14 0.36 0.20 0.56 0.96 1.75 0.52

Winter (January)

Summer (August)

Range

Median

Mean

DF (%)

Range

Median

Mean

DF (%)

bMQL-60 bMQL-100 0.93-68 1.35-93 ND ND ND-23 ND ND-7.5 ND-13 bMQL-43 1.4-190 0.55-100 2.2-99 4.6-710 3.8-830 ND-6.1 ND-12 1.6-23 2.4-320 ND-32 ND-3.9 ND-16 ND-2.7 ND-4.3 19.0-2,750

3.4 12 2.3 4.2 ND ND bMQL ND 0.93 ND 3.1 6.1 2.8 4.7 23.0 33 0.59 2.7 3.0 26 1.8 bMQL 1.3 bMQL 1.1 130

6.0 15 5.8 9.4 ND ND bMQL ND 1.2 bMQL 5.6 12 7.9 8.8 52 69 0.94 3.1 5.0 69 2.9 0.72 1.8 bMQL 1.2 275

100 100 100 100 ND ND 81 ND 95 5.0 100 100 100 100 100 100 92 97 100 100 76 68 97 47 97

ND-16 bMQL-12 ND-7.2 ND-5.6 ND ND ND ND ND-0.69 ND ND-9.4 ND-11 ND-8.7 ND-1.5 ND-29 ND-36 ND ND ND-15 ND-58 ND-9.2 ND-13 ND-0.99 ND-1.6 ND ND-235

ND bMQL bMQL 0.34 ND ND bMQL ND ND ND 1.8 ND bMQL ND 2.7 ND ND ND 3.9 5.3 0.73 ND ND bMQL ND 15

1.5 bMQL 1.0 0.94 bMQL bMQL bMQL bMQL bMQL ND 2.7 bMQL 0.84 bMQL 3.9 6.4 ND ND 4.9 12 1.6 2.3 bMQL bMQL ND 38

31 100 60 60 21 13 63 8.0 5.0 ND 97 10 50 10 90 39 ND ND 97 84 81 47 5.0 55 ND

Note: compounds without detection in either sampling campaign are not presented in this table; DF: detection frequency; ND: not detected; b MQL:concentration below MQL. ∑Chemicals: sum of the concentrations of the target PhACs.

(http://61.187.56.156/wap/index_sq.asp, in Chinese). Some studies reported higher concentrations after heavy rainfall since combined sewer overflows occur when precipitation exceeds the treatment capacity of WWTPs (Corada-Fernandez et al., 2017; Kim et al., 2016). However, in this study, the large dilution effect had a stronger impact than runoff from the point or non-point sources. The spatial map distribution of PhACs shows notable variation among different sampling sites (Fig. 1). The sampling sites located in Changsha were the most polluted, particularly at CS3 and CS4, followed by sampling sites located in Xiangtan City. This trend can be reasonably explained by the high population density and rapid development of the economy in these regions (Table S2). Sampling site CS4 is located at the confluence of the Xiangjiang River and a tributary (the Liuyang River), which received a large amount of domestic sewage. Other sites located at the confluence of tributaries also showed higher residue levels (YZ5, HY3, XT4, CS7, and YY2). The relative abundance of different pharmaceutical categories at each sampling site is shown in Fig. S1. In winter, NSAIDs were the predominant category in the middle of the study areas (Zhuzhou, Xiangtan, and Changsha sections). This trend can be explained by over-the-counter sales of these pharmaceuticals and the large population density in these regions (Table S2). However, βlactams were the predominant category in the upper and lower reaches, where there are less urbanization. People in less developed areas may lack professional knowledge of the proper use of medicine. Generally, they will choose the most accessible pharmaceuticals when they become ill. Moreover, doctors in rural hygiene hospitals may tend to prescribe newer and more potent antibiotics rather than the old generation. For example, it was reported that 28% of rural patients received prescriptions for cephalosporins and 15% for penicillin while the proportions were 24% and 14% respectively for urban patients (Currie et al., 2011). Sulfonamides showed no distinct spatial difference, which might indicate their wide use, or that they are easily transported in water environments. In summer, the relative abundance of the detected PhACs showed no trends. Due to the flood event, PhACs could

be introduced into the Xiangjiang River via runoff and leaching from soil, resulting in different PhAC distributions. 3.3. Correlation between PhACs and water quality parameters The monitored water quality parameters are shown in Table S3. The detected water quality parameters at least satisfied the standard limit values for Criteria III according to the Environmental Quality Standards for Surface Water (GB 3838-2002). Potential associations between the concentrations of investigated compounds and environmental parameters were analyzed through multivariate analysis. Detrended correspondence analysis (DCA) showed that the largest length value of the ordination gradient was b3 (1.839), hence the RDA model was adopted (Lepš and Šmilauer, 2003). Forward selection demonstrated that 58% of the variance could be explained by the water quality parameters. Temperature and NH3N strongly affected the PhAC profile and explained 33% (p = 0.002) and 12% (p = 0.002) of the variance in the PhAC distribution, respectively (Table S9). Ordination plots from RDA are presented in Fig. 2. Almost all the compounds exhibited a negative correlation with temperature (except for mefenamic acid). On the one hand, the temperature can influence the environmental behavior of PhACs; higher temperatures can promote biodegradation, photo-degradation, and adsorption (Hijosavalsero et al., 2011; Sui et al., 2011; Veach et al., 2012), resulting in relatively lower concentrations. On the other hand, the temperature will influence the usage pattern of some PhACs, since some diseases such as colds and arthrophlogosis occur more frequently in winter. Most PhACs were positively correlated with NH3-N, which was consistent with previous studies (Ferguson et al., 2013; Yang et al., 2013). It is well known that NH3-N generally comes from wastewater, and the positive correlation suggested that wastewater could be an important source of PhACs. As shown in Fig. 2b, the sampling sites are clearly separated based on the two seasons. Sampling sites with higher PhAC

980

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

Fig. 2. RDA ordination plots based on (a) species-each compound based on their concentrations and (b) samples-the PhACs concentrations at each sampling site with water quality parameters. For (b), the solid and hollow circles represent the different sampling locations in the summer and winter, respectively.

concentrations (collected during winter) are mainly distributed on the right side of the plot and were primarily influenced by NH3-N, whereas sampling sites with lower concentrations (collected during summer) are mainly distributed on the left side of the plot, and were heavily associated with temperature.

apportionment of the four identified principal components. The resulting regression equation was as following: ^SumPhACs ¼ Z

X

Bk FSk   ¼ 0:689FS1 þ 0:490FS2 þ 0:354FS3 þ 0:178FS4 R2 ¼ 0:871; pb0:001 ð5Þ

3.4. Source apportionment of PhACs using PCA-MLR To analyze the potential sources contributing to the PhAC burden, PCA-MLR was performed. This method, has been successfully employed in source identification of PPCPs in surface water (Dai et al., 2016; Jiang et al., 2016), polycyclic aromatic hydrocarbons in the atmosphere and sediments (Larsen and Baker, 2003), and heavy metals in road dust (Pan et al., 2017). The factor loadings and selected compounds are tabulated in Table 2. Four principal components (PC1, PC2, PC3, and PC4) were identified and cumulatively accounted for 76.9% of the variance. PC1 explained 30.0% of the total variance and was highly associated with trimethoprim, flumequine, erythromycin-H2O, roxithromycin, clarithromycin, azithromycin, and chloramphenicol, which were reported to have low removal efficiency in a typical WWTP in Changsha (Lin et al., 2017). Thus, PC1 can be highly indicative of ineffectively treated WWTP effluents. Other studies also indicated that WWTP effluent was a major pollutant source for the river system (Dai et al., 2016; Del Rosario et al., 2014). This is because these facilities were not designed to remove emerging contaminants like PhACs specifically, but primarily for nitrogen and phosphate removal (Kosma et al., 2010). PC2 explained 26.2% of the variance and was heavily related to sulfamethoxazole, erythromycin-H2O, amoxicillin, ibuprofen, and paracetamol, which are important chemicals used for human health, suggesting a domestic sewage origin. PC3 accounted for 11.1% of the total variance and was predominately composed of sulfamethazine and florfenicol. These antibiotics are animal-specific drugs and widely used in aquaculture and livestock to prevent and treat infectious diseases (Li et al., 2014; Sarmah et al., 2006), indicating the source was aquaculture and livestock activities around the Xiangjiang River. PC4 was correlated only with cefotaxime, which explained 9.6% of the total variance. Cefotaxime is one of the third-generation cephalosporins which is frequently used in hospitals for empirical and prophylactic therapy (Pinto et al., 2004). Thus PC4 may be assigned to hospital wastewater discharge. Stepwise multiple linear regression analysis with the factor score (FSκ) against the standard normalized deviate of the sum concentration ^SumPhACs ) was conducted to determine the mass of selected chemicals (Z

where Bκ represent the regression coefficients. The mean percentage contribution, which was defined as 100 × (Bκ/ ∑ Bκ), was 40% for domestic sewage (FS1), 29% for treated wastewater (FS2), 21% for hospital discharge (FS3), and 10% for both aquaculture and livestock activities (FS4). The contribution of source К to the total concentration can be calculated using Eq. (6).   Contribution of source К ng L−1  X  Bκ þ Bκ σ SumPhACs FSκ ¼ meanSumPhACs  Bκ =

ð6Þ

Table 2 Varimax-rotated component matrix following principal component analysis of all water samples from the two sampling campaigns. Variables

Rotated component number 1

Sulfamethazine Sulfamethoxazole Sulfadiazine Trimethoprim Flumequine Erythromycin-H2O Roxithromycin Clarithromycin Azithromycin Amoxicillin Cefotaxime Chloramphenicol Thiamphenicol Florfenicol Ibuprofen Paracetamol Antipyrine Percentage variance explained (%)

0.131 0.399 0.235 0.688 0.786 0.657 0.936 0.798 0.836 0.353 0.016 0.744 0.205 0.061 0.218 0.136 0.586 30.0

2 0.135 0.879 0.037 0.270 0.179 0.692 0.123 0.507 0.523 0.902 0.026 0.262 0.523 0.005 0.682 0.902 0.309 26.2

3

4 a

0.816 0.077 0.450 0.026 0.263 0.049 0.183 0.107 0.036 0.040 −0.017 0.203 0.073 0.805 0.251 −0.052 0.363 11.1

0.189 0.048 0.299 0.061 0.390 −0.037 0.044 −0.139 0.025 0.009 0.774 0.250 0.503 −0.266 0.064 0.243 0.527 9.6

a The bold values denote PCA loadings higher than 0.6. Extraction method: principal component analysis. Rotation method: varimax with Kaiser Normalization.

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

where meanSumPhACs (160 ng L−1) and σSumPhACs (240 ng L−1) are the mean concentration and standard deviation of the total PhAC concentration in all water samples, respectively. Fig. 3 illustrates the estimated contribution of each proposed source at each sampling site. Negative contributions are shown in some sampling locations, particularly in the summer. These negative values are physically impossible and may be the result of improper variable scaling inherent in eigenvaluebased methods, as previously described in Larsen and Baker (2003). Lower concentrations and detection frequency in summer resulted in greater uncertainty in this method. As shown in Fig. 3a, domestic sewage (FS1) was the main source in the upper reaches of the Xiangjiang River, as well as at some sampling sites located in Changsha (such as CS4). Less development of the wastewater treatment system in the south of the study region is likely to be responsible for the high contribution of domestic sewage to the pollutant load. Sample site CS4 is located at the confluence of the Xiangjiang River and the Liuyang River, which runs through many residential districts and receives a large amount of domestic sewage. Discharge from WWTP effluents (FS2) was an important source of PhACs in the middle reaches of the Xiangjiang River, since many WWTPs are located in these areas and the effluents may enter into the main stream. Hospital wastewater discharge (FS3) was seldom identified as a source in the upstream areas and mainly contributed to the pollutant load in the middle and downstream reaches. Pollutant source from aquaculture and livestock (FS4) accounted for a small proportion in most areas, except at some sites in Yongzhou (YZ5) and Yueyang (YY1), which have a large number of livestockand freshwater fisheries (Table S2).

981

3.5. Risk assessment Even trace levels of PhACs in the aquatic environment may engender adverse effects on aquatic organisms at different trophic levels as well as on human health (Hernando et al., 2006; Luo et al., 2011). Therefore, environmental risk assessment is essential. The individual RQs calculated from Eqs. (1) and (2) were shown in Fig. 4a and b, In winter, sulfamethoxazole, clarithromycin, and azithromycin exhibited a high risk (RQ N 1), indicating that these compounds were significantly harmful to algae. Sulfadiazine, erythromycin-H2O, and amoxicillin posed a medium risk to algae. Sulfamethoxazole, trimethoprim, and erythromycin-H2O showed a medium risk to invertebrates with RQs between 0.1 and 1. In summer, two compounds (sulfamethoxazole and clarithromycin) posed a medium risk to algae, and other compounds showed a low risk to invertebrates. However, pharmaceuticals in the environment never occur as isolated or single substances (Backhaus and Karlsson, 2014; Liu et al., 2015). Hence, ignoring potential mixture effects would underestimate the actual impact of these pollutants on aquatic organisms. In this study, two MRQ concepts (MRQMEC/PNEC and MRQSTU) were applied to comprehensively assess the mixture risks of PhACs (Eqs. (3) and (4)). These two approaches were based on the same input data and showed a remarkable positive correlation both in winter (R2 = 0.9972, p b 0.0001) and summer (R2 = 0.9901, p b 0.0001). This strong correlation was also observed in other studies (Liu et al., 2015; Yao et al., 2017). The estimated MRQs vary from 0.31 to 9.60 in winter and from 0.06 to 0.61 in summer (Table S10). In winter, the MRQs regularly exceed 0.1, and in

Fig. 3. PCA-MLR source contribution plots for all water samples collected during (a) winter and (b) summer. FS1: domestic sewage; FS2: insufficiently treated wastewater; FS3: hospital wastewater; FS4: aquaculture and livestock activities.

982

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

Fig. 4. Risk quotients calculated for investigated PhACs based on algae, invertebrates, and fish in (a) winter and (b) summer, and the estimated MRQSTU in winter (c) and summer (d). The red dotted line represents the value of RQ/MRQ is 1, and the green represents the value is 0.1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

some sampling sites, such as CS3 and CS4, were significantly higher than 1, indicating a potential risk for aquatic organisms. In summer, the MRQs at all sampling sites were b1 (Fig. 4c and d). From the above results, it is obvious that algae are the most sensitive trophic level, and among the three trophic levels, fish were least susceptible to PhACs, which is in agreement with previous studies (Li et al., 2014; Li et al., 2012). Since algae are at the bottom of the food chain, even tiny changes in the algal population may dis-equilibrate the aquatic system (Kummerer, 2009). However, the MRQ assessment approach can only serve as a “screening level” before more elaborate measures are implemented (Backhaus and Faust, 2012). It is important to analyze the effects of PhACs in these organisms after long-term exposure and bioaccumulation of pollutants in invertebrates and fish (Li et al., 2014).

limited professional knowledge or improper prescription habits, βlactams were the dominant compounds. Temperature and NH3-N were identified as playing important roles in the distribution of PhACs. Based on source apportionment, domestic sewage was regarded as the main source, followed by insufficient wastewater treatment. Therefore, it is important to control discharges of domestic sewage into the aquatic system and improve the infrastructure facilities and treatment technology in WWTPs. Risk assessment showed that sulfamethoxazole, clarithromycin, and azithromycin might present high risks to algae (RQ N 1). Moreover, the mixture effect of multiple-compounds showed higher potential risks to aquatic organisms than expected. However, the toxicity estimated in this study may not reflect the overall toxicity of pharmaceuticals on the river ecosystem, since not all contaminants were included. Mixture toxicity of pharmaceuticals to aquatic organisms or even humans still requires further investigation.

4. Conclusion Acknowledgement In this study, the first investigation of the seasonal variation and spatial distribution of 36 PhACs was conducted on the Xiangjiang River. Twenty-five compounds were detected with concentrations ranging from ND to 830 ng L−1 (cefotaxime). Pharmaceutical contaminant levels were moderate compared with previously reported data on a global scale. Higher temperature and the dilution effect of stream flow in summer could be responsible for the dramatic seasonal variation of most compounds (p b 0.05). Overall pharmaceutical contaminations were more serious in densely-populated areas. In rural areas, due to

This work was financially supported by the Special Fund for Agroscientific Research in the Public Interest (No. 201503108) and Science & Technology Project of Hunan Province (No. 2017WK2091). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2018.04.267.

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

References Aristilde, L., Melis, A., Sposito, G., 2010. Inhibition of photosynthesis by a fluoroquinolone antibiotic. Environ. Sci. Technol. 44, 1444–1450. Backhaus, T., Faust, M., 2012. Predictive environmental risk assessment of chemical mixtures: a conceptual framework. Environ. Sci. Technol. 46, 2564–2573. Backhaus, T., Karlsson, M., 2014. Screening level mixture risk assessment of pharmaceuticals in STP effluents. Water Res. 49, 157–165. Barceló, D., Petrovic, M., 2007. Pharmaceuticals and personal care products (PPCPs) in the environment. Anal. Bioanal. Chem. 387, 1141–1142. Blanchet, F.G., Legendre, P., Borcard, D., 2008. Forward selection of explanatory variables. Ecology 89, 2623–2632. Bureau, S. E. P, 2002. Methods of Monitoring and Analysis for Water and Wastewater. Chinese Environment Science Press, Beijing. Corada-Fernandez, C., Candela, L., Torres-Fuentes, N., Pintado-Herrera, M.G., Paniw, M., Gonzalez-Mazo, E., 2017. Effects of extreme rainfall events on the distribution of selected emerging contaminants in surface and groundwater: the Guadalete River basin (SW, Spain). Sci. Total Environ. 605, 770–783. Currie, J., Lin, W., Zhang, W., 2011. Patient knowledge and antibiotic abuse: evidence from an audit study in China. J. Health Econ. 30, 933–949. da Silva, B.F., Jelic, A., Lopez-Serna, R., Mozeto, A.A., Petrovic, M., Barcelo, D., 2011. Occurrence and distribution of pharmaceuticals in surface water, suspended solids and sediments of the Ebro river basin, Spain. Chemosphere (8), 1331–1339. Dai, G., Wang, B., Fu, C., Dong, R., Huang, J., Wang, Y., Deng, S., Yu, G., 2016. Pharmaceuticals and personal care products (PPCPs) in urban and suburban rivers of Beijing, China: occurrence, source apportionment and potential ecological risk. Environ. Sci.: Processes Impacts 18, 445–455. Del Rosario, K.L., Mitra, S., Humphrey, C.P., O'Driscoll, M.A., 2014. Detection of pharmaceuticals and other personal care products in groundwater beneath and adjacent to onsite wastewater treatment systems in a coastal plain shallow aquifer. Sci. Total Environ. 487, 216–223. Ding, H., Wu, Y., Zhang, W., Zhong, J., Lou, Q., Yang, P., Fang, Y., 2017. Occurrence, distribution, and risk assessment of antibiotics in the surface water of Poyang Lake, the largest freshwater lake in China. Chemosphere 184, 137–147. Du, J., Zhao, H., Liu, S., Xie, H., Wang, Y., Chen, J., 2017. Antibiotics in the coastal water of the South Yellow Sea in China: occurrence, distribution and ecological risks. Sci. Total Environ. 595, 521–527. Ferguson, P.J., Bernot, M.J., Doll, J.C., Lauer, T.E., 2013. Detection of pharmaceuticals and personal care products (PPCPs) in near-shore habitats of southern Lake Michigan. Sci. Total Environ. 458, 187–196. Hernando, M.D., Mezcua, M., Fernandez-Alba, A.R., Barcelo, D., 2006. Environmental risk assessment of pharmaceutical residues in wastewater effluents, surface waters and sediments. Talanta 69, 334–342. Hijosa-valsero, M., Matamoros, V., Sidrachcardona, R., Pedescoll, A., Martínvillacorta, J., García, J., Bayona, J.M., Becares, E., 2011. Influence of design, physico-chemical and environmental parameters on pharmaceuticals and fragrances removal by constructed wetlands. Water Sci. Technol. 63, 2527–2534. Jiang, L., Hu, X., Yin, D., Zhang, H., Yu, Z., 2011. Occurrence, distribution and seasonal variation of antibiotics in the Huangpu River, Shanghai, China. Chemosphere 82, 822–828. Jiang, L., Hu, X., Xu, T., Zhang, H., Sheng, D., Yin, D., 2013. Prevalence of antibiotic resistance genes and their relationship with antibiotics in the Huangpu River and the drinking water sources, Shanghai, China. Sci. Total Environ. 458–460, 267–272. Jiang, J.J., Lee, C.L., Brimblecombe, P., Vydrova, L., Fang, M.D., 2016. Source contributions and mass loadings for chemicals of emerging concern: Chemometric application of pharmaco-signature in different aquatic systems. Environ. Pollut. 208, 79–86. Kim, Y., Lee, K.B., Choi, K., 2016. Effect of runoff discharge on the environmental levels of 13 veterinary antibiotics: a case study of Han River and Kyungahn Stream, South Korea. Mar. Pollut. Bull. 107, 347–354. Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., Barber, L.B., et al., 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999–2000: a national reconnaissance. Environ. Sci. Technol. 36, 1202–1211. Kosma, C.I., Lambropoulou, D.A., Albanis, T.A., 2010. Occurrence and removal of PPCPs in municipal and hospital wastewaters in Greece. J. Hazard. Mater. 179, 804–817. Kummerer, K., 2009. Antibiotics in the aquatic environment-a review-part I. Chemosphere 75, 417–434. Lalumera, G.M., Calamari, D., Galli, P., Castiglioni, S., Crosa, G., Fanelli, R., 2004. Preliminary investigation on the environmental occurrence and effects of antibiotics used in aquaculture in Italy. Chemosphere 54, 661–668. Larsen, R.K., Baker, J.E., 2003. Source apportionment of polycyclic aromatic hydrocarbons in the urban atmosphere: a comparison of three methods. Environ. Sci. Technol. 37, 1873–1881. Lepš, J., Šmilauer, P., 2003. Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press. Li, W., Shi, Y., Gao, L., Liu, J., Cai, Y., 2012. Occurrence of antibiotics in water, sediments, aquatic plants, and animals from Baiyangdian Lake in North China. Chemosphere 89, 1307–1315. Li, N., Zhang, X., Wu, W., Zhao, X., 2014. Occurrence, seasonal variation and risk assessment of antibiotics in the reservoirs in North China. Chemosphere 111, 327–335. Lin, H., Li, H., Chen, L., Li, L., Yin, L., Lee, H., Yang, Z., 2017. Mass loading and emission of thirty-seven pharmaceuticals in a typical municipal wastewater treatment plant in Hunan Province, Southern China. Ecotoxicol. Environ. Saf. 147, 530–536. Liu, J.L., Wong, M.H., 2013. Pharmaceuticals and personal care products (PPCPs): a review on environmental contamination in China. Environ. Int. 59, 208–224. Liu, J., Lu, G., Xie, Z., Zhang, Z., Li, S., Yan, Z., 2015. Occurrence, bioaccumulation and risk assessment of lipophilic pharmaceutically active compounds in the downstream rivers of sewage treatment plants. Sci. Total Environ. 511, 54–62.

983

Long, Y., Dai, T., Wu, Q., 2013. Sources and distribution of polycyclic aromatic hydrocarbons in street dust from the Chang-Zhu-Tan Region, Hunan, China. Environ. Monit. Assess. 185, 1377–1390. Lopez-Serna, R., Petrovic, M., Barcelo, D., 2012. Occurrence and distribution of multi-class pharmaceuticals and their active metabolites and transformation products in the Ebro river basin (NE Spain). Sci. Total Environ. 440, 280–289. Luo, Y., Xu, L., Rysz, M., Wang, Y., Zhang, H., Alvarez, P.J., 2011. Occurrence and transport of tetracycline, sulfonamide, quinolone, and macrolide antibiotics in the Haihe River Basin, China. Environ. Sci. Technol. 45, 1827–1833. MAPRC, 2015. Decision on the Prohibition of Four Veterinary Drugs (Lomefloxacin, Pefloxacin, Ofloxacin and Norfloxacin) in Food Animals. Ministry of Agriculture of the People's Republic of China. Osorio, V., Larranaga, A., Acena, J., Perez, S., Barcelo, D., 2016. Concentration and risk of pharmaceuticals in freshwater systems are related to the population density and the livestock units in Iberian Rivers. Sci. Total Environ. 540, 267–277. Paiga, P., Santos, L.H., Ramos, S., Jorge, S., Silva, J.G., Delerue-Matos, C., 2016. Presence of pharmaceuticals in the Lis river (Portugal): sources, fate and seasonal variation. Sci. Total Environ. 573, 164–177. Pan, H., Lu, X., Lei, K., 2017. A comprehensive analysis of heavy metals in urban road dust of Xi'an, China: contamination, source apportionment and spatial distribution. Sci. Total Environ. 609, 1361–1369. Pinto, L.M.P., Phillips, M., Ramlal, H., Teemul, K., Prabhakar, P., 2004. Third generation cephalosporin use in a tertiary hospital in Port of Spain, Trinidad: need for an antibiotic policy. BMC Infect. Dis. 4, 59. Sarmah, A.K., Meyer, M.T., Boxall, A.B., 2006. A global perspective on the use, sales, exposure pathways, occurrence, fate and effects of veterinary antibiotics (VAs) in the environment. Chemosphere 65, 725–759. Singh, K.P., Rai, P., Singh, A.K., Verma, P., Gupta, S., 2014. Occurrence of pharmaceuticals in urban wastewater of north Indian cities and risk assessment. Environ. Monit. Assess. 186, 6663–6682. Sui, Q., Huang, J., Deng, S., Chen, W., Yu, G., 2011. Seasonal variation in the occurrence and removal of pharmaceuticals and personal care products in different biological wastewater treatment processes. Environ. Sci. Technol. 45, 3341–3348. Varela, A.R., Andre, S., Nunes, O.C., Manaia, C.M., 2014. Insights into the relationship between antimicrobial residues and bacterial populations in a hospital-urban wastewater treatment plant system. Water Res. 54, 327–336. Veach, A., Bernot, M.J., Mitchell, J.K., 2012. The influence of six pharmaceuticals on freshwater sediment microbial growth incubated at different temperatures and UV exposures. Biodegradation 23, 497–507. Vorosmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P., Glidden, S., Bunn, S.E., Sullivan, C.A., Liermann, C.R., Davies, P.M., 2010. Global threats to human water security and river biodiversity. Nature 467, 5561. Wu, C., Huang, X., Witter, J.D., Spongberg, A.L., Wang, K., Wang, D., et al., 2014. Occurrence of pharmaceuticals and personal care products and associated environmental risks in the central and lower Yangtze river, China. Ecotoxicol. Environ. Saf. 106, 19–26. Xie, W., 2016. Analysis on the current status and regulatory measures of water pollution in the Xiang River Basin of China. Nat. Environ. Pollut. Technol. 15, 1435. Xu, W.H., Zhang, G., Zou, S.C., Li, X.D., Liu, Y.C., 2007. Determination of selected antibiotics in the Victoria Harbour and the Pearl River, South China using high-performance liquid chromatography-electrospray ionization tandem mass spectrometry. Environ. Pollut. 145, 672–679. Xu, H., Xu, C.Y., Chen, H., Zhang, Z., Li, L., 2013. Assessing the influence of rain gauge density and distribution on hydrological model performance in a humid region of China. J. Hydrol. 505, 1–12. Yan, C., Yang, Y., Zhou, J., Nie, M., Liu, M., Hochella Jr., M.F., 2015. Selected emerging organic contaminants in the Yangtze Estuary, China: a comprehensive treatment of their association with aquatic colloids. J. Hazard. Mater. 283, 14–23. Yang, X., Chen, F., Meng, F., Xie, Y., Chen, H., Young, K., Luo, W., Ye, T., Fu, W., 2013. Occurrence and fate of PPCPs and correlations with water quality parameters in urban riverine waters of the Pearl River Delta, South China. Environ. Sci. Pollut. Res. 20, 5864–5875. Yang, Y.Y., Liu, W.R., Liu, Y.S., Zhao, J.L., Zhang, Q.Q., Zhang, M., et al., 2017. Suitability of pharmaceuticals and personal care products (PPCPs) and artificial sweeteners (ASs) as wastewater indicators in the Pearl River Delta, South China. Sci. Total Environ. 590, 611–619. Yao, L., Wang, Y., Tong, L., Deng, Y., Li, Y., Gan, Y., Guo, W., Dong, C., Duan, Y., Zhao, Y., 2017. Occurrence and risk assessment of antibiotics in surface water and groundwater from different depths of aquifers: a case study at Jianghan Plain, central China. Ecotoxicol. Environ. Saf. 135, 236–242. Yokota, H., Eguchi, S., Hasegawa, S., Okada, K., Yamamoto, F., Sunagawa, A., Tanaka, M., Yamamoto, R., Nakano, E., 2015. Assessment of in vitro antiovulatory activities of nonsteroidal anti-inflammatory drugs and comparison with in vivo reproductive toxicities of medaka (Oryzias latipes). Environ. Toxicol. 31, 1710–1719. Zhang, Z., Tao, F., Du, J., Shi, P., Yu, D., Meng, Y., Sun, Y., 2010. Surface water quality and its control in a river with intensive human impacts-a case study of the Xiangjiang River, China. J. Environ. Manag. 91, 2483–2490. Zhang, Q., Jia, A., Wan, Y., Liu, H., Wang, K., Peng, H., Dong, Z., Hu, J., 2014. Occurrences of three classes of antibiotics in a natural river basin: association with antibioticresistant Escherichia coli. Environ. Sci. Technol. 48, 14317–14325. Zhao, H., Cao, Z., Liu, X., Zhan, Y., Zhang, J., Xiao, X., et al., 2017. Seasonal variation, flux estimation, and source analysis of dissolved emerging organic contaminants in the Yangtze Estuary, China. Mar. Pollut. Bull. 125, 208–215.

984

H. Lin et al. / Science of the Total Environment 636 (2018) 975–984

Zhou, L.J., Ying, G.G., Liu, S., Zhao, J.L., Chen, F., Zhang, R.Q., Peng, F.Q., Zhang, Q.Q., 2012. Simultaneous determination of human and veterinary antibiotics in various environmental matrices by rapid resolution liquid chromatography-electrospray ionization tandem mass spectrometry. J. Chromatogr. A 1244, 123–138. Zhou, H., Ying, T., Wang, X., Liu, J., 2016. Occurrence and preliminarily environmental risk assessment of selected pharmaceuticals in the urban rivers, China. Sci. Rep. 6, 34928.

Zou, S., Xu, W., Zhang, R., Tang, J., Chen, Y., Zhang, G., 2011. Occurrence and distribution of antibiotics in coastal water of the Bohai Bay, China: impacts of river discharge and aquaculture activities. Environ. Pollut. 159, 2913–2920. Zuccato, E., Castiglioni, S., Bagnati, R., Melis, M., Fanelli, R., 2010. Source, occurrence and fate of antibiotics in the Italian aquatic environment. J. Hazard. Mater. 179, 1042–1048.