Combining Analytical Strategies to Provide Qualitative and Quantitative Analysis and Risk Assessment on Pharmaceuticals and Metabolites in Hospital Wastewaters

Abstract

The improper disposal of hospital wastewater (HWW) is a primary source of pharmaceutical pollution in aquatic systems. The complexity of the HWW matrix presents significant challenges for analytical chemists, necessitating meticulous sample preparation as the initial step for the analysis, followed by instrumental analysis. In the present study, a combination of dispersive solid phase extraction and solid phase extraction was evaluated for the preparation of HWW samples from two hospitals in Porto Alegre, Brazil, both for screening and quantitative analysis. The experiments performed by UHPLC-QTOF MS allowed the identification of 27 compounds and 23 suspected compounds. Furthermore, the UHPLC-QqQ-MS analysis enabled the quantification of 21 compounds, with concentrations ranging from 1.17 µg L⁻¹ to 213.33 µg L⁻¹. Notably, the pharmaceutical ciprofloxacin was detected at a concentration that exceeded the reported risk level for Microcystis aeruginosa. The environmental risk assessment revealed that the risk quotient (RQ) for several of the compounds quantified in the two HWW matrices exceeded 1, with the risk quotient of the mixture of compounds (RQ_mix) being approximately 30 × 10⁶ for Hospital A and 20 × 10⁶ for Hospital B. According to these findings, the two HWW systems exhibited risk levels for aquatic species and small rodents, thereby contributing to the persistence of pharmaceuticals in the environment.

1. Introduction

According to the World Health Organization, there has been an 8% increase in global life expectancy in recent years, largely attributable to advancements in pharmaceuticals for the prevention and treatment of infectious diseases [1]. However, this increase in access to pharmaceuticals and personal care products has led to significant environmental concerns. Also, wastewater treatment plants (WWTPs) are unable to adequately address the resulting challenges. These compounds have become virtually ubiquitous as they spread through different aquatic systems (such as rivers, lakes, and groundwater) [2,3,4,5,6].

According to the Atlas Esgotos, in Brazil, the majority of the generated sewage runs directly into water bodies, and only about 46% of it is treated in WWTPs [7]. The disposal of hospital wastewater (HWW) is a primary source of pharmaceutical pollution in aquatic systems. The magnitude of this pollution varies according to the size and occupancy of the hospital structure [8,9,10]. In recent years, various studies have been conducted to assess and monitor the compounds in HWW. The complexity of the HWW matrix varies from site to site, and it exhibits toxicity effects that could impact the environment depending on the concentration of compounds in each effluent [11,12,13,14,15,16,17,18].

A multitude of classes and compounds have been documented as quantifiable in HWW. In a particular study, ibuprofen, a nonsteroidal anti-inflammatory drug (NSAID), was reported with a range of 1.5 µg·L⁻¹ to 151 µg·L⁻¹, and acetaminophen, an analgesic and antipyretic drug, was reported with a range of 0.5 µg·L⁻¹ to 29 µg·L⁻¹ [15,19]. β-blockers, such as atenolol and metoprolol, were reported in concentrations ranging from 0.1 µL−1 to 122 µg·L⁻¹ and 0.6 to 2.0 µg·L⁻¹, respectively [10,15,19]. Antibiotics, such as azithromycin and erythromycin, were detected at concentrations as high as 7.3 µg·L⁻¹ and 83.0 µg·L⁻¹, respectively [10,14].

In this context, sample preparation constitutes a critical step aimed at mitigating matrix interferences, enhancing selectivity and sensitivity, and ensuring the robustness and reproducibility of the instrumental method [20,21,22,23]. The most frequently employed method is solid phase extraction (SPE), a technique that facilitates the elimination or reduction of interference in complex matrices and concomitantly provides the preconcentration of the analytes in the sample. A variety of SPE cartridges are employed in the analysis of pharmaceuticals from environmental samples, including polymeric reversed-phase sorbent (Oasis HLB), functionalized polymeric reversed-phase sorbent (Strata-X), polar embedded C18 stationary phase (Supelco C18), and non-polar polymeric phase (LiChrolut EN) [22,24,25,26,27].

Alternative methods: These include dispersive liquid–liquid microextraction, stir bar sorptive extraction, disposable extraction pipette, and electrocoagulation–flotation [28,29,30,31]. Dispersive solid phase extraction (DSPE) is used as a common cleanup step, widely employed in the QuEChERS method [32]. The quick, easy, cheap, effective, rugged, and safe (QuEChERS) method was originally developed for recovering pesticide residues from fruits and vegetables, but rapidly obtained popularity in the analysis of other analytes in different matrices. This method is a simple and straightforward extraction technique involving an initial partitioning followed by an extract clean-up using dispersive solid-phase extraction (DSPE) [32,33].

In DSPE, different sorbents, such as primary–secondary amine (PSA), octadecyl bonded silica (C18), graphitized carbon black (GCB), carbon nanotubes, and synthetic polymers are utilized in the extraction of pigments and the mitigation of matrix interferences from liquid samples, while maintaining minimal interference in the analyte’s concentration. C18 is employed for the removal of lipophilic interferents, while PSA has demonstrated efficacy in the removal of acidic interferences from the matrix [23,32,34,35].

The analysis of complex matrices, such as HWW, poses a significant challenge to analytical chemists. Typically, the screening of a broad scope of analytes is accomplished through qualitative analysis, employing high-resolution mass spectrometry (HRMS) instruments, such as quadrupole time-of-flight (QTOF) MS. This approach is particularly advantageous as it provides data on the combined full mass spectrum and high mass resolution and accuracy [36,37]. Consequently, it facilitates the differentiation of co-eluted compounds with highly similar masses, enabling retrospective analysis. The employment of purpose-built databases facilitates the investigation of hundreds of analytes for which no analytical standard is available. These databases contain information such as molecular mass, [M + H], and [M − H] ions, along with characteristic fragments generated by the parental ion [11,12,36,38,39,40].

The HRMS instruments are preferable for full scan analysis due to their higher selectivity; however, instruments with triple quadrupole (QqQ) analyzers are more suitable for the quantification of target analytes. Instruments for liquid chromatography coupled to QqQ analyzers (LC-QqQ) exhibit higher sensitivity, resulting in lower detection limits when compared with those obtained with HRMS instruments, when the same sample is analyzed by the two instrumental systems. This discrepancy can be attributed to the advancements made by companies in recent years, focusing on the interface and the ion optics of QqQ analyzers [36,38,41,42].

The environmental risk assessment of a wastewater can be calculated using the risk quotient (RQ) methodology. To calculate the RQ, the predicted environmental concentration (PEC) must be determined when quantification data are not available, or the measured environmental concentration (MEC) must be determined when quantification data are available. The RQ is then calculated as a ratio between the MEC (or PEC) and the predicted no-effect concentrations (PNEC), which is based on chronic ecotoxicity tests. When RQ > 1, an ecological risk is indicated. This methodology has been employed in numerous studies to assess the environmental risk of HWW. For instance, the RQ index was used to evaluate the environmental risk of HWW in Brazil [11], and the RQ was employed to assess the environmental risk of pharmaceuticals from HWW and its discharge on WWTP [18].

In the following study, hospital wastewater samples from two hospitals in Porto Alegre (Brazil) were prepared using a combination of DSPE and SPE techniques, and subsequently analyzed for both qualitative and quantitative purposes. To the best of our knowledge, there are no reported studies in the current literature regarding the combination of DSPE and SPE for sample preparation, which would provide a cleanup step prior to the SPE and might prevent the deposit of lipids and other materials on the upper portion of the solid phase, which could lead to the formation of preferential flow paths, and would remove some other constituents from the matrix, such as pigments and minor constituents. The qualitative analysis of target (confirmed) and suspect pharmaceuticals was performed by using a UHPLC-QTOF MS system with a custom-built database. The results of the screening analyses were then used to guide the development of the quantitative analytical method. The quantitative analyses of target compounds were performed using a UHPLC-QqQ MS system, and the results were applied to perform the environmental risk assessment of the HWW samples from both sites.

2. Materials and Methods

All pharmaceutical and metabolites standards (see Table S1, Supplementary Materials) used in this study were of analytical grade (>90% purity) and purchased from diverse suppliers. Isotopically labeled diazepam, used as an internal standard, was purchased from LGC Standards (Wesel, Germany). All solvents used in sample preparation and chromatography separation were LC-MS grade (Lichrosolv^®) acetonitrile (ACN), methanol (MeOH), and formic acid (purity = 98%), and were purchased from Merck (Darmstadt, Germany). Ultrapure water (18.2 MΩ·cm) was obtained with a MILLI-Q^® system (Millipore, Jaffrey, NH, USA). For simulated hospital wastewater (SWW), magnesium sulphate (MgSO₄·7H₂O) was purchased from Synth (Diadema, Brazil), calcium chloride (CaCl₂·2H₂O) and potassium phosphate (K₂HPO₄) were purchased from Dinamica (Indaiatuba, Brazil), and urea (CH₄N₂O), peptone bacteriological, and beef extract were purchased from KASVI (Pinhais, Brazil) were employed.

SPE cartridges OASIS HLB (200 mg and 60 mg) were purchased from Waters Corporation (Milford, MA, USA), while Bond ElutPlexa (500 mg and 200 mg) and Bond Elut PPL (1 mg) were purchased from Agilent Technologies (Santa Clara, CA, USA) and Strata-X (200 mg) was purchased from Phenomenex (Torrance, CA, USA). The PSA (40 µm) and C18 (40 µm) employed for the DSPE step were purchased from Supelco (Bellefonte, PA, USA) and GCB was purchased from Merck (Darmstadt, Germany).

The HWW samples were collected from two hospitals located in the city of Porto Alegre, Brazil (about 1.5 million inhabitants). Hospital A (HA) has a total capacity of 831 inpatients and the wastewater generated in this hospital is discharged directly into the public sewage system with no prior treatment. The samples from HA were collected directly from the hospital’s sewage out stream on six days spaced over two months throughout 2018 (12 February 2018, 9 April 2018, 11 June 2018, 13 August 2018, 8 October 2018, 10 December 2018). Aliquots of 1 L were collected once a month and then filtered in qualitative paper and stored in amber flasks under refrigeration (−20 °C). A composite sample from HA was prepared by mixing a 30 mL aliquot from each date and this composite sample was used as a representative of HA wastewater.

HB has a total capacity of 159 inpatients. In this hospital, the wastewater was collected from an equalizing fosse located on hospital grounds. As at HA, the effluent from HB is directly discharged into the public sewage system, without prior treatment. The sampling was performed on two dates (19 February 2020 and 10 March 2020) in which 2 L aliquots of wastewater were collected, filtered in qualitative paper and stored in amber flasks under refrigeration (−20 °C). There were more sampling dates scheduled but these were canceled due to the COVID-19 pandemic. One composite sample from HB was prepared by mixing a 30 mL aliquot from each date and this composite sample was used as a representative of HB wastewater.

The sample preparation protocol entailed a DSPE step, followed by SPE, then evaporation to dry the extract, and finally redissolving it with a H₂O/MeOH solution. The DSPE procedure involved the addition of a sample aliquot of 10 mL to a tube containing 50 mg of PSA and 50 mg of C18. The tube was then subjected to a 1 min vortex, after which the upper layer was loaded into the Strata-X SPE cartridge. Prior to use, the SPE cartridge was conditioned with 3 mL of ultrapure water (pH 7).

The sample was loaded and passed through the cartridge by gravity. Subsequently, the cartridge was dried under vacuum for 15 min, after which the analytes were eluted with 8 mL of MeOH. The extract was then evaporated to dryness using a gentle and regular flow of nitrogen at 40 °C, and redissolved with 1 mL of H₂O/MeOH (90:10, v/v), this proportion was used as it was the same composition for the chromatographic liquid phase. The final extract was filtered through a PTFE membrane (0.22 µm) into a vial, which was then used for the screening and quantification methods. The SPE step was adapted from a previous study [43]. Figure 1 represents a step-by-step sample preparation diagram. The pre-concentration factor in all sample preparation was 10 times.

A standard solution containing 91 pharmaceuticals and metabolites was prepared in ultrapure water. Simulated wastewater (SWW) was prepared according to Lumbaque et al. [44]. Two standard solutions with all pharmaceuticals were prepared at 20 µg L⁻¹ and 200 µg L⁻¹ by proper dilution of aliquots from the stock solution in SWW. These solutions were then analyzed directly using the UHPLC-QTOF-MS system, without undergoing any sample preparation steps, in order to generate a standard chromatogram for future reference.

During the screening evaluation, samples were analyzed using a UHPLC system (Shimadzu Nexera X2, São Paulo, Brazil) connected to a QTOF mass spectrometer (Bruker Daltonics Impact II, Billerica, MA, USA) operated with an electrospray ion source (ESI) in positive (ESI+) and negative (ESI-) modes. The raw data were processed with TASQ^® Software v.2.2 (Bruker). The analytical column utilized was a reverse-phase Hypersil Gold C18 analytical column (2.1 mm × 150 mm; 3 μm i.d.) with controlled temperature (30 °C), which is the same configuration used in previous studies that aimed to determine levels of pharmaceuticals in HWW [11,40,45,46,47,48]. The injection volume was set at 5 µL for both ionization modes. The database was constructed using previously available information, including MzCloud (www.mzcloud.org) and Mass Bank (www.massbank.eu). Further details regarding the screening methodology can be found in the Supplementary Materials. The screening step was implemented in order to determine the most effective sample treatment method and also as a preliminary result of the pharmaceutical content in the wastewaters. Isotopically labeled diazepam-d5 was utilized as an internal standard.

The quantitative analysis was conducted using a UHPLC-MS/MS system from Agilent (Santa Clara, CA, USA) connected to an Agilent G6470A TQ MS/MS triple quadrupole detector, with a column temperature controller and an autosampler. The UHPLC system was equipped with a Poroshell HPH C18 column (2.1 mm × 50 mm; 1.9 μm i.d.) from Agilent (Santa Clara, CA, USA). The instrument control and data processing were facilitated by the software MassHunter Quantitative Analysis v. 10.0 from Agilent (Santa Clara, CA, USA). The injection volume was set at 6 µL, the mobile phase was configured for a gradient with a flow rate of 0.400 mL min⁻¹, and the column temperature was maintained at 40 °C. Further details regarding the quantification methodology can be found in the Supplementary Materials. The internal standard, isotopically labeled diazepam-d5, was utilized to ensure the accuracy and precision of the analytical process.

The PNEC value was estimated based on toxicity screening by quantitative structure activity relationship (QSAR) obtained by the Toxicity Estimation Software Tool (TEST)—version 5.1 software. The lowest predicted value of LC₅₀ or LD₅₀ for each pharmaceutical was calculated according to Equation (1) [11,18,49,50]. Daphnia magna, fathead minnows, and rats were considered as the three ecotoxicological endpoints for the PNEC calculations.

$$PNEC={\displaystyle \frac{QSAR TEST LC50 or LD50}{1000}}$$

(1)

The risk quotient (RQ) of each pharmaceutical was calculated based on the MEC results according to Equation (2), and the risk quotient of the pharmaceuticals mixture (RQ_mix) was calculated according Equation (3) [18].

$$RQ={\displaystyle \frac{MEC}{PNEC}}$$

(2)

$${RQ}_{mix}={\sum }_i^n{\displaystyle \frac{{MEC}_i}{{PNEC}_i}}$$

(3)

3. Results

3.1. Sample Preparation Technique

The Strata-X 200 mg cartridge was selected for SPE experiments due to its higher capacity for extraction and its cost-effectiveness compared to the Bond ElutPlexa 200 mg. To assess the DSPE step, three distinct sorbents were examined: C₁₈, PSA, and GCB. These sorbents were utilized as a cleanup step prior to SPE. The sorbents were evaluated in both separate and combined modes. The optimal conditions, characterized by enhanced peak resolution and intensity, were identified through the use of a blend of 50/50 (w/w) PSA and C₁₈, with a total quantity of 100 mg (30 mL of sample). Consequently, the sample preparation entailed a DSPE step, in which a PSA/C₁₈ mixture was employed as the sorbent (1 min on vortex), followed by a SPE step utilizing a Strata-X 200 mg cartridge.

The initial item evaluated for the sample preparation was the type of SPE cartridge to be utilized, given that SPE functions as a “filter” for the preconcentration. It was hypothesized that some analytes would exhibit a stronger interaction with the solid phase and would be retained during the elution step. In

Table 1. List of cartridges used for SPE and the total amount of pharmaceuticals detected.

Cartridge	Sorbent Amount (mg)	Compounds Detected
Hydroxylated, amide-free surfaces and a nonpolar PS-DVB core (Bond Elut Plexa)	500	26
	200	32
Functionalized polymeric reversed-phase sorbent (Strata-X)	200	32
Styrene-divinylbenzene polymer modified with a proprietary nonpolar surface (Bond Elut PPL)	1	13
Polymeric reversed-phase sorbent (Oasis HLB)	200	29
	60	29

, the cartridges evaluated and the total number of analytes detected by screening analysis after the SPE are enumerated. To simulate hospital effluent, SWW was spiked with 20 µg L⁻¹ and 200 µg L⁻¹ by proper dilution of aliquots from the stock solution. This was carried out during the selection of the best cartridges for SPE, and to establish the best condition for the DSPE step.

3.2. Screening Analysis

A total of 50 compounds containing pharmaceuticals and metabolites were identified in the HWW samples evaluated through screening analysis. The screening results are displayed in Table S2 (ESI+) and Table S3 (ESI−). Figure 2 shows the comparison between the total ion count (TIC) results for acetaminophen (A) in simulated wastewater, (B) the standard in solvent at a concentration of 10 µg·L⁻¹, and (C) in real hospital wastewater. Compounds were classified as suspect or confirmed. Confirmed compounds were those that matched the peak information obtained from the corresponding analytical standard, detected in the analysis of spiked SWW.

The HA analysis identified 48 compounds, classified into 22 distinct pharmaceutical classes, while the HB analysis identified 44 compounds, categorized into 18 classes. Figure 3 presents the classes of compounds detected in each HWW and their respective relative amounts.

3.3. Compounds Quantification

A total of 23 compounds exhibited a satisfactory linear response (R² ≥ 0.995) among the analytes present in the mixture that was injected for quantification purposes. Recovery tests were conducted at three distinct levels (5.20 and 60 µg·L⁻¹), and the results ranged from 70% to 120%. In certain instances, a linear fitting was established for the calibration curve; however, the compound’s concentration was below the limit of quantification (LOQ). The final concentration was corrected by the respective dilution and the preconcentration factor (10). The recovery results, LOQ, limit of detection (LOD), and R² values are presented in

Table 2. Recovery results, LOD, LOQ, and R²; values obtained from the quantification method.

Analyte	Recovery (%)			LOD a (µg L−1)	LOQ a (µg L−1)	R2
	5 µg L−1	20 µg L−1	60 µg L−1
5-Hidroxythiabendazole	110	106	106	0.03	0.10	0.9967
Albendazole sulfoxide a	97	113	89	0.71	2.37	0.9954
Chloramphenicol	70	101	109	0.01	0.04	0.9981
Ciprofloxacin	120	108	79	1.13	3.77	0.9986
Clindamycin	73	120	84	0.92	3.09	0.9985
Diazepam	88	115	88	0.10	0.34	0.9960
Erythromycin	118	118	106	0.17	0.56	0.9985
Ethisterone	<LOQ	111	87	2.73	9.09	0.9959
Levamisole	97	105	119	0.06	0.21	0.9964
Levofloxacin	<LOQ	112	83	1.82	6.05	0.9964
Lincomycin	93	102	99	0.41	1.36	0.9991
Metoprolol	<LOQ	114	86	5.01	16.71	0.9995
Ofloxacin	<LOQ	119	83	1.77	5.89	0.9964
Oxolinic acid	112	112	92	0.54	1.79	0.9955
Praziquantel	70	105	99	0.01	0.01	0.9977
Progesterone	87	106	120	0.90	3.01	0.9967
Sulfadiazine	103	109	108	0.56	1.88	0.9989
Sulfadoxine	117	117	106	0.80	2.65	0.9969
Sulfamethazine	76	108	101	0.59	1.98	0.9952
Sulfathiazole	86	106	104	0.74	2.46	0.9963
Trimethoprim	<LOQ	113	92	1.85	6.17	0.9953

^a Calculated based on the signal to noise ratio.

. The matrix effect was evaluated by comparing the responses of spiked SWW with those of the standards of the same concentration. The final quantification results are presented in

Table 3. Measured environmental concentration of pharmaceuticals and metabolites in HWW.

Analyte	Concentration (µg L−1)
	HA	HB
5-Hidroxythiabendazole	<LOQ	1.32 ± 1.14
Albendazole sulfoxide	4.63 ± 0.73	1.17 ± 0.06
Chloramphenicol	1.56 ± 0.01	1.33 ± 0.01
Ciprofloxacin	172.50 ± 20.46	150.69 ± 2.32
Clindamycin	23.20 ± 1.05	35.50 ± 0.07
Diazepam	2.23 ± 0.95	68.61 ± 1.80
Erythromycin	<LOQ	<LOQ
Ethisterone	<LOQ	<LOQ
Levamisole	<LOQ	<LOQ
Levofloxacin	239.64 ± 7.86	3.45 ± 0.32
Lincomycin	3.00 ± 0.04	5.90 ± 0.28
Metoprolol	213.33 ± 0.72	164.04 ± 1.95
Ofloxacin	239.64 ± 8.58	3.45 ± 0.32
Oxolinic acid	<LOQ	<LOQ
Praziquantel	<LOQ	<LOQ
Progesterone	<LOQ	<LOQ
Sulfadiazine	<LOQ	1.87 ± 0.06
Sulfadoxine	<LOQ	<LOQ
Sulfamethazine	<LOQ	<LOQ
Sulfathiazole	<LOQ	<LOQ
Trimethoprim	50.17 ± 1.52	169.89 ± 4.38

.

3.4. Environmental Risk Assessment

The environmental risk assessment was performed for those analytes quantified in the HWW for which LC₅₀ and/or LD₅₀ values were available. The LC₅₀, LD₅₀, PNEC, and RQ values are available in Supplementary Materials Table S4. Figure 4 represents the compilation of RQ results for the pharmaceuticals in logarithmical scale (for better visualization). Results above 0 indicate suspected ecological risk (RQ > 1) and results below 0 indicate no suspected ecological risk (0 ≤ RQ < 1).

Figure 5 represents the RQ for each compound and each ecotoxicological level as heatmaps. RQ values above zero (which indicate RQ > 1) are represented in yellow to red scale (darker tones of red indicate higher RQ value and higher risk). RQ values below zero (which indicate 0 ≤ RQ < 1) are represented in yellow to green scale (lighter tones of green indicate lower RQ value and thus lower ecotoxicological risk).

4. Discussion

A predominant presence of antibiotics was observed in both HA and HB samples. The presence of such compounds may lead to a growth of antibiotic-resistant microorganisms as a result of genetic adaptation or through selective activities in water bodies. Moreover, their presence may have an impact on human and animal life as they are difficult to remove using urban WWTPs [51,52,53]. Amoxicillin, ampicillin, chloramphenicol, ciprofloxacin, ofloxacin, and trimethoprim are examples of antibiotics detected in the screening.

The presence of antibiotic residues in HWW are of environmental and public health concern as resistant bacteria may spread through superficial water bodies, soils, and therefore to potable water sources [54]. That spread creates paths for human and wildlife contact with resistant bacteria from healthcare centers. Even in low concentration, the presence of antibiotics in HWW may lead to selective pressure of bacteria, favoring the survival and proliferation of resistant strains, which may spread through the environment and spread the resistant genes, making the treatment of infectious diseases more difficult [55].

Antibiotic residues, as well as those from other pharmaceuticals, may disturb ecosystems and harm aquatic life. These compounds can affect the growth and behavior of diverse organisms, leading to the imbalance of aquatic systems. The prolonged exposure to low concentrations of antibiotics, in addition to favoring the development of resistant bacteria, can also lead to chronic adverse effects [56].

The second group most found in the screening was nonsteroidal anti-inflammatory drugs (NSAIDs), with diclofenac, ibuprofen, ketoprofen, naproxen, and nimesulide being examples of detected pharmaceuticals. The use of such compounds has grown over the years, with an estimated 300 million people making use of them daily [57]. Regarding the NSAIDs, different effects have been reported on aquatic organisms, such as cytopathological damage to trout by diclofenac [58], oxidative stress in zebra fish caused by ibuprofen [59], cellular damage caused to Daphnia magna by acetylsalicylic acid [60], and reported interference with chlorophyl activation and enzymatic activities of algae caused by naproxen [61].

Metabolites, such as 17-α-estradiol, 4-aminoantipiryne, and 17-α-ethinyl-estradiol, appeared as the third most common group in the samples analyzed. The by-products of metabolic activity or biodegradation have been reported to retain pharmaceutical activity from the parental compound. Furthermore, these by-products may be bioaccumulated in sediments or aquatic organisms, persisting in the environment and contributing to overall toxicity [62,63,64].

During the matrix effect studies, the responses of the spiked SWW were lower than 60% in comparison with the responses of the compounds in ultrapure water. This discrepancy can be attributed to co-eluted components that interfered with the ionization process in the mass spectrometer, as expected due to the complexity of the matrix. It was hypothesized that the wastewater from HA would contain a higher concentration of pharmaceuticals than that from HB, given that the former has a significantly larger capacity and mass flow than the latter. However, the samples from HA were collected directly from the effluent stream, while those from HB were collected from a fosse where pharmaceuticals and metabolites would concentrate between discharges.

The compounds with higher concentrations detected included three antibiotics (ciprofloxacin 175.5 µg L⁻¹, levofloxacin 239.6 µg L⁻¹, and ofloxacin 237.1 µg L⁻¹), and one β-blocker (metoprolol 213.3 µg L⁻¹). The two HWW streams evaluated presented concentrations of ciprofloxacin that exceeded reported risk levels for Microcystis aeruginosa [65]. In a recent study on the contamination of rivers around the globe [66], ciprofloxacin was determined with an average concentration of 22.1 ng L⁻¹, a concentration approximately 10,000 times lower than in the HWW. Diazepam, in the same study, was reported with an average concentration of 31 ng L⁻¹, a value about 100 times lower than the concentration in the HA effluent and 2000 times lower than in the effluent from HB. Lincomycin was reported with an average concentration of 795 ng L⁻¹, a value approximately four times lower than in HA effluent and 7.5 times lower than in HB effluent. Trimethoprim was determined in rivers of South America with an average concentration of 217 ng L⁻¹, 230 times lower than in HA effluent and 780 times lower than in HB effluent. Such data highlight the contribution of HWW to the pharmaceutical loading in the environment

For HA, only chloramphenicol showed no environmental risk regarding the LD₅₀ for rats. All the other compounds presented a risk for fathead minnow, Daphnia magna, and rats. For HB, the majority of compounds presented a risk for fathead minnow and Daphnia magna, and some compounds presented a risk for rats. The results of the RQ_mix for HA were in the order of 30 × 10⁶ and for HB on the order of 20 × 10⁶, so it is possible to state that the wastewater from HA had a higher environmental impact than that from HB. Regarding the determined pharmaceuticals that presented RQ values > 1, only diazepam and lidocaine had RQ values < 10 in HA and HB wastewater, levofloxacin, lincomycin, and progesterone, both for rats. All the other pharmaceuticals had higher RQ values.

It is important to highlight that just as the HWW matrix is heterogeneous regarding the presence of different classes of pharmaceuticals, so is the risk class associated with those compounds, either qualitatively or quantitatively determined. The risks and damage associated with the pharmaceuticals are diverse. Examples are genotoxicity (losartan) [67], mutagenicity (ranitidine) [68], hepatic (ketoprofen) [69] or neural risk (lidocaine) [70], cardiac disfunction (chloramphenicol) [71], and carcinogenicity (amoxicillin) [72]. The development and application of advanced oxidation processes have been studied as an alternative to remove pharmaceuticals in WWTPs [73]. Notably, the use of Fenton and photo-Fenton processes reduce the concentration of such pollutants in wastewaters and their use should be considered, in addition to the widely employed activated sludge process [74].

5. Conclusions

The combined DSPE and SPE sample preparation method presented in this study was a suitable fit for the screening and quantification of pharmaceutical compounds and metabolites present in hospital wastewaters, which form a highly complex matrix. Even though the sample preparation step was more time consuming then the single step SPE method, it enabled the determination of more compounds than the single step method (32 in comparison to 23). The screening method enabled the identification of five metabolites and 45 pharmaceuticals. The quantification method demonstrated a linear response to 21 analytes (1 metabolite and 20 active compounds), enabling the quantification of analytes present in the samples within a range of 1.17 µg L⁻¹ to 213.33 µg L⁻¹. Notably, the pharmaceutical ciprofloxacin was detected in a concentration exceeding the reported risk level for Microcystis aeruginosa. The RQ and RQ_mix results indicated that both wastewater samples posed a risk to aquatic species and small rodents. This is concerning because less than 50% of Brazilian sewage is treated in WWTPs, and the high values indicate a risk of environmental persistence and direct damage to fauna and flora. Consequently, the development of more efficient technologies to remove contaminants with such characteristics is imperative, particularly given the presence of numerous drugs, some of which are present in high concentrations.