Occurrence and risk evaluation of endocrine-disrupting chemicals in wastewater and surface water of Lahore, Pakistan

The current study highlights the occurrence, spatial distribution, and risk assessment of 16 endocrine-disrupting chemicals (EDCs) including their transformation products (TPs) in the wastewater and surface water of Lahore, Pakistan, using solid-phase extraction followed by liquid chromatography–mass spectrometry and gas chromatography–mass spectrometry. The parent EDCs include bisphenol A (BPA), triclosan (TCS), triclocarban (TCC), estrone (E1), estradiol (E2), estriol (E3), ethinylestradiol (EE2), 4-n-octylphenol (4n-OP), and 4-n-nonylphenol (4n-NP). The TPs include two TPs each of BPA, TCC, and estrogens along with a TP of TCS. Most EDCs showed 100% detection frequency in the wastewater with highest median concentration of 1310 ng/L for E3. In the surface water, the highest median concentration was, however, observed for BPA (54.6 ng/L). Spatial variations in terms of sum of concentration due to all EDCs and their TPs were observed at different sampling points which suggest contamination due to industrial waste from nearby industrial estate. Risk evaluation in terms of risk quotient (RQ) and estradiol equivalent factor (EEQ) showed that most of EDCs and their TPs could pose high risk and estrogenicity to the surrounding environment. From the results of the current study, it is observed that the environment of Pakistan is deteriorating and is potential risk for endocrine disruption. It is, therefore, recommended to take stringent measures to make it sustainable for current as well as for future generations.


Introduction
Endocrine-disrupting chemicals (EDCs) constitute an important group of emerging contaminants which are posing serious effects to animals and humans, such as interference in biosynthesis of hormones as well as their metabolism (Diamanti-Kandarakis et al., 2009), interference in functioning of adrenal, pituitary, and thyroid glands, birth defects, and tumor development (Balabanič et al., 2011). They may also cause effects on reproduction, neurobehavioral disorders, and other health disorders (Ibor et al., 2022). Among the listed EDCs, majority is produced synthetically (Burkhardt-Holm, 2010) which includes hormones, some chemicals used in pharmaceutical and personal care products, etc. (Le, 2012;Markey et al., 2002). The most common and ubiquitously detected EDCs in the environment include bisphenol A (BPA), alkylphenol ethoxylates, triclosan (TCS), triclocarban (TCC), and estrogens (Balabanič et al., 2011;Diamanti-Kandarakis et al., 2009;Lu et al., 2012).
BPA is manufactured in very high quantity for its use in the production of epoxy resins and polycarbonate plastics (Andra et al., 2015;Bhandari et al., 2015). It can be transformed to BPA monomethyl ether (BPA-MME) and bisphenol A dimethyl ether (BPA-DME) by Mycobacterium (McCormick et al., 2011), which exhibits more toxicity to developing zebrafish embryo than BPA itself. The estrogen family constitutes both natural [estrone (E1), estradiol (E2), and estriol (E3)] and synthetic [(ethinylestradiol (EE2)] estrogens. The natural estrogens can be transformed into each other in addition to transforming themselves into other transformation products (TPs). These estrogens may cause significant endocrinedisrupting effect in fish even at very low concentrations (Wang et al., 2018). The antimicrobials (TCC and TCS) are considered as the most ubiquitously detected contaminants in the world (Halden & Paull, 2005;Kolpin et al., 2002). TCC is transformed to dichlorocarbanilide (DCC) and carbanilide (NCC), whereas TCS into methyl triclosan (MeTCS), under different environmental conditions, which can also pose potential environmental risks (Armstrong et al., 2017;Lozano et al., 2013;Miller et al., 2008). In addition, the nonionic surfactants, such as alkylphenol ethoxylates which are commonly used in many household and industrial applications (Krogh et al., 2003), are transformed/degraded to many TPs such as 4-n-octylphenol (4n-OP) and 4-n-nonylphenol (4n-NP), which also show estrogenic activities. Considering their associated endocrine disruption risks, the investigation for the occurrence and distribution of these EDCs and their TPs in differential environmental matrices is of great importance.
Pakistan is a developing country with population around 220 million with poor environmental management system. Lahore is the second most populated city of Pakistan inhabiting about 11 million people. However, to the best of our knowledge, this city does not possess any type of wastewater treatment facilities, which lead to the direct discharge of wastewater into surface water. Several studies have been conducted throughout the world for the monitoring different EDCs in the wastewater treatment plants (WWTPs) as well as in surface water; however, most of these studies either include very few TPs or without TPs (Barber et al., 2015;Belhaj et al., 2015;Lozano et al., 2013;Tohidi & Cai, 2017). The most abundant EDCs detected in wastewater of different countries include E3 (398 ng/L) in Tunisia (Belhaj et al., 2015), E1 (187 ng/L) in Turkey (Komesli et al., 2015), E3 (586 ng/L), TCC (204 ng/L), TCS (184 ng/L), and DCC (3.0 ng/L) in China (Ashfaq et al., 2018a). Similarly, concentration detected in surface water includes BPA-MME (2.75 ng/L), BPA DME (20.7 ng/L), E1 (101 ng/L), NCC (66.5 ng/L), and MeTCS (165 ng/L) in China (Ashfaq et al.), TCS (62.1 ng/L) and BPA (1135 ng/L) in India (Saha et al., 2022). In addition, no study was conducted covering the environment of Pakistan. Thus, it was deemed necessary to conduct such studies in the environment of Pakistan in order to cover this knowledge gap.
The current study was designed to determine nine EDCs and seven of their TPs in the wastewater and surface water of Lahore, Pakistan, to investigate the spatial variation in their occurrence, and to evaluate their risk assessment. This is the first-ever study that reports the level of EDCs and their TPs in the environment of Lahore, Pakistan, and the results will create a baseline data for designing future strategies for sustainable environment. In addition, the results of the current study will open new horizons for investigating different sources which may enhance their concentrations along with the factors contributing transformation of different EDCs and effects of these TPs on human health.

Chemicals and reagents
Reference standards of all investigated chemicals except BPA-MME were purchased from Acros Organics and Sigma-Aldrich (USA). BPA-MME was synthesized in our laboratory as it was not available in the commercial market. The detailed procedure for the synthesis of BPA-MME along with its characterization details has been reported previously (Ashfaq et al., 2018). The stock solution of OP, NP, MeTCS, BPA, BPA-MME, and BPA-DME was prepared in acetone whereas all other EDCs were prepared in methanol (1 g/L each). Ultra-pure water was obtained from Milli-Q (USA) purification system. All other chemicals and reagents used were of either HPLC/ analytical grade or having highest purity available.

Sampling details
Lahore (31° 32′ 59″ N/74° 20′ 37″ E) is the densely populated and second largest city of Pakistan. It is the capital city of its largest province Punjab with population of around 11 million. Due to high urbanization and subsequent increase in population, there is increasing use of water resulting in large wastewater generation. There is no proper drainage system in the city, and the wastewater is discharged through open drains and their tributaries (Majeed et al., 2018) into the River Ravi (surface water). A canal (Lahore canal) carrying surface water passes through the heart of the city, used mainly for irrigation purpose and also mingle with River Ravi. The details of sampling have already been described in the previous studies (Ashfaq et al., 2019). Briefly, wastewater samples were collected from two wastewater drains (Cantonment Drain and Shama Drain) whereas surface water samples from Lahore canal at suitable locations identified through global positioning system (GPS) from October 27 to 29, 2015 ( Fig. 1). Five wastewater samples from each of two drains (W1-W5 for Cantonment Drain and S1-S5 for Shama Drain) and four samples of surface water from Lahore canal (C1-C4) were collected. Each site was sampled thrice on the same day in different times, and then, their composite was made. The wastewater drains receive wastewater from domestic and commercial sites with W3-W4 are most likely to receive waste from industrial sector due to the presence of industrial estate close to these points. The Lahore canal receives water from another canal called as Bambawali-Ravi-Bedian Canal (BRB canal), which itself originated from river Chenab. There is no population around C1, whereas C2 and C4 are the densely populated areas, while C3 has lesser population around it. After collection (1 L each), samples were immediately shifted to EPA Punjab, Pakistan, laboratory for sample preparation. The samples were filtered immediately, treated with disodium edetate, and then extracted within 2 days of sample collection. After sample preparation, samples were shipped to China for further analysis.

Sample preparation and analysis
Sample preparation was carried out using USEPA method 1694 (Englert, 2007) with slight modification as also used previously (Ashfaq et al.), the details of which are available in the previous study and in supplementary information (SI). The instrumental conditions for both GC-MS and LC-MS have been provided in the previous studies (Ashfaq et al., 2018a, 2018b) with details of precursor/ daughter ions along with qualitative and quantitative ions in Table S1 of SI.
Quality assurance and quality control (QA/QC) QA/QC procedures were followed for both qualitative and quantitative analyses of all investigated compounds in both GC-MS and LC-MS. During the analysis of each batch, blank spiked and matrix spiked solutions were analyzed in addition to instrumental blank and procedural blank. Recoveries were calculated by spiking known concentration of Map showing sampling locations of wastewater and surface water from different locations of Lahore, Pakistan. C1-C4, surface water from Lahore canal; W1-W5, wastewater from Cantonment Drain; and S1-S5, wastewater from Sharma Drain reference standard into sample matrix and ultra-pure water. Recoveries and quantification limits of all investigating compounds are provided in Table S1.

Risk evaluation
Theoretical risk evaluation of detected EDCs was conducted using conventional approach on the species living in the freshwater environment as also used in the previous studies (Ashfaq et al., 2017a(Ashfaq et al., , 2017b(Ashfaq et al., , 2019. The risk evaluation was expressed by risk quotient and was calculated by the following equation: where RQ is the risk quotient, MEC is the maximum measured concentration of each detected EDC in different sample types, and PNEC is the predicted no effect concentration. In addition, estrogenicity was also calculated in terms of estradiol equivalent concentrations (EEQs) by using following formula as used in the previous study (Vega-Morales et al., 2013).
where C i is the mean concentration of an individual EDC in a sample type, i.e., CD, SD, or Lahore canal, and EEF i is the estradiol equivalent factor of that EDC. The EEF i values for detected EDCs were taken from the previously reported study (Vega-Morales et al., 2013).

Occurrence of EDCs in wastewater drains
The range, mean, median concentration, and detection frequency of all the EDCs are presented in Table 1. Among the 16 EDCs, 14 EDCs were detected in the Cantonment Drain and 15 EDCs in the Shama Drain. The detection frequency of 14 EDCs in Cantonment Drain and 13 EDCs in Shama Drain was found 100%. The TP of estradiol, i.e., 4-OH E2 could not be detected in both the drains whereas TP of TCC, i.e., NCC was below the detection limits (BLD) in Cantonment Drain, whereas it showed low detection frequency (20%) in Shama Drain. Overall, E3 showed the highest median concentration (1310 ng/L) in Cantonment Drain and 800 ng/L in Shama Drain followed by OP with median concentration of 884 ng/L and 528 ng/L, respectively.
Among the alkylphenols, the median concentration of OP was found to be 884 ng/L in the Cantonment Drain and 528 ng/L in Shama Drain, whereas the median concentration of NP was found to be 448 ng/L and 400 ng/L, respectively. Comparing their concentration with the previous studies showed higher concentrations in the USA and China (Barber et al., 2015;Qiang et al., 2013;Yu et al., 2013) with comparable concentrations in the wastewater of Spain (Padilla-Sánchez et al., 2011).
The plasticizer BPA and its transformation product BPA-DME showed 100% detection frequency in both the drains. The detection frequency of BPA-MME was 100% in Cantonment Drain while 20% in Shama Drain. BPA was found with median concentration of 441 ng/L in Cantonment Drain and 452 ng/L in Shama Drain as shown in Figs. 2 and 3. Among the two TPs of BPA, the median concentration of BPA-MME was higher (62.2 ng/L) than the median concentration of BPA-DME (43.4 ng/L) in Cantonment Drain. In Shama Drain, BPA-DME was, however, found in very high concentration (median = 50.2 ng/L) compared to the concentration of BPA-MME (mean = 3.02 ng/L, median = BLD). The concentrations detected for both BPA-MME and BPA-DME (during the current study) were comparable with previous results which monitor the influent and effluent of wastewater treatment plant (Ashfaq et al., 2018). The concentration detected for BPA during this study was comparable to the concentration detected in the influent of most part of the world including our previous studies in China (Archer et al., 2017;Sun et al., 2017).
Among the antimicrobials and their TPs, the highest median concentration was observed for MeTCS in Cantonment Drain (53.6 ng/L) followed by TCS (34.2 ng/L) whereas the median concentration of TCS was little higher than MeTCS in Shama Drain. The transformation of TCS into MeTCS has also been reported in the previous studies with different degrees of transformation in different compartments of WWTP (Armstrong et al., 2017). The concentration Table 1 The concentration range, median concentration, detection frequencies, and method quantification limits ( of MeTCS observed during the current study was slightly higher than the previous studies monitoring antimicrobials and their TPs in the influent of WWTP of the USA and China (Ashfaq 2018a;Lozano et al., 2013). The concentration of TCS was, however, lower than most of the previous studies that measured its concentration in the influent and effluent of WWTPs (Ashfaq et al., 2018a(Ashfaq et al., , 2019Lozano et al., 2013;Tohidi & Cai, 2017). Photolysis is a common attenuation way for the variety of emerging contaminants which is influenced by both light intensity and water parameters such as pH, organic matter, or concentration of oxygen (Gmurek et al., 2017). Compared to WWTPs, the photolysis might be more pronounced in drains, which are less deep and hence photolysis might contribute more. The other antimicrobial agent, i.e., TCC was found higher in Cantonment Drain than in Shama Drain. However, the median concentration of DCC and NCC was almost equal in both the drains. Comparing the concentration of TCC with the published results, it was observed that the concentration in the current study was either comparable or low compared to its concentration in the influent and effluent of the previous studies (Lozano et al., 2013;Lv et al., 2014). The higher concentration of DCC than NCC indicates that the transformation from TCC to DCC was more likely than NCC, or DCC was more stable than NCC in the aquatic environment. The previous studies have shown that transformation of TCC to NCC and DCC is high, likely under anaerobic conditions (Chiaia-Hernandez et al., 2012;Miller et al., 2010;Venkatesan et al., 2012) compared to aerobic conditions (Kwon & Xia, 2012). The results of the current study were in good agreement with those reports where transformation from TCC to NCC was considered as rate limiting step (Souchier et al., 2015;Venkatesan et al., 2012) because of low/negligible detection of NCC. Among the estrogens, E3 was observed in the highest concentration in both the drains with median concentration of 1310 ng/L in Cantonment Drain and 800 ng/L in Shama Drain, followed by EE2 with median concentration of 174.2 ng/L and 161 ng/L, respectively. Among their TPs, 4-OH E2 could not be detected in both the drains, whereas 4-OH E1 was found with 100% detection frequency in both drains with median concentration of 107.8 ng/L and 94.2 ng/L, respectively. The median concentration of E2 was almost equal in both the drains whereas the median concentration of E1 was higher in Cantonment Drain. The concentration of estrogens detected during the current study showed either comparable or little greater concentration especially for EE2 than those detected worldwide in the influent and effluent of WWTPs (Atkinson et al., 2012;Behera et al., 2011;Blair et al., 2015;Gabet-Giraud et al., 2010;Komesli et al., 2015;Kumar et al., 2011;Manickum & John, 2014;Yu et al., 2013). Comparing the concentration of TPs, 4-OH E1 was found in very high concentration compared to BLD concentration in the previous study (Ashfaq 2018a). The concentration of E1 was higher than E2 during the current study as E2 could be converted to E1 as reported in the previous studies (Shi et al., 2014).
Among the sampling points of Cantonment Drain, the highest cumulative concentration due to all investigated EDCs was found at W3 (5065 ng/L) followed by W1 (4091 ng/L). In the Shama Drain, S3 showed highest cumulative concentration of EDCs followed by S4 (Fig. 1). The highest cumulative concentrations at W3 may be to the discharge of waste by nearby industrial estate whereas at S3 due to higher population.

Occurrence of EDCs in surface water
Among the 16 EDCs and their TPs, 10 compounds were detected at least once in the aqueous samples collected from Lahore canal (Fig. 4), with six of them showing 100% detection frequency. The concentration of most of these compounds was very low compared to their concentration in the drain wastewater. BPA was found with highest median concentration of 54.6 ng/L followed by NP (20.4 ng/L median concentration). E3 was below the detection limits (BLD) at all the sampling points although it was detected in highest concentration in the drain water. Although BPA was detected in highest concentration, both of its TPs were found below the detection limits. Similarly, the median value of MeTCS was BLD although the detection frequency of TCS was 100%. The concentrations of TCS and MeTCS of the current study were comparable with the previous study (Wang & Kelly, 2017). TCC and its TPs showed high detection frequency with comparable median concentration of NCC and DCC. The detection frequency and concentration of NCC was even higher than that observed in the drainage samples. This information suggests that transformation from TCC to NCC most likely occurred in surface water, or there might be other potential source of NCC in the surface water. The concentration detected in the current study was slightly higher than the concentration reported previously in Jiulong River and estuary of China (Ashfaq et al., 2019a(Ashfaq et al., , 2019b, the transformation behavior of TCC to NCC and DCC was, however, found almost similar. Among the two detected estrogens, only EE2 showed detection frequency of 100% with median value of 14.1 ng/L while detection frequency of E2 was 25% with mean concentration of 1.26 ng/L. The concentration of EE2 detected during the current study was higher than the previous studies whereas E2 was found very low in concentration compared to its concentration in surface water of China (Wang et al., 2018). EE2 is highly resistant to biodegradation compared to E2; thus, higher concentration of EE2 was detected as also reported previously (Jürgens et al., 2002).

Spatial variations
Spatial variation was evaluated by calculating the concentration of each EDC at different sampling points (Fig. 5). The distribution map showed that both the drains were highly contaminated with cumulative EDCs. No significant correlation was observed when population density around each sampling point was plotted against the sum of EDCs at that sampling point. The sum concentration of all EDCs in the Cantonment Drain was highest at W3 which was in contrary to population density as it was lower around W3 compared to around W2 and W4. A possible reason for higher concentration at W3 is the discharge of nearby industrial estate waste to W3. Transformation was confirmed when evaluating the concentration at each sampling point where TCC steadily decreased from W3 to W5 whereas DCC showed corresponding increase at those points. No correlation was observed among the concentration of BPA and its transformation products, which suggested that transformation of BPA was varied under the influence of different environmental factors. However, there was a decrease in the concentration for BPA-MME from W3 to W5 with BPA showing corresponding increase at these points. In Shama Drain, the sum of the concentration was almost comparable at all the sampling points with little fluctuation, and no correlation between population density and concentration was found. A steady increasing trend was, however, observed for E1 and E3 from S1 to S3 with corresponding decrease of E2. This corresponding decrease in E2 with subsequent increase in E1 and E3 was probably due to transformation of E2 to E1 and E3 (Belhaj et al., 2015). 4-OH E1 showed the similar behavior as shown by its parent compound, i.e., E1. In case of surface water, most of the detected EDCs showed almost comparable concentrations at all sampling points such as NP, TCC, DCC, and EE2. A significant positive correlation was, however, observed between population density and the concentration of TCC with value of Pearson correlation coefficient 0.9899 (p ˂ 0.05).

Risk evaluation
Risk evaluation of selected EDCs was conducted in two ways by calculating both risk quotient (RQ) and estrogenicity using Eqs. (1 and 2). RQ is generally evaluated to assess the dose of a contaminant which may be harmful to the freshwater species and is calculated by dividing the maximum measured environmental concentration of a particular contaminant with its predicted no effect concentration (Bouissou-Schurtz et al., 2014). RQ ˂0 .1 suggests no risk, 0.1 ≤ RQ ≤ 1.0 suggests medium risk, and RQ ˃ 1.0 suggests high risk (Verlicchi et al., 2012). The results are shown in Figs. 6 and 7. In the drain water, EE2 showed the highest RQ (600 in SD) against fish showing it to be highest risk posed EDC followed by E2 (RQ = 80) against aquatic species. All other EDCs showed highest risk with RQ values greater than 1 except BPA and TCC whose risk was found  medium. In the surface water samples, all EDCs showed either minimal or medium risk except E2 and MeTCS whose risk was found high with RQ of 12.5 and 1.40, respectively. It is to mention here that the actual risk calculated in the wastewater may be low due to dilution of wastewater when mixing with surface water. Based on the RQ values, high risk is expected in the current study area; however, the risk evaluation in terms of RQ calculation sometimes do not provide real risk picture because of their sublethal effects (Windsor et al., 2018). However, most of the EDCs were detected with values well below their sublethal concentrations except E2 whose very low concentration (1 ng/L) was found to induce oxidative stress in different parts common carp (Cyprinus carpio) (Gutiérrez-Gómez et al., 2016) in addition to disturb sexual characteristics Poecilia reticulata with sublethal value of 30 ng/L (Toft & Baatrup, 2003). Therefore, E2 may cause the sublethal effect in the area under study due to the detection of higher concentration than the sublethal concentration.
The risk in terms of EEQ was also evaluated according to Eq. (2). EE2 was highest estrogenic chemical in all the three sample types with mean EEQ of 210, 211, and 18.4, respectively, in CD, SD, and Lahore canal, respectively. All other estrogens (E1, E2, and E3) also showed high EEQ values compared to the previously reported values (Vega-Morales et al., 2013). The EEQ of OP and NP was low compared to the estrogens; however, they were considerable high compared to other studies (Vega-Morales et al., 2013).
The sum of estrogenicity evaluation at each sampling point (Fig. 7) revealed that all the sampling sites showed very high EEQ values. Among the sampling points of Cantonment Drain, W3 was highest estrogenic with EEQ value of 479 followed by W1 and W4 with EEQ value of 458 and 405, respectively. Among the sampling points of Shama Drain, S3 and S5 were almost equally estrogenic followed by S4, whereas in the Lahore canal, C1 and C4 showed almost equal estrogenicity followed by the other two points. The estrogenicity at all the sampling points of Lahore canal was solely due to EE2 except at C4 where E2 also contributed about 30% of the total estrogenicity. The higher EEQ values of almost all investigated EDCs and the higher sum EEQ values at every sampling point might cause higher endocrine disruption activity in the environment of Pakistan.

Conclusion
Nine EDCs and seven of their TPs were monitored in the wastewater and surface water of Lahore, Pakistan, along with their spatial variation and risk evaluation. The studies showed that the urban wastewater and surface water of Lahore was highly contaminated with EDCs and their TPs. The highest concentration in the wastewater was observed for estrogen E3, whereas in the surface water, the plasticizer BPA was found in highest concentration. The TPs were also abundant in the samples particularly in wastewater samples, suggesting that wastewater environment was most suitable for possible transformation. Spatial variations suggest contamination of wastewater through industrial source. Risk evaluation in terms of RQ and EEQ showed that most EDCs and their TPs could pose high risk and estrogenicity to the surrounding environment. The studies, therefore, recommend detailed studies regarding occurrence, transformation, and risk evaluation of all EDCs including currently reported in other environmental matrices also in order to generate baseline data for future strategic applications.