The Removal of Acidic Drugs from Domestic Wastewater Using an Innovative System of Constructed Wetlands/Stabilization Ponds in Series

Abstract

Nature-based solutions represent a decentralized wastewater treatment proposal, offering diverse mechanisms for effectively removing emerging contaminants, particularly acidic pharmaceuticals. This study evaluated the performance of acidic-drug (diclofenac, fenofibrate, ibuprofen, gemfibrozil, fenoprofen, naproxen, and indomethacin) removal from wastewater using a surface-flow constructed wetland with an organic bed (Eichhornia crassipes (Mart.) Solms, 18 ind/m²), and a horizontal subsurface-flow constructed wetland, divided into three sections. The process was complemented by two stabilization ponds and other horizontal subsurface-flow wetlands using papyrus (Cyperus papyrus L., 8–13 ind/m²) and tezontle as support media. The industrial-scale system (67.8 m²) was fed with wastewater at a rate of 1.33 m³/d with a hydraulic time retention of about 5.8 days. Drugs were quantified by gas chromatography. The results showed that gemfibrozil and indomethacin were completely removed (100%), while diclofenac (73%) and naproxen (94%) showed significant removals. Fenoprofen was not removed. Ibuprofen and fenofibrate showed increased concentrations, resulting in negative removals due to anoxic conditions (ibuprofen) and a slightly neutral pH (fenofibrate). These findings underscore the system’s ability to improve water quality by removing most acidic drugs, suggesting that the hybrid design is particularly effective in treating specific wastewater contaminants.

1. Introduction

Emerging Organic Contaminants (EOCs) are unregulated and regulated compound residues, which include pharmaceuticals, personal care products, pesticides, hormones, and nanoparticles, among others [1]. A wide range of pharmaceuticals, including antibiotics, anti-inflammatory drugs, and antiretrovirals, have been detected in rivers at concentrations ranging from ng/L to μg/L (i.e., pharmaceutical products: 0.3–146 ng/L) [2,3,4]. Their average consumption by developed countries is from 50 to 150 g/inhab-year, while in developing countries consumption is around 15 g/inhab-year [3]. Differences are related to the medical care/pharmaceutical investment made by developed countries (5–10% GDP, or gross domestic product) compared to developing countries (2–5% GDP) [5].

Acidic drugs (ADs) are pharmaceuticals from aromatic carboxylic acids, which have dissociation constants (pK_a) between 3.4 and 4.8, a high water-solubility (>100 mg/L), a wide ranges of adipose tissue affinity (log k_ow = 0.16–4.0), and a relatively low degradability (20–70%), favoring their environmental mobility [6,7,8,9]. ADs include several pharmaceutical products, such as analgesic/anti-inflammatory agents (ibuprofen, fenoprofen, naproxen, diclofenac, and indomethacin, among others) and lipid and cholesterol regulators (gemfibrozil and fenofibrate, among others) [10].

ADs are usually absorbed and metabolized by the human body. However, a fraction of their active principle or metabolites are excreted. Indeed, diclofenac is excreted mainly in the urine (65–70%) than in feces (20–30%) [11]. Gemfibrozil is excreted more in the urine (66%) than in feces (6%) [11]. Therefore, wastewater is the main AD route into the environment [11]. ADs could have sublethal effects on aquatic organisms, due to their ability to bioaccumulate in their adipose tissues [1,12]. The exposure of Oryzias latipes to ibuprofen (up to 0.1 mg/L) has resulted in reported impaired reproduction at 6 weeks of exposure [13]. Likewise, their exposure to diclofenac (1.0 µg/L) has resulted in cellular toxicity, genotoxicity (p53), and estrogenic effects [14]. Meanwhile, their exposure to gemfibrozil (3.7 mg/L) has resulted in a reduction in the testosterone levels (a hormone for oocyte maturation) in female fish, producing a decrease in fish hatchability [15]. Therefore, AD removal from wastewater is essential to protect aquatic ecosystems and human health.

ADs, such as naproxen, diclofenac, ibuprofen, indomethacin, gemfibrozil, and fenoprofen, have been removed at rates of between 15.0 and 99.7% using conventional activated sludge processes [16,17,18]. Nevertheless, some ADs (clofibric acid and carbamazepine, among others) are recalcitrant products, achieving low removal within conventional wastewater treatment plants (WWTPs) (e.g., carbamazepine: −122–58%) [17,19,20]. Different physical (e.g., reverse osmosis, ultrafiltration), chemical (e.g., ozonation, advanced oxidation processes) and biological/nature-based-solution (e.g., membrane bioreactors (MBRs), constructed wetlands (CWs)) technologies have been studied at the laboratory scale for AD removal [12,21]. However, biological treatment technologies are the most widely used for AD removal because of their lower investment/operational costs, requiring a lower energy demand than other technologies. For instance, the operational cost and energy consumption of MBRs are 88.6% and from 30 to 61.9% lower than for solar photo-Fenton technologies, so their large-scale application is prohibitive, especially in developing countries [22,23].

Constructed wetlands (CWs) and stabilization ponds (SPs) are nature-based solutions, being cost-effective (investment costs: ~0.25 USD/m³ treated wastewater) and sustainable alternative for wastewater treatment, particularly for rural/decentralized developing countries [24,25,26]. These technologies have reported the removal of some ADs, such as ibuprofen (CWs: 71–96%; SPs: 75–100%), naproxen (CWs: 50–100%; SPs: 75–100%), diclofenac (CWs: 80–100%; SPs: 65–90%) [10,20,27], and fenofibrate (68%) [28].

Several factors influence AD removal within CWs and SPs, such as the AD composition/concentration, design/operational factors (water depth, hydraulic retention time, and vegetation, among others), and environmental conditions (sunlight, temperature, redox conditions, and pH) [12,20,29,30]. In fact, for the long-term effective performance within CWs and SPs separately, AD removal is challenging [30]. Nevertheless, CWs and SPs in series could be a synergistic solution [31]. ADs have distinct physicochemical properties that make them more susceptible to specific processes occurring in CWs and SPs, such as photodegradation, sorption, and biodegradation [32]. Aerobic/anaerobic biodegradation is the main route for the removal of ibuprofen (53–79%), naproxen (63–75%), and gemfibrozil (45–58%) [25]. Thus, subsurface-flow CWs favor biodegradation under a mixed microenvironment, where aerobic and anaerobic bacteria coexist [24,33]. Meanwhile, diclofenac is photodegraded (67–77%), which is possible in SPs and surface-flow CWs, due to its large surface being exposed to sunlight [33,34]. Pharmaceutical removal within CWs and SPs has been well studied separately [12,26,29,34].

Bio-adsorbent materials, such as biochar, activated alumina, and MOFs, have shown high efficiency in removing pharmaceuticals [35,36,37,38]. These materials have been studied in both batch and continuous-flow systems in terms of their adsorption performance. Their effectiveness may be reduced by competing with solutes like salts, metal ions, and organic matter [39].

Moreover, Eichhornia crassipes (Mart.) Solms has been shown to have significant potential for removing pharmaceuticals (removal rates of up to 91.18% for diclofenac and ciprofloxacin) [40]. Its biomass can be utilized for various valuable by-products, including animal feed, energy generation, and composting [41], making it a promising solution for wastewater management and ecosystem restoration. Cyperus papyrus L. wetlands have achieved up to 84% ibuprofen removal, compared to 51% in unplanted systems [42].

The removal of acidic drugs by constructed wetlands or stabilization ponds has been studied independently, but their combined use in a hybrid system is rarely reported in the literature [33,43]. This study presents a real-life application of this system under semi-dry climatic conditions in Latin America. The hybrid configuration integrates constructed wetlands and stabilization ponds operated in series to treat domestic wastewater, targeting the removal of seven acidic pharmaceuticals: ibuprofen, fenoprofen, naproxen, diclofenac, indomethacin, gemfibrozil, and fenofibrate.

2. Materials and Methods

2.1. Wastewater Sampling Study Zone

Wastewater was collected from the Community Development Center located in Alpuyeca, Xochitepec, Mexico. The wastewater treated in this study was obtained from sources within the Community Development Center, including restrooms, kitchens, and cleaning areas associated with the kindergarten and adult facilities. There are no direct discharges or inputs from nearby rivers, streams, or springs. Therefore, the influent corresponds strictly to domestic wastewater generated on-site. The Xochitepec municipality is in the central-western zone of Mexico (18°42′ N, 99°11′ W) at an altitude of 1109 masl. Xochitepec has a semi-dry climate with rain in summer. The average daily temperature is 23 °C (T_max = 28 °C, T_min = 12.9 °C), the evaporation rate is 6.1 mm/d, and the average annual precipitation is between 750 and 840 mm. The main water bodies in Xochitepec are the rivers (Tetlama-Cuantepec, Salado, and Apatlaco), streams of permanent flow (El Sábado, El Tlazala, Agua Fría, Corralillo, and El Colotepec), and springs (San Ramón, Pablo Bolero, Real del Bridge, and the Sports Unit in Campo La Vega). The underground hydrology of Xochitepec is highly permeable since it is within the Cuernavaca aquifer area (recharge/extraction: 333 million m³/120 million m³). The water supply for drinking comes from deep wells (9475.2 m³) and springs (1123.92 m³). However, there is a lack of access to drinking water, so 12.4% of the population does not have piped water and other basic services (sewerage, electricity). To address local sanitation challenges, a decentralized wastewater treatment system was implemented at the center, allowing the treated wastewater to be reclaimed for irrigating green areas.

To help address the water issues in the Alpuyeca Community Development Center, a decentralized wastewater treatment plant (WWTP) was implemented. The treated wastewater is recycled for irrigating the center’s green areas.

2.2. Wastewater Treatment System

The wastewater treatment system was designed to treat 1.33 m³/d in a continuous process, with a total treatment area of 67.8 m². The treatment system (Figure 1) consists of the pre-treatment phase (grates and grease trap); a septic tank (V = 2.6 m³ and A = 3.3 m²); constructed wetland 1 (organic-bed constructed wetland = SF-CW₁); constructed wetland 2 (horizontal subsurface flow), divided into three sections with an average depth of 0.6 m (HSSF-CW_2a, HSSF-CW_2b, HSSF-CW_2c); stabilization pond 1 (SP₁); constructed wetland 3 (horizontal subsurface flow = HSSF-CW₃), stabilization pond 2 (SP₂); and constructed wetland 4 (horizontal subsurface flow = HSSF-CW₄). The CWs have a filter bed depth of 0.6 m (water depth = 0.5 m), while the maturation lagoons have a depth of 0.5 m. Each treatment unit has a length of 5.67 m and a width of 1.5 m. The hydraulic retention time in the MPs was twice that in the CWs, with a total hydraulic retention time in the system of 5.8 days.

The vegetation used in SF-CW₁ included water lilies (Eichhornia crassipes (Mart.) Solms), while the vegetation in HSSF-CW_2a, HSSF-CW_2b, HSSF-CW_2c, HSSF-CW₃, and HSSF-CW₄ included papyrus (Cyperus papyrus L.). The support medium was tezontle with a particle diameter of between 2 and 4 cm.

The treatment system was designed to comply with NOM-003-SEMARNAT-1997 [44] for reuse in public services with direct contact. This standard specifies the following discharge limits: biochemical oxygen demand over 5 days (BOD₅), 20 mg/L; total suspended solids, (TSS), 20 mg/L; oils and fats, 15 mg/L; fecal coliforms, 240 MPN/100 mL; and helminth eggs, ≤1 egg/L.

2.3. Analytical and Instrumental Methods

2.3.1. Sample Collection

The amber glass bottles (V = 4 L) were conditioned before taking the samples. Several washes were performed with hot and cold drinking water, distilled water, technical-grade acetone, and HPLC (High-Performance Liquid Chromatography)-grade methanol. Clean bottles were dried at 100 °C for 12 h and placed in clean coolers until sampling [45]. The sampling was conducted during the rainy season, which typically extends from May to October in this region. Water samples were collected from 0.5 to 1.0 m below the surface [46]. Two composite samples were taken, one in the morning and one in the afternoon. Each sample consisted of 4 L of water, collected over a 2 h period, with 1 L of water taken every 30 min [47]. The samples were conditioned at the sampling site before being stored in preconditioned amber glass bottles. These bottles were placed in coolers with ice and transported to the laboratory. Upon arrival, the samples were stored at 4 °C until analysis. Ten sampling points were taken, described as follows: (1) influent, (2) effluent from SF-CW₁, (3) water from HSSF-CW_2a, (4) water from HSSF-CW_2b, (5) effluent from HSSF-CW_2c, (6) effluent from SP₁, (7) effluent from HSS-CW₃, (8) effluent from SP₂, (9) effluent from HSSF-CW₄, and (10) WWTP effluent.

2.3.2. Sample Conditioning

The glassware used for analysis was washed sequentially with acetone and methanol of HPLC grade. Methanol from the final rinse was collected for use as a glass blank. Wastewater samples (1 L) were filtered using Whatman fiberglass filters (Cat. No. 1820-047). The filtered samples were acidified (pH ≤ 2), followed by solid-phase extraction (SPE) using HF Bond Elut-C18 cartridges (Cat. No. 730013) from Thermo Fisher Scientific. The cartridges were dried under a vacuum and, subsequently, at room temperature of 18 h. Samples were eluted with methanol of HPLC grade, and the eluate was collected in amber vials. The samples were concentrated using high-purity nitrogen gas, and 2 mL of methanol was added to the concentration. To separate the acidic drugs, present in the concentrate, 200 µL of pentafluorobenzyl bromide (2% v/v) and 2 µL of triethylamine were added and vigorously shaken in a Genie 2 vortex. The vials were then heated at 100 °C for 2 h, cooled to room temperature, dried under nitrogen gas, and stirred with 1 mL of HPLC-grade methanol. To verify the correctness of the procedure for quantifying acidic drugs, the following controls and samples were processed: a glass blank control (solvents used to wash glass materials), a blank control (deionized water), a duplicate of one sample, two fortified samples (sample + mixture of an acid standard solution at a concentration of 1 mg/L each), and two synthetic samples (deionized water + 100 µL of each standard at a concentration of 1 mg/L). This methodology was adapted from Sacher et al. [48], incorporating some modifications based on the specific characteristics of wastewater in Mexico.

2.3.3. Water Quality Characterization

Standard methods from APHA-AWWA-WEF [49] were used to analyze water quality parameters. The pH was measured using the electrometric method (4500-H⁺ B), and the temperature was recorded in situ (2550 B). Electrical conductivity was determined with a conductivity meter (2510 B), while BOD₅ was evaluated using the five-day incubation method (5210 B). Total suspended solids were assessed gravimetrically (2540 D). Total nitrogen was measured through persulfate digestion (4500-NC), and total phosphorus was measured using digestion followed by the ascorbic acid method (4500-PB). Oils and fats were quantified via the partition-gravimetric method (5520 B). Lastly, total coliforms were analyzed using multiple-tube fermentation (9215 B).

2.3.4. Acidic-Drug Quantification

Analytical standards for selected acidic drugs, each with a purity greater than 99.5%, were prepared using Sigma-Aldrich HPLC-grade reagents: diclofenac (DFC, CAS: 15307-86-5), ibuprofen (IBU, CAS: 15687-27-1), gemfibrozil (GMF, CAS: 25812-30-0), fenoprofen (FNP, CAS: 31879-05-7), naproxen (NPX, CAS: 22204-53-1), fenofibrate (FNF, CAS: 49562-28-9), and indomethacin (IND, CAS: 53-86-1). Chemical characteristics are presented in

Table 1. Chemical characteristics of acidic drugs.

Trade Name	Function	Molecular Structure (g/mol)	pka	Log Know	Log Koc	Chemical Structure	Reference
Diclofenac	Analgesic/anti-inflammatory	C14H11Cl2NO2	4.2	4.02	2.91		[50,51]
		296.1
Ibuprofen	Analgesic/anti-inflammatory	C8H10N4O2	4.9	3.71	3.5		[52,53]
		206.3
Gemfibrozil	Lipid regulator	C15H22O3	4.8	4.77	2.63		[53]
		250.3
Fenoprofen	Anti-inflammatory	C15H14O3	4.5	4.05	-		[54]
		242.3
Naproxen	Analgesic/anti-inflammatory	C14H14O3	4.2	3.18	2.54		[55,56]
		230.3
Indomethacin	Analgesic/anti-inflammatory	C19H16ClNO4	4.5	4.27	3.15		[57,58]
		357.8
Fenofibrate	Lipid regulator and metabolite	C20H21ClO4	-	5.19	3.58		[59,60]
		360.8

.

These standards were dissolved in HPLC-grade methanol to create 1000 µg/mL solutions of each individual drug. Dilutions ranging from 0.01 to 0.2 µg/mL were prepared to construct the calibration curve. These dilutions were processed following the same conditioning procedure as wastewater sample effluents post-elution. Quantification of the acidic drugs was conducted using a Shimadzu GCMS-TQ8040 gas chromatograph–triple quadrupole mass spectrometer [61].

The performance of the proposed method was evaluated in terms of limits of detection (LOD) and quantification (LOQ), specificity, linearity, and precision, following the guidelines of the United States Environmental Protection Agency (USEPA) [62] Method 1694 for the analysis of pharmaceuticals and personal care products in water, soil, sediment, and biosolids. The method demonstrated excellent linearity across the complete calibration curve (R² = 1) and exhibited a range for LOD/LOQ between 1.3 and 92 ng/L. The efficiencies of AD removal was calculated using Equation (1):

$$AD removal \left(\%\right)={\displaystyle \frac{Co-Ce}{Co}}\times 10$$

(1)

where Co represents the influent AD concentration of each CW/SP, and Ce represents the effluent AD concentration of each CW/SP.

2.3.5. Database Analysis

Descriptive statistical analysis was applied to the results. Specifically, average values and standard deviations were calculated for each parameter to summarize the concentrations and removal efficiencies at different stages of the treatment system. Non-additional statistical tests were performed.

3. Results and Discussion

3.1. Operational Behavior

The characterization of the influent/effluent is detailed in

Table 2. Physicochemical and microbiological characterization of raw and treated wastewater.

Parameter	Symbol	Units	Raw	Treated
Hydrogen potential	pH	-	7.3–7.6	7.9–8.1
Temperature	T	°C	23.0–26.0	20.0–23.0
Electrical conductivity	EC	µS/cm	1338.0–1623.00	244.0–1158.0
Biochemical Oxygen Demand	BOD5	mg/L	141.0–257.0	1.0–9.0
Suspended Solid Total	SST	mg/L	74–107.0	8.0
Total Nitrogen	TN	mg/L	84.6	18.0
Total Phosphorous	TP	mg/L	45.0–47.0	3.0–5.0
Oils and fats	O&F	mg/L	10.0–10.5	9.1–9.5
Total Coliforms	TC	NMP/100 mL	1.5 × 107–1.5 × 107	2.0–5.0

. The pH values at the different stages of the system showed fluctuations (7.28–7.74). These variations are characteristic of this type of system due to factors such as seasonal temperature variation, photosynthetic processes, plant type, and filter medium, among others [8]. Nevertheless, pH changes could impact microbial activity and nutrient cycling [63]. The influent pH was slightly alkaline (7.58) and remained practically constant from the septic tank (7.74). In the next stages, the pH had a different behavior; the stabilization ponds (SP₁ and SP₂) exhibited the most alkaline conditions (8.2 and 8.1, respectively), followed by the effluent pH of HSSF-CW₄ (pH = 8.06). The lowest pH values were present in the CWs, where HSSF-CW₃ and HSSF-CW₄ showed higher pH levels (7.55 and 7.96, respectively) than SF-CW₁ and HSSF-CW₂ (7.28 and 7.40, respectively). It is important to mention that the monitoring of this nature-based-solution system was carried out in August (summer–autumn). Constructed wetlands during summer usually increase their photosynthetic activity, which influences carbonate ionization by increasing hydroxyl ions. Indeed, the increase in hydrogen ions generated by the oxidation of carbonate is exchanged for other cations (e.g., Ca, Mg, K) by the filter medium [8]. Moreover, subsurface-flow wetlands generate reduced microenvironments (<400 mV) that promote an alkaline pH [64]. On the other hand, the pH increases in SPs due to the biological removal of carbon dioxide through photosynthesis and ammonia extraction. This increase in pH facilitates and makes nutrient removal more effective, particularly of nitrogen and phosphorus [65]. The slight decrease in pH from CW effluents is associated with producing organic substances that acidify the medium. These substances are generated within the CW during the growth and decomposition of plants and in the mineralization of organic matter. In the degradation of organic matter, heterotrophic bacteria produce acetic, butyric, and lactic acids that reduce the pH [66]. Although slight variations in pH were observed, these remained within a narrow range, demonstrating the significant buffering capacity of the treatment system [67].

The influent temperatures were slightly higher (23 and 26 °C) than the effluent temperatures (20 and 23 °C). Temperature is a key factor for efficient biota performance in CWs and SPs. For instance, nitrifying bacteria grow and reproduce optimally at temperatures between 25 and 35 °C [7,68]. Temperatures in the studied WWTP were within the range reported in several studies using CWs and SPs (15–35 °C). These environmental conditions favor the removal of organic matter, nutrients, and other organic pollutants, with efficiencies reaching up to 100% [68,69].

Electrical conductivity decreased throughout the treatment, from an average value of 1338 µS/cm in the influent to 735.9 (±378.4) µS/cm in the effluent. This indicates the effective removal of dissolved solids during the treatment, where plants play an important role, for example, in the reduction in nutrients in the form of NO₃⁻–N (TN removal = 78.6%), which resulted in a lower number of ions in the wastewater effluent [70]. This was evidenced by the growth of water lilies and papyrus, which reached apical heights of 53.3 ± 7.6 cm and 151.4 ± 14.5 to 249.6 ± 40.2 cm, respectively. Another important factor is the dilution effect caused by precipitation, as sampling was conducted during the rainy season (September 2019). Increased rainfall can lead to a decrease in electrical conductivity by introducing low-mineral-content water into the system, effectively diluting influent concentrations [71].

Additionally, high temperatures enhance microbial activity, which may further contribute to reducing electrical conductivity [68,69]. Therefore, the combined effects of dilution, plant uptake, and microbial activity played a key role in decreasing the electrical conductivity.

The system achieved a total-suspended-solid removal of approximately 93% and a BOD₅ removal efficiency of 94%. Constructed wetlands (CWs) are highly effective in removing organic matter, achieving removal efficiencies exceeding 90% [72]. However, Morató et al. [73] showed that their effectiveness in pathogen removal is more variable, with reported efficiencies ranging from 2% to 90%, highlighting the need for an additional polishing stage [74]. Despite the challenges associated with pathogen removal in CWs—such as limited light penetration due to the dense vegetation, hydraulic retention time, and redox potential—the studied system achieved complete E. coli removal (7-log reduction) [75,76].

MPs played a crucial role by extending the water’s exposure to sunlight, with a hydraulic retention time twice that of CWs. In this regard, the region’s climatic conditions—characterized by high temperatures and low humidity—enhanced light availability, further promoting pathogen inactivation [6]. Additionally, microbial interactions within CWs may have contributed to pathogen removal through mechanisms such as biofilm activity, predation by nematodes and protists, attack by lytic bacteria and viruses, and competition for limited nutrients or trace elements [73].

Root secretions from aquatic macrophytes may have also facilitated the removal of fecal coliforms and other pathogenic bacteria. Moreover, an increase in antibiotic-producing bacteria in the rhizosphere could have further contributed to coliform inactivation. The stems of aquatic macrophytes and certain attached algae likely aided in coliform entrapment, fixation, and subsequent removal [71].

Among the macrophyte species present in the system, the water lily stands out for its ability to achieve fecal coliform removal efficiencies exceeding 80% [67,71]. Meanwhile, papyrus mats allow for easy vertical wastewater penetration and provide a large surface area for coliform entrapment, fixation, and inactivation [67].

Another key factor influencing coliform removal is the type of support medium used in CWs. Studies have shown that using a fine granular medium (~3.5 mm) instead of larger particles enhances microbial inactivation by 1.0 to 2.0 log units for fecal coliform and somatic coliphage removal. In our system, tezontle (2–4 cm) was used as the support medium, with particle sizes consistent with those previously being reported as effective [73]. Similarly, CW depth influences pathogen removal. Shallower water depths promote more energy-efficient biochemical reactions, creating oxidizing conditions that enhance treatment efficiency.

Overall, the combination of an extended retention time, sunlight exposure, microbial interactions, plant contributions, and an optimized support medium created favorable conditions for high coliform removal efficiency. This synergy resulted in the complete removal of E. coli, with values characteristics of the combination of CWs and SPs [77]. Consistent with our findings, previous studies have reported similar trends, with total coliform removal efficiency increasing from 87.4% to 99.9% when an MP was incorporated after CW [78].

3.2. AD Removal Performance

Figure 2 shows the concentrations of DFC, IBU, GMF, FNP, NPX, and IND from the hybrid system, CWs–SPs, in the eight treatment stages. The concentrations of the influent varies as follows: NPX > DFC > IND > GMF > IBU > FNF > FNP, with NPX having the highest concentration (599.3 µg/L) and FNP the lowest concentration (4.5 µg/L). Meanwhile, DFC (179.8 µg/L), GMF (20.8 µg/L), IBU (9.4 µg/L), and FNF (6.9 µg/L) showed intermediate concentrations. Previous studies have shown that NPXs is among the most frequently detected drugs in municipal effluents, with concentrations higher than 1 µg/L [79], together with IBU (e.g., 1983 ng/L), DFC (131.61 ng/L), and GMF (67 ng/L). On the other hand, FNP is not commonly found in water due to its low excretion rate in humans (3%) [80]. Additionally, FNP has been detected in North America but not consistently in Africa, making it one of the least-monitored ADs [80,81]. IND is not always detected because it is easily photodegraded [56]. FNF is usually found in low concentrations (e.g., 23 ng/L) [82] and is sometimes not detected at all [83].

At the end stage of the WWTP (Figure 3), some concentrations were reduced: DFC (0.0493 mg/L), GMF (0.0493 mg/L), IND (0.0380 mg/L), and NPX (−0.0493 mg/L). Meanwhile, others increased: IBU (0.1119 mg/L) and FNF (0.0717 mg/L). Regarding removal efficiencies at the end stage of the WWTP, they are ranked as follows: GMF = IND > NPX > DFC. Thus, the best removals were for GMF and IND (100%). A removal efficiency of 94% was registered for NPX and of 73% for DCF, while FNP was not removed at all (0%). On the contrary, IBU and FNF had negative removals (−1096% and −944%, respectively), which refer to an increase in the concentration of certain compounds compared to the influent; these values may be influenced by desorption phenomena, particularly given the relatively long hydraulic retention time of the system [84]. The following sections describe the behavior of each contaminant in the different stages of the WWTP.

The DFC concentration decreased in SF-CW₁ (0.1380 mg/L), HSSF-CW_2a (0.0561 mg/L), HSSF-CW_2c (0.0946 mg/L), SP₂ (0.1281 mg/L), and HSSF-CW₄ (0.0587 mg/L), with the lowest concentration being that of HSSF-CW₃ (0.032 mg/L), while the WWTP effluent recorded 0.049 mg/L. Removal efficiency varied from 15 to 59%. Additionally, it presented negative removal percentages in HSSF-CW_2b (−99%) and SP₂ (−300%). DFC removal is related to photolysis, which might be explained by the fact that the C-N-C bonds are easily broken for further degradation under solar light [85]. Other studies have shown that photodegradation has been identified as the main mechanism of DFC degradation in surface waters within wetland systems [51]. Likewise, small-scale CWs with different design configurations have demonstrated higher removal efficiencies in SF-CWs, where wastewater exposed to strong sunlight enhances photolysis. On the contrary, the low DFC removal could be due to the presence of chlorine in its structure, which confers stability and hinders its degradation [10,86]. Another possible factor is the decomposition of conjugated drugs, which increases the concentrations of the original drugs, as indicated by Simazaki et al. [87].

In general, FNF showed low concentrations, ranging from 0.0023 mg/L in HSSF-CW_2a to 0.0717 mg/L in the effluent, with removal efficiencies between 22% in HSSF-CW_2c and 44% in HSSF-CW_2a. However, at the end of the treatment, FNF concentration increased to 0.072 mg/L, resulting in a growth of 30 times the influent concentration. Removal is explained by photodegradation, together with the subsequent biodegradation process [80]. Also, it is expected to be efficiently removed by the adsorption process due to high hydrophobicity (log K_ow = 5.28); however, it could be a competition with other contaminants of adsorption on activated sites [82,87] reducing the removal efficiency. In contrast, negative FNF removal could increase after wastewater treatment due to its metabolic conversion back to the parent compound [88]. The existing literature does not address FNF removal using wetlands or stabilization ponds, but it does discuss other treatments, such as high-rate algal ponds with similar removal rates (from 30 to 70%) and sedimentation in ponds with better removal (100%); consequently, more studies are required.

FNP registered similar concentrations in the effluent during each stage, ranging from 0.0045 to 0.0046 mg/L, where non-increase or a decrease in the drug was observed [81]. This exposed that FNP removal is affected by pH values, resulting in the ionization of the target compound, which impacts the binding at the active sites. In the mentioned study FNP recovery exceeded 60% in the acidic region (pH = 3–5), while recoveries dropped below 50% at a pH greater than 7. Also, FNP has been found to persist within WWTPs with removal efficiency ranging from 0% to 41% [80,81]. On the other hand, it shows a high degradation rate (46–90%) in sludge from four different conventional WWTPs (based on activated sludge), which is probably due to the sludge retention time (optimal: 14–20 days) [55].

GMF had high removal rates (100%) in SF-CW₁ and HSSF-CW_2c, reaching concentrations close to 0 mg/L. However, it presented negative concentrations in SF-CW₁ (−0.0168 mg/L), HSSF-CW_2a (−0.0401 mg/L), HSSF-CW_2c (−0.0234 mg/L), MP₁ (−0.0226 mg/L), HSSF-CW₃ (−0.0467 mg/L), and MP₂ (−0.0165 mg/L). In contrast, GMF concentrations increased in HSSF-CW_2b (0.0187 mg/L) and HSSF-CW₄ (0.0074 mg/L), even doubling the concentration value. High removal is associated with the one-ring structure of gemfibrozil which favors biodegradation, and it is further enhanced when plants are incorporated into the CW design [10,89]. Plants facilitate the removal of ADs through direct adsorption and assimilation. Moreover, they create optimal conditions for removal by providing habitats for rhizosphere microorganisms, enhancing substrate porosity, intercepting particulate matter, and mitigating low-temperature effects [80]. These results align with other studies, reinforcing the role of CWs in sustainable wastewater treatment. For instance, the authors in [72] established that GMF efficiency was higher in planted horizontal-subsurface-flow CWs (58 ± 18%) compared with unplanted ones (49 ± 13%). The same occurred with vertical-surface-flow CWs (planted: 50 ± 11%, unplanted: 40 ± 3%). Negative concentration values can be attributed to the presence of compounds consistently below the quantification limit [90]; therefore, these values are considered to bbe approx. 0 mg/L. On the other hand, the concentration increase is attributed to GMF persistence in water, with a degradation time exceeding 70 days [79]. The results are like those of a study developed in Singapore, which reported the negative removal of GMF (−0.2–−91.3%). It explains that the concentration increase occurs due to the cleavages of its conjugated forms into the original parent compound by microorganisms and/or the sampling methods used [91,92].

IBU presented an increase in concentration and negative removal after almost every stage (−703 to −5%), except for in HSSF-CW_2a (0.0189 mg/L) and HSSF-CW₃ (0.0535 mg/L), with efficiencies of 24 and 75%, respectively. This variation could be explained by the fact that some pharmaceuticals are best reduced under aerobic conditions, while the removal of others is favored under anaerobic conditions [10]. Furthermore, shallow wetlands (0.3 m) show a higher IBU removal efficiency (81%) than deeper ones (0.5 m, 48%), possibly due to less anaerobic conditions [56]. Moreover, oxygen dynamics significantly influence treatment processes, with submerged plants maintaining oxic conditions and floating plants creating an anoxic environment. Negative removals are consistent with the results of the authors in [93], who found that IBU and its metabolites are not efficiently removed under anoxic conditions, leading to their accumulation.

IND registered a decrease in concentration in SF-CW₁ (0.0720 mg/L), HSSF-CW_2c (0.1884 mg/L), SP₁ (0.1347 mg/L), HSSF-CW₃ (0.0390 mg/L), and HSSF-CW (0.0256 mg/L). Average removal rates were between 26% (HSSF-CW_2c) and 71% (HSSF-CW₃). They are contrasted with negative removals, which showed a rise in concentrations in HSSF-CW_2a (0.1176 mg/L, −63%), HSSF-CW_2b (0.2530 mg/L, −115%), and SP₂ (0.0571 mg/L, −47%). Positive removals are explained because CW systems promote the microbial degradation of ADs via the root release of certain organic acids, thus reducing the amount of ADs absorbed by the root [94] The literature does not report IND removal by adsorption processes with tezontle as a support medium. Still, other studies have established that other drugs (e.g., carbamazepine) can be removed within CWs in subtropical climates using hybrid systems planted with polyculture that includes ornamental species and the use of tezontle as a ground filter material [60]. On the other hand, the increase in concentration is explained by IND persistence in water for extended periods (2.5 weeks in rivers), even in the presence of sediment. Furthermore, its degradation products did not significantly adsorb to sediment due to the protection supplied by the sediment against sunlight irradiation and the presence of nutrients [57]. This aligns with other studies where some pharmaceuticals exhibited higher concentrations in WWTP effluents, resulting in negative removal efficiencies (e.g., IND −53%). This is attributed to the release of metabolized drugs, which may revert to their original forms during wastewater treatment [85].

NPX exhibited concentrations as follows: SF-CW₁ (0.1061 mg/L), HSSF-CW_2a (0.0856 mg/L), HSSF-CW_2b (0.2221 mg/L), HSSF-CW_2c (0.2059 mg/L), SP₁ (0.2282 mg/L), HSSF-CW₃ (0.0813 mg/L), SP₂ (0.02813 mg/L), and HSSF-CW₄ (0.0117 mg/L), with removal rates ranging from 7% in HSSF-CW_2c to 82% in SF-CW₁. Overall, the effluent of almost every stage showed a decrease in NPX concentrations compared to the influent, except for HSSF-CW₄ and SP₁, which presented negative removals of −215% and −11%, respectively. Removal of ionic compounds, such as NPX, occurred because they are primarily present in their anionic form at a pH of between 5 and 8, and they exhibit a trend of sorption decreasing while the pH rises [55]. It is reported that the NPX mechanism of removal from wastewater is probably anaerobic degradation in CWs [89]. Another study pointed out that NPX removal efficiencies were 55.6–82.3%, 51.6–74.9%, and 22.1–50.5% in summer, autumn, and winter, respectively. On the contrary, as mentioned previously the increase in the concentration of ADs is attributed to the release of metabolized drugs, which may revert to their original forms during wastewater treatment [95]. Finally, Figure 4 shows a summary of the mechanisms of the removal of all of the ADs described above.

4. Conclusions

The hybrid system composed of CWs and SPs, operated in series under real and full-scale conditions with a hydraulic retention time of 5.8 days, showed high efficiency in removing macro-contaminants—achieving 93% removal of suspended solids, 94% of biochemical oxygen demand (BOD), and 99.9% of fecal coliforms. The removal of pharmaceutical compounds was more variable. Thus, GMF and IND were completely removed (100%), while IBU and FNF showed negative removal percentages, possibly due to metabolite reconversion or compound accumulation. FNP remained stable, likely influenced by the buffering effect of the system, as reflected by the narrow and stable pH range (7.3–8.1) observed during the monitoring period.

CWs were particularly effective for organic-matter removal, showing some limitations in removing specific pathogens and certain pharmaceuticals. The hybrid nature-based system enabled diverse removal mechanisms. On the one hand, DCF was mainly removed through photolysis, while FNF was removed by photodegradation and biodegradation. Finally, GMF was removed by biodegradation and plant uptake, and IBU and NPX were removed under anoxic conditions. IND degradation appeared to be enhanced by microbial activity in the rhizosphere. These findings underscore the system’s capacity to treat a broad spectrum of contaminants under real operating conditions and highlight the importance of internal factors—such as the system’s buffering capacity—in maintaining stable conditions that support contaminant removal. Further research into metabolite behavior and seasonal influences may provide deeper insights into long-term system performance.

5. Patents

Armando Rivas holds patent MX 414294 B.