This study evaluates the removal efficiency of 15 estrogenic endocrine-disrupting compounds in two operational constructed wetlands with different designs: a hybrid system (constructed wetland A) and a horizontal system (constructed wetland B). The assessment involved analyzing composite water samples obtained from passive samplers through liquid chromatography-mass spectrometry coupled with yeast assays. Additionally, grab samples of sludge and sediment were examined to determine the endocrine-disrupting compound's adsorption efficacy. The application of the full logistic model enabled the discernment and ranking of the chemicals contributing to mixture toxicity. The findings revealed constructed wetland A's superior efficacy in the removal of individual endocrine-disrupting compounds (with an average efficiency of 94%) compared to constructed wetland B (60%). Furthermore, constructed wetland A displayed a higher estimated estrogenic activity removal efficiency (83%) relative to constructed wetland B (52%). Estrogenic activity was adequately accounted for (58–120%) in half of the analyzed samples, highlighting estrone as the primary estrogenic agent. The investigation underscores constructed wetlands’ effectiveness in purging endocrine-disrupting compounds, suggesting that their integration as secondary or tertiary treatment systems for such pollutants removal merits further exploration.

  • Hybrid-constructed wetland removes endocrine-disrupting chemicals (EDCs) in a more advanced way than horizontal constructed wetland (CW).

  • Vertical filters of hybrid CW exhibit the highest EDC removal due to aerobic conditions.

  • The removal of EDCs was proved to be driven by sorption followed by biodegradation.

  • Estrone was found as the main contributor to estrogenic activity in both CWs.

Endocrine-disrupting chemicals (EDCs) represent a diverse array of compounds that can interfere with the endocrine system of organisms, potentially altering hormonal balance. EDCs are notably found in a large variety of pharmaceutical and personal care products, products of everyday use, and industrial materials. Regulation of their presence in the environment is still limited, but some activities have already been carried out, for example, at the European Union level (Kättström et al. 2025). These compounds make their way into wastewater treatment facilities (WWTPs) where they can be found in concentrations of up to a few micrograms per liter (Han et al. 2024). WWTPs are not primarily engineered to filter out EDCs but rather to eliminate organic matter and pathogens. As a result, EDCs residuals are discharged into aquatic or terrestrial ecosystems through inadequately treated wastewater or are bound to sewage sludge, commonly used as a fertilizer in agriculture. Despite their seemingly low concentrations in these matrices – nanograms per liter in water bodies (Burkhardt-Holm 2010) and nanograms per gram in soil (Verlicchi & Zambello 2015) – their biological efficacy remains significant (Burkhardt-Holm 2010), making their presence in the environment undesired.

Innovative methods for EDC removal during wastewater treatment are under exploration, including advanced oxidation processes, membrane filtration, and specific adsorption materials (Azizi et al. 2022). Additionally, efforts are ongoing to enhance traditional treatment systems. Recent research has explored alternative wastewater treatment solutions, including constructed wetlands (CWs). These systems have shown potential as a complementary tertiary stage in conventional wastewater treatment or as standalone secondary treatments (Ilyas & van Hullebusch 2020). While such technologies can complement the traditional treatment process as a tertiary stage (He et al. 2018), CWs are also considered a viable standalone option (secondary treatment), offering comparable or better efficacy in removing certain bioactive pollutants (Hijosa-Valsero et al. 2010b; Qing et al. 2013; Nivala et al. 2018). CWs boast numerous benefits over standard WWTPs, including suitability for remote locations (Vystavna et al. 2017), reduced operational costs (Wu et al. 2015), enhanced phytoremediation through plant-root-associated microorganisms (Nguyen et al. 2019), and adaptability to fit the space available for wetland construction (Chen et al. 2021). However, their extensive spatial requirements may restrict their application in densely populated areas. Despite this, the design flexibility of CWs enables the simulation of various natural processes, augmenting the organic micropollutants' natural reduction. While studies have investigated EDC removal in CWs, knowledge gaps persist regarding specific mechanisms (Ilyas & van Hullebusch 2020). Research using both synthetic wastewater in lab experiments (Campos et al. 2019) and real wastewater in mesocosm studies (Herrera-Melian et al. 2018) has sought to identify key factors in EDCs elimination, including adsorption, uptake by plants, photodegradation, and biodegradation by flora, algae, and root-associated microbes (Song et al. 2011; Toro-Velez et al. 2016; Nguyen et al. 2019; Ilyas & van Hullebusch 2020). Despite CWs facilitating these removal processes, comprehensive research on EDCs in full-scale CWs is still lacking (Avila et al. 2015; Vymazal et al. 2015; Vystavna et al. 2017), with only a few studies comparing different CW types in terms of EDC removal efficiency. Furthermore, most of the existing research has been limited to analyzing grab water samples, with less focus on composite water and solid matrices, even though studies like those by Chen et al. (2021) and Hijosa-Valsero et al. (2016) have begun to address this gap.

This review highlights the ongoing need for a deeper understanding of CWs' capacity to eliminate EDCs and their estrogenic effects across the entirety of the treatment system. The present study monitors two operational full-scale CWs – a modern hybrid CW and a traditional horizontal CW – for their efficacy in reducing estrogenic activity and the 15 representative EDCs. The selected EDCs primarily include compounds with estrogen-like effects, such as natural estrogens, phytohormones, industrial chemicals, progestins, and xenoestrogens, based on their environmental prevalence and estrogen receptor affinity. The thorough examination of both CWs involved (1) contaminant sampling from wastewater over a 28-day period using Polar Organic Chemical Integrative Samplers (POCIS), (2) grab sampling of sludge and sediment across all treatment stages, and (3) chemical and bioassay analyses to identify the chemicals responsible for observed hormonal activities and to assess the formation of any estrogenic transformation products during treatment. The study's primary objective is to ascertain the relative effectiveness of these systems in removing target compounds and addressing specific toxicity, alongside exploring the mechanisms facilitating such removals in the observed settings.

Experimental sites

This research aimed to evaluate the efficacy of two distinct constructed wetland configurations in removing EDCs from wastewater. The first, a modern hybrid CW identified as CW A, is situated in Kostelec nad Ohří in the northwestern Czech Republic. This system is employed to process wastewater originating from a combination of a recreational facility, a hotel, and a brewery. The second, an older horizontal flow CW designated as CW B is located in Hostětín in the southeastern part of the country, primarily treating municipal sewage. These CWs were selected to assess the removal efficiency of EDCs under diverse operational conditions and system designs.

CW A consists of a combination of treatment steps: starting with a septic tank (A In), followed by a horizontal filter (A HF), a vertical filter (A VF), and concluding with a hybrid vertical-horizontal filter (A CF). A subsequent stabilization pond (A SP) not only provides tertiary treatment but also serves as a reservoir of irrigation water. In contrast, CW B's layout begins similarly with a septic tank (B In), but then utilizes a simple dual-series horizontal filter (B HF) setup, leading to a stabilization pond (B SP) dedicated to the system's final wastewater treatment.

CW A was put in operation in 2017 and exemplifies a modern hybrid system as characterized by Vymazal (2005), showcasing high efficiency in eliminating primary pollutants like chemical oxygen demand (COD), biochemical oxygen demand (BOD), total suspended solids (TSS), as well as total nitrogen (TN) and phosphorus (TP), as noted by Šereš et al. (2021). CW B, operational since 1997, represents an older horizontal flow model that, despite experiencing clogging issues, consistently achieves effluent quality within Czech regulatory standards for BOD, COD, and TSS.

Both CW systems' configurations and operational details are further showcased through schematic representations provided in Figure 1, which illustrates their comparative structural and functional dynamics for EDCs removal in differing environmental and operational contexts.
Figure 1

Schematic plans of the A and B constructed wetlands.

Figure 1

Schematic plans of the A and B constructed wetlands.

Close modal

CW A's design capacity is designed at 16.5 m3/day, aimed to serve a maximal community of 150 population equivalents (PE). Observations during site monitoring revealed the actual daily wastewater flow averaged 21.5 m3/day, indicating a 30% exceedance over its intended capacity. The wetland filters spread across 319 m2, allocating an area of 2.1 m2 for every PE. On the other hand, CW B was projected to handle 47.6 m3/day, suitable for servicing 300 PE. Nonetheless, the real-world application faced challenges due to an unexpectedly high inflow of ballast water, with flow rates soaring to 433 m3/day, nearly tenfold the designed capability. This system's horizontal filters cover 1,240 m2, with a planned provision of 4.1 m2 per PE. In both systems common reed (Phragmites australis) was predominantly utilized as cover vegetation. A comprehensive description of both CWs' configurations can be found in Table 1.

Table 1

Characteristics of the CW A and CW B

CWTreatment stagesEffective area (m2)Maximal depth (m)Effective volume (m3)HLRd (cm·day−1)HRTd (day)HLRm (cm·day−1)HRTm (day)
Hybrid CW A Septic tank (A In) 29.3 1.8 52.7 – 3.2 – 2.5 
Horizontal filter (A HF) 150.0 0.9 33.1 11.0 2.0 14.3 1.5 
Vertical filter (A VF) 81.0 1.3 26.1 20.4 1.6 26.5 1.2 
Combined filter (A CF) 88.0 0.8 + 0.6 34.6 18.8 2.1 24.4 1.6 
Stabilization pond (A SP) 148.2 2.0 42.6 11.1 2.6 14.5 2.0 
Total CW A 496.5  189.1 61.3 11.5 79.8 8.8 
Horizontal CW B Septic tank (B In) 36.3 3.5 125.2 – 2.6 – 0.3 
Horizontal filters 1 and 2 (B HF) 1,240.0 1.0 1,134.7 0.4 23.8 34.9 2.6 
Stabilization pond (B SP) 940.0 1.4 800.0 0.5 16.8 46.1 1.8 
Total CW B 2 216.3  2 059.9 0.9 43.3 81.0 4.8 
CWTreatment stagesEffective area (m2)Maximal depth (m)Effective volume (m3)HLRd (cm·day−1)HRTd (day)HLRm (cm·day−1)HRTm (day)
Hybrid CW A Septic tank (A In) 29.3 1.8 52.7 – 3.2 – 2.5 
Horizontal filter (A HF) 150.0 0.9 33.1 11.0 2.0 14.3 1.5 
Vertical filter (A VF) 81.0 1.3 26.1 20.4 1.6 26.5 1.2 
Combined filter (A CF) 88.0 0.8 + 0.6 34.6 18.8 2.1 24.4 1.6 
Stabilization pond (A SP) 148.2 2.0 42.6 11.1 2.6 14.5 2.0 
Total CW A 496.5  189.1 61.3 11.5 79.8 8.8 
Horizontal CW B Septic tank (B In) 36.3 3.5 125.2 – 2.6 – 0.3 
Horizontal filters 1 and 2 (B HF) 1,240.0 1.0 1,134.7 0.4 23.8 34.9 2.6 
Stabilization pond (B SP) 940.0 1.4 800.0 0.5 16.8 46.1 1.8 
Total CW B 2 216.3  2 059.9 0.9 43.3 81.0 4.8 

Abbreviations: CW, constructed wetland; HLRd, designed hydraulic loading rate; HRTd, designed hydraulic retention time; HLRm, measured hydraulic loading rate; HRTm, measured hydraulic retention time.

Operational efficiency and compliance with the regulatory parameters of both CWs were consistently evaluated throughout the study, and the findings are summarized in Table 2 for a detailed summary.

Table 2

Performance of the CW A and CW B during the study period (year 2020)

BOD
COD
TSS
TP
TN
Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)
A In 447.3 ± 86.8  759.5 ± 217.9  150.4 ± 59.5  10.7 ± 1.5  101.3 ± 18.8  
A CF 4.5 ± 4.8 98.9 ± 1.1 17 ± 5.4 97.5 ± 1.1 8.3 ± 10.3 93.7 ± 8 3.5 ± 2.5 69.5 ± 19.2 19.2 ± 18.3 82.7 ± 13.3 
A SP 16 ± 27.8 96.1 ± 6.7 38.8 ± 42.1 95.1 ± 5.3 25.1 ± 34 85.4 ± 16.3 3.3 ± 2.2 71.5 ± 16 13.8 ± 8.3 87.4 ± 6.5 
B In 31.8 ± 18.9  128.2 ± 133  32.7 ± 23.1  0.9 ± 0.9  10.5 ± 8.8  
B HF 3.1 ± 2.5 84.7 ± 16.2 20.1 ± 8.4 67.9 ± 27.6 7 ± 4.4 71.3 ± 18.7 1.1 ± 0.6 −77.4 ± 104.2 7.9 ± 3.5 −12.8 ± 67.9 
B SP 2 ± 0,7 91,1 ± 8,1 17,7 ± 4,9 72,9 ± 20,9 4,7 ± 2,2 78,3 ± 18,2 1,2 ± 0,5 −131,6 ± 150,9 7,4 ± 3,7 −7,9 ± 68,6 
BOD
COD
TSS
TP
TN
Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)Conc. (mg·L−1)RE (%)
A In 447.3 ± 86.8  759.5 ± 217.9  150.4 ± 59.5  10.7 ± 1.5  101.3 ± 18.8  
A CF 4.5 ± 4.8 98.9 ± 1.1 17 ± 5.4 97.5 ± 1.1 8.3 ± 10.3 93.7 ± 8 3.5 ± 2.5 69.5 ± 19.2 19.2 ± 18.3 82.7 ± 13.3 
A SP 16 ± 27.8 96.1 ± 6.7 38.8 ± 42.1 95.1 ± 5.3 25.1 ± 34 85.4 ± 16.3 3.3 ± 2.2 71.5 ± 16 13.8 ± 8.3 87.4 ± 6.5 
B In 31.8 ± 18.9  128.2 ± 133  32.7 ± 23.1  0.9 ± 0.9  10.5 ± 8.8  
B HF 3.1 ± 2.5 84.7 ± 16.2 20.1 ± 8.4 67.9 ± 27.6 7 ± 4.4 71.3 ± 18.7 1.1 ± 0.6 −77.4 ± 104.2 7.9 ± 3.5 −12.8 ± 67.9 
B SP 2 ± 0,7 91,1 ± 8,1 17,7 ± 4,9 72,9 ± 20,9 4,7 ± 2,2 78,3 ± 18,2 1,2 ± 0,5 −131,6 ± 150,9 7,4 ± 3,7 −7,9 ± 68,6 

Abbreviations: Conc. = concentration, RE = removal efficiency, A In = septic tank of CW A, A CF = outflow from CF of CW A, A SP = outflow from stabilization pond of CW A, B In = septic tank of CW B, B HF = outflow from HF of CW B, B SP = outflow from stabilization pond of CW B. RE for SPs is calculated as a comparison of concentrations at inlet (septic tanks) and outlets from SPs.

Sampling, sample preparation, and analysis

Sampling strategy

This research aimed at evaluating the presence of 15 specifically chosen EDCs detailed in Table 3, within the wastewater processed by two distinct CW systems. Given the minimal concentration levels of these EDCs in wastewater, the investigation utilized POCIS. These devices are helpful at collecting organic hydrophilic pollutants over prolonged periods, enabling the assessment of the average concentration of analytes over time, a necessity given the low and variable concentrations of these substances (Harman et al. 2012). Previous studies have demonstrated the efficacy of using POCIS in capturing EDCs, including research focused on CW systems (Vallejo et al. 2013; Vystavna et al. 2017).

Table 3

The estrogenic active compounds of interest

SubstanceAbbreviationOriginEC50 (nM per well)Rs (L·day−1)Reference for Rs
17α-estradiol aE2 Natural estrogen 1.65 0.239 Morin et al. (2013)  
17β-estradiol bE2 Natural estrogen 0.11 0.221 Morin et al. (2013)  
17α-ethinylestradiol EE2 Synthetic estrogen 0.10 0.221 Morin et al. (2013)  
Levonorgestrel NRG Synthetic progestin 4,076.35 0.346 Morin et al. (2013) a 
Norethindrone NORE Synthetic progestin 517.76 0.346 Morin et al. (2013) a 
Estrone E1 Natural estrogen 0.26 0.23 Morin et al. (2013)  
Estriol E3 Natural estrogen 33.38 0.185 Morin et al. (2013)  
Bisphenol S BPS Industrial chemical 214,288.94 0.245 Morin et al. (2013) b 
Bisphenol A BPA Industrial chemical 2,276.74 0.245 Morin et al. (2013)  
Bisphenol F BPF Industrial chemical 1,734.38 0.245 Morin et al. (2013) b 
Genistein GEN Phytohormone 399.37 0.2 Tousova et al. (2019) c 
Daidzein DAID Phytohormone 34,440.47 0.2 Tousova et al. (2019) c 
Equilin EQN Natural estrogen 0.37 0.052 Vallejo et al. (2013)  
Equol EQ Natural estrogen 329.92 0.2 Tousova et al. (2019) c 
α-zearalenol ZEA Mycoestrogen 3.73 0.454 Bartelt-Hunt et al. (2011)  
SubstanceAbbreviationOriginEC50 (nM per well)Rs (L·day−1)Reference for Rs
17α-estradiol aE2 Natural estrogen 1.65 0.239 Morin et al. (2013)  
17β-estradiol bE2 Natural estrogen 0.11 0.221 Morin et al. (2013)  
17α-ethinylestradiol EE2 Synthetic estrogen 0.10 0.221 Morin et al. (2013)  
Levonorgestrel NRG Synthetic progestin 4,076.35 0.346 Morin et al. (2013) a 
Norethindrone NORE Synthetic progestin 517.76 0.346 Morin et al. (2013) a 
Estrone E1 Natural estrogen 0.26 0.23 Morin et al. (2013)  
Estriol E3 Natural estrogen 33.38 0.185 Morin et al. (2013)  
Bisphenol S BPS Industrial chemical 214,288.94 0.245 Morin et al. (2013) b 
Bisphenol A BPA Industrial chemical 2,276.74 0.245 Morin et al. (2013)  
Bisphenol F BPF Industrial chemical 1,734.38 0.245 Morin et al. (2013) b 
Genistein GEN Phytohormone 399.37 0.2 Tousova et al. (2019) c 
Daidzein DAID Phytohormone 34,440.47 0.2 Tousova et al. (2019) c 
Equilin EQN Natural estrogen 0.37 0.052 Vallejo et al. (2013)  
Equol EQ Natural estrogen 329.92 0.2 Tousova et al. (2019) c 
α-zearalenol ZEA Mycoestrogen 3.73 0.454 Bartelt-Hunt et al. (2011)  

Note. The effective concentration that prompted a certain effect in 50% of assay organisms (EC50) was determined using the yeast assay. A review of the sampling rate (Rs) for POCIS samplers as determined in the literature.

aAnalytes are structurally similar to progesterone, therefore, the Rs value of the latter was used.

bAnalytes are structurally similar to bisphenol A, therefore, the Rs value of the latter was used.

cTo the best of our knowledge, the Rs for phytohormones are unavailable. The average value of Rs for estrogenic active compounds was applied according to Tousova et al. (2019). The effective concentration that prompted a certain effect in 50% of assay organisms (EC50) was determined using the yeast assay. A review of the sampling rate (Rs) for POCIS samplers as determined in the literature.

Utilizing POCIS equipped with Oasis HLB sorbent, each device contained 220 mg of sorbent. These samplers were strategically placed in pairs within control shafts at each treatment phase across both CWs. For the SPs, samplers were positioned 1 m beneath the water's surface. The collection phase spanned 28 days from September to October 2020, with CW A experiencing a total wastewater flow of 601 m3 and CW B a significantly higher flow of 12,101 m3.

Acknowledging the potential for micropollutants to adhere to solids within the water (e.g., sludge and sediment), the research also entailed examining EDC concentrations within these solid matrices. Simultaneously with the POCIS collection period's conclusion in October 2020, grab samples of sludge were obtained from each CW's multiple stages. This included gathering mixed sludge and sediment from the septic tanks (labeled A In S and B In S) and SP beds (A SP S and B SP S). Moreover, sludge samples were extracted from the wetland filters, involving three replicates of the filter medium mixture, each replicate weighing 1 kg and taken from a 0–50 cm depth at randomly chosen points within each filter (identified as A HF S, A VF S, A CF S, and B HF S).

POCIS extraction

Right after their retrieval, the POCIS devices were immediately frozen and preserved at a temperature of −20 °C. This preparatory step was crucial to maintain the integrity of the samples until extraction, which was scheduled within the shortest time sequence from the actual collection. Prior to the extraction, samplers were carefully detached from their enclosures and rinsed under a soft flow of tap water. This step was essential to avoid any potential contamination of the sorbent during the disassembly of the samplers. Subsequently, the POCIS devices were taken apart, and the sorbent within was transferred into a 6-mL polypropylene solid phase extraction (SPE) cartridge. This transfer was facilitated by a glass funnel and involved the use of Milli-Q water to ensure a clean move. Each sorbent was then securely sealed at both ends within the cartridge using polypropylene frits, after which the cartridges were affixed to an SPE manifold. A thorough wash with Milli-Q water (in two 4 mL volumes) was followed by a vacuum drying process. The analytes' extraction was finalized with methanol (two 4 mL volumes), and the methanol extract volume was then reduced to about 1 mL by evaporation. To ensure the reliability of the method, two unused POCIS devices underwent identical processing to act as method blanks.

Solid sample pretreatment and extraction

The sediment and sludge samples underwent freeze-drying, followed by manual homogenization and sieving. These samples, weighing between 1 and 3 g, were placed into stainless steel cells for extraction using an Accelerated Sample Extractor (ASE 200, Dionex). This extraction utilized a specific protocol: it included three cycles of extraction with methanol heated in advance to 150 °C, under a pressure of 1,500 psi, incorporating 5-min intervals of static extraction within each cycle. Post-extraction, the methanol-based solutions were reduced in volume to approximately 10 mL through evaporation. Subsequently, these concentrated extracts were subjected to centrifugation at 5,000 g for 5 min, enabling the separation of the liquid extracts from any solid residues formed during the process.

Chemical analysis applying LC-MS/MS

The liquid chromatography (LC) setup utilized for extract analysis consisted of an Agilent 1260 Infinity system, integrated with an Agilent 6470 triple quadrupole mass spectrometer. We analyzed a panel of 15 estrogen compounds utilizing a Poroshell EC-C18 column (2.7 μm particle size, 3 mm diameter × 100 mm length) paired with a guard column (2.7 μm, 3 mm × 5 mm), both maintained at 40 °C. Elution was performed using a gradient method involving phase A (0.5 mM ammonium fluoride in Milli-Q water) and phase B (methanol of LC-MS grade), following a detailed program of concentration and flow adjustments (time in minutes, %B, flow rate in mL·min−1): initiation at 0.0 with 35% B at 0.4 mL·min−1; increased to 100% B at 0.5 mL·min−1 by 6.0 min, maintained through 8.1 min; and reverted to 35% B at 0.5 mL·min−1 at 8.2 min. Each analytical run lasted 10 min with a 2 μL injection volume. To assess matrix effects, samples were analyzed with added standard concentrations of 0.5, 2.5, and 25 ng·mL−1. Optimization of the multiple reaction monitoring (MRM) transitions and mass spectrometer ion source settings was achieved using Agilent Technologies' MassHunter Workstation and Source Optimizers, both version 10.0, SR1. For detailed data on retention times, MRM transitions, and spectrometer configurations, refer to the study by Černá et al. (2022).

17β-estradiol equivalent calculation (EEQcal)

To identify the specific chemicals contributing to the estrogenic activity observed in the samples, we converted the measured concentrations into the anticipated effects generated by the mixture of n compounds (using the following equation). This conversion involved integrating the dose-response curve parameters – MAX (indicating the maximum effect), MIN (indicating the minimum effect), EC50 (the concentration at the curve's midpoint), and p (the curve's slope) – for each analyte and their respective concentrations (c1−n) into the full logistic model (as proposed by Ezechiáš & Cajthaml 2016). These parameters, essential for calculating the cumulative estrogenic effect (Emix), were derived from the yeast assay results, which are detailed subsequently. The EC50 values crucial for this calculation are documented in Table 3.

Estrogenic activity

Following the protocol established by Routledge & Sumpter (1996), this assay utilized a genetically modified Saccharomyces cerevisiae strain, engineered to produce β-galactosidase when exposed to compounds with estrogenic properties. Initially, the samples were dried under a mild nitrogen flow and subsequently reconstituted in 30% dimethyl sulfoxide (DMSO) to prepare them for the assay. Analytical standards, too, were prepared in 30% DMSO, which acted as the baseline control for the assay. To establish a positive control, various concentrations of 17β-estradiol (β-E2), ranging from 0.09 to 18.37 nM per well, were employed. For testing, the samples underwent a tenfold dilution in a yeast-containing medium within a 96-well plate, ensuring the DMSO concentration remained non-toxic to the organisms. After a 4-day incubation, absorbance readings at λ = 620 nm and λ = 540 nm were taken using an Infinite M200 PRO NanoQuant (Tecan) spectrophotometer. This allowed for the assessment of estrogenic activity via the chromogenic substrate and confirmed yeast cell viability. The estrogenic effect of the samples was quantified as 17β-estradiol equivalents (EEQ), mirroring the concentration of β-E2 that would induce a comparable response.

In this updated description, we further analyzed the EEQ measurements against calculated EEQ (EEQcal) based on the concentrations of detected EDCs, as outlined in the preceding section. This comparison aimed to quantify the toxicity attributable to the monitored compounds, with the explained estrogenic activity presented as a percentage ratio of EEQ to EEQcal.

Data evaluation

The calculation of the time-weighted average concentration (cTWA) of EDCs within the wastewater utilized the POCIS data, following the formula proposed by Vallejo et al. (2013):
where ms (ng) corresponds to the amount of analyte in the sorbent, Rs (L·day−1) to the sampling rate as listed in Table 3 and t (day) to the sampling period.
To facilitate a comparative analysis between the two CW systems, which have markedly diverse flow rates, the dataset was adjusted to reflect the equivalent surface area (1 m2) of each CW stage. This adjustment is a commonly accepted practice for comparing CW systems. The process begins with determining the mass loading (ML) of EDCs across all treatment stages:
where cTWA (ng·L−1) is the time-weighted average concentration of the individual EDCs at the corresponding sampling point and Qarea (L·m−2·day−1) is the wastewater discharge from the corresponding treatment stage with the consideration of its area. The equation used for the calculation of the annual mass loading MLa (g·a−1) of the whole CW was as follows:
The annual mass loading (MLa) for the entire CW system was calculated as follows:
where cTWA in (ng·L−1) is the time-weighted average concentration of the individual EDCs at the inlet and Qcwarea in (L·m−2·day−1) is the wastewater discharge from the whole of the CW considering its summary area.
Subsequently, the mass removal (MR) per treatment stage was computed:
where cTWAin (ng·L−1) is the time-weighted average concentration of the individual EDCs at the inlet to the corresponding sampling point, cTWAout (ng·L−1) is the time-weighted average concentration of the individual EDCs at the outlet from the corresponding sampling point, Qareain (L·m−2·day−1) is the wastewater inflow to the corresponding treatment stage considering its area and Qareaout (L·m−2·day−1) is the wastewater outflow from the corresponding treatment stage considering its area.
The formula for annual mass removal (MRa) across the CW was:
where cTWA in and cTWA out (ng·L−1) are the time-weighted average concentrations of the individual EDCs at the inlet and the outlet of the CW and Qarea in and Qarea out (L·m−2·day−1) represent the wastewater discharge from the whole of the CW considering its summary area.
Finally, the EDCs removal efficiency RE (%) of the individual treatment stages was calculated according to Vystavna et al. (2017):
where MLin (ng·m−2·day−1) is the mass load of the EDC at the corresponding treatment stage and MLout (ng·m−2·day−1) is the mass load of the EDC in the follow-up stage.

Summary values in Table 4 were derived from the difference between inlet and outlet concentrations for both CW A and CW B stages.

Table 4

Assessment of the annual loading and removal efficiency for constructed wetlands (CWs) A and B

SubstanceCW A
CW B
MLa (g·a−1)MRa (g·a−1)RE (%)MLa (g·a−1)MRa (g·a−1)RE (%)
17α-estradiol 0.011 ± 0.016 0.010 ± 0.012 87 0.246 ± 0.018 0.246 ± 0.315 100 
17β-estradiol 0.050 ± 0.013 0.047 ± 0.015 94 0.308 ± 0.032 0.308 ± 0.262 100 
Bisphenol A 4.891 ± 0.316 4.379 ± 0.335 90 6.736 ± 0.154 −1.576 ± 2.114 −23 
Bisphenol F 0.112 ± 0.001 0.109 ± 0.022 97 0.438 ± 0.004 0.309 ± 0.068 71 
Bisphenol S 0.342 ± 0.002 0.304 ± 0.151 89 3.033 ± 0.182 1.518 ± 0.039 50 
Daidzein 1.018 ± 0.034 0.995 ± 0.496 98 9.971 ± 0.210 9.696 ± 0.228 97 
Equilin 0.000 ± 0.000 −0.002 ± 0.000 – 0.000 ± 0.000 −0.168 ± 0.034 – 
Equol 0.681 ± 0.036 0.672 ± 0.397 99 7.991 ± 0.013 5.711 ± 0.508 71 
Estriol 0.099 ± 0.034 0.099 ± 0.011 100 0.231 ± 0.327 0.003 ± 0.493 
Estrone 0.147 ± 0.021 0.129 ± 0.046 87 0.930 ± 0.057 0.329 ± 0.057 35 
17α-ethinylestradiol n.d. n.d. – n.d. n.d. – 
Genistein 0.078 ± 0.009 0.078 ± 0.311 99 6.250 ± 1.275 6.160 ± 0.171 99 
Norethindrone n.d. n.d. – n.d. n.d. – 
Levonorgestrel n.d. n.d. – n.d. n.d. – 
α-zearalenol n.d. n.d. – n.d. n.d. – 
EEQ – – – 1.399 ± 0.058 – 91 
EEQcal 0.085 ± 0.007 0.071 ± 0.040 83 0.805 ± 0.015 0.417 ± 0.021 52 
SubstanceCW A
CW B
MLa (g·a−1)MRa (g·a−1)RE (%)MLa (g·a−1)MRa (g·a−1)RE (%)
17α-estradiol 0.011 ± 0.016 0.010 ± 0.012 87 0.246 ± 0.018 0.246 ± 0.315 100 
17β-estradiol 0.050 ± 0.013 0.047 ± 0.015 94 0.308 ± 0.032 0.308 ± 0.262 100 
Bisphenol A 4.891 ± 0.316 4.379 ± 0.335 90 6.736 ± 0.154 −1.576 ± 2.114 −23 
Bisphenol F 0.112 ± 0.001 0.109 ± 0.022 97 0.438 ± 0.004 0.309 ± 0.068 71 
Bisphenol S 0.342 ± 0.002 0.304 ± 0.151 89 3.033 ± 0.182 1.518 ± 0.039 50 
Daidzein 1.018 ± 0.034 0.995 ± 0.496 98 9.971 ± 0.210 9.696 ± 0.228 97 
Equilin 0.000 ± 0.000 −0.002 ± 0.000 – 0.000 ± 0.000 −0.168 ± 0.034 – 
Equol 0.681 ± 0.036 0.672 ± 0.397 99 7.991 ± 0.013 5.711 ± 0.508 71 
Estriol 0.099 ± 0.034 0.099 ± 0.011 100 0.231 ± 0.327 0.003 ± 0.493 
Estrone 0.147 ± 0.021 0.129 ± 0.046 87 0.930 ± 0.057 0.329 ± 0.057 35 
17α-ethinylestradiol n.d. n.d. – n.d. n.d. – 
Genistein 0.078 ± 0.009 0.078 ± 0.311 99 6.250 ± 1.275 6.160 ± 0.171 99 
Norethindrone n.d. n.d. – n.d. n.d. – 
Levonorgestrel n.d. n.d. – n.d. n.d. – 
α-zearalenol n.d. n.d. – n.d. n.d. – 
EEQ – – – 1.399 ± 0.058 – 91 
EEQcal 0.085 ± 0.007 0.071 ± 0.040 83 0.805 ± 0.015 0.417 ± 0.021 52 

Abbreviations: ML, mass loading; MR, mass removal; RE, removal efficiency; n.d.; not detected. The CW A inlet samples were cytotoxic, therefore, it was not possible to calculate any of the measured 17β- EEQ values.

Comparative analysis of inlet and outlet concentrations in CW A and CW B

Figure 2 provides a comparative analysis of the time-weighted average concentrations (cTWA) of 15 EDCs treated by CW A and CW B. The initial concentration of EDCs in CW A (947.1 ± 57.7 ng·L−1) was substantially higher than in CW B (229.1 ± 14.4 ng·L−1), highlighting a significant difference in pollutant loads. This variance underscores the diverse pollutant loads these systems are designed to manage. However, these concentrations are still lower than those that are orders of magnitude higher for WWTPs of larger agglomerations (Ngeno et al. 2023; Han et al. 2024).
Figure 2

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the time-weighted average concentration through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond. The analytes EE2, NORE, NRG, and ZEA were not detected in any sample.

Figure 2

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the time-weighted average concentration through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond. The analytes EE2, NORE, NRG, and ZEA were not detected in any sample.

Close modal

Bisphenol A (BPA) presented the highest concentration at CW A's inlet (623.5 ± 40.3 ng·L−1), while daidzein (DAID) led at CW B (63.2 ± 1.3 ng·L−1). The levels of 17α-estradiol (aE2), 17β-estradiol (bE2), estrone (E1), estriol (E3), and BPA align with those reported in other studies, including Yu et al. (2022), indicating a commonality in EDC presence across different wastewater treatment contexts. The study of Yu et al. (2022) showed the median concentration of BPA was a little less than 300 ng/L across the studied WWTPS, but similar to our study, it was still the most represented EDC in the wastewater. Notably, analytes such as 17α-ethinylestradiol (EE2), norethindrone (NORE), (D-)-levonorgestrel (NRG), and α-zearalenol (ZEA) were not detected in any samples.

Across both CWs, a significant reduction in cTWA was observed through the treatment stages, with CW A's HF notably eliminating aE2 and bE2 entirely. This significant removal efficiency, particularly for aE2 and bE2, to non-detectable levels from the HF, is proof of the system's effectiveness. Similarly, genistein (GEN) shows a substantial reduction in CW A, with a 96% decrease, showcasing the HF's capability in targeting specific EDCs. Conversely, CW B's HF showed a reduction exceeding 50% for DAID, GEN, and equol (EQ), which collectively represented 67% of its inlet EDCs concentration. This data, especially the notable increase in BPA and bisphenol S (BPS) concentrations post-treatment in CW B, corroborates the findings of Carranza-Diaz et al. (2014), emphasizing the challenge of mitigating certain EDCs through CW treatment.

The most profound cTWA reduction in CW A occurred in the vertical filter, with an overall decrease of over 89%, highlighting the VF's critical role in EDC removal, as supported by studies from Sklarz et al. (2009) and Venditti et al. (2022). This efficiency is further evidenced in the context of E1 removal, where Dai et al. (2016) identified vertical CWs as particularly effective. Additionally, CW A's combined filter (CF) showed limited reduction capacity, likely due to already diminished inlet concentrations, aligning with observations from Campos et al. (2019) that suggest higher inlet concentrations might yield better removal outcomes.

A notable distinction between CW A and CW B was the efficiency of their SPs, with CW A achieving a 93% overall decrease in EDC concentration, compared to the 62% reduction seen in CW B, as detailed in Table 4. The differences in SP efficiency could be attributed to variations in size, depth, and established wetland vegetation, which enhances photodegradation and phytodegradation processes, in line with findings from Hijosa-Valsero et al. (2010b) and Matamoros & Salvado (2012). This is particularly relevant for BPA, where its concentration increased post-treatment in CW A's SP, a phenomenon that Wirasnita et al. (2018) and Papaevangelou et al. (2016) suggest could be mitigated through enhanced sorption, plant uptake, and biodegradation processes. We assume that the increase in concentration is mainly related to the sorption of BPA onto the organic substrate at the bottom of the SP and to the release of BPA previously adsorbed onto organic substrates in the SP, which subsequently desorbs under changing conditions.

The superior performance of CW A may not be attributed solely to its longer hydraulic retention time (HRT) but also to the dynamic hybrid treatment system that integrates anaerobic and aerobic conditions across various filter media, as discussed by Hijosa-Valsero et al. (2010a). This is confirmed by Zhang et al. (2012), who suggest that a batch-feeding strategy, similar to the discontinuous VF filling system, could significantly increase pharmaceutical removal efficiency. In summary, the comparative analysis of CW A and CW B, based on examining inlet and outlet concentrations along with treatment efficiency, highlights the complexity of EDC removal in constructed wetlands. The integration of different filter types, alongside strategic operational approaches such as batch feeding, plays a crucial role in optimizing the removal of persistent pollutants like BPA, DAID, and GEN, contributing to the broader goal of enhancing water quality through advanced wastewater treatment technologies.

Comparative analysis of the mass loading and removal in CW A and CW B

In evaluating the performance of CW A and CW B for EDC removal, a detailed analysis was conducted to compare their efficacy according to the ML and MR metrics, reflecting each treatment stage's effectiveness within the respective CW systems, as delineated in Figures 3 and 4. These metrics offer insights into the handling and elimination capacities of the systems for the monitored EDCs, factoring in variations in wastewater flow, HRT, and wetland filter area, among other parameters.
Figure 3

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the mass loading through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond.

Figure 3

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the mass loading through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond.

Close modal
Figure 4

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the MR through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond.

Figure 4

EDCs (see Table 3 for the abbreviations) elimination data for constructed wetlands A and B expressed as the MR through the individual treatment units. In = influent, HF = horizontal filter, VF = vertical filter, CF = combined filter, and SP = stabilization pond.

Close modal

CW A's HF exhibited nearly double the ML for EDCs (135.7 ± 8.8 μg·m−2·day−1) compared to CW B's HF (79.8 ± 5.0 μg·m−2·day−1), indicating a superior capacity to process EDCs. Despite this, the MR between CW A's HF and CW B's HF did not show a marked difference (33.9 ± 8.8 μg·m−2·day−1 for CW A's HF vs. 29.5 ± 40.9 μg·m−2·day−1 for CW B's HF), suggesting that both HFs are somewhat comparable in their removal efficiencies for the sum of EDCs. However, distinctions emerge in the degradation outcomes for specific EDCs, with CW A's HF showing notable effectiveness against BPA (14.0 ± 5.8 μg·m−2·day−1), DAID (12.7 ± 0.6 μg·m−2·day−1), and EQ (8.7 ± 0.6 μg·m−2·day−1). In contrast, CW B's HF demonstrated a significant impact on DAID, GEN, and EQ, with DAID showing a remarkable MR of 21.0 ± 0.3 μg·m−2·day−1.

The A VF within CW A was particularly effective, with an MR of 161.2 ± 9.6 μg·m−2·day−1 for the sum of monitored substances, highlighting the vertical filter's role as a most important stage in EDCs removal. This high efficiency, especially in BPA removal (117.3 ± 7.1 μg·m−2·day−1), underscores the VF's critical function within the treatment process. The contribution of the CF to MR was minimal, likely due to the already reduced ML entering this stage.

A notable variance was observed in the SP's performance between the two CWs. CW A's SP displayed limited effectiveness, with a negative MR value (−1.1 ± 5.3 μg·m−2·day−1), suggesting possible challenges in EDCs removal or re-release phenomena. Conversely, CW B's SP demonstrated greater efficacy, contributing positively to the overall ED reduction, especially for BPA, BPS, and E1. This enhanced performance in CW B's SP, despite a shorter measured HRT compared to CW A's SP, highlights the SP's integral role in CW B's cleaning process.

The analysis of POCIS data over a year revealed CW A's VF as highly efficient, particularly in removing a broad spectrum of EDCs, including BPA. This aligns with findings from another study that reported a similar E1 MRa for a hybrid system, indicating CW A's high overall efficiency in EDC removal (Vystavna et al. 2017). In contrast, CW B, despite lower ML values, demonstrated significant removal capabilities, especially in its SP, showcasing the treatment efficacy of CW B.

In summary, for CW A, EDCs are best removed in the VF stage, whereas for CW B, the SP stage is essential. Thus, in general, it could be said that there is a unique contribution of each treatment stage to the overall wastewater treatment process.

Estrogenic activity and the effect-causative chemicals

Our investigation revealed a decrease in the anticipated estrogenic impacts (EEQcal) within both CW systems, achieving 83% in CW A and 52% in CW B. Notably, CW B showed a significant reduction in actual estrogenic impacts (EEQ) at 91%. However, the assay's cytotoxic influent extract in CW A prevented a clear determination of EEQ reduction efficiency, as outlined in Table 4. This occurrence suggests that cytotoxicity obscured the estrogenic impacts in CW A. Comparison of the efficacy of estrogen removal in both CWs remained unclear. Additionally, the accumulation of pollutants in the sediment near the passive samplers might explain the elevated estrogenic activity observed in the final samples of CW A's treatment step.

This study's findings align with previous research documenting estrogenic activity mitigation in both full-scale wastewater treatment plants (WWTPs) and CWs handling real wastewater. For instance, research by Neale et al. (2020) shows cytotoxicity reduction after the treatment in the range from 42 to 84%. Lundqvist et al. (2019) reported high removal efficiencies exceeding 90% for the studied WWTPs, mirroring the outcomes observed in our CWs.

A thorough evaluation of estrogenic activity, incorporating both in vitro assays for transformation products and LC-MS/MS analysis for estrogenic disruptors, facilitated the identification of key contributing chemicals, as detailed in Table 4. E1, an endogenous hormone, was identified as a significant contributor across most samples, with notable exceptions. The analytical findings for specific CW stages – A VF, B In, and B HF – were largely explained by the target analytes, yet A CF and A SP showed minimal explained activity, hinting at the potential presence of unaccounted endogenous hormones or other compounds.

Our sediment analysis highlighted the accumulation of these target substances, raising the possibility of more active metabolite formation in CW A. In line with the observations of Cai et al. (2012) and He et al. (2018), the effect of plant hormones on the level of estrogenic activity was observed. However, the overestimation of activity in B SP suggests a complex interplay of agonistic and antagonistic effects, possibly attributed to antiestrogenic compounds.

Reflecting on the role of root-associated bacteria, as suggested by Qing et al. (2013), our findings underscore the predominance of endogenous steroids in the activity profile, consistent with previous reports on conventional WWTPs. The important role of E1 in later treatment stages, as shown in studies like that by Shappell et al. (2007), emphasizes the necessity of considering both the analytical approach and the sorbent's selectivity in passive samplers for a holistic understanding of causative chemicals in estrogenic activity.

Solid samples evaluation

The investigation into the distribution and fate of estrogenic EDCs within constructed wetlands A and B highlights the intricate balance between aqueous and (bio)solid phases in the environmental behavior of these compounds, as underscored by Yost et al. (2013) and Hijosa-Valsero et al. (2016). This study's findings, identifying 10 of the 15 targeted EDCs in the sludge or sediment across various CW units, emphasize the significant presence of these compounds, notably BPA, which was found in concentrations as high as 124.5 ± 23.3 ng·g−1 in CW A's inlet sludge. Despite CW A experiencing almost double the mass loading compared to CW B, the concentrations of EDCs within the sludge and sediment of both systems were notably similar, suggesting a uniform capacity for EDC adsorption and degradation.

Particularly of interest is the detection of GEN, a known phytohormone, exclusively within the vegetation biomass, although its prevalence did not extend to the broader scope of this study. The absence of other EDCs within the biomass further narrows the focus on sediment and sludge as primary sites of EDC accumulation and potential mineralization.

Figure 5 illustrates the dynamic reduction of EDC concentrations through the CWs, aligning with the understanding that specific CW stages are more effective in transitioning EDCs from liquid to solid phases. However, the long-term accumulation of these compounds in CW A's SP sediment (A SP S) necessitates consideration for sustained management practices.
Figure 5

The concentration of pollutants (see Table 2 for explanations of the abbreviations) in the solid samples from constructed wetlands CW A and CW B. In S = sludge from the inlet, HF S = sludge from the horizontal filter, VF S = sludge from the vertical filter, CF S = sludge from the combined filter, and SP S = sludge from the stabilization pond. No EE2, EQN, NORE, NRG, and ZEA analytes were detected in any of the samples.

Figure 5

The concentration of pollutants (see Table 2 for explanations of the abbreviations) in the solid samples from constructed wetlands CW A and CW B. In S = sludge from the inlet, HF S = sludge from the horizontal filter, VF S = sludge from the vertical filter, CF S = sludge from the combined filter, and SP S = sludge from the stabilization pond. No EE2, EQN, NORE, NRG, and ZEA analytes were detected in any of the samples.

Close modal

Comparatively, the profile of EDC concentrations within these CWs mirrors that observed in activated sludge from conventional WWTPs, with bisphenols predominating over phytohormones and estrogens, as detailed by Černá et al. (2022). This similarity extends to the non-detection of certain EDCs such as EE2, equilin (EQN), norethindrone (NORE), norgestrel (NRG), and zearalenone (ZEA) across all samples, a finding corroborated by Chen et al. (2021) in their analysis of mesocosm CWs treating domestic wastewater.

Although the specific mechanisms leading to EDC removal were not studied in detail in this study, the results suggest that sorption followed by biodegradation plays an important role, especially under aerobic conditions, which are mainly present in vertical wetlands. This process, as suggested by Gikas et al. (2021), appears to be a critical pathway for EDC mitigation in CWs. The observed accumulation of EDCs in the solid phase, especially evident in the SPs, underscores the necessity for continuous monitoring and judicious management of CW sediment to mitigate the environmental release of these persistent compounds.

Our research investigated the dynamics of estrogenic EDCs and their effects in two different constructed wetlands: a modern hybrid-constructed wetland and a traditional horizontal flow constructed wetland. This comparative study reveals the differing efficacy and adaptability of these CWs in mitigating estrogenic EDCs and reveals crucial insights into the key roles played by their design and operational strategies.

Remarkably, CW A, subjected to higher pollutant loads, outperformed CW B in terms of pollutant removal efficiency. This underscores the superior capability of CW A's design to manage and mitigate higher concentrations of EDCs effectively. Notably, the horizontal filter in CW A (A HF) showcased proficiency in eliminating specific EDCs like 17α-estradiol (aE2) and 17β-estradiol (bE2), though it struggled with BPA removal. In contrast, CW B's horizontal filter (B HF) made significant progress in reducing various EDCs but failed with BPA and Bisphenol S (BPS), indicating distinct removal capabilities and challenges inherent to each system.

The standout performance of CW A's vertical filter (A VF) emerged as a key element in its EDCs mitigation strategy, achieving a remarkable reduction rate of over 89% for the detected EDCs. This efficiency not only exemplifies the vertical filter's exceptional role but also highlights the importance of incorporating multi-stage treatments that specifically target a broad spectrum of contaminants for enhanced removal efficiency.

The study revealed that both CWs significantly diminished estrogenic activities, achieving reductions of 83 and 52% for CW A and CW B, respectively. Such reductions are significant, given the environmental and health implications of estrogenic compounds. Similar to traditional wastewater treatment processes, the analysis confirmed that endogenous hormones, especially estrone, were predominant contributors to estrogenic activity, pointing to the presence of undetected estrogenic or antiestrogenic compounds in the system.

Further examination into mass loading and removal shed light on the distinct efficiencies of CW A and B, particularly lauding the VF of CW A for its unmatched efficacy. This finding emphasizes the necessity for customized CW designs that cater to specific wastewater compositions, highlighting the significance of each treatment stage in achieving optimal system performance.

Our study advocates for the comprehensive analysis of both aqueous and solid (sludge and sediment) phases in future research to thoroughly quantify the sorbed pollutants. Such an approach is crucial as it unveils the transient nature of sorption, which may only delay pollutant release until desorption conditions prevail, necessitating vigilant monitoring and management of micropollutants within CW systems.

In essence, this study not only demonstrates the comparable efficacy of two CW designs in addressing estrogenic EDCs but also serves as a challenge for the strategic planning and design of CW systems. It underlines the critical need for tailor-made configurations that optimize each stage's contribution to the overarching goal of mitigating environmental pollutants, ensuring a safer and healthier ecosystem.

The work was also supported by the Technology Agency of the Czech Republic (project numbers SS02030008, program Prostředí pro život and TJ04000322, program Zéta), and by the Johannes Amos Comenius Programme (P JAC) (project No. CZ.02.01.01/00/22_008/0004605, Natural and anthropogenic georisks).

M.Š. conceptualized the process, developed the methodology, wrote the original draft, rendered support in project administration; T. Č. wrote the original draft, rendered support in formal analysis, investigated the study; T. H. developed the methodology, validated the process; M. R. validated the process, provide the resources, rendered support in project administration; A. G. investigated the study and validated the process; J. S. investigated the study, wrote and validated the process, rendered support in project administration; K. N. investigated the study; T. C. conceptualized the work, validated the process, supervised the work.

All relevant data are included in the paper or its Supplementary Information, doi: 10.6084/m9.figshare.28608491.

The authors declare there is no conflict.

Avila
C.
,
Bayona
J. M.
,
Martin
I.
,
Salas
J. J.
&
Garcia
J.
(
2015
)
Emerging organic contaminant removal in a full-scale hybrid constructed wetland system for wastewater treatment and reuse
,
Ecological Engineering
,
80
,
108
116
.
Azizi
D.
,
Arif
A.
,
Blair
D.
,
Dionne
J.
,
Filion
Y.
,
Ouarda
Y.
,
Pazmino
A. G.
,
Pulicharla
R.
,
Rilstone
V.
,
Tiwari
B.
,
Vignale
L.
,
Brar
S. K.
,
Champagne
P.
,
Drogui
P.
,
Langlois
V. S.
&
Blais
J. F.
(
2022
)
A comprehensive review on current technologies for removal of endocrine disrupting chemicals from wastewaters
,
Environmental Research
,
207
,
112196
.
Bartelt-Hunt
S. L.
,
Snow
D. D.
,
Damon-Powell
T.
,
Brown
D. L.
,
Prasai
G.
,
Schwarz
M.
&
Kolok
A. S.
(
2011
)
Quantitative evaluation of laboratory uptake rates for pesticides, pharmaceuticals, and steroid hormones using POCIS
,
Environmental Toxicology and Chemistry
,
30
,
1412
1420
.
Burkhardt-Holm
P.
(
2010
)
Endocrine disruptors and water quality: a state-of-the-art review
,
International Journal of Water Resources Development
,
26
,
477
493
.
Cai
K.
,
Elliott
C. T.
,
Phillips
D. H.
,
Scippo
M. L.
,
Muller
M.
&
Connolly
L.
(
2012
)
Treatment of estrogens and androgens in dairy wastewater by a constructed wetland system
,
Water Research
,
46
,
2333
2343
.
Carranza-Diaz
O.
,
Schultze-Nobre
L.
,
Moeder
M.
,
Nivala
J.
,
Kuschk
P.
&
Koeser
H.
(
2014
)
Removal of selected organic micropollutants in planted and unplanted pilot-scale horizontal flow constructed wetlands under conditions of high organic load
,
Ecological Engineering
,
71
,
234
245
.
Černá
T.
,
Ezechiáš
M.
,
Semerád
J.
,
Grasserová
A.
&
Cajthaml
T.
(
2022
)
Evaluation of estrogenic and antiestrogenic activity in sludge and explanation of individual compound contributions
,
Journal of Hazardous Materials
,
423
,
127108
.
Chen
J.
,
Liu
S. S.
,
Wang
Y. J.
,
Li
J.
,
Liu
Y. S.
,
Yang
F.
&
Ying
G. G.
(
2021
)
Optimized constructed wetlands enhance the removal and reduce the risks of steroid hormones in domestic wastewater
,
Science of The Total Environment
,
757
,
143773
.
Dai
Y. N.
,
D
A.
,
Yang
Y.
,
Tam
N. F. Y.
,
Tai
Y. P.
&
Tang
X. Y.
(
2016
)
Factors affecting behavior of phenolic endocrine disruptors, estrone and estradiol, in constructed wetlands for domestic sewage treatment
,
Environmental Science & Technology
,
50
,
11844
11852
.
Gikas
G. D.
,
Papaevangelou
V. A.
,
Tsihrintzis
V. A.
,
Antonopoulou
M.
&
Konstantinou
I. K.
(
2021
)
Removal of emerging pollutants in horizontal subsurface flow and vertical flow pilot-scale constructed wetlands
,
Processes
,
9
(
12
), 1–18.
https://doi.org/10.3390/pr9122200
.
Han
Y.
,
Qi
C.
,
Niu
Z.
,
Li
N.
&
Tang
J.
(
2024
)
Occurrence and risk assessment of endocrine-disrupting chemicals in wastewater treatment plants in the Chaohu Lake Basin
,
Frontiers in Environmental Science
,
12
, 1–15.
https://doi.org/10.3389/fenvs.2024.1409011
.
Harman
C.
,
Allan
I. J.
&
Vermeirssen
E. L. M.
(
2012
)
Calibration and use of the polar organic chemical integrative sampler – a critical review
,
Environmental Toxicology and Chemistry
,
31
,
2724
2738
.
He
Y. J.
,
Nurul
S.
,
Schmitt
H.
,
Sutton
N. B.
,
Murk
T. A. J.
,
Blokland
M. H.
,
Rijnaarts
H. H. M.
&
Langenhoff
A. A. M.
(
2018
)
Evaluation of attenuation of pharmaceuticals, toxic potency, and antibiotic resistance genes in constructed wetlands treating wastewater effluents
,
Science of The Total Environment
,
631–632
,
1572
1581
.
Herrera-Melian
J. A.
,
Guedes-Alonso
R.
,
Borreguero-Fabelo
A.
,
Santana-Rodriguez
J. J.
&
Sosa-Ferrera
Z.
(
2018
)
Study on the removal of hormones from domestic wastewaters with lab-scale constructed wetlands with different substrates and flow directions
,
Environmental Science and Pollution Research
,
25
,
20374
20384
.
Hijosa-Valsero
M.
,
Matamoros
V.
,
Martin-Villacorta
J.
,
Becares
E.
&
Bayona
J. M.
(
2010a
)
Assessment of full-scale natural systems for the removal of PPCPs from wastewater in small communities
,
Water Research
,
44
,
1429
1439
.
Hijosa-Valsero
M.
,
Matamoros
V.
,
Sidrach-Cardona
R.
,
Martin-Villacorta
J.
,
Becares
E.
&
Bayona
J. M.
(
2010b
)
Comprehensive assessment of the design configuration of constructed wetlands for the removal of pharmaceuticals and personal care products from urban wastewaters
,
Water Research
,
44
,
3669
3678
.
Kättström
D.
,
Beronius
A.
,
Boije af Gennäs
U.
,
Rudén
C.
&
Ågerstrand
M.
(
2025
)
Impact of the new hazard classes in the CLP regulation on EU chemicals legislation
,
Environmental Sciences Europe
,
37
(
1
),
19
.
https://doi.org/10.1186/s12302-025-01054-4
.
Lundqvist
J.
,
Mandava
G.
,
Lungu-Mitea
S.
,
Lai
F. Y.
&
Ahrens
L.
(
2019
)
In vitro bioanalytical evaluation of removal efficiency for bioactive chemicals in Swedish wastewater treatment plants
,
Scientific Reports
,
9
,
7166
.
Neale
P. A.
,
O'Brien
J. W.
,
Glauch
L.
,
Konig
M.
,
Krauss
M.
,
Mueller
J. F.
,
Tscharke
B.
&
Escher
B. I.
(
2020
)
Wastewater treatment efficacy evaluated with in vitro bioassays
,
Water Research
,
9
,
100072
.
Ngeno
E.
,
Ongulu
R.
,
Orata
F.
,
Matovu
H.
,
Shikuku
V.
,
Onchiri
R.
,
Mayaka
A.
,
Majanga
E.
,
Getenga
Z.
,
Gichumbi
J.
&
Ssebugere
P.
(
2023
)
Endocrine disrupting chemicals in wastewater treatment plants in Kenya, East Africa: concentrations, removal efficiency, mass loading rates and ecological impacts
,
Environmental Research
,
237
, 1–12.
https://doi.org/10.1016/j.envres.2023.117076
.
Nguyen
P. M.
,
Afzal
M.
,
Ullah
I.
,
Shahid
N.
,
Baqar
M.
&
Arslan
M.
(
2019
)
Removal of pharmaceuticals and personal care products using constructed wetlands: effective plant-bacteria synergism may enhance degradation efficiency
,
Environmental Science and Pollution Research
,
26
,
21109
21126
.
Nivala
J.
,
Neale
P. A.
,
Haasis
T.
,
Kahl
S.
,
Konig
M.
,
Muller
R. A.
,
Reemtsma
T.
,
Schlichtinge
R.
&
Escher
B. I.
(
2018
)
Application of cell-based bioassays to evaluate treatment efficacy of conventional and intensified treatment wetlands
,
Environmental Science-Water Research & Technology
,
4
,
206
217
.
Papaevangelou
V. A.
,
Gikas
G. D.
,
Tsihrintzis
V. A.
,
Antonopoulou
M.
&
Konstantinou
I. K.
(
2016
)
Removal of endocrine disrupting chemicals in HSF and VF pilot-scale constructed wetlands
,
Chemical Engineering Journal
,
294
,
146
156
.
Routledge
E. J.
&
Sumpter
J. P.
(
1996
)
Estrogenic activity of surfactants and some of their degradation products assessed using a recombinant yeast screen
,
Environmental Toxicology and Chemistry
,
15
,
241
248
.
Šereš
M.
,
Innemanová
P.
,
Hnatková
T.
,
Rozkošný
M.
,
Stefanakis
A.
,
Semerád
J.
&
Cajthaml
T.
(
2021
)
Evaluation of hybrid constructed wetland performance and reuse of treated wastewater in agricultural irrigation
,
Water
,
13
,
1165
.
Shappell
N. W.
,
Billey
L. O.
,
Forbes
D.
,
Matheny
T. A.
,
Poach
M. E.
,
Reddy
G. B.
&
Hunt
P. G.
(
2007
)
Estrogenic activity and steroid hormones in swine wastewater through a lagoon constructed-wetland system
,
Environmental Science & Technology
,
41
,
444
450
.
Sklarz
M. Y.
,
Gross
A.
,
Yakirevich
A.
&
Soares
M. I. M.
(
2009
)
A recirculating vertical flow constructed wetland for the treatment of domestic wastewater
,
Desalination
,
246
,
617
624
.
Song
H. L.
,
Yang
X. L.
,
Nakano
K.
,
Nomura
M.
,
Nishimura
O.
&
Li
X. N.
(
2011
)
Elimination of estrogens and estrogenic activity from sewage treatment works effluents in subsurface and surface flow constructed wetlands
,
International Journal of Environmental Analytical Chemistry
,
91
,
600
614
.
Toro-Velez
A. F.
,
Madera-Parra
C. A.
,
Pena-Varon
M. R.
,
Lee
W. Y.
,
Bezares-Cruz
J. C.
,
Walker
W. S.
,
Cardenas-Henao
H.
,
Quesada-Calderon
S.
,
Garcia-Hernandez
H.
&
Lens
P. N. L.
(
2016
)
BPA and NP removal from municipal wastewater by tropical horizontal subsurface constructed wetlands
,
Science of The Total Environment
,
542
,
93
101
.
Tousova
Z.
,
Vrana
B.
,
Smutna
M.
,
Novak
J.
,
Klucarova
V.
,
Grabic
R.
,
Slobodnik
J.
,
Giesy
J. P.
&
Hilscherova
K.
(
2019
)
Analytical and bioanalytical assessments of organic micropollutants in the Bosna River using a combination of passive sampling, bioassays and multi-residue analysis
,
Science of The Total Environment
,
650
,
1599
1612
.
Vystavna
Y.
,
Frkova
Z.
,
Marchand
L.
,
Vergeles
Y.
&
Stolberg
F.
(
2017
)
Removal efficiency of pharmaceuticals in a full scale constructed wetland in East Ukraine
,
Ecological Engineering
,
108
,
50
58
.
Wu
H. M.
,
Zhang
J.
,
Ngo
H. H.
,
Guo
W. S.
,
Hu
Z.
,
Liang
S.
,
Fan
J. L.
&
Liu
H.
(
2015
)
A review on the sustainability of constructed wetlands for wastewater treatment: design and operation
,
Bioresource Technology
,
175
,
594
601
.
Yost
E. E.
,
Meyer
M. T.
,
Dietze
J. E.
,
Meissner
B. M.
,
Worley-Davis
L.
,
Williams
C. M.
,
Lee
B.
&
Kullman
S. W.
(
2013
)
Comprehensive assessment of hormones, phytoestrogens, and estrogenic activity in an anaerobic swine waste lagoon
,
Environmental Science & Technology
,
47
,
13781
13790
.
Yu
Q. M.
,
Yang
X. D.
,
Zhao
F. Z.
,
Hu
X. D.
,
Guan
L. C.
,
Ren
H. Q.
&
Geng
J. J.
(
2022
)
Spatiotemporal variation and removal of selected endocrine-disrupting chemicals in wastewater treatment plants across China: treatment process comparison
,
Science of The Total Environment
,
835
,
155374
.
Zhang
D. Q.
,
Gersberg
R. M.
,
Zhu
J. F.
,
Hua
T.
,
Jinadasa
K.
&
Tan
S. K.
(
2012
)
Batch versus continuous feeding strategies for pharmaceutical removal by subsurface flow constructed wetland
,
Environmental Pollution
,
167
,
124
131
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).